Value Activation with Vertical Annual Rings – Material, Production, Products Dick Sandberg

Value Activation with vertical annual D Sandberg Value Activation with vertical annual rings – material, production, products Dick Sandberg

TRITA-TRÄ R-98-36 ISSN 1104-2117 ISRN KTH/TRÄ/R--98/36--SE

Value Activation with vertical annual rings – material, production, products by Dick U. W. Sandberg

Stockholm Doctoral Thesis KTH-Royal Institute of Technology Dept of Manufacturing Systems 1998 Wood Technology and Processing SE-100 44 STOCKHOLM

Abstract This thesis is a contribution to the R&D-program Value Activation at KTH- Trä. Two main aspects can be distinguished in the thesis. (i) A documentation and description of star-sawing technology from the initial idea to the first industrial plant. (ii) A description of research results within the fields of wood science and production technology, obtained during the development of the star-sawing technique. The two aspects are strongly linked to each other. The relation between ideas, scientific basis of the ideas and development results, the development of production technology and systems, and the dissemination of knowledge and technology transfer from all levels to the industrial implementation are described. The way in which all these development stages are connected in this work is also discussed. The thesis presents a new sawing technique developed at KTH, Div. of Wood Technology and Processing (KTH-Trä). The star-sawing concept, i.e. the sawing of the log according to the star-sawing pattern and the further refinement of the sawn product into PrimWood are described in detail. Experimental results are presented concerning different material properties of timber in general and of timber with vertical annual rings in particular. The investigations refer primarily to distorsions and crack formation of sawn timber of different stages: before drying, after drying, and as a result of moisture variations. In addition, the results of a few initial tests are presented where the deformation and crack formation of the wood at the microscopic level are described on the basis of a micromechanical approach. Using a new approach, the effect of weathering on crack formation has been investigated. Test samples with tangential and radial surfaces have been exposed for a period of three years. After many years of development work, star-sawing can now be carried out on an industrial scale, but the most important feature of the research work has been that the sawing pattern has for a number of years constituted a concrete and reality-oriented basis within the R&D-program Value Activation. The timber obtained in star-sawing meets several of the most important property requirements of future wood products, including accuracy in size and geometry, no cracks, controlled moisture movement , strength and hardness, and aesthetic and tactile factors.

3 The realization of an idea

I have seen it at a close distance, during these last that which is usually called constructive wood years of the 20th century – a different type of doctoral protection and which is concerned with the design of study in a Division with special work forms. Five or six different components, the choice of wood and its years ago, KTH-Trä chose to let an important idea orientation in the structure. Scientific confirmation is penetrate its whole activity. The choice was VALUE often lacking here. Whether this depends on the fact ACTIVATION, the utilisation of those intrinsic that the experiments are difficult to design and properties of the wood material which were then evaluate or to the fact that it is not glamorous re- poorly utilised in the industrial system in our Swedish search field is difficult to say. It is therefore gratifying sawmills. The idea also incorporated a focus on pro- that this thesis adds an important piece to the large blem-based learning and a comprehensive approach to knowledge puzzle, information about the difference in the subject of wood technology; both the appearance crack susceptibility between a tangentially and and function of the material, its manufacture and radially sawn surface. marketing, theory and practice, internal studies and It is a well known fact that there is an under-repre- external cooperation with companies. Here, Dick sentation of personnel with a higher technical Sandberg now presents his doctoral thesis on this education in sawmills and the timber industry topic. compared with most other industrial branches. Today, The activity at the department has been focused on however, more and more companies are becoming those properties which can be attained with timber in familiar with the idea of employing graduate engine- which the annual rings are ”vertical”. That which ma- ers, who know wood and who have an analytical kes the work and the thesis different is that the task ability for problem solving. But what do you do with a has been attacked from the viewpoint that the aim is doctor in this field? This thesis actually straightens out industrial implementation and knowledge transfer - a the question mark into an exclamation mark. The possible interpretation of the university’s third task. knowledge and ability which are demonstrated here To attain the goal, traditional scientific methods have hits the mark with regard to the development of the been used to determine how ”it is”, and the ingenious traditional wood material into competitive and better engineer’s way of working has been followed to create paid products. And this is exactly what the branch ”that which did not exist before”. Since I have been needs! able to follow the work in detail, I know that there has been a continuous feedback between product, Stockholm, October 1998 market, material and production process – and where necessary, where there have been knowledge gaps, the Ingemar Ekdahl creation of a deeper scientific foundation; in the best sense, a modern approach to product development. Ingemar Ekdahl is information manager at Trätek- the Institute for Our understanding of how wood materials behave, Wood Technology Research. He participated in developing the e.g. when they are exposed outdoors, is often stamped business idea for KTH-Trä in the early nineties and has thereafter by traditions and beliefs rather than by knowledge. A e.g. assisted in the supervision of undergraduate students in their field which is particularly subject to vague beliefs is MSc thesis projects.

4 The author wishes to thank the Nils and Dorthi Troëdsson Foundation for financial support for the execution of the work described in this thesis.

Nils och Dorthi Troëdsson Foundation

The Nils and Dorthi Troëdsson Foundation has funded the research in the wood material field, and an advanced climate chamber at KTH-Trä: The Troëdsson Climate Chamber.

Contents 1. Introduction 11 III. Radially sawn-timber. 1.1 Background, aim and target group 11 The influence of annual ring orientation on 1.2 Value Activation 12 crack formation and deformation in water 1.3 The disposition and limitations of the thesis 13 soaked pine (Pinus silvestris L) and spruce 1.4 The arrangement of the thesis 15 (Picea abies Karst) timber 95

2. Star-sawing - from idea to industrial reality 19 IV. Radially sawn-timber. 2.1 Martin Wiklund, professor at KTH 19 Gluing of star-sawn triangular profiles into 2.2 Sawing of the log according to the star-sawing method 19 formstable products with vertical annual rings 109 2.3 Dissemination of results from Value Activation 22 2.4 The development company PrimWood AB 23 V. Weathering of radial and tangential wood 2.5 The first industrial plant 25 surfaces of pine and spruce. 1. Crack formation and colour changes in 3. The star-sawing concept – sawmilling and value exposed surfaces 123 added products of triangles 31 3.1 Sawing according to the star-sawing method 31 VI. Weathering of radial and tangential wood surfaces 3.2 PrimWood - boards glued from star-sawn triangular profiles 40 of pine and spruce. 2. Microscope studies 133

4. Some properties of star-sawn timber and the VII. Material damage due to electron beam during dependence of these properties on annual ring creep test in environmental scanning electron orientation and distance from the pith 47 microscopy (ESEM) 4.1 Definition of vertical annual rings 47 4.2 Distorsion and cracks in timber after sawing, drying VIII. The influence of pith and juvenile wood on and cycles of wetting and drying 50 proportion of cracks in sawn timber when kiln 4.3 The change in wood surface during outdoor usage dried and exposed to wetting cycles 151 above ground 55 4.4 Preliminary studies of the mechanisms for distorsion IX. The influence of annual ring orientation on and crack formation in wood 57 strength and dimensional changes during moisture variation in finger-joins 155 5. Conclusions 67 X . Bending strength of I-beams with webs of I. Radially sawn timber. wood-fibre board and flanges of star-sawn Star-sawing – a new method for producing timber triangular profiles 159 with vertical annual rings 71 Appendix 1. List of technical reports 163 II. Radially sawn-timber. Knots – number, type and size in star-sawn triangular Appendix 2. Division of the work in the papers 164 profiles of pine (Pinus silvestris L) and spruce (Picea abies Karst) 85

Chapter1 Introduction

IV. Sandberg and Holmberg Since the technical reports and 1 Introduction 1998: Radially sawn timber. Gluing articles to a certain extent describe In this chapter the arrangement of star-sawn triangular profiles into the same thing and since the thesis of the thesis, and the R&D-pro- form-stable products with vertical is written in English, I have chosen gram “Value Activation” at annual rings. Holz als Roh- und to include only the articles listed KTH-Trä, of which the thesis is Werkstoff 56 (3). above in the thesis itself. Appendix a part is described. V. Sandberg 1998: Weathering 1. presents a list of technical 1.1 Background, aim and of radial and tangential wood reports which are related to the target group surfaces of pine and spruce. 1. thesis. Crack formation and colour The aims of the thesis are as This thesis is a summary with changes in exposed surfaces. In follows: complementary documentation and press Holzforschung. (i) To describe and document the analyses. The thesis is based on VI. Sandberg 1998: Weathering origin and development of the so- the one hand on a number of of radial and tangential wood called star-sawing method as a technical reports from Wood surfaces of pine and spruce. 2. practical aspect of the R&D-pro- Technology and Processing at Microscope studies. In press gram “Value Activation” at KTH- KTH (KTH-Trä) which are largely Holzforschung Trä. written in Swedish, but with an VII. Kifetew and Sandberg 1998: (ii) To investigate some of the ma- English summary, and on the other Material damage due to electron terial properties, which are very hand on the following articles pub- beam during creep test in important for sawn timber in gene- lished in scientific journals: Environmental Scanning Electron ral and for star-sawn timber in I. Sandberg 1996: Radially sawn Microscopy (ESEM). Submitted to particular. The important material timber. Star-sawing - a new Wood and Fibre Science. properties here are primarily those method for producing timber with VIII. Sandberg 1996: The influ- properties, which are related to the vertical annual rings. Holz als Roh- ence of pith and juvenile wood on deformation and crack behaviour und Werkstoff 54(3). proportion of cracks in sawn of wood. II. Sandberg and Holmberg timber when kiln dried and (iii) To elucidate the process 1996: Radially sawn timber. Knots - exposed to wetting cycles. Holz als relating to the implementation of number, type and size in star-sawn Roh- und Werkstoff 54(3). the radically new star-sawing triangular profiles of pine (Pinus IX. Sandberg 1997: The manufacturing system from the silvestris L) and spruce (Picea abies influence of annual ring orientation university research to the wood- Karst). Holz als Roh- und on strength and dimensional working industry. Werkstoff 54(6). changes during moisture variation The thesis is directed primarily to: III. Sandberg 1997: Radially in finger joints. Holz als Roh- und (i) Industrial representatives sawn timber. The influence of Werkstoff 55(1). within the wood-working annual ring orientation on crack X. Sandberg and Stehr 1997: industries, sawmills, carpenters, formation and deformation in Bending strength of I-beams with furniture companies, wooden water soaked pine (Pinus silvestris webs of wood fibre board and house manufacturers, joinery L) and spruce (Picea abies Karst) flanges of star-sawn triangular industries and other wood refining timber. Holz als Roh- und profiles. Holz als Roh- und companies, who are affected by or Werkstoff 55(3). Werkstoff 55(5). participate in product development

11 and quality work directed towards basic education, research and characteristic appearance, which long-term profitability. graduate studies, and the dis- also gives a certain tactile sense. (ii) Representatives of the semination of technology. The Potential profits should also be sawmill industry who have realized research work has been carried greatest here. As an example, Wik- that “bulk-sawing” is not in the out as a problem- or idea-based lund (1992) mentions that the cost long term a durable strategy for process, and the basic education is about three times greater for the Swedish sawmill industry, and has become progressively more wood which has been prepared for possibly those who wish to renew integrated with the research and furniture and visible features in their sawing process to increase oriented towards problem- and carpentry than for the V-grade, the value of timber. idea-based learning. i.e. ordinary building timber. (iii) Researchers within the The Value Activation program Other fields in which wood will wood sector who wish to use the includes many areas within which have great competitive strength in techniques developed within Value undergraduates and graduate the future compared with other Activation, i.e. to create wood students actively participate to materials are those of cost-effective products and manufacturing solve different tasks. The projects constructions and systems, and systems which are based on the within Value Activation involve at composites. Value Activation nature of the wood, and where least one, but usually several diffe- includes the first two of these material properties are refined, rent fields of competence from product areas, viz. visible wood taking into consideration annual wood products and their market and cost-effective constructions ring orientation. via production systems to wood and systems. Within the Value (iv) Researchers and other per- structure and properties. The gen- Activation program, Wiklund sons who seek to transfer R&D- eral strategy of the program is to (1991) has chosen the properties, results to an industrial application. develop the uses, products and which he judged to be the most (v) Decision-makers who handle production techniques in order to important for future wood questions concerning the transfer fully utilize wood’s property prof- products. Wiklund (1991) arrived of research results to industrial ile. In other words, one tries to use at the following properties: application. the wood trunk in such a way that (i) Accuracy in size and the optimum material utilisation in geometry 1.2 Value Activation the growing tree is also utilised in (ii) Freedom from cracks Research at KTH-Trä has since final products and constructions. (iii) Controlled movement with the end of 1990 been focused on This strategy was anchored at an changing humidity an integrated R&D-program called early stage within KTH-Trä and (iv) Strength and hardness Value Activation (Wiklund 1991, outwards towards external (v) Aesthetic and tactile factors 1993). The program is based on a financiers. A basic idea within Value fundamental view of wood mate- Wiklund (1992) states that the Activation is to the ability to rial, namely that there are wood competitive strength of wood mate- produce, on an industrial scale, properties which are seldom used rial is greatest when it is used timber and boards with vertical today but could be useful with new visibly in constructions like annual rings for products in which procedures. carpentry and furniture. When e.g. small and controlled moisture Value Activation strives towards wood is used visibly, advantage is movements are important. An idea a strong integration of the three taken of the fact that wood is a that was developed at an early main tasks of a university, viz. natural material with a stage within Value Activation was 12 to saw timber with rectangular and 1.3 The disposition and always be as revolutionary as the triangular cross-sections in a limitations of the thesis star-sawing pattern to be import- pattern which gives a high volume The thesis is a contribution to the ant. The less ingenious ideas and yield. The method which gives R&D-program Value Activation at innovations are important for sol- timber with vertical annual rings is KTH-Trä. Two main aspects can ving the many small problems called star-sawing, (Figure 1). be distinguished in the thesis: which arise, and to push the work (i) A documentation and forwards towards a goal. description of star-sawing The ideas have been of different technology from the initial idea to character. Some have involved the first industrial plant. materials science or manufacturing (ii) A description of research systems. These ideas have been tested with the help of scientific results within the fields of wood methodology; i.e. I have sought a science and production technology scientific foundation. This was for which have been obtained during example the case when the the development of the star-sawing influence of the annual ring technology. orientation on crack formation These two aspects are strongly during the outdoor exposure of Figure 1. Star-sawing - a new sawing linked to each other. The relation pattern to produce timber with vertical wood was investigated (Sandberg between ideas, scientific basis for annual rings free from pith and most of 1998a, 1998b). the juvenile wood. ideas, development of results, and Other ideas have been more After many years of develop- the development of production related to production technology. mental work, star-sawing can now technology and systems has been These ideas have been tested with be carried out on an industrial important for the industrial imple- a smaller degree of scientific scale, but the most important fea- mentation of the R&D-results, and methodology than has been used in ture of the research work has been to a high degree also for the pro- the material-related investigations. the sawing pattern which has for a gress of my own work. I have also An iteration between idea and number of years constituted a related the dissemination of know- testing stage has then been the concrete and reality-oriented basis ledge and technology transfer from normal working procedure, on the within the R&D-program Value all levels to industrial implemen- one hand to verify the idea and on Activation. tation. Figure 2 illustrates my view the other hand to drive forward The timber obtained in star- of how the work has been carried the technological development. sawing meets several of the most out, i.e. the relation between the The relation between idea and important property requirements different development stages. An development has then been very of future wood products which example is given here of what I intensive and the actual develop- have been mentioned here. The wish to describe with Figure 2. ment work was subsequently not method has many advantages, but The ideas have been the basis as systematic and analytical as re- it has also given rise to many and driving force for the work. In search directed towards the mate- problems. The problems relate to the actual development of ideas, I rial properties of wood. The materials technology, materials wish to emphasize the importance development of a glue-press to join science and production of creativity and innovations for star-sawn triangular profiles into technology. This will be illustrated the research and development of blocks is a good example of this in more detail in this thesis. results. The innovations need not type of development work. 13 Testing ideas Analysis and verification

Adaptation of ideas

Ideas - Scientific foundation - Testing stage - Aim - the basis and driving force traditional scientific development of dissemination of knowledge and for the R&D work methodology production technology industrial implementation and systems

Feedback Feedback Iteration between idea and testing stage

Feedback of new ideas that may give new research task and new areas

Aim of the research- Within Value Activation, the aim is to introduce desired products and production systems for wood, where the value derived from a log is considerably higher than that which is obtained in today’s industrial system. In addition, the dissemination of knowledge that can be used by both the industry and research is very important.

Figure 2. The development of new products - different phases in the work and their interrelationships.

14 There is also a natural very important. There is thus a 1.4 The arrangement of the connection between the scientific strong interaction between ideas, thesis and development stages. Some of scientific basis and development, I have chosen the following arr- the ideas produced in the more on the one hand, and the angement for the thesis: creative development work have spreading of knowledge and Chapter 2 describes the been verified and analysed industrial implementation, on the development of the star-sawing afterwards, i.e. after the idea has other. method from first ideas to been tested in practice. An The feedback of views and ideas industrial implementation and the example is the influence of the from the industry is very important design of the first industrial plant. different planing parameters on for the development work, both to The description is based upon my volume yield, which is described in provide constructive criticism of own view of how the development chapter IV. In a similar way, an ongoing work and to share the has progressed and on the idea is first established, either competence and experience, which experience gathered at KTH-Trä. through existing knowledge or can be found in industry and Chapter 3. describes the star- through testing, to later be adapted which can never be created at a sawing method, i.e. the sawing of and applied in different production university. It should, however, be logs according to the star-sawing systems. The influence of juvenile recognized that the fact that these pattern and the further refinement wood on crack formation and dis- is little understanding of the ideas of the sawn product into Prim- torsion of timber is an illustrative presented need not mean that the Wood. The chapter deals on the example. It has been known for a ideas are wrong or inapplicable. one hand with technical solutions long time (see e.g. Koehler 1938, It is obvious that the stages and proposals for production Boutelje 1968) that moisture described above cannot be treated layouts, and on the other hand movements along the fiber thoroughly within a single doctoral with important production para- direction are greater in juvenile thesis. Analysis of the relation meters, e.g. volume yields, than in mature wood. Intuitively, it between the different stages in dimensions of timber etc. The can be realized that this can Figure 2 is itself a large task. It is basis for this description is influence the deformation of sawn therefore natural that certain parts primarily experience gathered in timber. Removing the juvenile of the thesis have been treated less test sawing and test manufacture at wood in the star-sawing pattern thoroughly than others. To the pilot plants for star-sawing and has eliminated this problem. describe in detail the industrial PrimWood manufacture. Parts of The generation of ideas in itself implementation of the star-sawing what is described in this chapter has no intrinsic value. There must technique would mean that all the are reported in scientific be a real goal, a target. Within work within Value Activation pro- publications. Value Activation, the aim is to gram would described in the thesis, Experimental results concerning introduce products and production but not even this would be enough different material properties of systems for wood, where the value to give a complete picture. timber in general and of timber from a log is considerably higher with vertical annual rings in than that which is obtained in particular are presented in chapter today’s industrial system. In addi- 4. The investigations refer tion, the dissemination of primarily to distorsion and crack knowledge that can be used by formation of sawn timber in diffe- both the industry and research is rent stages of usage: before drying,

15 after drying, as a result of moisture References Wiklund, M. 1992: Lönsam trä- variations, and during outdoor Boutelje, J. B. 1968: Juvenile tradition. Bedömningar ur ett exposure. In addition, the results wood, with particular reference to skogspolitiskt perspektiv av pro- of a few initial tests are given northern Spruce. Svensk Pappers- dukter och teknik i trämekanisk where deformation and crack for- tidning 71: 581-585 industri. Sammanställning för 1990 mation of the wood at the Koehler, A. 1938: Rapid growth års skogspolitiska kommitté. (A microscopic level are described on hazards usefulness of southern profitable wood tradition: the basis of a micromechanical pine. J. of Forestry 36:153-159 Evaluation of the political prospect approach. The chapter is a Sandberg, D. 1998a: Weathering of forest products and techniques summary of the most important of radial and tangential wood in wood working industries. results that I together with co- surfaces of pine and spruce. Prepared for 1990’s commision on authors have published or which 1. Crack formation and colour forest policy.), KTH, Wood are in press in scientific journals. changes in exposed surfaces. In Technology and Processing, Rep- Chapter 5 presents some of the press Holzforschung ort TRITA-TRT-1992-51, (in Swe- conclusions which can be drawn Sandberg, D. 1998b: Weathering dish) from the thesis. of radial and tangential wood Wiklund, M. 1993: Value surfaces of pine and spruce. Activation - a new sawing patterns 2. Microscope studies. In press give improved properties on Holzforschung wooden products. Paper at 11th Wiklund, M. 1991: Värde- International Wood Machining aktivering - ny forskningsinriktning Seminar, Honne, norway, 1993 vid Institutionen för träteknologi. (Value Activation - a new research area at the Department of Wood Technology and Processing.), KTH, Wood Technology and Processing, Report TRITA-TRT- 1991-45, (in Swedish)

16 Chapter Star-sawing - from idea to industrial 2 reality

2. Star-sawing - from idea to industrial reality

Within the R&D-program known would also be possible to good volume yield. The juvenile as Value Activation, an idea arose implement results in industrial wood would be removed because at an early stage of producing applications in the relatively short of its inferior properties. The star- timber with vertical annual rings term. sawing pattern took shape. with the help of a new sawing The activity received support KTH-Trä has subsequently pattern, star-sawing. This chapter from the sawmill industry and from presented ideas for products, prim- describes how the work on star- Nutek during a five-year period. arily based on triangular timber, sawing has been carried out from The support was channelled via which have been judged to have a the idea stage until the first Trätek. The activity at KTH-Trä high value for the customers. industrial company decided to increased dramatically when the build the first plant. Value Activation program was 2.2 Sawing of the log inaugurated. Several projects were according to the star- 2.1 Martin Wiklund, started e.g. to determine those sawing method professor at KTH wood properties that it is most To develop and test the star- Martin Wiklund, who previously essential to study regarding wood sawing technique and the different was Managing Director of the Swe- utilization. Several analyses on the products manufactured from star- dish Institute for Wood dependence of physical properties sawn timber, the project Production Technology Research, Trätek, of wood on e.g. the orientation of and testing of radially sawn timber became Professor and head of the fibres and annual rings were also (Sandberg 1996a) was started. The Department of Wood Technology started. At the same time, studies aim was to develop a small-scale and Processing at KTH (KTH-Trä) were initiated to solve technical saw for star-sawing of large logs in the middle of 1990. One of his problems of sawing timber with and to saw material for the first task was to create and launch vertical annual rings, a process manufacture of wood products program of the department. which appeared at an early stage to according to the Value Activation In the program formulation, the be highly desirable. concept. The products produced ambition was to give the depart- The work soon showed that have been used in different tests ment a profile with a coherent properties of wood products that and as references in marketing the R&D-program. The field of work are most highly appreciated are: ideas. should be so central for the whole shape stability, abcence of cracks, subject of wood technology that it strength, hardness and Organization and finance would be possible in a natural and appearance. It was shown that In 1992, it was realized at KTH-Trä constructive way to integrate basic these requirements are best that a plant was required to test education, graduate studies and accomplished by radially sawn certain production technology research constructively into the timber. ideas and to obtain material for the program. Another criterion was KTH-Trä’s idea was to achieve research work. After extensive that the R&D-program should this by sawing timber with contacts with different potential include interesting long-term re- rectangular and triangular cross- financiers, contact was established search, at the same time, as it sections in a pattern which gives a with the Arjeplog Group, which

19 undertook the responsibility of co- star-sawn timber and speeded up contract manufactured by a ordinating the total financing of the the exploitation of ideas produced mechanical workshop. project Production and testing of within the Value Activation pro- When the project started, the radially sawn timber. This was done gram. Both the sawing technique chosen sawing method was shown under the premise that the actual and the properties of timber and to be very suitable with respect to sawing was to be carried out in raw materials from different parts the early development phase of the Arjeplog. The different financiers, of Sweden were studied. A star-sawing technique (Sandberg all with a connection to the number of product prototypes 1996b). However, it was shown County of Norrbotten, who manufactured from star-sawn that the method was not suitable participated in the financing were timber with vertical annual rings for the rational and economically the Arjeplog Group, the Three- were produced. Property tests and profitable production of star-sawn County Delegation, the Norrbot- market-oriented investigations were timber (Alfheim and Andrée 1993, ten County Council, the Norrbot- carried out for these products. Sjöberg 1994, Sandberg 1996b). ten County Development Fund Results showed that the star- The drying of star-sawn timber and the Norrbotten Research sawing method in most cases gives was carried out together with A- Council. an increased value for timber and Trä’s normal production. A new The pilot plant was opened on products manufactured from this board stacking method was 24th August 1993 in connection timber. developped for the triangular profi- with a KTH-Trä seminar. les (Holmberg and Sandberg 1995a, Organizationally and admini- Results and experiences Sandberg 1996b). The drying of stratively, the project was managed The work within the project was the timber went very well without as follows. A company, Star- divided into three main areas; modification of the drying schedule sawing in Arjeplog AB, was for- sawing technique, quality and and further improvements of the med with Timber-Line & Co AB properties of raw material and drying technique was not conside- and A-Trä AB as main owners, sawn timber, and products from red to be of primary interest at this but the company was run solely to star-sawn timber. In the report stage. support KTH-Trä’s R&D-pro- Development and construction for Star-sawing is best suited for gram. Star-sawing in Arjeplog AB star-sawing (Brask 1994) a detailed large and excessively large logs purchased the star-sawing from description is given of the building (Sandberg 1996b). Test-sawing was Timber-Line & Co AB while A- of the pilot plant in Arjeplog and carried out with pine and spruce Trä AB carried out the drying and the performance of test-sawing logs obtained from different certain truck transport. until the first half of 1994. The regions of Sweden. Results from A total of more than 1000 m3 development which was thereafter this work showed that it was not timber was sawn in different carried out in Arjeplog consisted possible to use the same grading batches. The last test-sawing in primarily of improvements and rules for star sawn timber as those Arjeplog was performed during extension of peripheral equipment in current use. To produce high- spring 1996 and the project was for more rational handling of the quality products with vertical closed down shortly thereafter. timber after sawing. annual rings for the carpentry The project Production and testing The pilot plant was constructed industry, certain grading criteria of radially sawn timber contributed as a project within the under must be increased, e.g. allowable strongly to a rapid testing of the graduate program at KTH-Trä. compression wood, while other sawing technology and products of Thereafter, the equipment was grading criteria can be reduced,

20 e.g. allowable black knots. Sawing how this influences the yield from the products were completely new logs from different regions in star-sawing. We were well aware while others have been on the Sweden showed that star-sawing that variation in log quality within market for a long time, but have can be applied with advantage to the same region can be very large been given a higher value by being large logs from the whole country. and that too far-reaching manufactured with vertical annual However, it was shown that logs conclusions should not therefore rings. Results from the different from the mountainous regions be drawn from the test-sawings product tests are described here were less suitable for star-sawing which have been carried out. The briefly. because of a high frequency of knot appearance in the surface of The greater hardness of the internal defects such as cracks, star-sawn timber is different from radial wood surface and the compression wood and water-stain that of traditionally sawn timber increased shape stability obtained and to a certain extent microbial which means that the grading rules with vertical annual rings are used decay. used today for timber cannot be in solid wood floors. Several diffe- The clearest difference between applied (Holmberg and Sandberg rent types of floors have been timber from different parts of the 1995b). A method was adapted for manufactured and tested. The country was found in the volume grading of radially sawn timber and advantages of having vertical yield of knot-free boards and in has been introduced and applied in annual rings in these floors are the texture of the timber. The collaboration with customers. The unambiguous and the results are difference in texture consists different knot shapes made the directly applicable to industry primarily of different knot further treatment of the timber in (Heickerö and Wålinder 1992, appearance and growth rate certain cases impossible without Wålinder 1996). (annual ring width), while in the removal of the knots. Comparative tests with knot-free wood differences can be Investigations carried out on conventionally and star-sawn panel seen in colour nuance, waviness of star-sawn timber showed that rem- boards for outdoor use painted the annual ring and resin content, oval of the pith and the juvenile with the most common paint types and differences can be felt in wood was more important for the show that radially sawn panel weight, hardness and brittleness. final timber quality than had boards have clear advantages. The The differences in texture which previously been assumed. Clear paint adheres better and cupping can be found in wood from diffe- relationships have been found and crack formation are practically rent regions has, in my opinion, between the occurrence of juvenile eliminated (Sandberg 1994, Borgh received too subordinate a place in wood in sawn timber and the 1995). favour of strength properties and susceptibility to deformation and Within the building sector, light the visual sorting where knots and crack occurrence after drying beams of various designs have other so-called structural and (Sandberg 1992, 1995a, 1996d, been used for many years. A mate- manufacturing errors have been 1997a). rial-saving I-beam has been given decisive importance. It In order to investigate how well manufactured and tested with a should be pointed out, however, suited star-sawn timber was for use web of wood-fibre board and that test-sawing with logs from in different products, tests were flanges of star-sawn triangular different parts of Sweden was carried out on a selection of profiles (Stehr 1995, Sandberg and carried out to obtain information products where timber with Stehr 1997). on how different wood properties vertical annual rings is assumed to Windows with vertical annual vary between different regions and be especially favourable. Some of rings should have lower main-

21 tenance costs and a longer lifespan hardness tests and furniture tests. A 2.3 Dissemination of than ordinary wooden windows. new method has been presented to results from Value Tests with star-sawn timber show reduce the moisture movements in Activation that moisture movements in the the plane of the endwood surface Already in connection with the corner connections become mini- without chemical treatment first presentation of the R&D- mal. The durable heartwood of (Alfheim and Forsberg 1992, program Value Activation for the pine in radially sawn timber is one- Evansson and Brask 1992, Marelius representatives of the sawmilling dimensionally oriented and can et al 1993, Holmberg 1998). and the wood-working industries easily be turned outwards where A proposal for a production line we were completely convinced at the moisture load becomes extra for the manufacture of a glulam KTH-Trä that the realization of the large (Hjorth and Nylander 1992, beam from triangular profiles has new concept would take a long Sandberg 1996c). been produced. Beams have been time. One of the conditions for The possibility of using a triang- manufactured and tested. The any company to be interested in ular profile for mouldings has been beam has controlled moisture applying the ideas was that they evaluated. Technical, economic movements and internal stresses should receive legal protection of and aesthetic aspects have been which arise when moisture incorporeal right. A patent on the investigated. By using a triangular variations are minimized since the basic ideas was thus applied for. profile, vertical annual rings are beam has got vertical annual rings. This has then been followed up obtained which give more shape- The side surfaces are radial, which with patent applications on new stable mouldings. Test planing e.g. means slow moisture absorp- ideas within the basic concept. showed that a triangular profile tion and small susceptibility to Beginning in March 1991, KTH- can be used as basic material for cracks (Bryne et al 1995, Sandberg Trä has continuously informed the high quality mouldings to receive a 1998a). wood-working industries and its high yield (Envall and Stark- Triangular profiles are used to representatives about activities and hammar 1994). manufacture a solid wood panel results from the research program. A fence system for outdoor use with vertical annual rings. Because Besides distribution of reports has been produced. The geometry of the unfavourable knot shape in published, KTH-Trä has and shape-stability of triangular the timber, the knots were remo- participated with its own exhibition profiles is used to give a fence with ved. By finger-jointing the knot- cases at different national wood a long lifespan without using free pieces and simultaneously exhibitions, has arranged seminars chemical preservative agents taking into consideration annual and thematic days and has (Alexis et al 1994). ring orientation, heartwood pro- distributed information through The cross-section is both hard portion and annual ring width, it lectures and newspaper articles. and durable, which is of interest in has been possible to manufacture During the years 1991 to 1997, products such as e.g. floors, an aesthetically attractive solid KTH-Trä’s research was mentio- thresholds and staircases. Several wood panel with high shape stabil- ned in more than 100 articles in endwood surfaces have been ity (Carlsson 1996, Sandberg Swedish daily newspapers and manufactured from a triangular 1997b, Sandberg and Holmberg trade journals. In 1992, KTH-Trä profile and have been surface 1998a). won the region Innovation Cup treated in different ways. The final contest and in 1996, we were boards have thereafter undergone awarded the Elmia Trä prize. The

22 information has received a great 2.4 The development PrimWood AB was formed as a success within the wood-working company PrimWood AB company for the development of industries. The interest in Value In the course of trying to find production systems for ideas Activation with vertical annual companies which were willing to produced within Value Activation. rings has increased. At KTH-Trä, invest in star-sawing, it became In the first place, star-sawing and it was considered important to be increasingly clear that the change in PrimWood technology were to be able to show areas of utilization both production technology and developed. The development of and the final products from the product which the star-sawing production systems should mainly star-sawn timber. A large number concept offers was too revolution- be carried out by production of product ideas were presented in ary for any company to dare to companies to which PrimWood which the star-sawn timber was invest in. It was also difficult to find should give non-exclusive licences. used optimally. Many of the pro- companies that could cover the PrimWood can also have a certain ducts were also considered to have whole business context from log to part-ownership in the production great economic potential. final product that could invest in a companies. In the ongoing work within the new system without influencing It was judged to be essential that R&D-program, several products their own on going activity persons in the promotion group have been produced as prototypes, negatively. The companies which had knowledge and experience as seen above. Care has been ta- have the whole chain from raw which could be used in the ken however, not to disperse the material to final product however continued development of Prim- activities to much. This means that often lack the contractor spirit Wood AB. They would also have the greatest effort has been placed required to succeed with projects the time and interest in participa- on a few projects. Other ideas of this type. ting actively in PrimWood’s must wait until the capacity to At KTH-Trä, it was realized that continued development. Because further develop them becomes it was necessary to do more to of PrimWood’s limited economic available. The continued work was convince the industry that star- resources, it would also be directed towards the actual sawing sawing could become more profi- necessary for the part-owners to technology and gluing of the table. In 1993, the idea was work on voluntary basis to a large triangles into boards, so-called discussed of forming a development extent, which has also been the PrimWood. company to run the continued case. The persons who are today In 1993, a seemingly well- development and to promote the included in the promotion group founded concept was presented for industrial application of the star- are: star-sawing and PrimWood sawing concept. At KTH-Trä, Olle Lundqvist, who was a manufacturing of products with however, industrial and marketing retired technical director from vertical annual rings. Many experience was lacking which was STORA and as a member of the industrial representatives studied considered very important to combine management, is the presented concept with great succeed with industrialization. In interest and several industrial February 1994, PrimWood AB was responsible among other things for representatives also thought that formally founded and the part- the exploitation of new technology star-sawing had a future potential. owners were all employees at KTH- within STORA. In spite of this, everybody waited Trä as well as a group of Per Nydahl, part-owner of for someone else to take the first industrialists, the so-called promo- Neuman & Nydahl which is a step and invest in a plant. tion group. consultant company for company

23 management questions and mana- PrimWood’s High-quality knot-free furniture ger recruitment questions. work procedure and furnishing solid wood panels of Urban Sundberg, previously e.g. pine, PrimWood, were the products senior vice president in the forestry Within PrimWood AB, two which were considered to be the concern MoDo, board chairman in important fields of work were most suitable usage area for the Södra and in Trätek. identified. The manufacturing triangular profiles at an initial technology needed to be developed Kjell Wiklund, previously self- stage. KTH-Trä had shown that it further and the market needed to employed with sawmill, planing-mill was possible to manufacture such a be established for the products and timber dealer’s experience. He panel and that the properties of manufactured from star-sawn has also earlier participated as the panel were superior to a timber. To find areas of usage for advisor at the pilot plant for star- traditional solid wood panel. the triangular profiles was conside- sawing in Arjeplog, and is thereby However, experience was lacking red to be a key question. Besides very familiar with the problems of this, it was essential to maintain for the industrial production of star-sawing. good legal protection incorporeal PrimWood. Several production Rudolf Wiklund, previously MD right for the whole star-sawing ideas in the manufacture were for Geotronics AB and later princi- concept. untested and were considered to pal owner of C. E. Johansson AB Within PrimWood, it was be technically difficult to realise. and Dataliner AB in Eskilstuna. He understood early how the sawing During winter 1994/95, the idea has great experience of exploiting according to the star-sawing of building a pilot plant for Prim- new ideas for profitable products. pattern should be carried out. Wood manufacture to develop the Åke Lundqvist, previously MD The technology was not conside- technology and to produce Prim- in Ericson Radio AB where he red to be complicated, especially Wood for market tests was developed the mobile telephones not for the first plant where the discussed. The plant was erected within the concern. In 1994, he capacity probably should be on Åke Lundqvist’s farm Heda received KTH’s great prize for this relatively low. Layouts and outside Södertälje. On 30 August work. technical detail solutions were 1995, the first PrimWood board Maths O. Sundqvist, who e.g. is worked out through PrimWood’s was manufactured. However, a owner of Jämtlamell Industrier AB. agency. This basis was considered further year of intensive Maths became part-owner of Prim- to be very important for continued development and improvement Wood when Jämtlamell showed a contacts with people interested in work was needed before the great interest in becoming the first star-sawing, and with machine equipment was in a state that Prim- company to invest in star-sawing. suppliers. In an initial stage, many Wood could be manufactured with PrimWood has a collaboration machine suppliers were very the high quality requirements contract with KTH during the sceptical of the whole star-sawing which had been adopted. The formation of the company. We concept, both from a technical and plant was simple in its design but were and are very anxious that market perspective. It was finally showed how each single KTH in the future would become shown that the ideas for produc- production stage could be carried part-owner in the company, which tion layouts presented by Prim- out. was not possible at the time of Wood at an early stage could also The pilot plant in Heda has PrimWood’s initiation. be applied in the first production been of decisive importance for the plants for star-sawing and Prim- commercialization of the star- Wood manufacture. sawing technology. In Heda, one

24 has been able to show in practice a county governor Sven Heurngren ers and other companies. The work final product in the form of Prim- and a number of industry consisted for example in thoroughly Wood and has proven its technical representatives. The interest in evaluating, in co-operation with and aesthetic properties. Although star-sawing and the manufacture of machine suppliers, the proposals for the manufacture in Heda has been PrimWood was great but on, production layouts produced by more or less manual, it has been Jämtlamell’s part, it was perceived PrimWood AB, and to introduce possible to show that an industrial that many questions around raw production technology for the production in full scale is possible. material, production technology, treatment stages where standard The reservations which many per- economy and market remained. It machines could not be used. Bes- sons had towards the usage of was essential to look closer to the ides this, three test sawings were triangles were thereby removed. purely industrial and economic carried out to evaluate the sawing This, in combination with the part before it was decided to go technique and the yield and quality production of materials for market further. of the sawn products. The sawn tests, has probably been more Discussions with PrimWood AB material was used for PrimWood important for the industrialization were carried further to answer the manufacture in Heda and for process than the technical advan- questions which had arisen. A visit market tests of both timber and ces made in Heda. The develop- was carried out to the pilot plant PrimWood. KTH-Trä and Prim- ment work carried out within Prim- for star-sawing in Arjeplog. The Wood AB were strongly involved Wood AB has given rise to several interest increased and the next step in the project, as specialists and to problems which have been solved was decided. The intention was to carry out test-sawings and in minor R&D-projects. To study investigate conditions for running a manufacture of PrimWood in Heda. these problems, PrimWood AB has star-sawing and PrimWood line In May 1998, most of the financed a number of M. Sc within Backe Trä AB. Backe Trä questions had been resolved, the projects at KTH-Trä. These with sawmills in Backe and Tåg- financing was arranged and it was projects have been much sjöberg was at this time owned by decided to begin the industrial appreciated by the students since Jämtlamell, SCA Timber, Ramsele phase of the star-sawing techno- they have concerned future areas Skogsägareförening and the logy. In Tågsjöberg, 100 km north within the wood-working industry. Westerlund family. To push the of Sollefteå, a sawmill is built for star-sawing idea further, the star-sawing and in the adjacent 2.5 The first industrial interested parties created a deve- town Junsele, a plant is built for plant lopment project within Träinnova. the manufacture of PrimWood. As mentioned earlier, KTH-Trä Träinnova is an extensive and long- The plants will be operated within and in a later stage also PrimWood term regional development project the company NovaWood AB AB took contact with different which in cooperation with about 20 which is owned equally by Jämt- wood-refining companies to inte- wood-treating companies is aimed lamell, SCA and Ramsele rest them in beginning to apply the at increasing further refinement in Skogsägareförening. star-sawing method. A company Jämtland and Västernorrland. After all this long-term work, with which early contact was made The aim of the project was to the situation is currently that one was Jämtlamell Industrier AB. The produce industrially adapted company, NovaWood AB will first contact was taken in the equipment for star-sawing and build a full-scale industrial plant autumn of 1994 in connection with PrimWood manufacture in based on KTH-Trä’s concept. a visit to KTH-Trä by Jämtland’s collaboration with machine suppli-

25 Internal report No. 14/94, (in Hjorth, T.; Nylander A. 1992: Swedish) Fönster gjorda av trä med stående References Bryne, L.-E.; Gustafsson, S.; årsringar. (Timber with vertical Hach, V.; Jansson, J.; Neubauer, P.; annual rings in windows.), KTH, Alexis, M.; Andersson, K.; Berg- Wolff, M. 1995: Trebalk - limträ- Wood Technology and Processing, man, L.; Marsch, C.; Thorslund, E. balk av trekantprofil. (Trebalk - Internal report No. 4/92, (in Swe- 1994: Staket med stående års- Glue lam beam from star-sawn dish) ringar. (Timber fence with vertical triangular profiles.), KTH, Wood Holmberg, H. 1998: Influence of annual rings.), KTH, Wood Technology and Processing, grain angle on Brinell hardness of Technology and Processing, Internal report No. 37/95, (in Swe- Scots pine (Pinus silvestris L.) Holz Internal report No. 17/94, (in Swe- dish) als Roh- und Werkstoff 56(5) dish) Carlsson, S. 1996: Produktions- Holmberg, H.; Sandberg, D. Alfheim, M.; Forsberg, K. 1992: system för tillverkning av limfog av 1995a: Utvärdering av en ny VIMSCARD för ändträgolv. stjärnsågad trekantprofil. ( Manu- ströläggningsmetod för stjärnsågad (VIMSCARD - end-grain floo- facturing system for production of trekantsprofil. (Evaluation of a new ring.), KTH, Wood Technology solid wood panels from star-sawn stacking method for star-sawn and Processing, Internal report triangular profiles.), KTH, Wood triangular profiles.), KTH, Wood No. 5/92, (in Swedish) Technology and Processing, Rep- Technology and Processing, Alfheim, M.; Andrée, P. 1993: ort Examensarbete No. 29/96, (in Internal report No. 31/95, (in Swe- Ekonomisk utvärdering av stjärn- Swedish) dish) sågning samt limfog-, golv- och Envall, T.; Starkhammar, J. 1994: Holmberg, H.; Sandberg, D. lättbalktillverkning. (Economical Panel och lister ur trekantprofil. 1995b: Kvistars frekvens, form och evaluation of star-sawing, solid (In-door panels and mouldings beskaffenhet i trekantprofil av furu wood panel, flooring and light from star-sawn triangular profiles.), (Pinus silvestris L) och gran (Picea weight beam.), KTH, Wood KTH, Wood Technology and abies Karst). (Frequency and Technology and Processing, Rep- Processing, Internal report No. 15/ character of knots in triangular ort Examensarbete No. 13/93, (in 94, (in Swedish) profiles of pine (Pinus silvestris L) Swedish) Evansson, J.; Brask, P. 1992: and spruce (Picea abies Karst).), Borgh, A. 1995: Årsrings- Ändträ som slityta. (End-grain KTH, Wood Technology and orienteringens inverkan på resistance for wearing.), KTH, Processing, Report TRITA-TRÄ målarfärgers varaktighet. (The Wood Technology and Processing, R-95-12, (in Swedish) influence of annual ring orientation Internal report No. 2/92, (in Swe- Marelius, M.; Winther, E.; on paints performance.), KTH, dish) Winther, J. 1993: Tregolv. (Tregolv Wood Technology and Processing, Heickerö, R.; Wålinder, M. 1992: - flooring with star-sawn triangular Report TRITA-TRÄ R-95-15, (in Massivt trägolv med stående års- profiles.), KTH, Wood Technology Swedish) ringar. (Solid wood flooring with and Processing, Internal report Brask, P. 1994: Utveckling av vertical annual rings.), KTH, No. 12/93, (in Swedish) Wood Technology and Processing, anläggning för stjärnsågning. Sandberg, D. 1992: Jämförelse Report TRITA-TRT-1992-46, (in (Development and construction for mellan genomsågning, fyrsågning Swedish) star-sawing.), KTH, Wood och stjärnsågning med avseende på Technology and Processing, torkningsdeformationer. (Deforma-

26 tion after drying in timber from Sandberg, D. 1996c: Förbättrade profiles. Holz als Roh- und through-and-through sawing, block träfönster. Årsringsorientering, Werkstoff 55(5):292 sawing and star-sawing - a sprickor och kärnved - tre viktiga Sandberg, D.; Holmberg, H. comparsion study.), KTH, Wood faktorer för fönstervirkets bestän- 1998a: Radially sawn timber. Technology and Processing, Report dighet. (Improved wooden Gluing of star-sawn triangular Examensarbete No. 8/92, (in Swe- windows. Annual ring orientation, profiles into form-stable products dish) cracks and heart wood - three with vertical annual rings. Holz als Sandberg, D. 1994: Nedbryt- important factors influencing the Roh- und Werkstoff 56(3):171-177 ningsmekanismer hos trä. - 1. Fo- durability of timber used in Sjöberg, L. 1994: Maskiner och tokemisk nedbrytning. (degrada- windows.), KTH, Wood materialflöden för stjärnsågning - tion mechanisms of wood. 1. Technology and Processing, Rep- utvärdering av olika alternativ. Photochemical degradation. A ort TRITA-TRÄ R-96-23, (in Swe- (Machines and material flow in a literature survey.), KTH, Wood dish) sawmill for star-sawing - evaluation Technology and Processing, Rep- Sandberg, D. 1996d: The influ- of different alternatives.), KTH, ort TRITA-TRÄ R-94-10, (in Swe- ence of pith and juvenile wood on Wood Technology and Processing, dish) proportion of cracks in sawn Report Examensarbete No. 18/94, Sandberg, D. 1995a: Årsrings- timber when kiln dried and (in Swedish) orienteringens betydelse för sprick- exposed to wetting cycles. Holz als Stehr, M. 1995: Lättbalk med bildning och deformation hos Roh- und Werkstoff 54(3):152 triangulära flänsar. (Triangular vattenbegjutet virke av furu (Pinus Sandberg, D. 1997a: The influ- profiles from star-sawing as flanges silvestris L) och gran (Picea abies ence of annual ring orientation on for light-weight beams.), KTH, Karst). (Influence of annual ring crack formation and deformation Wood Technology and Processing, orientation on crack formation and in water soaked pine (Pinus Report TRITA-TRÄ R-95-14, (in deformation in water soaked pine silvestris L) and spruce (Picea abies Swedish) (Pinus silvestris L) and spruce Karst) timber. Holz als Roh- und Wålinder, M. 1996: On (Picea abies Karst) timber.), KTH, Werkstoff 55(3):175-182 Characterization of Surfaces Wood Technology and Processing, Sandberg, D. 1997b: The influ- Parallel to Grain in Scotch Pine Report TRITA-TRÄ R-95-16, (in ence of annual ring orientation on (Pinus silvestris L.), KTH, Wood Swedish) strength and dimensional changes Technology and Processing, Rep- Sandberg, D. 1996a: Produktion during moisture variation in finger ort TRITA-TRÄ R-96-20 och provning av radialsågat trä. joints. Holz als Roh- und (Production and testing of radially Werkstoff 55(1):50 sawn timber.), KTH, Wood Sandberg, D. 1998a: Weathering Technology and Processing, of radial and tangential wood Internal report No. 42/96, (in Swe- surfaces of pine and spruce. 1. dish) Crack formation and colour Sandberg, D. 1996b: Radially changes in exposed surfaces. In sawn timber. Star-sawing - a new press Holzforschung method for producing timber with Sandberg, D.; Stehr, M. 1997: vertical annual rings. Holz als Roh- Bending strength of I-beams with und Werkstoff 54(3):145-151 webs of wood-fibre board and flanges of star-sawn triangular

Chapter The star-sawing concept – sawmilling 3 and value added products of triangles

3. The star-sawing concept – sawmilling and value added products of triangles

This chapter presents a simplified geometrical study of the star- sawing pattern to illustrate the parameters which influence the final dimensions of sawn timber. The sawing methods which have been evaluated in the trials and a b c the volume and quality yields obtained are also described. A few Figure 3. Different patterns of star-sawing. Sawing patterns with (a) the same comments are then given number of pieces with a rectangular cross-section as with a triangular cross-section, (b) regarding the log quality twice the number of rectangular pieces, (c) twice the number of rectangular pieces which include the pith, centre pieces. requirements associated with the star-sawing. Finally, a description is given boards with a rectangular cross- Figure 4. Ideal case of the star-sawing of the principle of gluing the section are sawn, (Figure 3). pattern for estimation of the dimension b yield according to formulae 1-3. s triangles obtained by star-sawing If the sawing pattern in Figure 4 into solid wood panels or beams. is regarded as an ideal geometric case, i.e. an inscribed circle with diameter D, the following can be 3.1 Sawing according to established regarding the sawn a the star-sawing method timber dimensions: a) The maximum timber width, The geometry of the b, which can be sawn is: sawing pattern for the centre pieces: s The star-sawing pattern is intended b for the industrial production of (1) c D timber with vertical annual rings free from pith and from most of for the side pieces: the juvenile wood. The sawkerfs are more or less radially oriented (2) from the surface of the log inwards towards the pith area and they are where m is the width of the pith- cutter, t is the thickness of the placed so that six timber pieces c centre pieces, t is the thickness of with equilateral triangular cross- s sections are obtained. Between the the side pieces (raw measure) and s is the width of the sawkerf. In this bc t s triangular timber, one or several m 31 case, t and t are equal and are c s Table 1. Examples of dimensions in star-sawing calculated according to formulae 1-3. hereafter called t. The notations The dimensions refer to dried and sharp-edged timber. In the calculations, the width are illustrated in Figure 4. of the pith-cutter was 50 mm and the sawkerf thickness 3 mm. D = the top diameter of the log, t = the thickness of the timber, b and b = the widths of the centre and b) The relationship between the c s side pieces respectively, a = the length of the sides of the triangles. In the calculations, triangular timber dimension, a, and the degree of shrinkage has been set to 3% and 6% in the radial and tangential the thickness of the rectangular directions, respectively. timber, t, for a given log diameter, D, and sawkerf thickness, s, is given by: Dt bc bs a (cm) (mm) (mm) (mm) (mm) (3) From these expressions, the 20 50 63 44 45 following conclusions can be 24 50 90 63 65 drawn: 28 50 100 75 85 (i) Besides the log diameter, the width of the rectangular centre 32 50 125 100 105 pieces is determined by the amount 36 50 150 125 125 of juvenile wood which is removed, 40 50 163 150 145 and by the thickness of the pieces. 44 50 180 160 165 (ii) The width of the side pieces 48 50 200 175 185 for a given log diameter is determined by the thickness of the middle pieces, the thickness of the 20 2 75 63 90 side pieces and the width of the 24 22 90 75 90 sawkerf. 28 45 110 90 90 (iii) At a given log diameter, the 32 63 125 95 90 width of the rectangular timber is 36 85 150 95 90 much smaller than in conventional sawing methods, e.g. the main 40 105 163 95 90 yield in block-sawing. 44 125 180 95 90 (iv) There is a clear and simple 48 145 200 95 90 relationship, at a given log diameter, D, between the dimensions of 20 32 63 63 60 the triangles and the thickness of the rectangular timber. 24 44 90 75 70 To further demonstrate the 28 44 100 90 90 dimensions achieved in star-sawing, 32 50 125 100 105 Table 1, gives some examples of 36 63 150 115 110 dimensions for the sawn timber, 40 75 163 125 120 calculated with the help of 44 100 180 115 120 formulae 1-3, for some different log diameters. The uppermost part 48 100 200 125 130

32 of the table shows what happens if part of the cross-section and thus panels. Too large triangles become the thickness of the rectangular weaken the timber. difficult to handle and to process. timber is held constant at 50 mm When the thickness of the while the diameter increases. In a rectangular timber increases, the corresponding way, the middle angle between the flat side of the part of the table shows what timber and the annual ring near happens when the dimension of the surface decreases, i.e. the the triangular timber is fixed at 90 degree of vertical annual rings mm. The lower most part of the decreases. This is described in table shows a few patterns which more detail in chapter 4.1. give realistic dimensions for all If consideration is given to the timber. question of which dimensions are In the calculations, the width of suitable from a quality handling the pith-cutter was assumed to be and manufacturing viewpoint, the 50 mm, which has been found in sawing patterns may be as shown the test-sawings to be a suitable in the lower part of Table 1. dimension for removing the pith However, it is doubtful whether Figure 5. Splay knots in star-sawn pine. and most of the juvenile wood in logs with a top diameter less than logs from northern and central If the centre and side pieces are 25 cm should be star-sawn Sweden. This also means that the sawn to the same thickness and the according to the basic pattern, thickness of the middle piece dimension of the pith-cutter is 50 x (Figure 3a). should not be less than 50 mm to 50 mm, the centre pieces are In the test sawings, dimensions avoid the risk of getting pith in the always wider than the side pieces. which deviate from those side pieces. With this dimension of This “extra” edging of the side calculated were observed. The the pith-cutter, i.e. greater than 50 pieces is always carried out on the deviations depend primarily on the pith side of the timber, which gives x 50 mm, most of the small knots cross-section of the logs which had an extra margin to the pith. in the juvenile wood are also remo- a more or less oval shape and were If the thickness of the rectang- ved, which is an advantage in crooked in the longitudinal ular timber is kept constant, the producing knot-free components. direction. Besides this, damage and dimensions of the triangles increase On the basis of Table 1, it can growth stresses influence edging. with increasing log diameter. In a be established that log diameters similar way, the thickness of the Large amounts of reaction wood in around 20 cm give small rectangular timber varies with log the log often made sawing more dimensions for all the timber. In diameter if the dimension of the difficult and gave a poorer yield. test sawings it has been difficult to triangles is fixed. handle small dimensions because When the triangles are glued Sawing methods the timber easily breaks at large into blocks, it has been found that In order to produce star-sawn knots. A large proportion of the suitable dimensions of the triangles timber with rectangular and knots are in fact cleaved longitudi- are between 60 and 130 mm. Too triangular cross-sections, different nal and so-called splay knots are small triangles give handling sawing methods can be utilized. It obtained in the timber surface, problems and too thin blocks, and is essential that the sawn timber (Figure 5). In timber of small this reduces the flexibility in have vertical annual rings (Sand- dimensions, these occupy a large dividing the blocks into solid wood berg 1995b) along the whole length 33 and that the pith with the surrounding juvenile wood is removed. Two sawing methods have been applied. At an early stage in the development of the star-sawing method, a pilot plant was built in Arjeplog to develop and test star- sawing technology (Sandberg 1996a, Sandberg 1996b). At the pilot plant in Arjeplog, the log was clamped at the ends between two dowels so that it could be rotated. Mantle wood was sawn away and a hexagonal block was obtained. This block was then loosened from the dowels and sawn into rectangular and triangular timber, (Figure 6). The method was shown to be good for its purpose, but was not suitable for commercial production. With the help of experience obtained at the pilot plant, a new method of sawing was developed (Sandberg 1997c). This method is intended for operation on an industrial level and all sawing steps can after certain modification be carried out using conventional machines available on the market. The sawing procedure is shown schematically in Figure 7. The different steps are: I. The log is positioned with consideration taken to its crook and ovality so that the pith can with certainty be enclosed within the centre plank. II. Two parallel sawkerfs are made along the periphery of the Figure 6. Sawing steps according to the Arjeplog model. log so that they just touch the mantle wood at the top end of the

34 log. Two further sawkerfs divide the IX. The triangular profiles rectangular timber. Since it has log into a centre board and two so- which will be used for knot-free been possible to carry out drying called “coffin lids”. products will to a great extent be without any problems, improv- III. The centre board is cleaved cut free from defects and will be ements in the drying technique into two, as close to the pith as finger-jointed. For these triangular have not been considered to be of possible. profiles, a slightly different edging primary interest. However, the IV. The two pieces obtained procedure has been used. The shape of the triangular profiles has from the centre board are triangular profiles are divided at meant that it has been necessary to thereafter taper edged, i.e. the approximately half their length modify stacking. Four different sawkerfs are placed parallel to the close to a knot or other defect ways of stacking the triangles has wane sides of the timber. Through which has to be removed at a later been tested; simple stacking, in this procedure, higher straight- stage. The two parts are thereafter blocks, in groups of three, and grained timber is obtained and edged separately. With this unedged triangles in groups of more of the high-quality sapwood method, the tapering of the log can three, (Figure 8). than in conventional edging is be utilized to a greater extent than Simple stacking means that the achieved. Furthermore, the pith if the triangular profiles are edged triangular profiles are arranged one and the juvenile wood are remo- in one piece. by one with a space of about 2 cm ved. The taper edging thereby Test sawings suggest that the between, no consideration being increases the certainty that the pith method offers a high volume yield given to annual ring orientation. is removed at the butt end of the at the same time as quality-raising The method was shown to have a timber, since the pieces which are advantages are obtained with the number of disadvantages, e.g. edged away on the pith side sawn product. The quality compression damage arises on the become wedge shaped. improvement is a consequence of profile corner which lies against V. The “coffin lids” are edged the fact that the rectangular timber the row above (Brask 1994; Holm- to obtain two control surfaces for becomes more straight-grained and berg and Sandberg 1995a). This the subsequent sawing. The edging that the pith and surrounding is done very sparingly and the juvenile wood can be removed with sawcuts touch the sides of the greater certainty than was possible coffin lid at the top end. with the earlier sawing method. VI. The “coffin lids” are tilted 60° and a board and a triangular Drying profile are sawn. One alternative is It has been very easy to dry the a b to tilt the band saw 30° from the sawn timber from the different test vertical plane. sawings without any drying dam- VII. The remaining, rhomboid age. It has been dried in the same piece is sawn in the same way as in drying kiln as conventionally sawn VI. timber and both triangular and VIII. The boards from the rectangular profiles have been d “coffin lids” are taper-edged in the dried in the same drying batch. c same way as in IV, except that an The drying schedules were the same Figure 8. Arrangement of triangular profiles; (a) simple stacking, (b) in blocks, (c) in extra degree of edging to remove as those used for the drying of groups of three, (d) unedged triangles in groups of three. the pith is normally not necessary.

35 IV sapwood heartwood III top end butt end juvenile wood pith centre board

coffin lids top end

I - II

V VI VII

VIII IX

top end butt end top end butt end pith side pith side

wane wane saw kerf

Figure 7. A schematic presentation of the new method of star-sawing.

36 becomes particularly obvious if the triangles before drying was conside- This yield is shown as an unbroken drying package is placed far down red to be impracticable and the line in Figure 9. This figure also in the drying batch. control surfaces which were sawn shows average values for volume In order to obtain a more even on the “coffin lids” (see Figure 7) yields from test sawings and load-distribution between stickers also meant that drying batches can simulations which have been and triangular profiles, the profiles be placed with unedged triangles. carried out so far. There is good were placed in blocks across the To dry the triangles with the edge agreement between the estimated whole width of the package. The surface intact and without any curve for the total yield and later tangential sides were turned against cracks was not considered to be test sawings. The total yield for each other to reduce moisture any problem with modern kilns. logs with a top diameter above 25 evaporation. It was shown, However, there is a risk that very cm is as an average slightly less however, that the moisture content small cracks arise on the edge sur- than 70 % with respect to the top variation between the profiles was face. By edging the triangles after cylinder volume. large with this method. Problems drying, these cracks could be The yield of timber with rec- also arose on some occasions with removed. The method has been tangular cross-section in the test blue stain and mould growth on tested and the results showed a sawings was slightly lower than that contacting surfaces. low degree of crack formation and shown by the unbroken curve in When the profiles were placed 1.5 % higher volume yield than if Figure 9. This was due mainly to in groups of three with the tangen- edging had been carried out in a the fact that, in the test sawings, tial sides inwards in each group, single stage (Sandberg 1997c). The we have tried to keep the volume the moisture evaporation from the primary disadvantages are that of triangles at a high level to tangential surfaces was reduced at circa 15% larger volume is dried produce as much material as the same time as the air and that the volume which is edged possible for the manufacture of percolation was increased so that away cannot be used as pulp chips. prototype products from the tri- the risk of moulding and blue stain was reduced. The method gave Volume yield few cracks and little compression The volume yield for star-sawing damage, and blue stain was very has been determined through unusual. The method has been simulations and test sawings (Nord- used for most of the triangles fors and Sandberg 1993, Beckman Total which have been dried. and Stehr 1993, Sandberg 1996b, As the sawing method was Holmberg and Sandberg 1996, developed, the idea of edging the Sandberg 1997c, Sandberg 1998c). Rectangular triangles after drying arose. This Results show that the total volume should be done for two reasons. yield, i.e. both rectangular and On the one hand, it was desirable triangular profiles, is high in to try to crosscut the triangles comparison with conventional before edging to increase the sawing methods. volume yield. On the other hand, Sandberg (1996b) on the basis of it was desirable to reduce the risk simulations and test sawings Figure 9. Volume yield for star-sawing from simulations and test sawings between the of crack formation on the tangen- carried out until 1995 proposed a years 1993 and 1998. Unbroken curves are estimated volume yield according to tial surface. To crosscut the general yield curve for star-sawing. Sandberg (1996b). 37 angles. The total volume yield is yield from block-sawing. In the knots and deformations, the influenced to a negligible extent by grading rules for quality assortment regulations are based on Gröna the proportion of triangles sawn in of pine and spruce, e.g. Gröna Boken (1976) but with certain each setting. In the lower unbrok- Boken (1976) and Nordiskt Trä additions. Grade SI corresponds to en curve of Figure 9, it was assum- (1994), the classification of splay class I-III in Gröna Boken, grade ed that the proportion by volume knots is severe, i.e. timber with SII corresponds to grade IV and of triangles in each pattern in the splay knots is often down-graded most of grade V in Gröna Boken. future would be about 40 %. In the to a lower quality. In the early The remaining timber falls into test sawings, on the other hand, sorting of star-sawn timber grade SIII. In grade SI, as opposed the proportion has been nearly according to Gröna Boken, a very to Gröna Boken, one to two large 60 %. It is thus the choice of pat- skew quality distribution was edge or face knots are accepted if tern which controls the proportion obtained compared with what the timber otherwise lies within this by volume of triangles obtained, square-sawn timber from the same grade. Other defects may occur as (see equation 3). Figure 9 also logs would have given. A large indicated in Table 2. The timber is shows that the yield decreases with proportion of the star-sawn timber intended to be sawn sharp-edged, decreasing log diameter. This is a was classified as grade better U/S but this is of course a question of consequence of the fact that the or VI. A small proportion of the usage area for the timber. A pith-cutter removes a larger part of timber was in the intermediate certain allowable wane can often the log volume, the smaller the log classes. This distribution depended give a better timber usage. diameter. This is also true of the mainly on the knot appearance of The grading instructions for other residual volumes which are the timber. star-sawn timber with a rectangular obtained. Star-sawn timber is intended in cross-section have been tested with the first place for carpentry, two sawings of pine from Jämtland Quality yield furniture and furnishing compo- (Holmberg and Sandberg 1996, When the log is split radially, knots nents where the quality of the Sandberg 1998c) and one sawing of will be cleaved lengthwise and whole length of the timber is often pine from Västmanland (Sandberg appear as splay knots or spike less important than how a large 1997c). In all the investigations, the knots in the timber surface. In part of the piece can be used for log was of woods scale with a top star-sawing, as in all other radial different components. As an diameter greater than 26 cm. Table sawing methods, splay knots occur example, a few “ugly” knots on an 3 shows the quality distribution considerably more frequently than otherwise knot-free board of wood from the three investigations. in e.g. square-sawn boards, where are less important for users of For the triangles, no general they in principle occur only in the knot-free timber than if the knots grading instruction has been central section. The radial sawing are small, fresh, and evenly produced because there is as yet methods also frequently use large distributed in the timber. no market for this timber. For the logs and in such logs, parts of the In an attempt to produce a use of triangles in defect-free solid knots are often dry or decayed grading instruction for star-sawn wood panels, initial tests have towards the periphery of the log timber, especially for timber with begun by dividing the triangles into (Holmberg and Sandberg 1995b). rectangular cross-section, the sawn classes to see how long defect-free This part of the knot will always be timber was divided into three boards can be obtained from each found in star-sawn timber, but this grades, SI-SIII. Table 2 shows the triangle. need not be the case in the main criteria for these grades. Regarding 38 Table 2. Quality classification of star-sawn timber with rectangular cross-section.

Defect Quality class S I S II S III Knots and distorsion GB I-III* GB IV-V GB V-VI Cracks must not occur must not occur may occur Pitch pockets single cases may occur ** single cases may occur ** may occur Bark pockets may occur to a lesser extent** may occur to a lesser extent ** may occur Compression wood must not occur must not occur may occur Interlocked grain must not occur may occur to a lesser extent ** may occur Cross grain must not occur may occur to a lesser extent ** may occur Water-wood must not occur must not occur may occur Cankers may occur to a lesser extent** may occur to a lesser extent** may occur Rot must not occur must not occur may occur to a lesser extent Insect damages must not occur must not occur may occur to a lesser extent Blue stain must not occur must not occur may occur to a lesser extent

GB I-III Gröna Boken quality class I - III * 1-2 very large edge or face knots may occur if the piece of wood otherwise falls within this quality class. ** If the piece of wood otherwise falls within this quality class and is without other defects. Table 3. Quality distribution (volume proportion in %) of star-sawn timber with rectangular cross-section from pine timber of woods scale. The quality estimation has been carried out in accordance with the criteria shown in Table 2. Logs for star-sawing As shown above, star-sawing in its Timber felling Top diameter Quality basic design (Figure 3) is best area on log (cm) class suited for relatively large logs, with S I S II S III an approximate top diameter below Jämtland (1) 32 42 37 20 bark greater than 25 cm. To make a very rough estimate of how much (1) Jämtland 28 26 52 21 such logs exists in Sweden, the Jämtland (2 28 26 60 14 statistics over annual fellings can Västmanland (3) 29 38 44 18 be regarded. In Sweden, about 65 million cubic metres forest is felled 1) Holmberg and Sandberg 1996; 2) Sandberg 1998c; 3) Sandberg 1997c 39 annually and about 45% of this rings, (Figure 10). It can be noted that the production would be becomes saw logs, (Statstical Year- that PrimWood is also the name of rational and industrially applicable. book of Forestry 1994). A rough the development company describ- It was therefore decided to cut the estimateion, based on private com- ed in chapter 2.4. fingers parallel with the tangential munication with different sawmill When PrimWood began to be surface of the triangles, (Figure managers show that at least 15 % developed into a product, a num- 11a). This means that in the sur- of the saw logs has a top diameter ber of technical, aesthetic and face of the PrimWood, the finger- which exceeds 25 cm. This means marketing problems arose. One of joint will be visible as a zigzag that about 4 million cubic metres these was how to treat the knots. pattern, (Figure 11b). The primary of logs are felled annually which It was shown that the knots in the reason for this direction of the from a dimensional aspect are timber used for star-sawing were finger-joint was that the block suitable for star-sawing. However, not acceptable in the high-quality glued from triangles can be divided in order to support the above material which was desirable. This into an arbitrary number of boards estimation a more detailed investi- depended, on the one hand, on without the joint in the surfaces of gation need to be done to see the the knot appearance in the timber the boards becoming crooked. diameter distribution of logs in after the block had been divided If the cutting is instead carried different regions in Sweden. and, on the other hand, on the out parallel to one of the radial fact that the knots were often dry, surfaces, there is a risk that an 3.2 PrimWood - boards partly with bark and/or rot inclined joint is obtained in the glued from star-sawn damage (Sandberg and Holmberg surface, (Figure 11c). The inclined triangular profiles 1996). It was therefore decided to joint can be avoided by making an Triangles are a product which manufacture PrimWood completely end joint in the cutting at the base cannot normally be found on the without knots and defects such as and top of the triangle and at the market and consequently, the pitch pockets. The work was height where the blocks shall be areas of use for this product are concentrated primarily towards divided, (Figure 11d). limited. In an early stage of producing PrimWood from pine, A further question remained and development of the star-sawing but tests have also been carried that was which pieces should be method, a number of proposals out with spruce. joined together. In pine, there is a were produced for products from When it was decided to remove great difference between sapwood triangles, (see e.g. chapter 2.2). all knots and defects from the and heartwood, and it is important One of the products which was triangles, finger-jointing was the that the proportion of heartwood is considered to be especially interes- natural solution to the question of practically the same in the joined ting from a market perspective and how it would be possible to manu- pieces in order to obtain a good fit which one wished to develop facture long lengths. with regard to colour, (Figure 12a). further was solid wood panels, so- Finger-jointing can be carried This can probably be solved with called PrimWood. PrimWood is out in different ways resulting in vision system and an advanced triangles glued together into a different appearances in the final handling system for the clear block so that the annual rings in product. The intention with Prim- pieces, but in our case the problem the cross-section of the block form Wood was that it would be was solved as follows: When a knot an S-curve. This block can then be aesthetically attractive, of high or defect was removed, the pieces divided according to thickness into quality with regard to shape stabil- on each side of the defect were boards with almost vertical annual ity and tightness of glue line, and monitored so that these pieces

40 a b

c d

Figure 11. Finger-jointing of the triangles can be carried out in different ways giving a different appearance of the join in the visible surface. Finger cutting parallel to the tangential surface of the triangles (a) gives a zigzag pattern in the surface when the triangles have been glued together (b). If the finger cutting is carried out parallel to the radial surfaces, the risk is great that the join becomes inclined (c). This can be avoided if the cutting tool is designed with an end joint in the surface of the board (d).

could be joined together. This results in an almost perfect pattern a fitting, i.e. the proportion of heartwood in the joined pieces is practically the same and the annual ring pattern on the radial surfaces coincides between the joined pieces, (Figure 12b). The annual ring width and the texture of the timber on both sides of the joint are also similar. This is only valid b however when pieces from the same triangle are joined. If a Figure 12. If triangles of pine with different proportions of heartwood (a) are triangular board is changed, a joined together, a poorer pattern fitting is slightly poorer pattern fitting is of obtained regarding colour than if the course obtained because the texture pieces have practically the same Figure 10. PrimWood, defect-free triangles glued together into a block. heartwood proportion (b).

41 varies between different boards. dependent on the person who which are not to be finger joined This effect is reduced the more carries out the work. are selected more or less at random, homogeneous the material is with In order to obtain good pattern it will be difficult to obtain a good regard to texture. fitting and to carry out the cutting pattern fitting of the remaining In order to obtain an even texture of the fingers without error, it is pieces. This is a problem which in the whole PrimWood panel, very important that fibre results in the manufacture of three consideration must also be given to disturbances around the knots are varieties of PrimWood with regard the texture of the single triangle removed. to pattern fitting. These three when glued together into a block, In star-sawing, a certain volume varieties are all knot- and defect- so-called pattern fitting in the of clear pieces will be obtained free: width of the block, (Figure 13). which are relatively long, i.e. I. PrimWood without finger-joint This pattern fitting is carried out longer than 1 metre (Sandberg and with a certain pattern fitting in visually when the triangles are 1996, 1998a). It is not desirable to the width of the block. placed together in a block and the finger-joint these boards, but II. PrimWood with a pattern- pattern fitting is to a high degree instead to manufacture PrimWood fitted finger-joint and a certain without a finger-joint. If long pieces pattern fitting in the width of the block. III. PrimWood with a finger-joint where no consideration has been given to pattern fitting. A rough calculation made on the basis of empirical investigations (Sandberg and Holmberg 1996, 1998a) suggests that the distribution between the different varieties will be approximately 30 % of I, 50 % of II and 20 % of III. In order to manufacture these three varieties of PrimWood, the triangles should be divided into two groups with respect to the frequency of defects before the removal of defects; one group for the manufacture of PrimWood I and III, and one group for the manufacture of PrimWood II. Other factors which influence the appearance of PrimWood are e.g. the choice of glue, the number of finger-joints per unit surface and the length of the fingers. These Figure 13. Pattern fitting between the triangles which are glued together into a block is important to obtain a uniform texture in the PrimWood panel. parameters have been studied to a 42 certain extent but not investigated of the star-sawing pattern, long Star-sawn timber can be planed A great advantage in using in greater detail in this study. lengths of timber are obtained with lesser losses than conven- triangles in the manufacture of which are knot-free. The propor- tionally sawn timber because panels is the flexibility which is Production steps and tion of knot-free material that can timber with vertical annual rings obtained since the height of the volume yields be obtained from star-sawn pine lacks cupping. The height of the block is determined by the height and spruce timber has been triangle, and thus the height of the of the triangles, which in turn can The manufacture of PrimWood investigated for timber from diffe- glued block, must agree exactly be determined in the planing takes place in six steps: rent parts of Sweden (Sandberg with the board dimensions into without great volume losses. 1. Sorting of triangles into two groups and Holmberg 1996). In the which it is desirable to divide the Flexibility is thereby also created in with respect to the frequency of investigated material around 25 % block, so that there are no the sawmill where rectangular defects. knot-free pine with a length greater excessive volume losses. By deci- timber can be sawn according to 2. Removal of defects and sorting of than 2 metres was obtained from ding already in the planing stage demand and triangles suitable for defect-free pieces which will not be southern and central Sweden. For which thicknesses are to be PrimWood manufacture can be joined. the wood from northern Sweden, extracted from the finished block, produced at the same time. 3. Finger joining the corresponding figure was 5 %. these losses can be avoided. 4. Planing In this case, the fibre disturbance 5. Assembling and gluing into blocks around the knots had also been 6. Division of blocks into boards and removed. possible dimension adjustment and For the same raw material, a sanding. larger volume and longer knot-free The volume yield for the diffe- timber are obtained with star- Branch wohrl rent operations has been sawing than can be obtained from investigated (Sandberg and Holm- the main yield in block-sawing berg 1996, 1998a). The average (Neubauer 1998). This is a value for the volume yield from a consequence of the fact that the Branch wohrl dried triangle of pine to a sanded main yield from block-sawing PrimWood board was 53.8 % contain knots from practically (Sandberg and Holmberg 1998a). every branch whorl in the tree, This is a high volume yield for whereas in star-sawing, the knots Branch wohrl defect-free pine board. It can be in one branch whorl will not be established that three operations found in all the pieces, Figure 14. are decisive for the yield, the rem- Neubauer shows through oval of defects, the planing and simulations from the pine stem division of blocks, and the sanding bank (Grundberg et al 1995) that of the panels. from block-sawn timber, knot-free Branch wohrl In the investigation by Sandberg boards with a length of up to 600 and Holmberg (1998a), the losses mm can be cut and that the knot in removal of defects constituted removal constitutes 13 % of the less than 15 % of the volume of the volume. When the same logs are b. Main yield from block-sawing triangles. The knots were star-sawn, the corresponding a. Star-sawn timber responsible for most of the cutting values become 2000 mm and 9 %. Figure14. The knots in each branch whorl are found with different frequencies in the sawn timber depending on the appearance of the sawing pattern, (a) star-sawing, spillage. Because of the appearance (b) block-sawing. 43 (Pinus silvestris L) och gran (Picea (Pinus silvestris L) and spruce profiles into form-stable products References abies Karst). (Frequency and (Picea abies Karst).), KTH, Wood with vertical annual rings. Holz als Beckman, E.; Stehr, M. 1993: character of knots in triangular Technology and Processing, Rep- Roh- und Werkstoff 56(3):171-177 Stjärnutbyte – Volymutbyte vid profiles of pine (Pinus silvestris L) ort TRITA-TRÄ R-95-13, (in Swe- Statstical Yearbook of Forestry stjärnsågning (Stärnutbyte - volume and spruce (Picea abies Karst).), dish) 1994: Statstical Yearbook of yield when sawing according to the KTH, Wood Technology and Sandberg, D. 1996a: Produktion Forestry 1994, National Board of star-sawing pattern.), KTH, Wood Processing, Report TRITA-TRÄ och provning av radialsågat trä – Forestry, Sweden, ISSN 0491-7847 Technology and Processing, R-95-12, (in Swedish) slutrapport. (Production and testing D. Sandberg 98-11-03 12:57 Internal report No. 9/3 (in Swe- Holmberg, H.; Sandberg, D. 1996: of radially sawn timber.), KTH, dish) Volym- och kvalitetsutbyte vid Wood Technology and Processing, Brask, P. 1994: Utveckling av stjärnsågning. (Volume yield and Internal report No. 42/96, (in Swe- anläggning för stjärnsågning. quality of timber when using Star- dish) (Development and consruction for sawing technique.), KTH, Wood Sandberg, D. 1996b: Radially star-sawing.), KTH, Wood Technology and Processing, Rep- sawn timber. Star-sawing – a new Technology and Processing, ort TRITA-TRÄ R-96-22, (in Swe- method for producing timber with Internal report No. 14/94, (in Swe- dish) vertical annual rings. Holz als Roh- dish) Nordfors, L.; Sandberg, D. 1993: und Werkstoff 54(3):145-151 Grundberg, S.; Grönlund, A.; Utbytessimulering av sönderdelning Sandberg, D. 1997c: Utvärde- Grönlund, U. 1995: The Swedish enligt stjärnsågningsmetoden. ring av ny sönderdelningsmetod för stem bank – a database for diffe- (Model for simulation of the stjärnsågning. (Evaluation of a rent silvicultural and wood volume yield when sawing acording conversion method for Star- properties. Luleå University of to the new sawing pattern ”Star- sawing.), KTH, Wood Technology Technology, Div. Wood sawing”.), KTH, Wood and Processing, Report TRITA- Technology, Report TULEA No. Technology and Processing, Rep- TRÄ R-97-26, (in Swedish) 1995:31, ort TRITA-TRT-1993-53, (in Swe- Sandberg, D. 1998c: Stjärn- Gröna Boken 1976: Guiding dish) sågning - volym- och kvalitets- principles for grading of Swedish Nordiskt Trä 1994: Guiding utbyte vid en provsågning med sawn timber, second edition. The principles for grading of Swedish furu. (Star-sawing – Volume yield Swedish Timber and Wood Pulp sawn timber of pine and spruce. and quality from one test sawing of Journal Föreningen Svenska Sågverksmän, pine.), KTH, Wood Technology Holmberg, H.; Sandberg, D. ISBN: 91-7322-175-9 and Processing, Report TRITA- 1995a: Utvärdering av en ny Neubauer, P. 1998: Yield of knot TRÄ R-98-33, (in Swedish) ströläggningsmetod för stjärnsågad free boards from Pinus silvestris - Sandberg, D.; Holmberg, H. 1996: trekantprofil. (Evaluation of a new A comparsion study between Radially sawn timber. Knots stacking method for star-sawn square sawing and star-sawing. number, type and size in star-sawn triangular profiles.), KTH, Wood KTH, Wood Technology and triangular profiles of pine (Pinus Technology and Processing, Processing, Report Examensarbete silvestris L) and spruce (Picea abies Internal report No. 31/95, (in Swe- No. 41/98 Karst). Holz als Roh- und dish) Sandberg, D. 1995b: Stående Werkstoff 54(5):369-376 Holmberg, H.; Sandberg, D. årsringar hos furu (Pinus silvestris Sandberg, D.; Holmberg, H. 1995b: Kvistars frekvens, form och L) och gran (Picea abies Karst). 1998a: Radially sawn timber. beskaffenhet i trekantprofil av furu (Vertical annual rings in pine Gluing of star-sawn triangular

44 Some properties of star-sawn timber Chapter and the dependence of these properties on annual ring orientation and 4 distance from the pith

4. Some properties of star-sawn timber and the dependence of these properties on annual ring orientation and distance from the pith

Star-sawing gives timber with reds of years. The term also implies annual ring orientation in the cross- vertical annual rings which is how timber should be removed from section of the timber is changed. free from pith and juvenile wood. logs to make the best use of the The texture is thus dependent on Traditionally, this timber has inherent properties of the wood. the angle at which the annual ring been considered to have proper- Timber has vertical annual rings if it meets the surface of the timber, ties which are favourable in the is sawn or cleaved radially from the (Figure 16). Timber with vertical advanced use of timber. In the log and the degree to which the annual rings has the radial texture of investigations reported in this annual rings are vertical is reflected the wood on the flat side, with a chapter, the truth of this state- in cross-sections of the timber, striped pattern which is crossed by ment has been examined. (Figure 15). rays in certain species. In contrast to The implication of the vertical the radial surface, the tangential annual ring concept is first surface of the timber has a consid- described. Thereafter, the erably more mottled surface in influence of the position of the which the opening of the rays can be timber in the log cross-section on 90° seen as small longitudinal flecks in distortion and crack-formation 60° during drying, moisture cycling and outdoor exposure is discussed. Finally, some results are described from a number of fundamental material investigations Figure 15. The annual ring orientation is determined from the cross-section of the where attempts have been made, timber. When the annual rings are at macroscopic and microscopic practically perpendicular to the flat side levels to find explanations and of the timber the timber is considered to have vertical annual rings. mechanisms for the importance of annual ring orientation for the distortion and crack susceptibility Two properties of timber are of of timber at the macroscopic primary interest when vertical (a) (b) level. annual rings are referred to, namely texture and shape stability. 4.1 Definition of vertical Texture refers to the annual rings appearance of line and colour in Vertical annual rings are a quality the timber surface as a designation of timber which has consequence of the annual ring Figure 16. The texture of timber is dependent on the angle between the annual rings and the surface of the timber. (a) Vertical annual rings have a striped texture, and (b) horizontal been used by craftsmen for hund- pattern. This varies when the annual rings have a more mottled texture.

47 certain species. The radial and tang- This concerns on the one hand the shows the change in shape on drying ential textures of pine are shown in distortion of the timber in the from a moisture content higher than Figure 16. longitudinal direction, i.e. crook, fibre saturation, but changes in With texture it is difficult to bow and twist and, on the other shape will also take place during determine when deviation from hand the change in shape of the moisture changes which occur vertical annual rings are so great timber in the cross-section, e.g. during the normal usage of timber that the texture does not meet the cupping. Vertical annual rings, and after timber has been further requirements with regard to refers primarily to the shape stabil- treated. pattern which are referred to as ity in the cross-section of the Figure 18a shows a cross-section vertical annual rings. Figure 17 timber. taken close to the pith so that half shows the texture of pine with Regarding shape stability in the the annual ring circumference can some different annual ring cross-section of the timber, there be found in the cross-section. orientations. It is evident that the are two factors which must be During drying, this timber will change in texture when annual ring taken into consideration, namely change shape drastically, the orientation is changed from 90 to annual ring curvature, or the timber becomes cupped. Even if 60 degrees is moderate, and that distance from the pith, and the such a piece is planed after drying, the texture becomes mottled when inclination of the annual ring in the annual ring orientation is relation to the timber surface. To changed from 60 to 30 degrees and illustrate what happens to the begins then to be more similar to shape of cross-sections with the texture of the tangential sur- change in moisture content, Figure face, i.e. at 0 degrees. 18 shows a few different timber Shape stability refers to the pieces sawn at different distances ability of the timber to maintain a from the pith and with different (a) given geometric shape during use. annual ring orientations. The figure

(b)

(c)

(d) 0° 30° 60° 90°

Figure 18. Timber pieces sawn at different distances from the pith and with different Figure 17. The texture of pine at some different annual ring orientations. annual ring orientations.

48 the moisture movement will affect so timber. A certain variation in ring curvature, i.e. the distance from that the cupping will be seen. thickness is obtained from the pith the pith, and the timber thicknesses The timber in Figure 18b has side towards the sapwood side are also included. Figure 19 shows been sawn more or less radially because of variations in annual ring how the greatest annual ring from the log the annual rings are curvature and density. inclination in the cross-section strongly curved and the closer to When the annual ring orien- varies with distance from the pith the periphery of the log, the more tation of the timber is changed for different timber thickness. In the vertical the annual rings become. from completely vertical (90 calculation, an idealized case has During drying, the thickness degrees), the shrinkage and been assumed, that the timber has shrinkage on the pith side (radially) swelling of the timber width been sawn as in Figure 19b and that will be about half that on the increases. Sandberg (1995b) has the greatest annual ring inclination F sapwood side (tangentially). This defined vertical annual rings by is found closest to the pith. Figure means that the cross-sectional maximizing this movement increase 19 shows how the increase in p shape of this timber will not be to 10 % of the purely radial thickness of the timber that is sawn maintained if the moisture changes. moisture movement. For pine and according to Figure 19b increases t In order to prevent changes in spruce, the requirement for the amount of wood which must be shape caused by the annual ring vertical annual rings is then that removed from the pith side of the m curvature and because of the the annual ring inclination defined timber for the whole cross-section to annual ring orientation in relation as the angle between a tangent to have vertical annual rings. F - greatest annual ring orientation t - thickness of the timber to the surface, timber should be the annual ring and the flat side of It is evident in figure 19a that, for m - distance from pith sawn free from the area around the timber should be between 60 vertical annual rings in the whole p - pith the pith and with the annual rings and 90 degrees. Sandberg (1995b) piece of timber, the distance to the parallel to two of the sides of the states that this requirement should pith must be at least 25 mm at a Figure 19. (a) The influence of the thickness of the timber and the distance timber, (Figures 18c and 18d). The be met after sawing and before thickness of 30 mm and 90 mm if from the pith on the annual ring piece of timber in Figure 18c must drying. This agrees with the Ger- the timber is sawn with a thickness orientation in the timber surface. (b) The be sawn so far from the pith that man concept of ”rift” which is used of 100 mm. This means that the idealized case which has been used in the calculations. the annual rings become practically particularly on hardwood and width of the pith cutter must be straight in the cross-section. The which is now also the standard twice the distance to the pith, i.e. 50 closer to the pith this piece is definition for vertical annual rings and 180 mm, respectively. sawn, the greater is the tendency (CEN 1995). For rift, the angle It is realized that the width of the towards cupping. The cross- between the flat side of the timber pith cutter will be unrealistically section in Figure 18d shows what is and the rays should be between 0 large for thick timber and that it is usually meant by vertical annual and 30 degrees. Since the rays are necessary in practice, not least for rings. The cross-section shape oriented radially in the wood, this economic reasons, to abandon the remains practically unchanged will agree with the fact that the requirement of vertical annual rings when the moisture content varies. angle between the flat side of the in the whole width of the timber in The shrinkage and swelling timber and a tangent to the annual some cases. movements are approximately ring in the surface of the timber is Considering the crack formation twice as large in the thickness 60 to 90 degrees. on the pith side of the timber direction of the timber as in its By determining annual ring during drying, Sandberg (1995a, width, i.e. tangential in thickness orientation on the surface of the 1996d, 1997a) has shown that, if and radial in the width of the timber, the influence of the annual the timber is sawn at least 30 mm

49 from the pith, then crack formation outdoor exposure, radial surfaces are practically useless. To reduce the is radically reduced. In these also considerably more resistant problems of distorsions and measurements, some consideration than tangential surfaces, as shown cracks, work has been carried out has been given to the irregular in 4.3. for many years to improve and path of the pith in the log. develop new drying methods, to Characteristic for timber with 4.2 Distorsion and cracks in avoid the build-up of large stress vertical annual rings is the shape- timber after sawing, concentrations in the timber. The stability of the timber in the cross- drying and cycles of result has been that it is possible section during moisture changes, wetting and drying today to deliver timber from many that the moisture movement is When speaking of distorsion in sawmills which is fairly free from minimal in critical directions, i.e. in timber, one refers to the change in cracks and is also conditioned so the width of the timber, and that shape of the timber in an unloaded that it is relatively stress-free. the timber has a radial texture on state. Four different forms of distor- In the continued handling of the at least two sides, usually the flat sion are recognized: cup, crook, bow timber after it has left the sawmill, sides. To fulfil these requirements and twist, Figure 20. it will in most cases be exposed to or characteristics, the pith and Distorsion in timber can be the a very varying climate, which leads most of the juvenile wood are result of stresses in the living tree, to changes in the moisture content. removed, which means that the so-called growth stresses, or they Already at the end of the 1950s, timber becomes less susceptible to can arise during manufacture, Armstrong and Kingston (1960) cracking and that the shape stabil- primarily during drying, as a discovered that, if small loaded ity in the longitudinal direction consequence of the inherent specimens of wood were exposed (crook, bow and twist) is properties of the wood or of the to different humidities, the defor- improved. Radial surfaces are also actual refining process. In the same mation increased drastically harder and more resistant to wear way, cracks can already exist in the compared with when the climate than the corresponding tangential trunk or can arise in the refining was constant. In the same way, the surfaces (Thunell and Perem 1948, process. Cracks and distorsions recovery after unloading was accel- Perem 1949, Kühne 1955, Holmberg cause losses, and in some areas of erated by a cyclically varying and Wålinder 1997). During use the timber can become moisture content. This phenomenon is usually called mechanosorption. On an everyday level, this behaviour can be recognized at a building site and where timber which has been out in rain and nasty weather without being a covered. The timber is often b twisted and crooked, although it was probably straight when it left the sawmill. Within this field, I have gone further back in the chain, really to d c the actual wood material, to investi- Figure 20. Definition and measuring of warp: (a) crook; (b) twist; (c) bow; and (d) cup. gate the causes for the cracking and

50 distorsion in timber. I have also gone results, in order to make the cupping, as has also been shown by forward slightly in the chain to see description more accessible for the Sandberg and Holmberg (1998b). what may happen to the timber reader. The complete analyses of Figure 21 shows the average degree when it reaches the final user. the test results are available in of the cupping for about 300 planks Two parameters have been given published reports and articles, of pine and 400 of spruce with diffe- a more central role in the (Sandberg 1995a, 1996b, 1997a, rent annual ring orientations (Sand- investigation, viz. the annual ring Sandberg and Holmberg 1998b). berg and Holmberg 1998b). The orientation in the cross-section of cupping is normalized and the timber and the distance of the Cup determined on the concave flat side timber from the pith. Both para- Cupping is the change in shape of the timber after the timber has meters are strongly related to the which is clearest in its nature and been dried to about 18 % moisture star-sawing pattern. Thus to has to a certain extent already content. investigate the influence of annual been discussed in relation to Figure 21 shows that the cupping ring orientation seems to be natu- vertical annual rings in chapter 4.1. is greatest for timber which is sawn ral. The distance from the pith has The cupping of timber has two close to the pith and with horizontal been chosen because it is an easily causes: a) The difference in annual rings (F in Figure 21). If the measurable parameter and has a shrinkage and swelling in the tang- timber is sawn further from the pith certain relation to the juvenile ential and radial directions, i.e. the with increasingly more vertical wood. The concept of juvenile shrinkage and swelling anisotropy, annual rings, the cupping decreases. wood is really undefined in this and b) the curvature of the annual Timber containing pith often cracks context. In these investigations, rings in the cross-section of the consideration has also been given timber, when the radius of to other relevant parameters which curvature is small. may influence distorsion and crack If both these conditions are met formation. The influence of drying at the same time, timber which is methods has intentionally received not loaded so that the cupping is less attention, because consider- counteracted will become cupped able research resources are spent with any change in moisture con- in this field in other places and this tent. The smaller the difference knowledge will also be applied to between the radial and tangential the timber investigated here. moisture movements and the Most of the studies are greater the radius of curvature of empirical and of a comparative the annual rings in the cross- nature. Here, I try briefly to section of the timber, the smaller is illustrate the influence of a few the cupping. The cupping will of factors on distorsion and crack course also depend on the formation. This means that in this magnitude of the moisture ABCD presentation of the results, the movements, i.e. there is a moisture P P P influence of the other factors is content and density dependence, Figure 21. The average value for the cup of EFGHP and on the width of the timber. dried timber of pine and spruce with diffe- given a minor role. I have also rent annual ring orientations in the cross- intentionally avoided introducing a From these arguments, it follows P section of the timber, (Sandberg and Holm- P P P statistical analysis around the that undried timber will not exhibit P - pith berg 1998b).

51 during drying, and this reduces the 23a). In the same way, there is a stresses in the centre of the trunk in cupping. clear increase in crooking during the longitudinal direction which are The cupping of timber can be drying when juvenile wood is pre- gradually transformed into tensile reduced if the timber is dried un- sent along one of the edge sides, (A stresses in the radial direction der load. This cupping reduction is in Figure 24). For most of the pieces towards the periphery (Archer in general not permanent. Sandberg (ca 70 %), the crook and bow during 1986). Corresponding stress fields (1997a) has shown that the cupping drying are directed towards the pith are found in the radial and tangen- increases in the case of timber which side of the timber, i.e. where the tial directions. Growth stresses are has been dried under load if the juvenile wood is placed, (see defini- dependent on the tree species, timber is moistened. Figure 22 shows tion of warp in Figure 20). growing conditions et c. Com- the increase in cupping of dried Timber sawn further from the pression wood is an example of timber of pine and spruce which has pith consequently shows less distor- wood with often high internal been exposed to three cycles of sion during drying, (Figures 23, 24). wetting and drying cycles down to The timber which has the pith practically the same moisture con- enclosed shows relatively little bow 1,6 tent as the timber had before the and crook. This can depend on a 1,4 moisture cycling (Sandberg 1995a, certain symmetry around the pith 1,2 1997a). If the timber initially which balances the drying stresses 1,0 0,8 Pine before the w etting cycle exhibited cupping, the moisture which arise, to the hindrance of a 0,6 cycling meant that the cupping bow or crook due to load, or to 0,4 Pine after the w etting increased in all cases. A deviation the fact that the shape change 0,2 cycle 0,0 from vertical annual rings at the appears as twist (see Figure 29). BCEF butt end of certain timber pieces of The radially sawn timber with Annual ring orientation pine in group B (Figure 22a) meant rectangular cross-section exhibits a that this timber showed cupping. large percentage increase in bowing during drying (A and B in Bow and crook Figure 23). One reason for this is One of the reasons for removing probably that this timber has a the juvenile wood in star-sawing large proportion of splay knots on b already when sawing the log is that the flat side (Sandberg and Holm- the longitudinal shrinkage in the berg 1996) and that the shrinkage juvenile wood is greater than in the due to the knots leads to an mature wood and that the increased bow during drying, presence of the juvenile wood thus Figure 25. For all groups (A-H) in Figures deforms the timber during drying. BCE F 23 and 24, the timber exhibits bow Figure 23a, which shows bowing P before and after drying of pine, and crook already before the P illustrates clearly how the bowing drying. This distorsion comes from P P - pith P increases drastically during drying stresses in the living tree, so-called of the pieces in which juvenile growth stresses. In general, the Figure 22. Average values for the cup of dried timber of pine (a) growth stresses in the tree trunk are and spruce (b) exposed to 3 wetting and drying cycles. The cross- wood is present along one of the sectional area of the timber was 50 x 100 mm, (Sandberg flat sides of the timber (F in Figure such that there are compressive 1995a, 1997a).

52 stresses and also a large longitudinal are released or balanced and the 4

shrinkage, Figure 26. triangles become practically straight. ) The star-sawn triangles (H in Timber which has been loaded 3

Figure 23, 24) have a relatively large during drying has in the unloaded 2 bow before drying which in 70% of state been exposed to three wetting Pine before drying Pine after drying the cases is directed towards the and drying cycles to study the 1 sapwood. The star-sawing pattern change in crook and bow (Sand- Mean value of crook (mm means that the triangles are sawn far berg 1995a, 1997a). Regardless of 0 A BCDEFGH

from the pith and a large part of the how the timber had been sawn in 4 Annual ring orientation volume of the triangles lies in the the log, the distorsion increased ) periphery of the trunk where the when the timber was exposed to 3 longitudinal tensile stresses domi- moistening cycles (Figure 27, 28). Spruce before drying nate. During drying, these stresses In the case of pine, it is evident 2 Spruce after drying that the absolute deformation 1 increase was greater in the pieces (mm value of crook Mean 6 sawn close to the pith (C and F in 0 5 Figure 27, 28), but in the case of ABCDEFGH Annual ring orientation 4 spruce, the difference is not so ABCD 3 clear. P Pine before drying P P P 2 Pine after drying EFGH

Mean value of bow (mm) bow of value Mean Twist 1 Twist is almost non-existent before P P P P 0 ABCDEFGH drying (Figure 29). Only the timber P - pith Annual ring orientation with enclosed pith and the triangles Figure 24. Average values for crook in green and dried 6 show a certain degree of twisting timber of pine and spruce with different annual ring orientations in the cross-section of the timber, (Sand- 5 before the drying. This suggests berg and Holmberg 1998b). that the twisting is not influenced 4 Spruce before drying to any great extent by stresses in Spruce after drying 3 the timber. Several earlier 2 investigations have shown that the Figure 25. Large splay knots on the flat side of radially Mean value of bow (mm) 1 twisting after drying is largely sawn timber means that the timber is deformed around these knots during drying. 0 dependent on the fibre angle and ABCDEFGH Annual ring orientation the annual ring curvature in the timber (see e.g. Stevens and Johns- ABCD P P P ton 1960, Balodis 1972, Forsberg EFGHP 1997). This is also reflected in the results shown in Figure 29. Timber P sawn close to the pith (A, C, D, F Figure 26. Compression wood of spruce. P P P The compression wood has large internal P - pith in Figure 29), where the fibre stresses and a large longitudinal Figure 23. Average values for bow in green and dried inclination is great, consequently shrinkage, which can lead to large wood of pine and spruce with different annual ring exhibits a large twist. The timber deformations and possibly crack forma- orientations in the cross-section of the timber, tion in timber containing compression (Sandberg and Holmberg 1998b). which is sawn further from the pith wood.

53 12 14

annual ring orientation in the cross- k 12 10 section of the timber. 10 8 8 Pine before the w etting Pine before the w etting

6 (mm) 6 cycle cycle Crack-formation 4 4 Pine after the w etting Pine after the w etting cycle Figure 31 shows the relative crack Mean value of croo 2 2 cycle 0 Mean value of bow (mm) of bow value Mean 0 length in the timber before and BCEF BCEF after drying, where the relative Annual ring orientation Annual ring orientation 14 12 crack length is calculated as the 12 10 total crack length in each plank Spruce before the Spruce before the 10 w etting cycle 8 w etting cycle divided by the plank length. It is 8 Spruce after the w etting

6 Spruce after the w etting evident that the mean relative (mm) 6 cycle cycle 4 crack length is highest in the 4

Mean value of crook 2 Figure 28. Average values for 2 timber sawn close to the pith (A, 0 crook in dried timber of pine Mean value of bow (mm) Mean value 0 C, D and F in Figure 31). Cracks BCEF (a) and (b) spruce exposed to BCEF Annual ring orientation 3 wetting and drying cycles. Annual ring orientation are present before the timber is The cross-sectional area of dried, and a closer analysis reveals BCE Fthe timber was 50 x 100 mm, P BCE F P (Sandberg 1995a, 1997a). P that they occur to a great extent P adjacent to the pith (Sandberg P - pith P P P P - pith P 1995a, 1996d, 1997a, Sandberg and Figure 27. Average values for bow in of green and Holmberg 1998b). dried timber of pine (a) and spruce (b) exposed to 3 wetting and drying cycles. The cross-sectional These cracks are so-called ”pith 12

area of the timber was 50 x 100 mm, (Sandberg shakes” which radiate out from the 10 1995a, 1997a). pith and become narrower towards 8 the periphery of the trunk (Manual 6 in Forest Technology 1922, König Pine before drying 1957, Grönlund and Hägglund 4 Pine after drying Mean value of twist (mm) twist of value Mean 1987). When the timber is dried, 2 these pith shakes create sites for 0 shows slightly less twisting, whereas ABCDEFGH further cracking, which is Annual ring orientation 12 the star-sawn timber is the least consequently greatest in timber twisted. The triangles show the sawn close to the pith (Figure 31). 10 earlier pattern with less distorsion 8 The fact that timber in groups B, Spruce before drying during drying. Spruce after drying E, G, H (Figure 31) do have cracks 6 Several investigations have before drying depends to a certain 4

shown that the twist decreases if (mm) twist of value Mean extent on the fact that the path of 2 the timber is loaded during drying, the pith through the trunk is not 0 (Johnston 1957). In moisture straight and may come too close to ABCDEFGH cycling of unloaded timber which Annual ring orientation Figure 29. Average values for the timber during sawing. ABCD has been dried under load, the P twist of dried timber of pine When the dried timber is remoisten- P P (a) and spruce (b) with diffe- twisting also increases to a certain ted, the crack length increases EFGHP rent annual ring orientations extent (Figure 30). An increase in slightly (Figure 32). This increase in the cross-section in green P and the timber, (Sandberg twisting occurs regardless of the P P P P - pith and Holmberg 1998b).

54 12 intended for star-sawing, i.e. 1,2 10 h consisted of large butt logs and 1,0 8 Pine before the w etting middle logs, and this means that it 6 0,8 cycle 4 is inadvisable to draw general con- Pine after the w etting 0,6 2 cycle clusions from the results. Never- Pine before drying

Mean value of twist (mm) twist of Mean value 0 0,4 Pine after drying BCEF theless it can be said that, if large-

Annual ring orientation sized logs are to be sawn, then 0,2 star-sawing or some similar method Mean value of relative crack lengt 0,0 ABCDEFGH 12 of sawing is a suitable technique if Annual ring orientation 10 straight and crack-free timber is Spruce before the 8 w etting cycle desired. It may also be interesting 6 Spruce after the w etting to note that the star-sawn triangles cycle 4 show extremely little distorsion and 2 crack formation compared with Mean valueMean of twist (mm) 0 BCEF other sawn timber. Annual ring orientation The increase in distorsion dur- BCE Fing humidity cycling can appear to P P lack an upper limit, but this is not true. The upper limit for how P - pith P P much the distorsion can increase Figure 30. Average values for twist in dried timber during moisture cycling is presum- ABCD P of pine (a) and spruce (b) after exposure to 3 P P ably that which is observed1,2 if the moistening and drying cycles. The cross-sectional P area of the timber was 50 x 100 mm, (Sandberg timber is dried without any load EFGH 1995a, 1997a). 1,0 and so that the stresses in the P 0,8 P P P timber are reduced to a minimum. P - pith Spruce before drying The increase in crack formation0,6 Spruce after drying during humidity cycling should Figure 31. Average relative crack length in 0,4 green and dried timber of pine (a) and then be reduced. spruce (b) with different annual ring 0,2

is greatest in timber sawn close to Mean value of relative crack length orientations in the cross-section of the the pith. 4.3 The change in0,0 wood timber (Sandberg and Holmberg 1998b). ABCDEFGH surface during outdoorAnnual ring orientation Comments on the above usage above ground results In Sweden, the ”permanence” of and the wood’s intrinsic strength, The above results show unam- timber is related primarily to the the amount of nourishing substrate biguously that the orientation of ability of the timber to resist attack for micro-organisms, the chemical the annual rings and the position from different fungi. Insect attack nature of the cell wall and the of the timber in relation to the pith of timber in constructions and the temperature which are the real are important factors affecting the change in appearance of the timber criteria for the resistance of timber distorsion and cracking of the in outdoor use are of less import- to decomposition (Öqvist 1988). timber. It should however be noted ance. It is the interaction between For timber in ground contact, there that the timber which was used for the ability of the timber to maintain are classification rules regarding the the investigation was primarily the moisture content at a low level degree of permanence. A similar

55 1,0 timber (Boutelje and Nilsson 1985), pressure impregnated with a CCA- 0,8 annual ring width (Rydell 1981; agent, some were coated with 0,6 Grönlund and Rydell 1983; linseed oil and others were left length 0,4 Pine before the w etting cycle Pine after the w etting cycle Boutelje and Nilsson 1985), density untreated. 0,2

Mean value of relative crack 0,0 (Grönlund and Rydell 1983; Steffen The results after about three BCE, GA, D, F et al 1990), artificial drying versus years’ weathering showed clear Annual ring orientation 1,0 air drying (Träteknikcentrum differences in crack formation, 0,8 1987). The raw material for most (Figure 33). The radial surfaces Spruce before the w etting cycle 0,6 Spruce after the w etting cycle investigations of this type has been had shorter and a considerably length 0,4 timber produced by conventional smaller number of cracks than the 0,2 methods, i.e. square-sawn timber. tangential surfaces, (Figure 34). On Mean value of relative crack 0,0 BCE, GA, D, F This has meant that the radial surfaces, there was no clear Annual ring orientation importance of the annual ring difference between pine and

ABCD orientation for the permanence of spruce. On the tangential surfaces, P P P the timber in outdoor exposure has on the other hand, pine had more EFGHP been neglected. cracks than spruce. The investig- In order to investigate whether ation showed that neither P P P P annual ring orientation of the treatment of the surface with linseed P - pith exposed surface has any influence oil nor impregnation influenced the Figure 32 Average relative crack length in dried timber on how timber surfaces without occurrence of cracks. No relation of (a) pine and spruce (b) after exposure to 3 wetting surface treatment change on and drying cycles. The cross-sectional area of the timber was 50 x 100 mm (Sandberg 1995a, 1996d, outdoor exposure, a field trial was 1997a). carried out with pine and spruce timber (Sandberg 1997d, 1998a, 1998b). Star-sawn triangles were classification is lacking for timber used as test material, which meant which is used outdoors, but is not in two important advantages. On the ground contact. one hand, test specimens with One problem is that the radial or tangential surfaces could permanence of timber above gro- be produced from largely identical und in outdoor situations cannot material and, on the other hand, be measured in a useful practical the star-sawing pattern meant that way. On several occasions, sub- the pith and the adjacent area were stitute criteria have been resorted excluded. It is known that in the to such as e.g. annual ring width. area around the pith, particularly As an example, attempts have been in pine, so-called pith shakes may made in the case of wooden occur which can distort the test windows to find relationships result (Manual in Forest Techno- between the following factors and logy 1922; Durst 1955; König 1957; permanence without finding any Grönlund and Hägglund 1987). To Figure 33. Radial (a) and tangential (b) surfaces of pine after 33 months’ clear relationship: differences investigate the influence of weathering and subsequent conditioning between summer and winter felled impregnation, some samples were to a moisture content of 12%.

56 30 the influence of sunlight. This 4.4 Preliminary studies of 25 hypothesis must, however be the mechanisms for Pine investigated more thoroughly. distorsion and crack 20 Spruce formation in wood 15 The mechanisms and thus also the () 10 reasons why timber is deformed, 5 and in certain cases becomes cracked, must be searched for at a 0 lower structural level of the wood. Crack length per unit area (m/m²) Crack length per unit area Tangential Radial In order fully to understand the Exposed surface deformation behaviour of the timber, e.g. when it undergoes a 600 moisture change and is at the same 500 time subjected to a mechanical Pine 400 Spruce load, it is certainly necessary to start from its molecular structure. 300 The absolute forefront of wood 200 material research is today

100 approaching the molecular level. In this final section of the summary 0 part of this thesis, I wish briefly to Tangential Radial Number of cracks per square metre Number of cracks per square Exposed surface present some investigations in Figure 34. Average crack length (a) and which the behaviour of the wood number of cracks per area unit (b) for at a microscopic level has been pine and spruce after 33 months’ studied. The aim of the weathering and subsequent conditioning to a moisture content of 12%. investigations has primarily been to build up my own knowledge within between the density of the timber this field, but I hope also to and the crack formation was proven. contribute with a small piece of the In order to find explanations for puzzle to help research forward. the clear differences in crack Wood is a hygroscopic material, occurrence between radial and and this influences to a great tangential surfaces, a microscope extent its deformation under load. study of the test material exposed The deformation of wood during outdoors was carried out (Sand- loading can be divided into diffe- berg 1997d, 1998b). Differences rent components: a part which is Figure 35. Degradation of the weathered elastic, a part which depends on were also evident at the micro- wood on a microscopic level. In the radial scopic level, (Figure 35). One rea- surfaces cracks occur most frequently at the shrinkage and swelling of the son for the variations in crack for- the annual ring border (a), spruce at 175 wood, and a part which is called times magnification. In the tangential creep. In speaking of creep in mation may be differences in the surfaces cracks occur both in the shrinkage and swelling of the early- earlywood and latewood (b), spruce at wood, one distinguishes between and latewood in combination with 300 times magnification. creep when the moisture content

57 in the material is constant recreated while the material has samples where the external load was (viscoelastic) and creep when the been deformed, i.e. sliding has applied in radial direction developed moisture content is varying taken place between the molecules. a torsion which increased as a result (mechanosorptive). All these This means that the material must of the moisture cycling, as shown in components are dependent on the be free from micro-cracks if the Figures 39. moisture and temperature state of deformation is to be regarded as Torsion is here defined as the the wood material, but this will not purely plastic. angular difference between the be treated here in detail. ends of the samples, for the angle Figure 36 shows in general what formed by the vertical side surfaces the deformation can look like in a of the sample with the horizontal wood sample in loading and plane which coincides with the unloading at a constant and at a plane in which the support points varying moisture content. In can be found, (Figure 38). loading, a purely elastic deformation (a) is first observed and this is followed by a viscoelastic deformation. When the humidity is c Figure 36. General description of the L changed, the deformation follows deformation of a wood sample in bending a the characteristic sawtooth pattern as a function of time in different loading shown in Firgure 36. In unloading, modes, (Kingston 1962). L the elastic deformation (d) recovers L d instantaneously and this is followed Creep tests under a bending load b by a delayed recovery (e). If the have been carried out on clear L atmospheric humidity is varied in small samples of pine (10 x 10 x 360 the same way as when the load was mm) in order to investigate the L applied, further recovery occurs reasons of for the ”plastic” defor- (f). Figure 36 shows the recovery Figure 37. The direction of the load (L) mation (g in Figure 36) and to and the location of the support points on during wetting and drying cycles, investigate whether the annual ring the test samples under (a) radial and (b) with the characteristic, sawtooth- orientation in the cross-section of tangential load. The radially loaded shaped recovery curve. When the samples were under tension on either (c) the wood influences the creep the sapwood side or (d) the heartwood sample has been exposed to a high behaviour (Kuukkanen and side. load, the deformation may not Nordqvist 1992, Sandberg and recover completely, and the sample Johansson 1995, Sandberg 1998d), then exhibits a so-called permanent (Figure 37). deformation (g). The permanent The annual ring orientation deformation is usually classified as seems to have no decisive a plastic deformation if no visible importance for the deformation of failure has occurred in the mate- the samples (Sandberg 1998d). The rial. From a strictly materials sci- deformation or bowing showed a ence viewpoint, plastic deformation pattern similar to that observed in means that secondary bonds in the similar earlier tests. It was, material have been broken and however, observed that those

58 25,0 Torsion (degree)

20,0 RSD 15,0 RKD T 10,0

5,0 α α butt 0,0 top 0 50 100 150 200 250 -5,0 Time (24 hours) -10,0

-15,0 Figure 38. Determination of the torsion -20,0 of test bodies exposed to a bending load. -25,0 The angles α butt and α top are a determined at the butt and top ends 25,0 respectively and the torsion is calculated RSD as the difference between these two angles. 20,0 RKD T 15,0

Figure 39 shows a clear difference 10,0 between tangentially and radially 5,0 loaded samples. The tangentially 0,0 loaded samples show a small 0 50 100 150 200 250 degree of twisting, whereas the -5,0 radially loaded samples are twisted -10,0 considerably more. The relative -15,0 humidity was varied between 23 % -20,0 and 85 % at a constant temperature of 200C. The time for one -25,0 b cycle with a dry and humid climate was ca. 30 days. 25,0 When the samples were loaded, 20,0 RSD Figure 39. Torsion of samples of pine (10 x 10 x 360 mm) under a they showed no tendency to twist, RKD 15,0 bending load subjected to a and after 14 days in a constant dry T cyclically varying relative climate the twisting tendency was 10,0 humidity (T = 200C), (Sandberg still low. When the humidity was 5,0 1998d). The samples have been loaded at levels of (a) 5 %, (b) 20 increased, a drastic increase in the 0,0 % and (c) 40 % of the bending twisting took place in the case of 0 50 100 150 200 250 -5,0 failure load at a moisture the radially loaded samples. Figure content of 12 %. See Figure 37 39 shows that the torsion increases -10,0 regarding loading directions. -15,0 RSD – Radial load, sapwood side with increasing load, which is natu- under tension. ral. It is also evident that the tor- -20,0 RKD – Radial load, heartwood sion takes place in different direc- side under tension. -25,0 c tions in the case of the radially T – Tangential load.

59 loaded samples, depending on level. Creep tests under a tensile Figure 36), speculations arose whether the heartwood side of the load were therefore carried out regarding the extent to which sample is loaded under tension or using the ESEM technique wood can be deformed if the wood under compression. In absolute (Kifetew and Sandberg 1998). The after subsequent moistening is to figures, however, the torsion is tests were found to be very resume its original shape (Sandberg practically the same. difficult to perform and a lot of and Johansson 1995). This One hypothesis is that the fibre effort was spent on the technical phenomenon is also discussed in inclination on the tangential performance of the tests. chapter 4.2 in relation to the dis- surfaces of the samples influences torsion of timber which has been the torsion. dried under a load. The samples with the highest In order to test the extent to load failed in an early stage of the which wood can be deformed and tests, which means that the torsion still resume its original shape after curves (Figure 39c) are very moistening, tests were performed irregular. with isostatically compressed The samples were unloaded and wood. Isostatic compression, moisture-cycled further. The de- which means that the wood is formation and the torsion compressed under the same recovered to a certain degree but pressure from all directions, was not 100 %, which is a well-known chosen because this gives large phenomenon. deformations at the same time as Samples used for creep tests little external damage is visible on have, after the test was completed, Figure 40. Damage in the cell-wall after the wood. Tests were carried out the creep test was completed. been studied in a microscope Magnification 750 x. at the macroscopic level and the (Sandberg and Johansson 1995, analysis was done using a Sandberg 1998d). The aim has The initial investigations were microscope. been to investigate whether struc- devoted to investigating how the Figure 41 shows the cross- tural changes had occurred in the electron beam influences the test section of compressed pine where wood material during the creep material. Within this field, a lot of the density has been increased to tests. In these studies, damage was knowledge can be obtained from ca 1200 kg/m3, (Sandberg 1998e). observed in the cell-walls, Figure other materials fields, but in this The isostatic compression was 40. This damage can to a certain case it was also important to build carried out at room temperature extent explain the non-reversible up our own knowledge and skill and the samples had a moisture deformation (g in Figure 36). with the instrument. The results content of ca 10 %. In the investigations of creep showed that the influence of the It can be seen in Figure 41 that deformation, it was suggested that electron beam was considerable, the cells are very deformed. When a creep test might be carried out due to the high electrode voltage these samples were again directly under a microscope. The with which we worked initially. moistened, the cells resumed their idea was to obtain a picture of how This is discussed in an article by original shape (Figure 42). material at the cell-wall level Kifetew and Sandberg (1998). deforms and how and where In the work on the permanent damage is initiated at the micro- deformation after a creep test (g in

60 When the cell structure was studied more closely, very little damage was found, Figure 43. In certain areas, however, folds or the beginnings of cracks can be seen in the actual cell wall, Figure 44. This study shows examples of the incredible ability of the wood material to resume its natural and original shape after a very large deformation. In order to prevent this from taking place, the glass transition temperature of the wood polymers Figure 41. Cross-section of isostatically Figure 43. Isostatically compressed pine must be exceeded when the wood is compressed earlywood of pine. after moistening. Magnification 2100 x. deformed. The results can also be Magnification 245 x. related to the question of whether it is possible in today’s drying plants in sawmills to load the timber during drying so that the distorsion is reduced. The timber from the sawmills will in fact be exposed to countless moistenings and dryings before it reaches the final user.

Figure 42. Cross-section of isostatically Figure 44. Folds or indications of cracks in compressed earlywood of pine after isostatically compressed pine. Magnification moistening. Magnification 900 x. 2000 x.

61 (Analysis of decayed windows.), Kühne H. 1955: Über den Einfluß References Swedish Institute for Wood von Wassergehalt, Raumgewicht, Archer, R. R. 1986: Growth stresses Technology Research, Report No. Faserstellung und Jahrringstellung and strains in trees. Springer-Ver- 23, (in Swedish) auf die Festigkeit und lag, Berlin Grönlund, A.; Hägglund, A. 1987: Verformbarkeit schweizerischen Armstrong, L. D.; Kingston, R. S. Kärnved till fönster. Hög Fichten-, Tannen-, Lärchen- T. 1960: Effect of moisture kärnvedsandel genom ändamålsan- Rotbuchen- und Eichenholzes. changes on creep in wood. Nature passad sågning. (Heartwood for Eidgenössische Materilaprüfungs- 185: 862 windows. Sawing for high und Versuchanstalt für Industrie, Balodis, V. 1972: Influence of heartwood yield.) Swedish Insti- Bauwesen und Gewerbe, Zürich, (in grain angle on twist in seasoned tute for Wood Technology Re- German) boards. Wood Science 5(1): 44-50 search, Report No. P 8705037´, (in Kuukkanen, R., Nordqvist, M. Boutelje, J.; Nilsson, T. 1985: Swedish) 1992: Den mekanosorptiva effek- Inverkan av avverkningstid och - Johnston, D. D. 1957: Effect of tens beroende av belastnings- plats, våtlagring, träslag och densi- mechanical restraint during drying riktningen. (The influence of load tet på virkets beständighet. Del I. on the subsequent distorsion of direction on the mechano-sorptive Beständighet mot röta. (The effects timber. Timber Technology 65: effect.), KTH, Wood Technology of felling time and felling area, wet 453-455 and Processing, Report TRITA- storage, tree species and wood Kifetew and Sandberg 1998: Mate- TRT-1992-52, (in Swedish) density on the durability of wood. rial damage due to electron beam Perem, E. 1949: Försågnings- Part I: Durability against fungal during creep test in Environmental sättets inverkan på avnötningen dacay.), Swedish Institute for Scanning Electron Microscopy hos golv. (The influence of sawing Wood Technology Research, Rep- (ESEM). Submitted to Wood and pattern on the resistance of ort No. 73 Fibre Science wooden floorings.), Föreningen CEN 1995: European Standard Kingstone R: S: T: 1962: Creep, Svenska Sågverksmän FSS, 7:e EN 844-3: Round and sawn timber relaxation and failure of wood. årsmötet 11-12 mars 1949, Stock- - Terminology - Part 3: General Research Applied in Industri 15: holm, (in Swedish) terms relating to sawn timber. Ref. 164-170 Sandberg, D. 1995a: Årsrings- No. EN 844-3:1995 König, E. 1957: Fehler des orienteringens betydelse för sprick- Durst, J. 1955: Taschenbuch der Holzes. Holz-Zentralblatt verlags- bildning och deformation hos Fehler und Schäden des Holzes. GMBH, Stuttgart vattenbegjutet virke av furu (Pinus Fachbuchverlag Leipzig Manual in Forest Technology 1922: silvestris L) och gran (Picea abies Forsberg, D. 1997: Shape stabil- Handbok i skogsteknologi, (in Karst). (Influence of annual ring ity of sawn wood of Norway Swedish) orientation on crack formation and spruce in relation to site para- Holmberg, H.; Wålinder, M. deformation in water soaked pine meters, wood characteristics and 1997: Comparsion of resistance to (Pinus silvestris L) and spruce market requirements. The Swedish permanent indentation on radial (Picea abies Karst) timber.), KTH, University of Agricultural Science, and tangential sections in pine Wood Technology and Processing, Dept of Forest-Industry-Market (Pinus silvestris L.) using Brinell Report TRITA-TRÄ R-95-16, (in Studies, Report No. 45 hardness test. KTH, Wood Swedish) Grönlund, A.; Rydell, R. 1983: Technology and Processing, Rep- Sandberg, D. 1995b: Stående Analys av rötskadade fönster. ort TRITA-TRÄ R-97-28 årsringar hos furu (Pinus silvestris L)

62 och gran (Picea abies Karst). Sandberg, D. 1998d: Årsrings- Biological, chemical, and physical (Vertical annual rings in pine orienteringens inverkan på properties of Scots pine from (Pinus silvestris L) and spruce deformationer hos små felfria pro- Sweden under special conside- (Picea abies Karst). KTH, Wood ver av furu under böjbelastning ration of crown defoliation. The Technology and Processing, Rep- och samtidig fuktvariation. (The Swedish University of Agricultural ort TRITA-TRÄ R-95-13, (in Swe- influence of annual ring orientation Science, Dept. For. Prod. Report dish) on deformation of clear specimens No. 217 Sandberg, D. 1996d: The of Scots pine in bending and Stevens, W. C., Johnston, D. D. influence of pith and juvenile wood simultaneous variation in moisture 1960: Distorsion caused by spiral- on proportion of cracks in sawn content.), KTH, Wood Techno- led grain. Timber Technology timber when kiln dried and logy and Processing, Report 68:217-218 exposed to wetting cycles. Holz als TRITA-TRÄ R-97-32, (in Swedish) Thunell, B.; Perem, E. 1948: Un- Roh- und Werkstoff 54(3):152 Sandberg, D. 1998e: Inverkan av dersökning av avnötning hos Sandberg, D. 1997a: The isostatisk komprimering på cell- olika trägolv. (Investigation of the influence of annual ring orientation strukturen. (Influnce of isostatic resistance of wooden floorings.) on crack formation and deforma- compression on the cell Svenska Träforskningsinstitutet tion in water soaked pine (Pinus structure.), KTH, Wood Techno- STFI, Trätekniska avdelningen, silvestris L) and spruce (Picea abies logy and Processing, Report meddelande 19, (in Swedish) Karst) timber. Holz als Roh- und TRITA-TRÄ R-97-35, (in Swedish) TräteknikCentrum 1987: Fönster - Werkstoff 55(3):175-182 Sandberg, D.; Holmberg, H. fönster och utveckling inom Sandberg, D. 1997d: 1998b: Deformationer och TräteknikCentrum. (Windows - Nedbrytningsmekanismer hos trä. sprickförekomst hos virke före och Windovs and development at the 2. Utomhusexponering av radiella efter torkning. - Inverkan av Swedish Institute for Wood och tangentiella ytor av furu och virkets läge i stocken. (Distorsion Technology Research.), Swedish gran. (Degradation mechanism of and cracks in timber after sawing, Institute for Wood Technology wood. 2. Outdoor exposed radial drying and cycles of wetting and Research, Report No. P 8703023, and tangential surfaces of Scots drying. – Influense of location in (in Swedish) pine and Norway spruce.), KTH, the log.), KTH, Wood Technology Öqvist, H. 1988: Utomhusvirkets Wood Technology and Processing, and Processing, Report TRITA- beständighet. Fältförsök: Ovan- Report TRITA-TRÄ R-97-24, (in TRÄ R-98-34, (in Swedish) jordexponering av träpaneler. (The Swedish) Sandberg, D.; Johansson, I. 1995: resistance of timber in outdoor use Sandberg, D. 1998a: Weathering Kvarstående deformation hos furu above ground exposed wood of radial and tangential wood efter böjbelastning under samtidig panels.), The Swedish University surfaces of pine and spruce. 1. fuktvariation. (Non-recoverable of Agricultural Science, Dept. For. Crack formation and colour deformation in Scotch pine after Prod. Report No. 204, (in Swedish) changes in exposed surfaces. In bending and simultaneous variation press Holzforschung in moisture content.), KTH, Wood Sandberg, D. 1998b: Weathering Technology and Processing, Rep- of radial and tangential wood ort TRITA-TRÄ R-95-17, (in Swe- surfaces of pine and spruce. 2. dish) Microscope studies. In press Steffen, A.; Frühwald, A.; Holzforschung Tamminen, Z.; Puls, J. 1990:

Chapter5 Conclusions

5. Conclusions

Research at KTH-Trä has since the research to the wood-working deformation and crack behaviour of end of 1990 been focused on an industry i.e. a practical aspect of the wood. The results show that star- integrated R&D-program called R&D-program Value Activation at sawing gives timber that have an Value Activation. KTH-Trä. accuracy in size and geometry, is free Value Activation strives The relation between ideas, from cracks and has a controlled towards a strong integration of scientific basis for ideas and movement with changing humidity. the three main tasks of a development results, and the The Value Activation program university, viz. basic education, development of production has so far shown that there are research and graduate studies, technology and systems has been great possibilities of utilizing the and the dissemination of important for the industrial properties of wood in a better way technology. The research work implementation of the R&D- than our conventional wood has been carried out as a pro- results, and to a high degree also production concept can achieve. blem- or idea-based process, and for the progress of the work New wood products with desirable the basic education has become presented in this thesis. I have also properties can be developed. Most progressively more integrated with related the dissemination of of these products are expected to the research and oriented towards knowledge and technology transfer give a greater added value to the problem- and idea-based learning. from all levels to the industrial wood. The R&D-program will A new sawing pattern, star- implementation. continue with the further sawing, passes through both In the thesis some of the mate- development of improved products education and research as a rial properties which are very and also with the development of continuous thread. important for timber in general and cost-efficient production systems One of the aims of this thesis is for star-sawn timber in particular for implementing the ideas from to elucidate the process relating to have been investigated. The our R&D in profitable production the implementation of the important material properties here units. radically new star-sawing referred to are primarily those production system from university properties which are related to the

Radially sawn timber Chapter Star-sawing – a new method for producing timber with vertical annual rings I D. Sandberg

anatomical primary directions in will increase, as will the changes in Summary the wood, the radial and tangential shape. To prevent a greater degree This work presents a new directions. The distortions and of annual ring inclination from method of sawing to produce angular changes in the cross- causing any perceptible deterio- pith-free timber with vertical section become more evident the ration from the timber properties annual rings and without greater the difference between the associated with vertical annual juvenile wood. The method, radial and tangential moisture rings, the angle between the flat which is called star-sawing, gives movements (Schwab 1978). side of the timber and the annual sawn timber with two different It has been known for a long rings shall lie between 60 and 90 shapes, viz, conventional timber time that if timber is sawn radially degrees. This applies to pine (Pinus with a rectangular cross-section from the log, no or very small silvestris L) and spruce (Picea abies and timber with a triangular changes in shape occur. The Karst) (Sandberg 1995a). cross-section. The volume yield timber then has vertical annual There are several methods of for star-sawing has been rings. The dimensional changes producing timber with vertical determined with the help of associated with moisture changes annual rings. The methods which computerized simulations and are still present, of course, but the give timber with a rectangular test sawings. The results show radial moisture movement occurs cross-section adopt in most cases that star-sawing is best suited for only along the width of the timber the traditional quarter-sawing logs with a top diameter greater and the tangential changes only in pattern, as in Figure 1a. With other than 200 millimetres. Both a its thickness direction. The sides of methods which give timber with high volume yield and suitable the timber are thus parallel to the vertical annual rings, sectors are dimensions of the sawn timber main directions of the wood and sawn with a more or less triangular were then obtained. The volume the movement as a result of any shape, Figure 1b. A compilation of yield of timber with vertical moisture change can be predicted. some of these methods has been annual rings in star-sawing is If the annual rings are not comple- made by Polaczek (1990). about 0.70 calculated with regard tely vertical, the moisture expan- It is well known that the juvenile to the top cylinder volume under sion across the width of the timber wood, the annual rings closest to bark. 1. Introduction The shrinkage anisotropy of the wood means that timber exposed to moisture changes will in most cases change both in its dimension and in its shape, i.e. deviations will occur from the original, usually right-angled parallelepipedal shape. The changes in shape appear as distortions and angular changes in a b the cross-section and are due mainly to the fact that the sides of Figure 1. Example of two sawing patterns which give timber with vertical the timber are not parallel to the annual rings. a) quarter sawing, b) sector sawing.

71 the pith, differs considerably from with today’s industrial technology the wood located in the periphery are dormant (Wiklund 1993). of the trunk. In the case of spruce (Picea abies Karst) it has been 2. Star-sawing found for instance that the fibres It seems obvious that if the of the juvenile wood are shorter timber is sawn radially from the and have a lower stiffness and log, without pith and juvenile density than other wood (Boutelje wood, properties are obtained 1968). It has also been found that which are much better than those the fibril angle is greater in the of conventionally sawn timber. juvenile wood than in the mature Star-sawing is a new method for wood, which means that timber producing such timber. The yield which contains juvenile wood will from star-sawing consists of timber be deformed during drying (Dan- sections with both rectangular and borg 1990). It has also been shown triangular cross-sections, all with that if timber contains pith or if Figure 2. Star-sawing, new sawing pattern vertical annual rings. Figure 2 to produce timber on an industrial scale any part of the timber has been shows the star-sawing pattern. with vertical annual rings. sawn close to the pith, the timber has a greater tendency to crack (Sandberg 1995b). These properties of the juvenile wood mean that it should be removed already at the time of sawing in order to obtain straight timber free from cracks. At the Royal Institute of Technology, KTH, Division of Wood Technology and Processing, a new sawing method to produce timber with vertical annual rings has been developed during recent years. The method is called ”star- sawing” and is part of an extensive R&D-program, ”Value Activa- tion”, which is aimed at producing timber and materials with vertical annual rings on an industrial scale and with good profitability for products in which e.g. small and controlled moisture movements are valued highly. The object is thus to activate values in timber which

72 2.1 The adaptation of the wood, for use in products where the tively is to place two parallel saw- sawing pattern to the properties of the reaction wood are kerfs on each side of the pith at a log geometry particularly valuable. distance which corresponds to the The star-sawing pattern is well Removing the pith and the thickness of the rectangular adapted to the circular cross- juvenile wood is one of the most timber. If the log is crooked, the section of the log. The sawing critical steps in star-sawing. The log is rotated until the plane in pattern in its basic design gives six path of the pith in the length which the crookedness lies is pieces with a triangular cross- direction of the log is often very parallel to the sawkerfs. The section and six pieces with a irregular. Top fractures and other timber section containing the pith rectangular cross-section but, types of damage which frequently is then sawn with a pith-catcher depending on the desired timber arise during the growth of the tree which can be adapted to the exten- thickness and the dimensions of often cause a fairly large displace- sion of the pith and the juvenile the triangular profiles, several ment of the location of the pith. If wood. more rectangular pieces can be the log is crooked, the pith will sawn to maintain a high volume follow this crookedness, at the 2.2 Dimensions of the yield, Figure 3a. For non-circular same time as the pith becomes sawn timber logs, it is also possible to adapt the displaced from the geometrical The appearance of the star-sawing sawing pattern to the shape of the centre of the log cross-section. pattern means that the width of the log, Figure 3b. Non-circular logs This means that high demands timber becomes considerably smal- usually contain reaction wood in must be made on the position of ler than in e.g. square sawing, for parts of the cross-section. In star- the log so that the pith and the the same log diameter. This means sawing, the log can be positioned surrounding juvenile wood can be that star-sawing is best suited for during insertion so that the reject ”captured” and separated from the coarse logs where the top diameter as a consequence of the reaction other wood. is greater than 200 millimetres. The wood is minimized. It is also A method which has proven dimensions of the timber with a possible to saw timber pieces with useful to remove the pith and the rectangular cross-section have a large proportion of reaction surrounding juvenile wood effec- been chosen so that they agree with the standardized dimensions used in e.g. Sweden (SIS 1970). In the case of triangular profiles, there are no standardized dimensions and in the test sawings carried out, triangular profiles with sides of 80, 100 or 120 millimetres have been obtained. Besides the timber from the actual star-sawing, sideboards are also obtained from b a the slabs. These boards have horizontal annual rings, however, but Figure 3. Variations of star-sawing to achieve optimum utilisation of the log volume. with a uniform annual ring a) Sawing pattern with twice the number of timber pieces with rectangular cross- sections. orientation and they are often b) Sawing pattern adapted to the non-circular log cross-section. knot-free and of very good quality.

73 a e bcd

Figure 4. Sawing procedure in the pilot plant for star-sawing. The pictures c) The six-sided block is loosened from the pegs and lowered onto the sawing bed. A show log and timber cross-sections. pith plank and two blocks, the so-called coffin lids, are sawn. a) The log is positioned and fastened between two pegs so that it can be d) The ”coffin lids” are turned and fastened, after which two timber sections are rotated. sawn from each block. b) The suspended log is turned and sawn into a six-sided block. e) The remaining blocks with rhombic cross-section are placed in a fixture and sawn.

2.3 Pilot plant for star- Figure 4b. The log is loosened from 2.4 Drying of star-sawn sawing the pegs and lowered onto the timber In order to develop a system for sawing bed. A pith plank is sawn The timber sawn in the pilot plant the star-sawing technology and to after which two more blocks are has been dried in a chamber drier test the properties of the sawn obtained, the so-called coffin lids, at a sawmill. The kiln schedules timber and of the products in Figure 4c. The coffin lids are turned used are the same as those used to which the timber is used, a pilot and fastened together with hydrau- dry conventional timber at the plant for star-sawing has been lic grab devices and two timber sawmill. In the drying, the built. The sawing is carried out sections are sawn from each block, triangular profiles have been mainly in a horizontal bandsaw of Figure 4d. After this sawing step, stacked according to two different the ”Forestor” type which has been two blocks with rhombic cross- methods, completely block stacked rebuilt to cope with star-sawing. sections are obtained. These are and in groups of three, Figure 5. In Figure 4 shows the sawing turned and placed in a fixture the stacking of the triangular profi- procedure used in the plant. This before the sawing continues, les, the tangential surface, from method of star-sawing is neither Figure 4e. The rhombs stand freely which evaporation is considerably the most optimum nor the only in the fixture, i.e. without being more rapid than from the radial method, but it was shown in this fastened. surface, is always placed against first phase of development to be very suitable. The log is lifted into the band saw and is centered with the help of underlying lifting arms and is then fastened at the ends between two pegs, Figure 4a. The outer part of the log is sawn away ab after which the log is rotated in steps of 60 degrees and sawn, so Figure 5. Board stacking methods tested during the drying of triangular profiles. a) that a six-sided block is obtained, Tightly packed profiles. b) Triangular profiles placed in groups of three.

74 some other surface to ensure some not necessarily be maximized. From fifty ”real” trees of pine, the degree of deceleration of the an economic viewpoint and for internal and external geometries of moisture evaporation. Preliminary environmental reasons, however, it which have been measured and tests show that timber with vertical is very important that as large a digitalized. With these data as a annual rings and without pith and part of the log as possible is used. base, the trunks have then been juvenile wood can be dried with a The yield values reported here refer reconstructed in a computer and a conventional drying program with only to volume yields for timber trunk database has been establis- a good result and without cracks. with vertical annual rings. hed. In addition, the system This is also true of the triangular consists of software for simulated profiles. The greatest difference in 3.1 Simulation of the sawing and calculation of the drying between the two stacking volume yield quality and value of the stems. For methods for the triangular profiles Star-sawing has been simulated in a a deeper study of the OPTSAW- has been found to be that the free computer with the help of the so- system, Drake and Johansson profile corners of the grouped called OPTSAW-system. This sys- (1987) and Liljeblad et al (1988) are profiles become too severely dried tem is a computer-based simulation referred to. The OPTSAW-system (Sandberg and Holmberg 1995). aid, developed at the Swedish Insti- can simulate the sawing methods tute for Wood Technology Re- square sawing and through-and- 3 Volume yield in star- search, with which it is possible to through sawing. In order to sawing analyse quantitative and economic simulate star-sawing, new software The basic idea of star-sawing is to relationships between the quality of has been developed (Nordfors and obtain products with the highest logs and the quality of sawn Sandberg 1993). possible value from a given raw timber. The OPTSAW-system material. This means that the consists of three main components: 3.1.1 Input data for the volume yield in the sawing need A log bank consisting of more than simulation The simulated sawing differs from the real sawing at the pilot plant in Table 1. Sawing patterns used in the simulation. that no six-sided block is sawn. Top diameter Sawing Thickness of the timber with rectangular cross-section Instead, all sawn timber are edged pattern (mm) (no) (mm) after they have been sawn free from the log. This gives a greater 121-140 1-5 12 16 19 22 25 flexibility in the choice of 141-160 6-9 16 19 22 25 dimensions for an individual piece 161-180 10-13 16 19 22 25 of timber. The width and thickness 181-200 14-18 16 19 22 25 32 of the rectangular timber have in 201-220 19-24 16 19 22 25 32 38 the choice of sawing patterns followed standardized dimensions. 221-240 25-29 19 22 25 32 38 All timber sections have been sawn 241-260 30-35 19 22 25 32 38 50 without wane and no volume 261-280 36-40 22 25 32 38 50 optimization has been carried out 281-300 41-46 22 25 32 38 50 63 with regard to the length/width 301-320 47-51 25 32 38 50 63 ratio during the edgeing. As in the pilot plant, a timber section

75 containing the pith has been sawn, not affect the calculation of volume means that only realistic timber after which a so-called pith-catcher yield for the timber with vertical dimensions will be sawn in the has been used to remove the pith annual rings. simulation. Table 1 shows alter- and the juvenile wood. In the simu- The logs from the stem bank native simulated sawing patterns. lation, the width of the pith- used in the simulation have a top Only the log diameter and timber catcher has in all cases been 15 diameter of between 140 and 305 thickness determine the sawing millimetres and the width of the millimetres and an average length pattern. The dimensions of the sawkerf 4 millimetres. The of 4.7 metres (standard deviation triangular profiles will be deter- shrinkage during drying of the 0.8 m). The logs have been divided mined in the subsequent edgeing rectangular section has been set to into diameter classes with an operation. It is possible either to 4 per cent in the tangential interval of 20 millimetres and for saw the same dimension on all direction and 2 per cent in the each class a few different sawing triangular profiles or to determine radial direction, and the volume pattern alternatives have been individually the dimensions of each shrinkage of the triangular profiles determined, with the constraint of individual profile. In the simula- has been set to 4 per cent. The a condition for the relationship tion, we have chosen to saw sharp- longitudinal shrinkage has been between the cross-sectional areas edged triangular profiles with the neglected. This shrinkage corre- of the triangular profiles and of the side length in a 5 millimetres sponds to a final moisture content rectangular timber, for each diame- module. Only sawing patterns in of about 15 per cent after drying. ter class. The area ratio, i.e. the which six rectangular and six In the simulation, no consideration ratio of the cross-sectional area of triangular timber sections are sawn has been given to the yield of the rectangular timber to that of have been used in the simulation. chips, shavings and sideboards, the triangular profiles, was found which is very important for the to lie between 0.40 and 2.0 for all 3.1.2 Results value yield in the sawing, but does log diameters. The condition Figures 6 and 7 show the volume yield, calculated with regard to the top cylinder volume under bark, as a function of the top diameter. Figure 6 shows the total yield and the corresponding yield for the timber with rectangular cross- section for the sawing patterns Volume yield Volume yield which gave the highest total yield. The logs have been divided into diameter classes according to the top diameter in intervals of 10 Topdiameter (mm) Topdiameter (mm) millimetres and the average value Figure 6. Volume yield in simulation of Figure 7. Volume yield in simulation of has been calculated within the star-sawing, calculated with regard to the star-sawing, calculated with regard to the classes for the sawing patterns top cylinder volume and as a function of top cylinder volume and as a function of the top diameter. The diagram shows the top diameter. The diagram shows the which gave the highest total yield. average value for the sawing patterns average for the sawing patterns which Figure 7 shows the sawing patterns which gave the highest total yield within gave the highest yield of timber with which within each diameter class each diameter class. rectangular cross-section within each diameter class.

76 gave the highest yield of timber 3.2 Test sawings when the methods for insertion and with a rectangular cross-section. Since the start, ca 600 cubic metres sawing of the log were optimized to Table 2 shows examples of of timber have been sawn in the the full extent that was considered dimensions of the timber from star-sawing pilot plant. Contin- possible in the pilot plant. sawing patterns which gave the uously during the test sawings, logs highest total yield. Table 3 shows have been chosen for which the 3.2.1 Test material and timber dimensions in the corre- quality and volume yield have been choice of sawing sponding diameter classes for the determined. On the basis of the pattern sawing patterns which gave the results obtained, the position and 50 logs of pine have been randomly highest yield of timber with a sawing methods have been refined. chosen from a log catch of 250 rectangular cross-section. The volume yield calculations logs. The top diameter under bark reported here apply at a stage varied between 325 and 440

Table 2. Examples of dimensions and volume yields in simulated star-sawing. Sawing patterns which gave the highest total yield in each diameter class have been chosen.

Rectangular* Triangular profile Volume yield Top diameter Dimension Number Side length Number Total Rectangular* (mm) (mm) of pieces (mm) of pieces

150 25 x 63 2 25 1 0,70 0,42

25 x 50 2 35 2

25 x 38 1 50 1

25 x 32 1 55 1

60 1

200 32 x 75 2 50 2 0,73 0,40

32 x 63 3 60 1

32 x 50 1 65 1

75 1

80 1

250 50 x 100 2 45 2 0,74 0,48

50 x 75 1 70 2

50 x 63 3 90 2

300 38 x 125 2 85 2 0,68 0,33

38 x 115 1 90 1

38 x 100 2 110 2

38 x 75 1 130 1

* Refers to the timber with rectangular cross-section.

77 millimetres. The average length to a moisture content of 15 per yields for four different sawing was 4.5 metres (standard deviation cent. patterns. 0.4 metres). All logs were made up of butt logs of good quality, and 3.2.2 Results 3.3 Discussion crookedness and eccentricity only Figure 8 shows the volume yield as The simulations of star-sawings occurred in the logs to a very a function of the top diameter for and the tests sawings show that the limited extent. The width of the all test logs. The yield is calculated yield in star-sawing is considerably pith catcher has for all sawing with regard to the top cylinder higher than for traditional sawing patterns been at least 50 milli- volume under bark i.e. in the same methods, e.g. through-and-through metres. The width of the sawkerf way as in the simulated yield calcu- sawing, square sawing and quarter- was 3 millimetres and, after the lations. Table 4 shows examples of sawing. The simulation results subdividision, the timber was dried sawn dimensions and corresponding show that the yield with respect to

Table 3. Examples of dimensions and volume yields in simulated star-sawing. Sawing patterns which gave the highest total yield in of timber with rectangular cross-section in each diameter class have been chosen.

Rectangular * Triangular profile Volume yield

Top diameter Dimension Number Side length Number Total Rectan-gular * (mm) (mm) of pieces (mm) of pieces

150 25 x 63 2 25 1 0,70 0,42

25 x 50 2 35 2

25 x 38 1 50 1

25 x 32 1 55 1

60 1

200 38 x 75 2 40 2 0,69 0,43

38 x 63 4 55 1

60 1

70 1

80 1

250 50 x 100 2 45 2 0,74 0,48

50 x 75 1 70 2

50 x 63 3 90 2

300 38 x 125 2 50 2 0,67 0,46

38 x 75 3 65 1

38 x 50 1 90 1

38 x 75 1 95 1

120 1

* Refers to the timber with rectangular cross-section.

78 the top cylinder volume varies The yield decreases strongly for the timber, especially in small between 0.50 and 0.75. small log diameters, especially dimensions, that they constitute a Approximately half the timber when the log diameter is less than great risk of failure already in an volume is made up of timber with a 200 millimetres. There are several unloaded state. This means that rectangular cross-section and the reasons for this, but the most from a production technical remaining part of timber with a important is that the standardized viewpoint and with consideration triangular profile. Boards are also module measurements for to the volume yield, star-sawing in obtained from the slabs, but this rectangular timber agree poorly the form stated in this article yield has not been included in this with the sawn dimensions. This should be carried out on large investigation. The total yield is not causes large waste in the edgeing logs, e.g. with a top diameter affected noticeably by whether the stage. Tables 2 and 3 show that the greater than 200 millimetres. The Figure 8. Volume yield in the star-sawing yield of timber with rectangular dimensions for both the lower limit for the log diameter trials, calculated with regard to the top cross-section is maximized or not, rectangular and triangular timber must of course be determined with cylinder volume and plotted as a function of the top diameter. as is evident if Figures 6 and 7 are are very small. From a handling consideration given to the products compared. On the other hand, the and allocation viewpoint, small for which the timber shall be used. scatter in the yield of rectangular timber dimensions are usually not On the basis of the simulation timber becomes much smaller preferable to coarse dimensions. results, it was considered suitable when the yield is maximized. The knots in radially sawn timber in the pilot plant for star-sawing to can also have such an extension in saw only logs with a top diameter

Table 4. Examples of dimensions and volume yields in the test sawing.

Rectangular* Triangular profile Volume yield Top diameter Dimension Number Side length Number Total Rectangular* (mm) (mm) of pieces (mm) of pieces

325 50 x 125 1 100 6 0,64 0,33

50 x 100 2

50 x 75 3

340 63 x 150 2 100 6 0,75 0,47

63 x 100 3

63 x 75 1

380 63 x 150 1 120 6 0,72 0,39

63 x 125 1

63 x 115 2

63 x 100 2

440 50 x 175 1 100 6 0,64 0,47

50 x 150 2

50 x 125 2

50 x 100 1

* Refers to the timber with rectangular cross-section.

79 greater than 300 millimetres. Figure 8 shows that the volume yield in the test sawings lies at the same level as in the simulation, i.e. about 0.70 for the total yield and 0.35 for the rectangular timber. Figures 6 and 7 show that a slightly higher total yield can be expected than that obtained in the test sawings. This is due to two essential differences in sawing manner between the simulation and the test sawing. Firstly, the width of the pith-catcher has been increased from 15 millimetres in the simulation to at least 50 millimetres in the test sawings. It was found in practice that it is very difficult to remove the pith and the juvenile wood with the help of a pith- catcher which is smaller than 50 Figure 9. Yield curves for star-sawing assessed with the help of the results obtained millimetres. Secondly, only three from test sawings and simulations of the star-sawing method. dimensions of triangular profile were sawn in the test sawing, 80 , On the basis of the simulated yield rectangular cross-section and 100 and 120 millimeter sides, and results and experience from the timber with a triangular cross- only one dimension per sawing sawings carried out in the pilot section. The sawing pattern is well pattern. In the simulation, the di- plant, a yield curve according to adapted to the more or less mension was determined Figure 9 can be expected circular shape of the log cross- individually for each triangular section and permits a large variety profile in the sawing pattern, which 4. Conclusions of different patterns depending on meant that the yield could be Star-sawing is a new sawing the shape of the log cross-section. increased slightly. In general, this method to produce timber on an This means that the volume yield means that the yield in the test industrial scale with vertical annual in star-sawing becomes high. The sawing was slightly less than that in rings and free from pith and yield of timber with vertical annual the simulation. The tendency for juvenile wood. This timber has rings for star-sawing is about 0.70 the total yield to decrease in Figure several advantages compared with and approximately half this yield is 8 is also a consequence of the fact conventionally sawn timber, among made up of triangular profiles. that only three sizes of triangular other things very small changes in Besides this, high-quality timber profile were sawn. An increase in shape as a result of moisture con- with horizontal annual rings is also the number of sawing pattern tent changes. The star-sawing obtained from the slabs. alternatives probably causes the method gives two shapes of the total yield to stabilize at about 0.70. sawn timber,viz. timber with a

80 References Sandberg, D. 1995a: Vertical Boutelje, J. 1968: Juvenile Wood, annual rings in pine (Pinus silvestris with particular Reference to L) and spruce (Picea abies Karst), Northern Spruce. Svensk Pappers- Royal Institue of Technology, tidning 71: 581 Wood Technology and Processing, Danborg, F. 1990: The effect of Report TRITA-TRÄ R-95-13 silvicultural practice on the amount Sandberg, D. 1995b: Influence of and quality of juvenile wood in annual ring orientation on crack Norway spruce. Paper presented at formation and deformation in the IUFRO-World Congress Mont- water soaked pine (Pinus silvestris real L) and spruce (Picea abies Karst) Drake, E.; Johansson, L. G. 1987: timber. Royal Institue of OPTSAW - simulation models for Technology, Wood Technology positioning and sawing in sawmills, and Processing, Report TRITA- TRÄTEK, Swedish Institute for TRÄ R-95-16 Wood Technology Research, Rep- Sandberg D; Holmberg H. 1995: ort I 8701001 Utvärdering av en ny ströläggnings- Liljeblad, Å.; Johansson, L.-G.; metod för stjärnsågad trekant- Drake, E. 1988: Quality simulation profil. Royal Institue of of saw logs - methods for Technology, Wood Technology reconstruction and disintegration and Processing, Report No. 31/95 of stems, TRÄTEK, Swedish Insti- Schwab, Von E. 1978: Das tute for Wood Technology Re- Stehvermögen von Holz. Holz- search, Report I 8812081 Zentralblatt, Stuttgart Nr.6 Polaczek, Von P. 1990: Stehende SIS 1970: Swedish Standard SIS Jahrringe und Stehvermögen. 23 27 11: Sawn timber, Maßnahmen gegen das Arbeiten Dimensions. Central body for des Holzes – Neue industrielle standardization in Sweden möglichkeiten bei Massivholz und Wiklund, M. 1993: Value Furnieren. Holz-Zentralblatt, Stutt- activation – new sawing pattern gart Nr. 150 give improved properties on Nordfors, L. Sandberg, D; 1993: wooden products. Paper at the Model for the Simulation of the 11th International Wood Volume yield when sawing Machining Seminar, Honne according to the new sawing pattern ”star-sawing”. Royal Institue of Technology, Wood Technology and Processing, Rep- ort TRITA-TRT-1993-53

81 82 Radially sawn timber. Knots – number, type and size Chapter in star-sawn triangular profiles of pine (Pinus silvestris L) and spruce (Picea abies Karst) II D. Sandberg, H. Holmberg

Summary 1. Introduction R&D-program ”Value Activation” The number of knots in timber which is aimed at producing timber This paper describes the and their location, shape and na- and boards with vertical annual appearance and number of knots ture are very important for e.g. the rings on an industrial scale and in star-sawn triangular profiles of appearance, strength properties with good profitability for products pine (Pinus silvestris L) and spruce and susceptibility to deformation of in which e.g. small and controlled (Picea abies Karst). The emphasis is the timber. The knots often play a moisture movements are highly placed on obtaining the volume decisive role for the price-setting of valued. It is thus desirable to yield of boards with a given knot the timber and for its aesthetic activate values in timber which appearance, together with the sizes value. Depending on the location with today’s industrial technology of the fibre disturbances around are dormant. The R&D-program the knots and how they effect the and the sawing method are yield. The test material has been described in greater detail in two taken from three regions in articles by Wiklund (1993) and Sweden and is made up of butt Sandberg (1996). logs of pine and of normal and The star-sawing pattern gives fast-growing spruce. timber with both rectangular and The results show that a large triangular cross-sections. All the proportion of the knots found in timber have vertical annual rings the triangular profiles from butt and are free from juvenile wood. logs are unacceptable in further The term ”vertical annual rings” refinement of the timber and must here means that the angle between be removed. Very few healthy the annual rings and the flat side of knots can be found. the timber is greater than 60° To produce knot-free boards, a Figure 1. Sawing pattern for star-sawing. (Sandberg 1995). In star-sawing, as volume reduction of on average 8 in all radial sawing methods, the and 21 percent is obtained for of the sawkerf in relation to the radial sawkerfs cleave the knots pine and spruce respectively, longitudinal axis of the knot, the longitudinally and a larger propor- calculated with respect to the origi- section through the knot can be tion of splay knots are obtained nal profile volume. If the fibre round, oval or more or less than in through-and through disturbance around the knots is triangular. The choice of sawing sawing and block-sawing. These also removed, a further volume pattern will thus strongly influence knots are often unacceptable from loss of 5 percentage for pine and 6 the geometrical appearance of the a strength viewpoint and must be percentage for spruce is obtained. knot in the sawn timber surface. removed. Star-sawing is especially The boards produced have a A new sawing method for suitable for large log dimensions. broad spectrum of lengths between producing radially sawn timber is The top diameter should be 2 and 500 centimetres for pine and under development at the Division greater than 200 millimetres to between 2 and 400 centimetres for of Wood Technology and Proces- ensure that the sawn timber do not spruce. sing at the Royal Institute of become too narrow (Sandberg Technology in Stockholm, KTH, 1996). A large proportion of the Figure 1. The method is called timber which assumes these star-sawing and is a part of the proportions is made up of butt logs

85 and the knots in these butt logs are catch of a local sawmill. The logs is metres and the thickness was the often of such a nature that for presumed to be representative of same as for the rectangular timber. aesthetic reasons they should be oversized logs in this region. The After the sawing, the timber was removed. It has been found among logs from northern Sweden is stacked and dried according to a other things that butt logs of pine made up of normally grown pine conventional drying schedule in a have a low proportion of healthy and spruce from two sawmills and drying chamber. Table 1 presents a knots in comparison with middle the same selection method has compilation of the extent of the and top logs (Nylinder et al 1995). been used for the logs from central test material. Sweden. All logs have been quality 2 Aim graded according to the regulations 3.3 Measurement method The purpose of this investigation is applied in Sweden by an author- For each triangular profile, the to determine the nature and ized log-grader before they were relative positions and extensions of number of knots in star-sawn sawn (Timber Measurement the knots were determined on the triangular profiles from butt logs of Council 1990). three sides of the profiles. In addition, the lengths of the profile pine and spruce. On the basis of 3.2 Sawing and drying these basic data, the yield is which must be removed to obtain, calculated for the production of All logs were sawn according to on the one hand, knot-free timber knot-free triangular boards. the star-sawing method. Sawing boards where only the knot is patterns were chosen so that the removed and, on the other hand, 3 Materials and methods sides of the triangular profiles were boards where the fibre disturbance 120 millimetres (dried measure- around the knot is also removed, 3.1 Test material ment) and the rectangular timber were determined. Each knot was had a thickness of 50 or 63 milli- classified according to shape and The test material has been taken metres (dried measurement). In type according to the assessment from three regions: southern, cen- each sawing pattern, a pith-catcher rules applied in the Nordic grading tral and northern Sweden. The was sawn away. The width of the rules for sawn timber goods logs is so-called oversized which in pith-catcher was at least 50 milli- (Timber Grading Committee 1976). this case means that the top diameter exceeds 35 centimeters and the Table 1. The extent of the test material and the lengths of the profiles. selection has been made so that Test group Total Number of Average Number of only timber of U/S and fifth length (m) profiles length (m) butt logs quality (V) has been included in pine central Sweden 221 46 4.8 8 the study material. The logs from southern Sweden was felled during pine northern Sweden 189 39 4.9 9 January 1994 and was made up of pine southern Sweden 109 24 4.5 3 pine and normal and fast-growing total pine 519 109 4.7 20 spruce. ”Fast-growing” here means spruce central Sweden 176 35 4.9 7 an average annual ring width spruce northern Sweden 129 24 5.0 5 greater than 10 millimetres. The spruce southern Sweden fast* 167 38 4.4 5 logs from central Sweden was made up of normally grown spruce southern Sweden 47 12 4.4 7 oversized pine and spruce which total spruce 519 109 4.7 24 was randomly taken from the logs *fast = fast-growing spruce with an average annual ring width greater than 10 mm 86 4. Results and discussion healthy knots in U/S. The maxi- the proportion of each type of knot mum number of knots in both is indicated for each test group with 4.1 Log quality classes is equal. Besides this, knots regard to the total number of knot with rot must not occur in U/S, surfaces. For each knot type, the The quality of the logs used in the whilst these knots can occur in proportions of the total number of investigation is shown in Table 2. fifths. The assessment of the logs knot surfaces having a certain na- For the logs from central and has been made with the assump- ture is also indicated. The sides of northern Sweden, the distribution tion that they will be block-sawn. the triangular profile are observed between U/S and fifths is very separately, which means that the similar, whilst the normally grown 4.2 Distribution of knot same knot can appear simultaneou- pine and spruce timber from types and their nature, sly on three sides of the profile and southern Sweden has an over- dimensions and thus be observed as three different representation of fifths. Compared accompanying fibre knots. Also indicated are the with the official statistics over the disturbance. average length of the removals distribution of logs grades, the logs Tables 4 and 5 show the distribu- required to remove only knots (L from northern and central Sweden tion between knot types, their knot), the average value of the has a higher proportion of U/S. nature and the extent of fibre dist- extension of the fibre disturbance Table 3 gives the quality urbance in the longitudinal direc- in the length direction of the distributions for pine and spruce tion of the profiles. For the tangen- timber on the root side (L root) logs according to the Swedish tial surface of the triangular profile and top side (L top) of the knots, Forestry Statistics for 1994 (Natio- (a) and its radial surfaces (b, c), and the removal required to obtain nal Board of Forestry 1994). The down-grading of the logs from the Table 3. Quality distribution of logs in Sweden according to official best quality to fifths is related statistics (National Board of Forestry 1994). mainly to the knot distribution in the logs. The difference between Region Pine Spruce U/S and fifths from a knot U/S V Others U/S V Others viewpoint is primarily that the northern Sweden 0.28 0.68 0.04 0.38 0.61 0.01 healthy knots can be 15-20 central Sweden 0.34 0.62 0.04 0.69 0.30 0.01

millimetres larger in fifths than the southern Sweden 0.25 0.67 0.08 0.69 0.27 0.04 Table 2. Quality of the logs in the investigation. Test group Quality of log 12345678 pine central Sweden U/S U/S V U/S V V V U/S pine northern Sweden V U/S V V U/S U/S U/S pine southern Sweden V V V V V spruce central Sweden U/S V U/S U/S U/S U/S V spruce northern Sweden U/S V U/S U/S V V U/S V spruce southern Sweden V V U/S V V V V spruce southern Sweden fast* U/S V U/S

*fast = fast-growing spruce with an average annual ring width greater than 10 mm

87 Table 4. Knot distribution in pine. Test group Proportion Number of knots with a certain nature Knot extension (mm)

barkringed knot healthy knot loose knot rotten knot dead knot Lknot L root L top L tot Knot type a b,c a b,c a b,c a b,c a b,c a b,c

central spike knot 0.15 0.04 0.01 0.09 37 15 18 70 Sweden knot cluster 0.03 0.06 0.01 0.02 0.04 0.01 208 15 11 234 splay knot 0.10 0.04 0.05 49 19 16 84 edge knot 0.02 0.24 0.01 0.08 0.04 0.01 0.01 0.05 0.05 35 15 24 74 oval knot 0.15 0.04 0.02 0.01 0.07 30 12 18 60 pearl knot 0.03 0.01 0.02 85821 round knot 0.16 0.05 0.05 0.01 0.11 0.03 18 11 16 45

northern spike knot 0.11 0.01 0.02 0.01 0.06 0.01 33 19 21 73 Sweden knot cluster 0.06 0.09 0.02 0.01 0.01 0.01 0.04 0.03 0.02 205 15 19 239 splay knot 0.06 0.01 0.01 0.02 0.01 64 24 23 111 edge knot 0.03 0.20 0.03 0.01 0.01 0.01 0.05 0.01 0.11 30 18 21 69 oval knot 0.03 0.21 0.06 0.01 0.01 0.11 0.03 0.03 26 18 18 62 pearl knot 0.03 0.01 0.02 0.01 0.01 6 14 19 39 round knot 0.14 0.01 0.03 0.08 0.02 20 9 13 42

southern spike knot 0.09 0.02 0.02 0.01 0.03 0.01 51 27 30 108 Sweden knot cluster 0.01 0.02 0.01 0.01 0.01 0.01 0.01 264 11 9 284 splay knot 0.15 0.05 0.02 0.01 0.06 0.02 70 42 30 142 edge knot 0.05 0.24 0.02 0.04 0.01 0.08 0.01 0.02 0.02 0.01 0.10 44 40 45 129 oval knot 0.05 0.14 0.03 0.08 0.04 0.01 0.02 0.02 36 27 37 100 pearl knot 0.03 0.02 0.01 9 12 13 34

knot-free timber boards without surfaces. The location of the cross- The explanation of why this is not fibre disturbances (L tot). For the cut is determined by the total ex- the case is obtained if the whole measurement of the lengths L tension of the knot in the profile extension of the knot is observed. knot, L root, L top and L tot, the volume, with or without taking the Closest to the pith, the knot is extension of the knots in the whole fibre disturbance into conside- often healthy and well grown into triangular volume has been taken ration. Table 6 shows volume the surrounding wood material. into account, i.e. consideration has yields for both knot-free material Further from the pith, the knot been given to the knot extension and material with healthy knots. often becomes dry, bark-drawing or on all sides of the profile. Where Tables 4 and 5 show that a large damaged by rot as a consequence of values are lacking in the table, the proportion of the knots in this the natural pruning. In the proportion is less than 0.005. timber are either loose, dry or triangular profiles, this corresponds contain rot. Few healthy knots can to healthy edge knots on the radial 4.3 Yield of boards with be found. Table 6 shows that there surfaces and a dry, loose or bark- healthy knots and is little difference in volume yield drawing round or oval knot on the knot-free boards. between knot-free material and tangential surface of the profile. When producing the triangular material with healthy knots. A The change from healthy to dry profiles to boards with a certain certain increase in the yield could knot is often clearly visible on splay knot appearance, consideration be expected, however, if the pro- knots which are usually healthy must be given to the shape and portion of healthy knots were con- closest to the pith but deteriorate nature of the knot on all profile sidered, as shown in Tables 4 and 5. towards the sap-wood side. Tables 4

88 and 5 only take into consideration only a small difference in volume tion without fibre disturbance. the nature of the knots on one side yield between knot-free boards and Figures 2 and 3 show that pine has of the profile, independently of the boards with healthy knots. a larger volume proportion of appearance on the other sides. This When the knots are removed, a knot-free boards over a given means that we overestimate wide spectrum of boards of diffe- length of the corresponding boards healthy knots in Tables 4 and 5 rent lengths is obtained. Figures 2 than spruce. As an example, the when we wish to produce boards and 3 show the cumulative volume proportion by volume (with regard with healthy knots on all sides. distribution for the knot-free to the original triangular profile The yields produced in Table 6 lengths. Figure 2 shows the distri- volume) of the knot-free pine ma- take into consideration the nature bution for knot-free material where terial without fibre disturbance of the knots on all sides in the the fibre disturbance around the from central Sweden and shorter determination of the length of the knot has not been removed. Figure than 25 centimetres is 0.10. The removed section, so that there is 3 shows the corresponding distribu- corresponding proportion for spruce

Table 5. Knot distribution in spruce. Test group Proportion Number of knots with a certain nature Knot extension (mm)

barkringed knot healthy knot loose knot rotten knot dead knot Lknot L root Ltop L tot Knot type a b,c a b,c a b,c a b,c a b,c a b,c central spike knot 0.08 0.02 0.01 0.01 0.03 0.01 26 11 14 51 Sweden knot cluster 0.11 0.16 0.05 0.05 0.01 0.01 0.05 0.04 0.01 0.05 235 10 14 259 splay knot 0.06 0.03 0.01 0.02 36 12 18 66 edge knot 0.04 0.13 0.01 0.02 0.01 0.01 0.01 0.01 0.01 0.07 25 11 16 52 oval knot 0.01 0.15 0.01 0.05 0.02 0.01 0.05 0.02 29 12 16 57 pearl knot 0.02 0.02 0.01 0.01 0.01 0.01 8 7 8 23 round knot 0.19 0.02 0.05 0.01 0.01 0.10 0.01 0.03 18 9 11 38 northern spike knot 0.03 0.01 0.02 41 26 29 96 Sweden knot cluster 0.15 0.20 0.03 0.03 0.01 0.04 0.07 0.06 0.08 262 13 18 293 splay knot 0.08 0.01 0.02 0.01 0.02 0.04 47 20 21 88 edge knot 0.05 0.13 0.01 0.01 0.02 0.02 0.03 0.10 31 22 24 77 oval knot 0.03 0.17 0.03 0.01 0.06 0.02 0.07 36 19 17 72 pearl knot 0.02 0.02 0.01 0.01 36 6 6 48 round knot 0.12 0.02 0.01 0.05 0.04 13 10 10 33 southern spike knot 0.14 0.03 0.01 0.02 0.07 18 11 18 47 Sweden knot cluster 0.12 0.18 0.03 0.04 0.06 0.01 0.01 0.04 0.03 0.04 0.04 286 17 22 325 fast* splay knot 0.02 0.01 0.01 0.01 0.01 40 18 20 78 edge knot 0.02 0.10 0.01 0.03 0.03 0.01 0.01 0.01 0.01 0.01 0.02 22 14 16 52 oval knot 0.02 0.12 0.02 0.05 0.03 0.01 0.01 0.01 0.01 24 15 23 62 pearl knot 0.03 0.03 7 7 13 27 round knot 0.22 0.02 0.09 0.01 0.01 0.01 0.02 0.01 0.08 0.02 19 9 13 41 southern spike knot 0.07 0.02 0.02 0.02 0.01 34 17 18 69 Sweden knot cluster 0.12 0.16 0.01 0.02 0.02 0.01 0.02 0.05 0.04 0.04 0.06 207 13 19 239 splay knot 0.12 0.01 0.03 0.01 0.01 0.06 60 25 29 114 edge knot 0.07 0.14 0.01 0.01 0.01 0.03 0.01 0.01 0.03 0.09 41 20 26 87 oval knot 0.01 0.09 0.01 0.01 0.03 0.02 0.01 0.03 35 18 20 73 pearl knot 0.02 0.01 0.01 0.02 13 9 11 33 round knot 0.18 0.01 0.02 0.01 0.02 0.05 0.08 35 17 20 72

*fast = fast-growing spruce with an average annual ring width greater than 10 mm

89 Table 6. Volume yields for boards of triangular profiles. that a small dimension goes

Test group Share of total profile volume through a branch whorl than it is for a larger dimension. This must be Knot-free boards Boards with heal- Knot-free boards Boards with healthy with fibre disturb- thy knots and without fibre knots without fibre investigated more closely, however. ance fibre disturbance disturbance disturbance

pine central Sweden 0.90 0.90 0.85 0.85 5. Conclusions pine northern Sweden 0.92 0.93 0.89 0.89 Triangular profiles produced with pine southern Sweden 0.94 0.95 0.89 0.89 the star-sawing method have vertical annual rings and are free spruce central Sweden 0.84 0.84 0.79 0.79 from juvenile wood. This means spruce northern Sweden 0.74 0.74 0.69 0.69 that the timber becomes straighter spruce southern Sweden fast* 0.74 0.74 0.69 0.69 and flatter, has considerably more spruce southern Sweden 0.83 0.83 0.77 0.78 uniform dimensions and has a average pine 0.92 0.92 0.87 0.88 more predictable movement as a average spruce 0.79 0.79 0.73 0.74 result of moisture variation. From a knot point of view, the triangular *fast = fast-growing spruce with an average annual ring width greater than 10 mm profile has the advantage compared with conventionally sawn timber that there can be no comp- from central Sweden is 0.20. millimetre side. The dimension of letely hidden knots in the profile Further, the proportion by volume the profile will probably influence which will appear at a later refine- of knot-free boards of pine with a both the nature and frequency of ment stage. A knot in a triangular length greater than 200 centimetres knots in the surfaces of the profi- profile will always be visible on one is 0.27, but only 0.02 for spruce. For les. The dimension investigated of the three sides of the profile. the same species, however, there is here is the largest which is now The two radial surfaces of the no clear difference in length distri- sawn and it is improbable that any profile can, however, contain splay bution for the knot-free boards considerably larger dimension will knots which are unacceptable from between the different regions. It be sawn. Compared with profiles a strength and aesthetic viewpoint. may seem natural to assume that with a side of 120 millimetres, smal- Judging from these results, it the proportion by volume of knot- ler profiles will show a higher pro- seems evident that triangular profi- free boards will increase with portion of loose and dry knots. les of pine and spruce from butt increasing growth rate of the tree This is a consequence of the fact logs constitute a good starting ma- because of an increasing distance that these profiles will be sawn terial for the manufacture of knot- between the branch whorl. For further from the pith, i.e. the free boards with very good prop- spruce, this does not seem to be smallest distance between the pith erties, which in a later refinement correct since the yield for fast- and the profile will increase if the step can either be joined into growing spruce from southern same log diameter is used for the greater lengths or be used directly Sweden is of the same magnitude as different sawing patterns. On the in manufacture. that for spruce from northern other hand, the consequence of a Sweden, Table 6. smaller profile dimension may be The investigation has been that the distance between the knots carried out with only one dimen- increases because it is less probable sion of triangular profile, a 120 90 Figure 2. Cumulative volume distribution Pine from central Sweden with fibre disturbance Pine from northern Sweden with fibre disturbance Pine from southern Sweden with fibre of knot-free boards with fibre disturbance disturbances from knots. 1,0 1,0 1,0 0,8 0,8 0,8 0,6 0,6 0,6 0,4 0,4 0,4 0,2 0,2 0,2

0,0 Cumulative volume proportion 0,0 0,0 Cumulative volume proportion Cumulative volume proportion 0 100 200 300 400 500 600 0 100 200 300 400 500 600 0 100 200 300 400 500 600

Knot-free board length (cm) Knot-free board length (cm) Knot-free board length (cm)

Spruce from central Sweden with fibre disturbance Spruce from northern Sweden with fibre Spruce from southern Sweden fast-growing Spruce from southern Sweden with fibre disturbance with fibre disturbance disturbance 1,0 0,8 1,0 1,0 1,0 0,8 0,8 0,8 0,6 0,6 0,6 0,6 0,4 0,4 0,4 0,4 0,2 0,2 0,2 0,2 0,0 0,0 0,0 0,0 Cumulative volume proportion Cumulative volume proportion Cumulative volume proportion 0 100 200 300 400 0 100 200 300 400 0 100 200 300 400 Cumulative volume proportion 0 100 200 300 400 Knot-free board length (cm) Knot-free board length (cm) Knot-free board length (cm) Knot-free board length (cm)

Pine from southern Sweden Pine from northern Sweden Pine from central Sweden Figure 3. Cumulative volume distribution without fibre disturbance without fibre disturbance without fibre disturbance of knot-free boards without fibre disturbances from knots.

1,0 1,0 1,0 0,8 0,8 0,8 0,6 0,6 0,6 0,4 0,4 0,4 0,2 0,2 0,2 Cumulative volume proportion

Cumulative volume proportion 0,0 Cumulative volume proportion 0,0 0,0 0 100 200 300 400 500 600 0 100 200 300 400 500 600 0 100 200 300 400 500 600 Knot-free board length (cm) Knot-free board length (cm) Knot-free board length (cm)

Spruce from southern Sweden Fast-growing spruce from southern Sweden Spruce from northern Sweden Spruce from central Sweden without fibre disturbance without fibre disturbance without fibre disturbance without fibre disturbance

1,0 1,0 1,0 1,0 0,8 0,8 0,8 0,8 0,6 0,6 0,6 0,6 0,4 0,4 0,4 0,4 0,2 0,2 0,2 0,2 0,0 0,0 0,0 0,0 Cumulative volume proportion Cumulative volume proportion Cumulative volume proportion Cumulative volume proportion 0 100 200 300 400 0 100 200 300 400 0 100 200 300 400 0 100 200 300 400 Knot-free board length (cm) Knot-free board length (cm) Knot-free board length (cm) Knot-free board length (cm)

91 References. Holmberg, H.; Sandberg, D. Sandberg, D. 1996: Radially 1995: Frequency and Character of sawn timber. Star-sawing - a new Knots in Triangular Profiles of Pine method for producing timber with (Pinus silvestris L) and Spruce vertical annual rings. Holz als Roh- (Picea abies Karst). Royal Institue und Werkstoff 54(3): 145-151 of Technology, Wood Technology Timber Grading Committee 1976: and Processing, Report TRITA- Guiding Principles for Grading of TRÄ R-95-12 (in Swedish) Swedish Sawn Timber, Second Nylinder, M.; Grace L. A.; Jons- edition, The Swedish Timber and son L. 1995: Identifying pine logs Wood Pulp Journal yielding lumber with sound knots. Timber Measurement Council Swedish University of Agricultural 1990: Instructions for Wood Science, Department of Forest measurement, No. 1-87 Products, Report No. 245 Wiklund, M. 1993: Value National Board of Forestry 1994: Activation - new sawing pattern Statistical Yearbook of Forestry give improved properties on 1994, Published by the National wooden products. Paper at the Board of Forestry, Sweden, 11th International Wood Sandberg, D. 1995: Vertical Machining Seminar, Honne Annual Rings in Pine (Pinus silvestris L) and Spruce (Picea abies Karst), Royal Institue of Technology, Division of Wood Technology and Processing, Rep- ort TRITA-TRÄ R-95-13

92 Radially sawn timber. The influence of annual ring orientation on crack Chapter formation and deformation in water soaked pine (Pinus silvestris L) and spruce (Picea abies Karst) timber III D. Sandberg

Summary surrounding juvenile wood is of load and exposed to cyclic humidity vital significance for cracking. variations, the deformation This work describes deformation During moisture cycling, the increased dramatically compared to and crack formation in sawn timber of pine and spruce after amount of boards that develop when constant climatic conditions first drying and subsequent cycles cracks increased irrespective of were used. Similarly recovery after of resoaking in water and drying. their prior location in the cross unloading was accelerated by The influence of annual ring section of the stem. Timber sawn varying humidity. This phenom- orientation and the occurrence of from near the pith or distinctly enon is usually referred to as juvenile wood was determined. In containing pith has a higher mechanosorption. addition, the influence of compres- relative crack length compared to The influence of juvenile wood sion wood, annual ring orientation timber sawn away from and lacking on warp has been investigated by at the edges of the cross section, pith. In timber exposed to repeated several researchers, including Hall- the position of the board surface in cycles of wetting and drying the ock (1965), Balodis (1972), Dan- relation to the pith, and the con- crack length increases irrespective borg (1990) and Perstorper et al. dition of the board surface i.e. of its prior location in the stem. (1995). Their results show that whether wet or dry during resoak- occurrence of juvenile wood nor- ing were studied. 1. Introduction mally increases warp in softwood. When timber is exposed to Cracks and major distorsion can Timber with vertical annual repeated cycles of wetting and make timber unusable for some rings, i.e. radially sawn timber, drying, warp, viz spring, bow, twist applications. Various kinds of de- undergoes no, or very small and cup, increases and is greater formation in sawn timber, usually changes in shape because the sides after the first cycle. The influence referred to as warp, can be a of the timber are parallel to the of annual ring orientation on consequence of primary stresses in main directions of the wood. spring, bow and twist depends on the log, so-called growth stresses, A new method of sawing, Star- the type of deformation and on the or they may arise during proces- sawing, to produce industrial pith- kind of wood. Generally, the sing, especially when drying free timber with vertical annual results indicated that timber with timber. If these stresses, which are rings and without juvenile wood vertical and semi (half) vertical has been proposed (Sandberg annual rings show less deformation a consequence of shrinkage, 1996a). It has been shown that (mean values) than plain sawn become greater than the fracture timber produced with this method timber and timber containing pith. strength of the wood material, is less subjected to cracking, Cup is mainly caused by transverse cracks will develop in the timber. especially when the timber is exp- anisotropy and is strongly influen- The anisotropic, hygroscopic ced by the radius of the annual and non-homogeneous nature of osed to moisture variation, than ring. Therefore, timber with ver- wood creates feasible conditions timber produced with traditional tical annual rings do not show any for warp. The most troublesome sawing methods (Sandberg 1996b). cup. Spring, twist and, especially characteristic of wood in structural The objectives of this work was bow are strongly influenced by applications and furniture to study the influence of annual compression wood. Large amounts manufacture, is its hygroscopicity. ring orientation, pith and juvenile of compression wood in sawn During the late 1950’s Arm- wood on warp and to investigate timber increases such deformation. strong and Kingston (1960) macroscopic cracks induced in The distance between sawn observed that when small wood timber exposed to moisture timber in the log and the pith with samples are subjected to flexural changes.

95 2. Material and method Star-sawing pattern, (Figure 1). sections were cut from each board, The boards were kiln dried using a one at each end and one in the 2.1 Log selection and conventional drying program to a middle of the board. These speci- sawing moisture content of 18±3 %. All mens were used to determine the the boards were sawn with a cross density and the moisture content. Six butt logs of Scots pine (Pinus section dimension of 50 by 100 The specimens were free from silvestris L) and seven of Norway mm. knots and wood near knots, but spruce (Picea abies Karst) were compression wood may be present randomly selected from a local 2.2 Density and moisture in some specimens. The moisture sawmill in northern Sweden. The content determination content was defined as the ratio: of logs were oversized which in this the mass of water in the specimen case means that the top diameter When the test was finished, to the dry weight of the specimen. exceeded 35 centimeters. Only three two-centimeter-long cross timber of U/S (the best quality) 2.3 Wetting cycles and fifth quality (V) was included in the study. The down grading of The boards were exposed to the timber from U/S to quality (V) three cycles, lasting 20 days and is mainly due to knot distribution nights, of wetting and drying. in the log. The logs were presumed Before the first wetting, the weight to be representative for oversized of the boards was determined. In logs in this region. Before sawing, the wetting phase the boards the cross section of each log was floated freely for 30 minutes in a coloured in an area with a radius tray with water. Then, before of 30 millimeters around the pith. being weighed, the boards were This coloured area was later used placed in a frame for 15 minutes to detect the distance from the pith with the wet side up to let free water on the surface penetrate into for each board, and was chosen as Figure 1. The sawing pattern used in the a borderline between juvenile and investigation. Half the log was sawn the wood or evaporate. mature wood as far as slowly through-and-through and the other half Thereafter, the boards were dried using a star-sawing pattern. grown wood from northern Sweden is concerned. To obtain boards with different annual ring orientations and different distances from the pith, the logs were sawn according to both a ”through-and-through” sawing and b ”star-sawing”. Triangular profiles were not included in this investigation. To avoid one annual ring orientation being sawn from only one, or only a few logs, half c d of each log was sawn through-and- through and the other half in the Figure 2. Definition and measuring of warp: crook (a), twist (b), bow (c), and cup (d). 96 for 20 days and nights at a tempera- the proportion of compression ture of 21±1°C and 39±5 % RH. wood and how to find a fast and The same surface was soaked in reliable measure. One of the best each wetting phase. Before the test measures of compression wood is was started the boards were con- through the combination of volume ditioned for two weeks at the same proportion and location of climate as mentioned above. compression wood in the board. Such methods are however very 2.4 Warp measurement timeconsuming. It is more Warp was measured after every Figure 3. Examples of different annual convienient to select parts of the drying cycle. Cup was measured at ring orientations in the cross sections of board where the occurrence of five locations along the length of boards. compression wood may be the board: five cm from each end determined. In doing so, it is length of the board. The other of the board and at three important to select samples which three groups contained boards with equidistant locations between these are representative of the board. horizontal, semi-vertical and two end measuring points. Crook, When warp is investigated, it is vertical annual rings. Figure 3 bow and twist values were also important to avoid small local illustrates examples of timber with measured over the entire length of areas of compression wood, for different annual ring orientations. the board and the values were later example around knots, that may be Vertical annual rings exist in the reduced to equal lengths of three of minor importance for the global cross section when the angle bet- metres according to Morén (1987). deformation of the board. It ween a tangent to an annual ring at Figure 2 illustrates the principles should, however, be noted that half of the thickness of the board for measuring warp used in this shrinkage of knoted wood and by and the face side is between 60 and investigation. wood surrounding knots may to a 90 degrees (Sandberg 1995), For analysis, the boards were large extent influence the total (Figure 4). Similarly, boards with divided into four groups according warp of the board. In this horizontal annual rings exist when to the annual ring orientation in investigation, the specimens used the same angle was between 0 and the cross section. The first group to determine the density have also 30 degrees. When the angle was contains boards in which pith was been used to detect the presence between 30 and 60 degrees the enclosed in the cross section, in of compression wood. boards are said to have semi-ver- the whole or in a part of the total The boards were classified into tical annual rings. four classes, C0 - C3, with respect to the occurrence of compression 2.5 Determination of wood in three cross sections cut compression wood from each board. Class C0 has no Compression wood has a much compression wood in the cross greater longitudinal shrinkage than sections; class C1 has visible normal wood (Timell 1986). This compression wood in one of the will cause more warp in boards three cross sections; class C2 in containing a greater proportion of two and class C3 in all three cross compression wood. A common Figure 4. Definition of annual ring sections. orientation. complication is how to determine

97 As mentioned earlier, the speci- 3. Results and discussion 3.2 Moisture cycling mens were cut free of knots and Figure 5 shows the mean values local compression wood areas 3.1 Density of the moisture content variation in beneath knots were avoided as far Table 1 shows the mean values for the boards during the test. The as possible. dry density of the boards and the moisture content was determined for three cross sections after the 2.6 Crack measurement density separated into mean values for root, middle and top parts of last wetting cycle. It is assumed The total length of visible cracks the boards. The density values are that the mean value of the on each surface of the boards was quite normal for butt logs from moisture content for these three determined after the described northern Sweden. It should be cross sections was equal to the conditioning period and after the noted however that the density of mean value of the moisture content last wetting cycle. pine is the least at top end and of the board. The moisture con- increases towards the butt end of tent was then determined before 2.7 Statistical methods the log. In spruce this density and after each wetting phase. A common complication when difference is negligible. As shown in Figure 5, the analysing experimental results in Normally, compression wood moisture content during the wood is that the values do not has a higher density than normal wetting phases increases only a few have a normal distribution. Söder- wood (Timell 1986). When speci- percent. It should be noted that ström (1990) has shown that the mens of compression wood were the variation in moisture content in assumption of normal distribution excluded from the samples of nor- those parts of the boards which in when analysing lengths and areas mal wood, no difference in the contact with water was much of cracks is poor. The results of density values occurred. This fact higher. the present investigation have also suggests that the fraction of indicated that warp is not compression wood in the speci- normally distributed. For this rea- mens was relatively low. Therefore son, a non parametric test was the influence of compression wood used when the results for the test with its higher density was slight. groups were evaluated as mean values (Montgomery 1991). In this statistical method, normally distributed test parameters are not Figure 5. Variation of moisture content in the boards during the test. necessary. The tendency of a test group (boards with different annual ring orientations and distances from the pith) to show a Table 1. Density (Kg/m3) of boards in the investigation. Root, special property (warp or cracking) middle and top reflects the location of the specimens in the board. was tested using a homogenity test The mean value of these three measurements is also given. (Blom 1991). Unless otherwise mentioned, a confidence interval of Root Middle Top Mean 0.95 has been used in all the tests. Pine 476 412 389 426 Spruce 353 350 351 351

98 3.3 Influence of annual ring orientation and wetting on warp Figure 6 shows the measurements of warp such as crook, bow, twist and cup for the boards during three cycles of wetting and drying. According to the annual ring orientation in the cross section, the boards are divided into four groups. Group A has the pith enclosed in the cross section in all of, or in a part of, the length of the board. Groups B, C and D correspond to the horizontal, semi-vertical and vertical annual rings, respectively. Diagrams in Figure 6 have four bars for each group. The first bar represents the warp before the wetting cycle, while the second to fourth bars correspond to warp after one to three cycles of wetting and drying. 3.3.1 Crook In both pine and spruce except for spruce group A, crook increases for each cycle of wetting and drying, independent of the annual ring orientation. Crook for pine with vertical annual rings (group

Figure 6. The influence of annual ring orientation (A-D) on warp in boards of pine and spruce during three cycles of wetting and drying. The first bar reflects warp before wetting, the second bar warp after the first cycle etc. A - pith enclosed in the cross section, B - horizontal annual rings, C - semi-vertical annual rings, D - vertical annual rings.

99 D) seems to be lower than for the D) was significantly lower than face showed an increase of bow other groups, but this difference is those with horizontal annual rings during moisture cycling. On the not significant. Due to great and pith. Spruce did not show any other hand the boards with a non- scattering of the crook values, no significant differences between wetted face showed a decrease of significant differences were found groups A to D. When both pine bow. This means that the deforma- between groups A to D. The and spruce are considered, the tion measured as bow always relatively low crook for spruce in only difference was found in group increased on the wetted face. group A may reflect the high D, where spruce has significantly number of cracks in these boards more bow than pine. 3.3.3 Twist (see section 3.5). When the boards were resoaked In the same way as for crook In both pine and spruce, no in each cycle, the same face was in and bow, twist increases when the correlation was found between the contact with water. The influence boards are exposed to cycles of distance from the pith and the of resoaking on bow is illustrated wetting and drying. The increase in crook. This was expected, as no in Figure 7. Before resoaking, the twist was greatest for boards in consideration was given on how bow was of the same size for the group A, which may be considered the edges of the boards were wetted and the non-wetted face, natural as these boards contain oriented in relation to the pith. In for both pine and spruce. During pith. The level of twist was spruce, boards with crook the wetting cycles the behaviour expected to be greater for boards (measured as shown in Figure 2) of the boards was completely diffe- in group A compared to groups B- on the edge oriented towards the rent for pine and spruce. Pine D. The relative low twist in the pith demonstrated significantly demonstrates a slight increase of boards belonging to group A may more crook than boards where the bow, regardless if it occurs on the be a consequence of a higher crook was oriented in other wetted surface or not. There was number of cracks (see section 3.5) directions. This indicates that no significant difference in bow in these boards. Several research- juvenile wood has an influence on between the boards as far as the ers (e.g Mishiro and Booker 1988; warp. The greater crook for pith- wetted and the non-wetted face Perstorper et al. 1995) found that associated boards is consistent with was concerned. In spruce, the twist was more pronounced in several studies, for example Kloot boards with bow on the wetted boards taken near the pith than in and Page (1959) and Perstorper et al. (1995). 3.3.2 Bow Independent of the annual ring orientation of the boards, bow increased in both pine and spruce during moisture cycling. Before the first cycle of wetting and drying no difference in bow between the groups in pine was found. When boards were resoaked, bow for the Figure 7. Mean values for bow in pine (a) and spruce (b) when consideration has been boards with vertical and semi- taken of the wetted surface. Measure 0 was before wetting and measure 1 - 3 after vertical annual rings (groups C and the first to the third wetting cycle.

100 boards sawn distant from the pith. 3.3.4 Cup the width of the boards was 100 This study, however, shows no Cup was measured at five mm, the cup was small and only a significant differences between locations in each board. Here, the couple of mm. The change of cup twist in groups A-D or between greatest of these five measurements during moisture variations was pine and spruce. was used to measure cup. Since even smaller and was hard to detect with the measurement accuracy used. As shown in Figure 6, cup was highest for boards in groups A and B as a consequence of the anisotropic nature of the wood material and the curvature radius of the annual rings. Thus boards with horizontal annual rings sawn near the pith usually have more cup than boards with the same annual ring orientation sawn at the periphery of the log. Pine with vertical annual rings also showed cup. That may be explained by the annual ring orientation at the butt end of some of the boards which was not vertical due to buttress. When the maximal value were used for cup, the cup at the butt end will represent the board. 3.4 Influence of compression wood on warp In Figure 8 crook, bow and twist are shown for the four different levels (C0-C3) determined for compression wood. In pine the number of boards in levels C1-C3 was low (3-4 boards), which means that too far-reaching conclusions should not be drawn for a separate level. Figure 8. Influence of compression wood on warp according to four levels C0 - C3 of The effect of compression wood compression wood. on bow was more pronounced

101 than on crook and twist (see Figure A Boards with pith enclosed in the 3.5.1 Presence of cracks and 8). For both pine and spruce, cross section in all of, or in the relative crack length boards with compression wood part of the length. This group in the boards (levels C1-C3 as one test group) includes the same boards as in Figure 9 shows, the relative length have significantly more bow than group A, previously described of cracks before and after the boards without compression wood. in section 3.3. wetting cycles. The influence of A significant difference regarding E Boards sawn at a distance less than pith and juvenile wood on crack crook and twist was found in 30 mm between the pith and any length in the boards is obvious. spruce, where the amount of crook part of the board. In this group Boards sawn at a distance greater and twist in level C3 were higher the annual ring orientation varied than 30 mm from the pith (groups than in the other groups. from 0 to 90 degrees. Boards in F and G) have significantly lower group A are not included in this crack lengths than boards sawn 3.5 Cracks group. nearer the pith. This applies to The results from crack measuring F Boards sawn at a distance greater both pine and spruce. are presented as the relative crack than 30 mm between the pith and Between groups A and E, and length, and the ratio between the any part of the board. The annual groups F and G, there is no total length of the cracks and the ring orientation in the boards from significant difference regarding length of the board for both pine this group was between 0 and 60 crack length. Between pine and and spruce. degrees. spruce, differences regarding crack The boards are divided into G Boards sawn at a distance greater length was significant only for four groups with regard to annual than 30 mm between the pith and group F. ring orientation and the distance any part of the board. The boards As shown in Figure 10, the pro- between the pith and boards. The in this group have vertical annual portion of boards without cracks four groups are: rings, i.e. 60 to 90 degrees. relative to the total number of

Figure 9. The relative length of cracks before and after the wetting cycles. The boards are divided into four groups (A, E, F, G) according to annual ring orientation and the distance between the pith and the board.

102 boards in each group is higher for 3.5.2 The influence of pith This applies both before and after spruce than for pine. This was on crack formation the wetting cycles. Homogeneity probably due to the higher density The previous analysis of cracks was tests show that pith surfaces tend of pine. based on a board as a unit. Here to crack to a larger extent than Pine has a greater tendency to each surface will be examined other surfaces. It should be noted form new cracks than spruce when separately. Figure 11 shows the that the develpement of cracks on the boards are exposed to cycles of crack length for pith surfaces ver- the bark surface in boards with wetting and drying. After the sus non-pith surfaces and the horizontal annual rings during wetting cycles, spruce has almost changes regarding crack length drying is common, especially when the same proportion of boards during the wetting cycles. The pith the annual ring curvature are without cracks as it had before, surface means that the surface (or small, i.e. the boards are sawn (see Figure 10). It should be noted surfaces) of the board was oriented near the pith. that none of the boards with the towards the pith. The boards in When the crack lengths of each pith enclosed in the cross section group A are excluded in the analy- group (E-G) are analysed have been dried without cracks, sis as these boards do not have separately, the variation makes it (see group A in figure 10). Further- pith surfaces. In both pine and difficult to draw any distinct more, the amount of boards with spruce, where boards from groups conclusions. However, spruce in cracks was 3-4 times greater in E, F and G have been analysed as group E had significantly higher boards sawn near the pith (group one group, the pith surfaces show crack lengths on the pith surfaces, E) than in boards sawn away from crack lengths that are significantly both before and after wetting, (see the pith (groups F and G). higher compared to other surfaces. Figure 11). The increase of crack

Figure 10. The proportion of boards without cracks, related to the total number of boards in each group. The boards are divided into four groups (A, E, F, G) according to the annual ring orientation and the distance between pith and the board.

103 length during wetting was also length between the surfaces. Pine from the pith indicates that the higher on the pith surfaces. Pine in group F, on the other hand, had pith with the surrounding juvenile did not show these results. This a pith surface with greater crack wood has a great influence on the may be a result of the trend for length before the wetting cycles, proportion of cracks in sawn pine to crack more easily than and in group G both before and timber. spruce, and that the cracks after the wetting cycles. The In pine, the influence of pith initiated on the pith surfaces in development of cracks in pine, and juvenile wood on the propor- pine develop into the adjoining group F, also indicates that cracks tion of cracks ranges over a great- surfaces, especially when the pith initiated of the pith surfaces may er distance from the pith compared is located close to the edge propagate to other surfaces during to spruce. Thus, in pine, the dif- between two surfaces. the wetting cycles. ference in crack length between Spruce boards sawn at a The distinct difference in crack boards sawn at a distance greater, distance greater than 30 mm from length on spruce pith surfaces or less than 30 mm from the pith, the pith (groups F and G) did not between boards sawn at a distance is not as clear as in spruce. show any difference in crack of less and greater than 30 mm

Figure 11. Crack length for pith surfaces contra non-pith surfaces. The boards are divided into three groups (E, F, G) according to annual ring orientation and the distance between the pith and the board.

104 3.5.3 Influence of other with horizontal annual rings and spruce. The relative crack length in factors on crack boards with the pith enclosed. As boards sawn near the pith was propagation a consequence of the great varia- greater than in boards sawn away The influence of the annual ring tion in warp, it was difficult to from the pith. During wetting and orientation on crack formation was detect any significant (confidence drying, the crack lengths increase also investigated on both wet and 0.95) differences between boards regardless of the location of the non-wetted surfaces. Crack forma- with different annual ring board in relation to the pith. tion in the boards was not depend- orientations. Where differences in When the four surfaces of the ent on annual ring orientation. warp were found, the warp was board were examined separately the For pine and spruce the mean significantly lower in boards with results show that pith surfaces have value of moisture content at the semi-vertical and vertical annual greater relative crack lengths and a end of each wetting cycle was rings. greater tendency to crack than the lower than that at the beginning of Cup is greatly influenced by the other surfaces, especially when the the test, (see Figure 5). Decrease curvature radius of the annual boards are sawn near the pith. of the moisture content may result rings as a consequence of the in an increase of cracks in the anisotropic nature of wood. This boards when compared to a cons- means that boards with horizontal tant moisture content at the end of annual rings, sawn near the pith, each cycle. have the greatest cup. Consequently, boards with vertical 4. Conclusions annual rings show no cup. The results show that if boards are Compression wood has the sawn radially from the log, without greatest influence on the increase pith and juvenile wood, properties of warp. Bow is more sensitive to are obtained which are better than compression wood than crook and those of conventionally sawn twist. timber. This study, however, Warp in terms of crook, bow, demonstrates that the number of twist and cup as well as number of cracks in boards was clearly cracks in sawn timber from both related to the juvenile wood and pine and spuce increase when the distance from the pith in sawn timber is exposed to repeated timber. The number of boards cycles of wetting and drying. Warp with cracks was 3 to 4 times increases in each wetting cycle. greater in timber sawn near the The greatest increase occurs after pith, that is when the distance the first cycle. between the board and the pith The influence of annual ring was less than 30 mm, compared to orientation on warp, except for timber sawn further away from the cup, was not clear. Generally, the pith. When the boards were boards with semi-vertical and exposed to wetting and drying vertical annual rings have lower cycles, the number of boards with mean values for warp than boards cracks increased in both pine and

105 References Armstrong, L. D.; Kingston, R. S. Perstorper, M.; Pellicane, P. J.; T. 1960: Effect of moisture Kliger, I. R.; Johansson, G. 1995: changes on creep in wood. Nature Quality of timber products from 185:862-62 Norway spruce. Part 2. Influence Balodis, V. 1972: Influence of of spatial position and growth grain angle on twist in seasoned characteristics on warp. Wood boards. Wood Science , 5(1): 44-50 Science and Technology 29:339-352 Blom, G. 1989: Sannolikhetsteori Sandberg, D. 1995: Stående års- och statistikteori med tillämp- ringar hos furu (Pinus silvestris L) ningar. 4:e upplagan, Student- och gran (Picea abies Karst). Royal litteratur Institute of Technology, Wood Danborg, F. 1990: The effect of Technology and Processing, Rep- silvicultural practice on the amount ort TRITA-TRÄ R-95-13 and quality of juvenile wood in Sandberg, D. 1996a: Radially Norway spruce. Paper presented at sawn timber. Star-sawing - a new the IUFRO-World Congress, method for producing timber with Montreal vertical annual rings. Holz als Roh- Hallock, H. 1965: Sawing to und Werkstoff 54(3):145-151 reduce warp of loblolly pine studs. Sandberg, D. 1996b: The USDA Forest service, Forest. influence of pith and juvenile wood Prod. Lab. Research paper FPL- on proportion of cracks in sawn 51, Madison USA timber when kiln dried and Kloot, N. H.; Page, M. W. 1959: exposed to wetting cycles. Holz als A study of distortion in Radiata Roh- und Werkstoff 54(3):152 pine scantlings. CSIRO, Div. of Söderström, O. 1990: Torknings- For. Prod., Technological Paper sprickor och slutfuktkvoter. The no 7, Melbourne, Australia Swedish Institute for Wood Mishiro, A.; Booker, R. E. 1988: Technology Research, TRÄTEK, Warping in new crop Radiata pine Report I 9011062 100x50 mm boards. Bull. Tokyo Timell, T. E. 1986: Compression Univ. Forests 80:37-68 Wood in Gymnosperms. Springer- Montgomery, D. C. 1991: Design Verlag, Berlin and analysis of experiments. Third edition, John Wiley & sons, Inc. Morén, T. 1987: Vidaretorkning och konditionering av furuvirke. Royal Institute of Technology, Stockholm, Wood Technology and Processing, Report TRITA-TRT- 37

106 Radially sawn timber. Chapter Gluing of star-sawn triangular profiles into form- stable products with vertical annual rings IV D. Sandberg, H. Holmberg

and defects. The average length of However, the longitudinal moisture Summary the remaining pieces used for movement of the wood can be A method is presented for the finger-jointed units was 0.41 m. considerable in reaction wood, gluing of star-sawn triangular profi- juvenile wood, and in areas with les of pine (Pinus silvestris L) into 1 Introduction severe fibre disturbances. For pine form-stable wood products with and spruce, it is usually assumed The function and properties of vertical annual rings. The that the tangential movement is wood products are greatly influen- triangular profiles are free from approximately twice as large as the ced by the natural properties of pith and from most of the juvenile radial movement. This means that the wood and how these are used wood. changes in shape appear as in the product concerned. There The method is based on dried distortions and angular changes in are four main conditions which are and conditioned triangular profiles. the cross-section if the side characteristic of wood. It is an Undesired defects are removed surfaces of the timber are not anisotropic and heterogeneous and the wood is finger-jointed into parallel with the main directions of material, a material with great long lengths. In the finger-jointing, the wood. variability and a material with consideration is given to the The importance of the annual pronounced hygroscopicity. All appearance and annual ring ring orientation for moisture this is related to the fact that the orientation of the joined materials. movement of furniture boards has wood material is built up in an After the joining, the triangular been investigated by several optimum way to meet the specific profiles are planed and glued into researchers, e.g. Peck and Selbo requirements of the living tree. rectangular blocks with vertical (1950), Selbo (1952), Perry (1953), When the log is divided into planks annual rings. These blocks can Marian and Suchland (1955). They and boards to be used in different then e.g. be used as construction state that in order to obtain a products, complications can arise if beams or be divided into solid dimensionally stable furniture one does not use the available wood panels. board, it shall be constructed of knowledge. By starting from an A pilot plant has been built for timber with vertical annual rings, understanding of the properties of the manufacture of knot-free solid i.e. the annual rings shall be largely the wood material, it is in many wood panels based on the perpendicular to the surface of the cases possible to manufacture a proposed method. Results show a timber . product which meets the specific total volume yield of 53.8%. Three Sandberg (1996a, 1997) has demands of different uses. critical production stages can be shown that for pine and spruce, For timber, such requirements distinguished: removal of knots there is a clear connection between are often related to accuracy in and defects, planing, and division the occurrence of cracks in timber measurements and geometry, of blocks into boards. These three from butt logs and the distance crack-freeness and controlled operations are together responsible from the pith from which this movement during moisture for more than 93% of the total timber has been sawn. Timber changes. Changes in the moisture losses in the manufacture. containing pith, and wood without content below the fibre saturation The removal of knots and pith where the shortest distance point cause changes in shape and defects meant a volume loss of between the pith and the timber is dimensions. The wood swells and 12.8%. 13% of the manufactured less than 30 mm, have considerably shrinks, and this takes place mainly units were 2.1 m long without more cracks than timber sawn at a across the fibre direction. finger-jointing and free from knots greater distance from the pith.

109 Star-sawing, is a sawing method for block-sawing. These knots are often the production of timber with unacceptable aesthetically and from vertical annual rings which is free a strength viewpoint, and they from pith and most of the juvenile must be removed. Sandberg and wood (Sandberg 1996b), (Figure 1). Holmberg (1996) have mapped the The yield from star-sawing consists knot appearance and knot of timber with both rectangular frequency in star-sawn triangular and triangular cross-sections, all profiles of pine and spruce. with vertical annual rings. In star- In the present investigation, a sawing, as in all radial sawing method is presented for gluing methods, the radial sawkerfs cleave together star-sawn triangular profi- the knots in the length direction les into form-stable products with and a greater proportion of so- vertical annual rings and radial called splay knots is obtained than surfaces. in through-and-through sawing and

Figure 1. Star-sawing, sawing pattern for producing timber with vertical annual rings.

110 2 Principles for gluing profile, primarily from butt logs, are used in the finger-joint triangular profiles into a have proved to be less suitable for depending on whether the joined block with a rectangular further refinement of the wood timber is to be used for cross-section and must be removed (Sandberg construction purposes or in The further refining of the and Holmberg 1996). A great applications where the strength of triangular profiles takes place in advantage of a triangular profile the join does not have high prior- five main steps: over conventionally sawn timber ity. The timber can either be 1. Removal of unacceptable knots when knots are to be removed, is joined continuously and then cut and defects. that the knots in the triangular into desired lengths or joined 2. Finger-jointing of clear pieces. profile are always visible on at least directly into predetermined length 3. Planing of triangular profiles. one of the surfaces. This means modules. In each join one of the 4. Gluing of triangular profiles into that, after the knots which are joined pieces shall be displaced half blocks. visible on the surface have been the division of the fingers to avoid 5. Further treatment of blocks removed, no knots remain hidden displacement in the join which in through planing and sanding, or in the profile to appear in subsequent treatment operations division of the blocks into subsequent operations. can cause volume losses. boards or other components. Fibre disturbances around knots In the finger-jointing, one reduce the strength and can cause should try as far as possible to 2.1 Removal of defects pick-up when finger-jointing and keep together the timber sections planing. In the manufacture of e.g. cut from the same triangular prof- Depending on what final product is high-quality furniture, fibre ile to obtain a homogeneous to be produced, the extent of disturbances can be troublesome annual ring structure (pattern defects, e.g. knots, pitch pockets, because they locally give different fitting) across the finger-joint. In reaction wood, allowed in the moisture movements, and can lead order also to minimize stresses product will vary. to undesirable visual changes in the which arise in the join because of A large proportion of the knots surface. For certain applications, it different swelling and shrinking found in a star-sawn triangular can therefore be advantageous to movements in the different remove the surrounding fibre directions of the timber, the annual disturbance together with the ring orientation shall coincide in knots. the joined pieces (Sandberg 1996c). Figure 2 shows examples of a 2.2 Finger-jointing to obtain finger-joint where two pieces from long lengths the timber wood section have been Finger-jointing of a triangular prof- joined after an intervening knot ile is in principle carried out with has been removed. In general, it is the same methods as are used in more favourable to have radial the finger-jointing of conventional surfaces than tangential surfaces Figure 2. The annual ring orientation in timber. Certain modifications of when pattern fitting is desired jointed boards is very important for the the equipment are necessary to between the joined pieces. appearance of the join and for moisture adapt e.g. transport and holding The finger-joint can be oriented movements in the join. Two pieces from the same piece of timber with vertical devices to the triangular shape of in two different directions in rela- annual rings finger-joined (see arrow). the timber. Different finger lengths tion to the annual ring orientation

111 in the cross-section of the trian- to an uneven pressing pressure, gular profile: parallel to the tangen- Figure 4. tial surface of the triangular profile The orientation of the finger- or parallel to one of its radial joint parallel to the surface in the surfaces, Figure 3. In principle, the triangular profile which will later orientation of the join can be be glued into the block means that chosen arbitrarily if this is proven the characteristic zigzag pattern of to be advantageous from a finger-joints becomes visible on the technical or aesthetic point of surfaces of the block, Figure 5. view. This is not discussed in greater detail here, however. 2.3 Planing of triangular In joining tests, it has been profiles Figure 4. Waviness in a finger-joint because of uneven distribution of pressure in the finger-joining. found advantageous to orient the The finger-jointed triangular profi- finger-joint parallel with the surface les are planed on all sides. In order which in subsequent treatment to facilitate the subsequent pres- steps is to be glued into the con- sing of the profiles into blocks and struction, Figure 3a. This is, on the to determine the height of the one hand, because errors develop glued block, one of the edges of in the surface in the transition the profile is also planed off. The between the joined wood pieces triangular profile is oriented before and, on the other hand, because the planing so that the surface the join can become wave-shaped obtained when the edge is removed in the surface because a lack of full can be used to achieve the correct parallelism between the direction annual ring orientation in the glued of the fingers and the surface leads block. Figure 5. Block glued from finger-jointed triangular profiles. The finger-joints are oriented so that the characteristic zig-zag pattern (see arrows in figure) becomes visible in the surface of the block.

a b

Figure 3a and b. The direction of the finger cutting in relation to the annual ring orientation in a triangular profile. (a) Parallel with the tangential wood surface. (b) Figure 6. Block with uniform annual ring orientation Parallel with one of the radial wood surfaces. DFJ - direction of finger-joint, AR - glued from star-sawn triangular profiles. annual ring

112 2.4. Gluing of triangular a coloured glue, visible glue-line are osed to the tangential surface, lies profiles into blocks with obtained. This can be used to in the length direction of the a rectangular cross- create aesthetically attractive wood, almost invisible repairs can section patterns in the glued products. be carried out. These repairs are After the planing, glue is applied to Because of the shape of the also relatively simple and can be two of the four surfaces of the triangular profiles, the glue-line made rationally. triangular profile and the profiles forms an angle towards the side are placed together into a block surfaces of the glued block, Figure 2.6 Pilot plant for further with the same annual ring 6, which means that the glue sur- refinement of a orientation, Figure 6. The tangen- face becomes larger than if the triangular profile tial surfaces of the triangular profi- glue-line were perpendicular to the A pilot plant has been built to les are glued into the block. surface. The larger glue surface remove defects from triangular means that the stress on the glue- profiles and for joining, planing and 2.4.1 The choice of glue line is reduced. gluing of these profiles into blocks. and the design of the The aim is to create on a small sca- glue-line 2.5 Treatment of blocks and le new production technology for repair of pitch pockets The choice of glue is, in the first the gluing of a triangular profile. place, dependent on the function The glued blocks can after gluing Properties of different products ma- of the final product and on the be either directly planed and sand- nufactured by the proposed method environment in which it will be ed or split into sheets or boards. will also be tested. used. In supporting building Pitch pockets can be very constructions, high demands are troublesome, especially in furniture 2.6.1 Design of the pilot made on the strength and and furnishings. In contrast to plant permanence of the glue-line. In knots, pitch pockets can be con- The triangular profiles, dried to a these contexts, resorcinol- cealed inside the triangular profile moisture content of 8 to 10 per- formaldehyde-based glues are often without being visible on any of its cent, arrive from the sawmill or used. In the gluing of components surfaces. This can mean that the storage. At the crosscut saw, for indoor use, where the loads on pitch pockets in subsequent treat- undesired knots and defects are the glue-line are considerably less ment become visible, with accom- removed. The defect-free pieces than in building constructions, panying trouble. Discarding the are then joined together PVAc-glue can be used. product in a late refinement stage continuously to lengths of 2.10 Depending on the production rate involves high costs, and repair is metres in order, as far as possible, required in the gluing, a suitable often the most economic solution. to achieve good pattern fitting method of curing for the glue must The extension of the pitch between the joined pieces. be chosen, e.g. pre-heating of the pocket in the tree means that in Triangular profiles without defects glue surfaces or high-frequency the radial surface of the sawn and which need not be joined to curing. timber, the pitch pocket will attain a length of two metres are A further aspect regarding the appear as an oblong, elliptical cav- placed aside for the manufacture choice of glue is whether or not ity, extending between two annual of products without finger-joints. the glue shall give a coloured joint. rings (Weslien 1993). Since the The smallest length which can be With a colourless glue, an almost main extension of the pitch pocket used in finger joining is 80 mm. invisible glue-line is obtained. With in the radial wood surface, as opp-

113 The fingers are cut parallel to the active pressure is applied on all together with different annual ring tangential surface of the triangular surfaces of the block. orientation. profile. After being pressed In the present study, star-sawn together, the joined boards are 3 Volume yield in the triangular profiles have been used stored until the glue has hardened. manufacture of a pine for the manufacture of a knotfree Planing takes place in a five- solid wood panels from a solid wood panel with vertical spindle universal planer (Weinig star-sawn triangular annual rings. The volume yield has 17215, 1960). In order to manage profile been studied in the manufacture of the planing of the triangular profile In general, board materials two grades of material, one with a in one step, the press and feed manufactured by gluing together finger-joint and one without a join. have been redesigned. One of the pieces of timber side by side are edges of the triangular profile is called solid wood panels. A solid 3.1 Materials and methods planed in the top cutter. After the wood panel is usually manufactur- Star-sawn triangular profiles of planing, glue is placed on two of ed from planks which are cleaved saw-falling qualities have been the four surfaces of the profile and in the length direction into lamellae used. The timber from which the a number of profiles are placed and are planed before being glued triangular profiles were sawn came together into a block. In the press together into a board. This method from Jämtland in Sweden and had used, only rectangular blocks can of manufacture gives large volume a top diameter of between 260 and be pressed, which means that one losses and lamellae are glued 340 mm (Holmberg and Holmberg of the profiles must be cleaved into 1996). Altogether, 7.8 m³ triangular two halves and be used as side pieces on the block. The manufactured blocks have an area of 2.10 x 0.35 m² and the thickness can be varied between 50 and 100 millimetres, depending on the dimensions of the triangular profiles included. The pressing of the triangular profiles into blocks differs considerably from the conventional gluing of timber with a rectangular cross- section since a large pressure is required on all of the surfaces of the block. In the prototype press used, active pressure is applied to the edge sides of the block, whilst the thickness is fixed, Figure 7. Because of friction, this means that Figure 7. Side view of a press module for gluing of blocks of a star-sawn triangular the pressure is not constant over profile. The press consists of five modules which are connected by two press plates. A, B - locking device for part E that can be opened, C - beam for distributing the press the width direction of the block. force along the side surfaces of the block, D - hydraulic cylinder for lateral press force, This problem is avoided if an E - part that can be opened, F - block of triangular profiles, G - upper press plate.

114 profiles with side measurements III. Random samples have been 3.2 Results and discussion between 60 and 110 mm have been taken for the measurement of made into solid wood panel blocks. dimensions before and after pla- 3.2.1 Yield The wood quality for the ning. The yield is shown for each part triangular profiles used has not IV. For each finished solid operation, i.e. the analysis of each been determined, but the quality wood panel block, the thickness, operation is based on 100% input was estimated by determining the width and adjusted length have material to that operation, table 2. quality of adjacent rectangular been determined and also the In the investigation, the length of boards from the same log. dimensions of the triangular profi- the jointed boards has varied According to Table 1, quality clas- les included. depending on the number of joints. ses S1, S2 and S3 are specially The glued blocks have not been A large number of joints gives a adapted for star-sawn timber, but divided into finished boards in this shorter length than when the correspond roughly to the quality investigation. Losses in the division number of joints is few. The losses classes U/S, V and VI in the of the blocks into boards have in the final adjustment of the block Nordic grading regulations for been determined on the basis of length have therefore been greater sawn-timber (Timber Grading experience of values for sawkerfs than can be expected in an Committee 1976). A detailed and sanding losses. industrially adapted plant. description of the quality classes The yields of the final for star-sawn timber is given by production stages, division and Holmberg and Sandberg. (1996). sanding, are based on the following Measurements have been carried out manually at four different stages in the production: I. The dimensions of all incoming triangular profiles have been determined with regard to the Table 2. The yield for each production operation and the operation’s largest and smallest side and share of the total loss. length. Operation Yield (%) Proportion of total loss (%) II. After cutting, the cut pieces have been placed continuously Cutting 87.2 27.7 together into modules with a length Finger jointing 98.3 3.0 of 2.10 m. Planing1) 76.6 43.5 Gluing of blocks 100 0

End cutting in blocks2) 97.4 3.7

Division, sanding3) 84.2 22.1

Total of all operations 53.8 100 Table 1. Quality distribution for the rectangular boards, (Holm- berg and Sandberg 1996). 1) The yield in the planer is based on the total loss for all planed units, cf. Table 3.

Top diameter Proportion of the total number of sawn boards 2) The variation in the length of the triangular profiles included in the blocks has interval of the been much greater than can be expected in an industrial plant. logs Quality S1 Quality S2 Quality S3 3) Yield calculation based on: blocks with a height h²60 mm give two boards, blocks with h<_ 60 mm give three boards. Sawkerf 3 mm and sanding loss 1 mm 260-340 mm 0.35 0.45 0.20 per surface.

115 assumptions: if the block has a 3.2.2 Defect removal and of 1.7%. This loss can be reduced if height which is greater than 60 joining shorter fingers are used. The short- mm, three boards are produced In the defect removal stage, all est length of the pieces which has from this block, otherwise only knots with surrounding fibre been joined is 80 mm. two boards are produced. Each disturbances, and defects which sawcut is assumed to remove 3 mm affect the appearance of the finis- 3.2.3 Planing and the sanding loss is assumed to hed board have been removed. The yield for the material which be 1 mm per board surface, i.e. 2 This gave a loss of 12.8%. has been measured before and mm/board. The losses in format The length yield in the cutting is after planing is shown in Table 4, cutting of the boards and in edge shown in Table 3 and Figure 8. Of with regard to the smallest and sanding have not been included in the total incoming volume, 13% is largest sides and with regard to the yield calculations. defect-free and unjoined 2.1 m their average value. As shown in the Table 2, the units. If shorter lengths than 2.1 m The yield varies between 75 and most demanding steps are the rem- are extracted, the volume of mate- 84%, which suggests that the oval of defects, planing, division rial to be joined decreases. This dimensional accuracy of the and sanding of the blocks into will be natural in practice. triangular profiles was very poor. If finished boards. Altogether, these The finger-joining causes a loss the accuracy can be improved, the production steps cover more than which depends on the finger yield in the planing will probably 93% of the total loss in the solid length, in this case 10 mm. In total, be improved. wood panel manufacture. the finger cutting has given a loss The planing of the triangular profiles is the operation which gives the greatest volume loss in the solid wood panel manufacture. It is therefore very important that the planing is carried out with as high precision as possible. Table 3. Number of joints after knots and other defects have been removed and the average length of the cut pieces.

Number of Average length of joins per metre pieces per unit

Including unjoined boards 1.73 0.45

Excluding unjoined boards 1.99 0.41

Table 4. Yield (%) in planing based on the different dimensions of the joined pieces before planing. Number Yield in the planer based on: of units smallest largest average dimension dimension dimension

138 83.7 75.0 79.2 Figure 8. Distribution of the number of joins per 2.1 m length.

116 Figure 9 shows the yield in the mm. When the edge of the cross-section becomes rectangular. planing for different planing triangular profile is planed, only 1- The volume loss in this division is depths. The planing depth is very 2% of the volume is lost. When less than 1% and has not been important for the yield, especially if instead the thickness of the block included in the account of the the triangular profiles are small. is reduced, the losses are at least volume yield, Table 2. The average dimension of the twice this and can in some cases The losses which arise in the triangular profiles in this investig- amount to 10% of the volume of division of the block into boards ation was 90 mm and the yield in the block. depend on how many boards are the planing was 76.6%, as shown in In the drying, the cross-sections to be extracted from the block, Table 2. This suggests, Figure 9, of the triangular profiles become whether the division of the block that the average depth of the pla- deformed slightly so that the edge works out evenly, the thickness of ning was more than 3 mm. angles deviate from the angle the sawkerfs and the sanding los- In the planing, one of the edges which they had after sawing, i.e. ses. In the present investigation, of the triangular profile is also normally 60°. Angular deviations the blocks have not been divided. removed, primarily to guarantee a can also occur as a consequence of From experience, it is known satisfactory gluing of the blocks incorrect sawing and these errors that the sawcut requires 2.5-3.0 and to fix the thickness of the are often considerably greater than mm and that a sanding loss of 1 block. The cutting of this edge the angular errors arising in the mm is sufficient to give a fine sur- causes no great volume loss. Figure drying. face. 10 shows the theoretical yield when Angular errors mean that the Figure 12 shows the volume one of the edges of the planed side surfaces of the triangular yield in the division of blocks into triangular profile is removed. The profiles must be treated further in 2, 3 or 4 boards. ”1 board” means loss is 1-2%, which is neutralized the planing so that surfaces that the block is sanded only on by the advantages which are suitable for gluing are obtained. It the side surfaces. Division of the obtained. can be seen in Figure 11 that this blocks and the sanding normally The thickness of the block is loss is independent of the reduce the yield by between 10 and determined by the heights of the dimensions of the triangular profile 20% but the losses increase triangular profiles. If the height of and that large angular errors give drastically if many boards are sawn the triangular profiles is limited considerable volume losses. from thin blocks. already in the planing, large Angular errors which arise in the volume losses can be avoided drying are of the order of 1°, 4 Conclusions compared with when the thickness which according to Figure 11 gives The triangular profiles obtained in of the block is adjusted on the a volume loss of 2%. star-sawing constitute a high- finished glued block. quality raw material for the The theoretical yield is also 3.2.4 Gluing of triangles manufacture of wood products. influenced when the thickness of into blocks, and Triangular profiles have vertical the block is reduced by ”h” mm. division of the blocks annual rings and are without pith This can be achived on the one into boards and most of the juvenile wood. hand by planing the edge of the In the gluing of the blocks, one of This means that the profiles are triangular profile, or on the other the triangular profiles has been stable in shape, but the appearance hand by reducing the thickness of divided lengthwise to be used as and nature of the knots are often the finished glued block by ”h” sides of the block, so that the such that they cannot be accepted

117 in the final product and must be removed. The wood pieces obtained after knots and other defects have been removed are finger-joined into long lengths. By taking the annual ring orientation into consideration in the finger-joining, the stresses Figure 9. The yield in planer based on unilateral which arise when the moisture triangular profile planing with the same planing allowance, k, on all sides. ratio of the wood changes are minimized. The uniform annual ring orientation in the finger join also facilitates pattern fitting between the joined pieces regarding the structure of the wood surface. The method presented for the gluing of triangular profiles into blocks shows how triangular profi- Figure 10. Volume yield in removal of only one edge on les can be further refined into each triangular profile. The width of the bevel form-stable products on an obtained when the edge is removed is designated ”b”. industrial scale. The tests which are now being carried out in a pilot plant show that knotless solid wood panels can be manufactured according to this method and with a high volume yield.

Figure 11. Volume yield after correction of the angular error, α, in planing triangular profiles.

Figure 12. The dependence of the volume yield on the number of extracted boards, n, and the block thickness.

118 5 References Holmberg, H.; Sandberg, D. 1996: Sandberg, D. 1996c: The Volume yield and quality of timber influence of annual ring orientation when using Star-sawing technique. on strength and dimensional Royal Institute of Technology, changes during moisture content Wood Technology and Processing, changes in finger joints. Holz Roh- Report TRITA-TRÄ R-96-22 (in Werkstoff 55:50 Swedish) Sandberg, D.; Holmberg, H. 1996: Marian, J. E.; Suchsland, O. Radially sawn timber. Knots - 1956: Waviness of lumber- and number, type and size in star-sawn chipcore board an investigation of triangular profiles of pine (Pinus the factors affecting surface stabil- silvestris L) and spruce (Picea abies ity. Paper and Timber no. 6-7 Karst). Holz Roh- Werkstoff Peck, E.C.; Selbo, M. L. 1950: 54:369-376 Comparison of redwood and flat- Sandberg, D. 1997: Radially sawn grained yellow-poplar for cores in timber. The influence of annual furniture panels. Forest Products ring orientation on crack formation Lab. Madison, Report No. 1785 and deformation in water soaked Perry, T. D. 1953: Making pine (Pinus silvestris L) and spruce Lumber Cores for Plywood. Wood (Picea abies Karst) timber. Holz and Wood Products 61 (2) Roh- Werkstoff 55:175-182 Sandberg, D. 1995: Vertical Selbo, M. L. 1952: Effectiveness annual rings in pine (Pinus silvestris of different conditioning schedules L) and spruce (Picea abies Karst), in reducing sunken joints in edge Royal Institute of Technology, glued lumber panels. J. For. Prod. Wood Technology and Processing, Soc. 2 (1) Report TRITA-TRÄ R-95-13 (in Timber Grading Committee 1976: Swedish) Guiding principles for grading of Sandberg, D. 1996a: The Swedish sawn timber, Second edi- influence of pith and juvenile wood tion, The Swedish Timber and on proportion of cracks in sawn Wood Pulp Journal timber when kiln dried and Weslien, H. 1993: Estimating size exposed to wetting cycles. Holz and amount of pitch-pockets in Roh-Werkstoff 54:152 roundwood. The Swedish Univer- Sandberg. D. 1996b: Radially sity of Agricultural Science. Dept. sawn timber. Star-sawing - a new of For. Prod. Research Notes No. method for producing timber with 171 vertical annual rings. Holz Roh- Werkstoff 54:145-151

119 120 Weathering of radial and tangential wood surfaces Chapter of pine and spruce. 1. Crack formation and colour changes in exposed surfaces V Dick Sandberg

pine have 13 times more total crack mechanical degradation caused Summary length per unit area than the largely by shrinkage and swelling The development of cracks and corresponding radial surfaces. In with changing moisture content; changes in appearance have been spruce, the total crack length on biological degradation due to investigated on radial and tangen- the tangential surfaces is 6 times fungal and insect attack and tial surfaces of pine (Pinus silvestris greater than on the radial surfaces. bacteria. L) and spruce (Picea abies Karst) Tangential surfaces of both pine The photochemical degradation which have been exposed outdoors and spruce have a greater number is a very slow process which during for 33 months. Untreated samples, of cracks per unit area and wider a decade degrades only a few samples impregnated with a CCA- cracks than the corresponding millimetres of the wood surface agent and samples surface treated radial surfaces. and leaves the underlying wood with linseed oil have been tested. Tangential and radial surfaces practically unaffected (Hon et al The annual ring orientation is show the same colour change in 1978). The combined effect of the most important factor for the surface as a result of water and sunlight degrades the crack development on weathering. weathering. main components of the wood and The type of wood, impregnation transforms the wood surface into a treatment and surface treatment Introduction network of weakly connected with linseed oil have only a margi- Wood exposed to weather and cellulose fibres which are strongly nal effect on the crack wind without being protected with contaminated by spores from development. No relation has been some kind of surface layer, e.g. a microorganisms (Sell et al 1969). found between the density of the paint layer, degrades as a result of Mechanical degradation also samples and the crack three different processes: takes place relatively slowly and development. photochemical degradation results primarily in an aesthetic After 33 months of outdoor depending primarily on the degradation of the wood surface in exposure, tangential surfaces of ultraviolet radiation in sunlight; the form of cracks and loose wood fragments on the surface. On the other hand, biological attack can in tangential surface a short time lead to large damage to the wood material if the radial surface temperature, moisture level, and supply of oxygen and nutrients are favourable. On a macroscopic level, the a colour change of an untreated wood surface is one of the first and perhaps the clearest sign of the degradation of the wood during outdoor exposure. Visible light and UV-radiation alter the Figure 1. The star-sawing pattern showing (a) wood with a rectangular cross-section colour of the wood to a darker or and triangular profiles, (b) production of samples from the triangular profiles. For test lighter shade, depending on the piece a1, a3, a5, .... one of the radial surfaces has been exposed and for test piece a2, a4 .... the tangential surface has been exposed. The length ai=300 mm, k=knot. type of wood (Sandermann et al

123 1962, Fengel et al 1984). After a Table 1. The extent of the test material. Number of test pieces with different long period of outdoor use, all treatments. types of wood develop a greyish Treatment of test body Pine Spruce appearance (FRN 1966, Sell et al Radial* Tangential* Radial* Tangential* 1971). This depends on the fact that water-soluble decomposition Untreated 6 5 16 14 products are removed and the Treated with linseed oil 5 6 5 5 more or less delignified fibres are Impregnated with CCA-agent 13 13 4 4 exposed. If, on the other hand, the Impregnated and treated with wood surface is protected against linseed oil 12 13 19 18 rain, it develops a dark red-brown Total number of samples 36 37 44 41 surface (Browne 1959). * Refers to the annual ring orientation of the surface exposed at an inclination of 45° in a southerly direction, Visible cracks arise in the wood see also Figure 1. surface during outdoor exposure because of the growth of microcracks formed during the drying of the wood, photochemical was sawn from one pine and two were brushed with linseed oil before reactions or moisture-induced spruce logs. After sawing and the outdoor exposure began. stress fields (Coupe et al 1967). drying, some of the triangular Table 1 presents a summary of Stamm (1965a) considers that wood profiles have been pressure the test material. The samples were for outdoor use should have impregnated with a CCA-agent. exposed in Stockholm for 33 vertical annual rings. This From each triangular profile, months (September 1993 - May minimizes the risk of cracks as a between 7 and 14 knot-free and 1996) at an inclination of 45 consequence of anisotropic defect-free test pieces with a length degrees towards the south. moisture movements. Cracks in of 300 mm have been prepared. After the outdoor exposure had the radial surface are also smaller Figure 1 shows how the samples been completed, all samples were than in the corresponding tangen- have been produced. conditioned for two months at a tial surfaces (Browne 1960, Stamm After being cut into lengths of temperature of 24°C and a relative 1965a, 1965b). 300 mm, the samples were planed humidity of 66 %. Thereafter, the The aim of the present on all surfaces to facilitate the de- lengths of all cracks with a crack investigation is to demonstrate termination of crack lengths. The width greater than 0.25 mm, which possible differences in the samples were planed with the top was the smallest crack size which degradation process between radial end in the planing direction to a could be measured in practice, and tangential wood surfaces of depth of 0.5 mm. The final width were determined. Cracks with a Scots pine and Norway spruce of the side varied between 75 and width less than 0.25 mm were exposed outdoors above ground. 125 mm. The dry density was assessed visually. In the statistical determined for all samples: pine evaluation of the results, a ”non- Material and method 475±25 kg/m³ and spruce 416±25 parametric test” has been used Star-sawn timber of pine and kg/m³. The end-wood surfaces with 0.95 confidence limits for the spruce with a triangular profile has were sealed with an oil alkyd group mean values (Montgomery been used in the investigation primer and a silicone-based sealing 1991, Sandberg 1995). (Sandberg 1996). The test material compound. Some of the samples

124 Results Table 2. Average of the total crack length per unit area (m/m2) after 33 months’ Cracks wider than 0.25 mm on outdoor exposure and subsequent conditioning to a moisture content of 12 %. . the exposed surfaces Treatment of test body Pine Spruce Table 2 shows mean values for Radial* Tangential* Radial* Tangential* the sum of the crack length per unit area (total crack length on the Untreated 2.1 25.5 2.9 27.3 surface divided by the area of the Treated with linseed oil 4.1 28.6 6.7 27.2 surface) on the exposed surfaces Impregnated with CCA-agent 1.1 29.8 0.1 31.0 after 33 months’ outdoor exposure Impregnated and treated with and subsequent conditioning. linseed oil 2.7 27.3 6.0 23.9 The results show a clear All samples in the group 2.2 28.1 4.4 26.1 difference between radial and tangential surfaces, regardless of * Refers to the annual ring orientation of the surface exposed at an inclination of 45° in a southerly direction, see impregnation or surface treatment. also Figure 1. The difference between radial and tangential surfaces is statistically length than spruce on the tangen- area which are wider than 0.25 significant. On the other hand, it is tial surfaces, but a shorter crack mm. Figure 3b shows the average not possible to find any influence length on the radial surfaces. The value of the greatest crack width on crack length from impregnation statistical analysis shows, however, for the samples, i.e. the maximum or linseed oil treatment. that there is no significant crack width has been determined Figure 2 shows the average of difference in mean crack length for each test piece. Both quantities the total crack length for the between pine and spruce, either on have been determined on the exposed radial and tangential the tangential or on the radial exposed surface. surfaces in pine and spruce. Pine surfaces. The radial surfaces have a shows a slightly greater mean crack Figure 3a shows the mean value significantly smaller number of of the number of cracks per unit cracks than the tangential surfaces and the maximum crack width is 30 also significantly smaller. In the case of the radial

) 25 2 Pine surfaces, it is not possible to show 20 Spruce any significant difference between pine and spruce, either for the 15 number of cracks or the crack 10 width. On the other hand, tangential pine surfaces have significantly 5 more cracks and a smaller maxi-

Crack length per unit area (m/m 0 mum crack width than spruce Tangential Radial surfaces. No influence on the crack for- Exposed surface mation of either surface treatment Figure 2. Average of the total crack length per unit area (m/m2) for pine and with linseed oil or impregnation spruce after 33 months' outdoor exposure and subsequent conditioning to a moisture content of 12 %. has been found.

125 Table 3. Number of samples with small cracks (crack width less than 0.25 mm) on the Cracks wider than 0.25 mm exposed surface after 33 months’ outdoor exposure and subsequent conditioning to a on the non-exposed surfaces moisture content of 12 %. Besides the surface exposed in the Main crack orientation Pine Spruce southerly direction, each test piece Radial* Tangential* Radial* Tangential* has two adjacent sides which have

The annual ring border 33 2 40 1 not to the same degree been exposed to sunlight. These two Earlywood 29 1 42 2 surfaces have degraded more Latewood 1 11 2 18 slowly than the exposed surfaces. The whole exposed area 3 25 2 23 Figure 4 shows the crack length Total number of samples 36 37 44 41 for the non-exposed surfaces. In * Refers to the annual ring orientation of the surface exposed at an inclination of 45° in a southerly the same way as for the exposed direction, see also Figure 1. surfaces, the average of the total crack length on the radial surfaces the naked eye on the exposed sur- is significantly smaller than that on Cracks with a width less face and where on the surfaces the tangential surfaces. As can be than 0.25 mm on the these cracks occur most seen in Figure 2, the average of the exposed surfaces frequently. total crack length on the non- Cracks with a width less than 0.25 On the radial surfaces, the exposed surfaces is smaller than on mm are usually short, only a few cracks were primarily found in the the exposed surfaces. For the ra- mm, and many. In practice it has annual ring border and in the dial pine surfaces, the average of not therefore been possible to earlywood. On the tangential the total crack length on the non- determine the crack length for surfaces, the cracks appeared in exposed surfaces is 55 times smal- these. the latewood and across the whole ler and on the tangential surfaces 6 Table 3 shows the number of of the exposed surface, i.e. in both times smaller than on the exposed samples with small cracks visible to the earlywood and latewood.

Figure 3. Average values for (a) the number of cracks per unit area and (b) the maximum crack width on the exposed surfaces after 33 months' outdoor exposure and subsequent conditioning to a moisture content of 12 %.

126 5 Changes in the appearance Figure 5 also shows an example

4 of the surfaces of a colour change in samples after Pine )

2 outdoor exposure. This colour 3 Spruce Cracks develop in wood surfaces during outdoor exposure are, as change is particularly evident on 2 already shown, particularly visible the exposed surface. The surface 1 changes colour to silver-grey,

Crack length per unit area (m/m on tangential surfaces. Figure 5 0 shows the characteristic regardless of whether or not the Tangential Radial appearance of a radial and a tang- wood is impregnated, Figure 5. Exposed surface ential surface from the present At the upper edge of the Fig. 4. Total crack length per unit area investigation. On the tangential samples in Figures 5b and 5d it is (m/m²). The average value on the non- surface (Figure 5d), cracks are evident that the part of the surface exposed surfaces after 33 months' out- which has been protected against door exposure and subsequent condition- clearly visible. On both the radial ing to a moisture content of 12 %. and the tangential surfaces, a large sunlight but exposed to rain and number of small cracks can be snow has a colour which is diffe- found which are, however, difficult rent from that of the remaining to see in Figure 5. surface. The influence of the surfaces. The corresponding values The degradation follows the sunlight on the colour change is for spruce are 23 and 9 times. annual ring pattern in such a way clarified if the backs of the samples Cracks with a width less than that the earlywood degrades more are studied. Practically no colour 0.25 mm on the non-exposed rapidly than the latewood and change has occurred on these surfaces valleys develop in the wood sur- surfaces. Table 4 shows the number of face. The surfaces develop a samples with small cracks on the corrugated appearance, which is Density non-exposed surfaces. On these particularly clear on the radial Figure 6 shows the average of the surfaces, small cracks have not surfaces. total crack length as a function of occurred to the same extent as on density for all the samples. No the exposed surfaces. relationship has been found On the radial surfaces, small between crack formation and the cracks in the annual ring border density of the samples. dominate and, on the tangential surfaces, the cracks are most Discussion frequent in the latewood. Table 4 The samples included in this shows that the cracks are not investigation have mainly been evenly distributed across the whole exposed to simultaneous surface of any test piece, but that photochemical and mechanical the crack formation has begun on degradation. parts of surfaces. The tendency is The chemical reactions in the the same as for the exposed wood surface initiated by sunlight surfaces, however, namely that it is lead to a change in colour of the on the tangential surfaces that the Figure 6. The average total crack length surface. This process is similar for cracks become visible across the on the exposed surface of pine and spruce radial and tangential surfaces. whole surface. as a function of dry density, R2 = 0.002.

127 a b

Figure 5. Colour change of wood surfaces after 33 months' outdoor exposure. Radial surface of spruce (a) before and (b) after exposure. Tangential surface of unimpregna- ted pine (c) before and (d) after exposure. cd

128 The photochemical degradation the test pieces and its internal reg- tangential direction, the early- and takes place in a thin layer of the ion, i.e. mechanical degradation. latewood are parallel, which means wood surface and this means that The shrinkage and swelling in that the moisture movement is the strength of this layer is reduced the tangential direction are about limited by the individual layers and (Derbyshire et al 1981, 1995, twice as large as the radial mois- that stresses arise in and between Raczkowski 1980). When the wood ture movement. Tangential sur- these layers, see e.g. Kifetew material is exposed to moisture faces thus move more than radial (1996). This is a reason why tang- variations, stresses develop in the surfaces. This means that the ential surfaces crack more than cell walls and between cells, and stresses become higher in the tang- radial surfaces. these lead to damage and to the ential direction than in the radial On the radial surfaces, cracks propagation of existing cracks. The direction when the moisture move- occur first at the annual ring bor- photochemical reactions which ment in the surface layer is limited der (Table 4) and thereafter in the take place in the wood surface by underlying wood which does earlywood (Table 2). On the tang- during outdoor exposure acceler- not have the same moisture con- ential surfaces, cracks appear first ate this process, but they are not tent as the wood in the surface in the latewood (Table 4) and then the main reason for the great layer. This relationship applies in develop across the whole of the difference in numbers of cracks the case of rapid moisture changes exposed surface (Table 2). This between radial and tangential wood in the surface, e.g. when the sur- agrees with results reported by surfaces. face is alternately exposed to rain Coupe et al (1967). The fact that If the crack length ratio between and is dried out in strong sunlight. cracks do arise at the annual ring tangential and radial surfaces is Moisture gradients then arise border on the radial surfaces is a studied for the exposed surfaces between the surface area and the consequence of the large and ab- (Figure 2) and for the non-exposed underlying wood material. This rupt change in density which takes surfaces (Figure 4), it can be esta- results in the formation of more place in the transition from late- blished that this ratio is smaller for cracks on tangential surfaces than wood to earlywood. It seems that the exposed surfaces, which imp- on radial surfaces. cracks also arise at an early stage lies that the difference in crack The latewood shrinks and swells in the latewood on the radial development between radial and more than the earlywood. In the surfaces, but these cracks are tangential surfaces is not a conse- radial direction, the moisture difficult to observe because the late quence of photochemical degrad- movement in the latewood is about wood strips are very thin, about ation. On the other hand, strong 3 times higher than in the early- 0.25 mm. exposure to sunlight and rain wood and, in the tangential direc- The strains to which the wood means that the size and frequency tion, about 1.5 times greater in the material is exposed in pressure of cracks increase strongly. latewood (Boutelje 1962, Vintila impregnation can cause damage to The difference in crack 1939). When the moisture content the material, which can mean that susceptibility between radial and in the wood changes, the early- impregnated wood would be more tangential surfaces is mainly the and latewood thus move to diffe- susceptible to cracking during result of stresses which arise in the rent extents. In the radial direc- weathering than wood which is not wood as a consequence of tion, the layers are oriented in impregnated. This hypothesis is anisotropic moisture movements of series and the early- and latewood not supported by the result of the the wood material and moisture layers can move practically inde- now reported investigation. gradients between the surface of pendently of each other. In the

129 FRN 1966: Wood finishing: Sandermann, W.; Schlumbom, F. Conclusions References Weathering of wood. US. For. 1962: On the effect of filtered The annual ring orientation in the Boutelje, J. 1962: The Res. Note FPL-0135 ultraviolet light on wood. Part II: wood surface has a great influence relationship of structure to Hon, D. N. -S.; Ifju, G. 1978: kind and magnitude of colour on the susceptibility of the surface transverse anisotropy in wood with Measuring penetration of light into difference on wood surfacees. Holz to crack. To avoid cracks reference to shrinkage and elasti- wood by detection of photo- als Roh- und Werkstoff 20(8):285- occurring on wood used outdoor, city. Holzforschung 16(2):33-46 induced free radicals. Wood Sci. 291 wood pieces with annual ring Browne, F. L. 1959: Understan- 11:118-127 Sell, J.; Leukens, U. 1971: orientation perpendicular to the ding of the mechanism of Kifetew, G. 1996: Some aspects Investigations on weathered wood exposed wood surface should be deteriation of house paint. Forest on the deformation behaviour of surface - Part II: weathering selected. This is especially Product Journal 11:417-427 wood in relation to its structure. phenomena on unprotected wood important if the wood lacks a Browne, F. L. 1960: Wood siding KTH, Royal Institute of species. Holz als Roh- und covering surface layer, e.g. paint, left to weather naturally. Southern Technology, Wood Technology Werkstoff 29(1):23-31 or if the surfaces are strongly Lumberman, December 15:141-143 and Processing, Stockholm, Report Sell, J.; Wälchli, O. 1969: exposed to weather and wind. Coupe, C.; Watson, R. W. 1967: TRITA-TRÄ R-96-19 Changes in the surface texture of For wood of pine or spruce Fundamental aspects of Montgomery, D. C. 1991: Design weather exposed wood. Material exposed outdoors without a weathering. Proc. Ann. Conv. Br. and analysis of experiments. Third and Organismen 4(2):81-87 covering surface layer, the annual Wood Preservation Ass. Pp. 37-49 edition, John Wiley & Sons, Inc. Stamm, A. J. 1965a: Wood and ring orientation in the wood sur- Derbyshire, H.; Miller, E. R. Raczkowski, J. 1980: Seasonal cellulose - water relationships and face is more important for the 1981: The photodegradation of effects on the atmospheric their effect upon finishes. Official crack development than the dens- wood during solar irradiation. Part corrosion of spruce micro-sections. Digest, June pp. 654-669 ity of the wood or the type of I: Effects on the structural integrity Holz als Roh- und Werkstoff Stamm, A. J. 1965b: wood. Impregnation treatment with of thin wood strips. Holz als Roh- 38:231-234 Modification of wood for improved a CCA-agent or surface treatment und Werkstoff 39:341-350 Sandberg, D. 1995: Influence of finishing. Official Digest, June pp. with linseed oil has no more than a Derbyshire, H.; Miller, E. R.; annual ring orientation on crack 707-717 marginal influence on the crack Turkulin, H. 1995: Investigations formation and deformation in Vintila, E. 1939: formation. into the photodegradation of wood water soaked pine (Pinus silvestris Untersuchungen über using microtensile testing. Part 1: L) and spruce (Picea abies Karst) Raumgewicht und Schwindmass The application of microtensile timber. KTH, Royal Institute of von Früh- und Spätholz bei testing to measurement of Technology, Wood Technology Nadelhölzern. Holz als Roh- und photodegradation rates. Holz als and Processing, Stockholm, Report Werkstoff 2:345-357 Roh- und Werkstoff 53:339-345 TRITA-TRÄ R-95-16 (In Swedish) Fengel, D.; Wegener, G. 1984: Sandberg, D. 1996: Radially sawn Wood - chemistry, ultrastructure, timber. Star-sawing - a new reactions. Walter de Gruyter, Ber- method for producing timber with lin vertical annual rings. Holz als Roh- und Werkstoff 54(3): 145-151

130 Weathering of radial and tangential Chapter wood surfaces of pine and spruce. 2. Microscope studies VI Dick Sandberg

Summary Introduction 1975, Derbyshire et al 1981), and The sensitivity of the wood to strength (Derbyshire et al 1995, The degradation of radial and degradation is one of its greatest Raczkowski 1980). Of the whole tangential surfaces of pine (Pinus weaknesses in outdoor usage. Like electromagnetic spectrum, it is only silvestris L) and spruce (Picea abies all biological materials, wood the shortwave and energy-rich Karst) has been studied at the decomposes under the influence of region which has a measurable micro-level. The surfaces were the surrounding environment. influence on wood and which is untreated, impregnated with a When wood is exposed outdoors thus of technical interest. As a CCA-agent or treated with linseed above ground, a complex consequence, a large number of oil. decomposition process continues studies have been carried out Tangential surfaces have more in the material as a consequence of within this field and summaries of and deeper cracks than radial chemical, biological, mechanical earlier results have been published surfaces. The cracks on the tang- and light energy related factors. A by e.g. Kenaga et al (1959), Desai ential surfaces occur frequently in common name for this process is (1968), Kringstad (1969) and Hon both earlywood and latewood. On ”weathering” (Feist 1982). et al (1979, 1991). Sandberg (1998) radial surfaces, cracks occur The factors which are generally has shown a clear difference in primarily at the annual ring bord- considered to cause changes in crack occurrence on a ers, but to a certain extent also in wood surfaces on weathering are: macroscopic level between radial the earlywood. sunlight (UV, visible and infrared and tangential surfaces in pine and The radial cell wall of the early- radiation), moisture (dew, rain, spruce weathered over a period of wood has a large number of pores snow), temperature and oxygen ca 3 years. Cracks on the tangen- which are degraded at an early (Hon 1983). tial surfaces were shown to be stage. Decomposition of the cell Because of the limited ability of more and also larger than cracks wall takes place on both radial and light to penetrate into wood on the radial surfaces, for tangential surfaces. Cracks arise (Browne et al 1957), the effect of untreated, impregnated and which follow the S2 fibril the weathering is limited to a thin linseed-oil-treated samples. The orientation in the different cellwall surface layer and the erosion is aim of this investigation was to use layers. Delamination in the middle slow, 5-12 mm per 100 years (Feist the test material from the lamella is especially noticeable in et al 1978). investigation reported by Sandberg the latewood on tangential Investigations of the effect of (1998) and on a microlevel surfaces. No differences have been weathering on wood, carried out characterize differences in crack observed regarding linseed oil by different researchers, have dealt pattern between radial and tangen- treatment, impregnation or type of with several aspects, e.g. colour tial surfaces. wood. change (Fengel et al 1984, Sandermann et al 1962, Sell et al Material and method 1971), erosion (Arnold et al 1992, The test material used in this Feist et al 1978, 1984), free radicals investigation has been taken from (Hon et al 1980, 1981, 1991), sur- an investigation into crack forma- face wetting characteristics tion in weathered wood surfaces (Kalnins et al 1992, 1993), (Sandberg 1998). Machineplaned anatomical changes (Miniutti 1967, radial and tangential surfaces of Borgin 1970, 1971, Borgin et al pine (Pinus silvestris L) and spruce

133 (Picea abies Karst) were exposed at Microscopic changes in the wood two subsequent degradation levels an angle of 45° in a southern surfaces were investigated by occur in both radial and tangential direction in Stockholm during 33 ESEM (Philips ElectroScan 2020). surfaces, with slightly different months (September 1993 - May processes, however, which will be 1996). After the exposure, the Observations and illustrated here. In the analysis of samples were conditioned for two comments the different surfaces, no months at 24°C and 66% RH. The In the investigation of the exposed significant differences have been test material consisted of untreated surfaces, the analysis has taken found regarding linseed oil samples, samples which had been place at three structural levels: treatment, impregnation or type of pressure impregnated with a CCA- – macroscopic cracks which wood, and the influence of these agent, samples which had been propagate along practically the parameters will not be further coated with linseed oil and samples whole length (300 mm) of the test discussed. which had been both impregnated body and with a relatively large Figure 1a shows an example of a and linseed oil treated. 40 samples depth. planed and unexposed wood sur- with an approximate size of 10 x 50 – small short cracks, normally face. The surface in the lower part x 10 mm3 (WxLxT) were selected with a depth of within one or a few of the picture has been treated by randomly from 158 weathered annual rings. UV-laser ablation to remove the samples. To investigate the cross – decomposition of the actual layer of crushed fibres which was a section, the cross section was wood fibre and middle lamella. result of planing. Figure 1b shows prepared with the help of UV-laser The first level of degradation a corresponding surface after ablation to minimize the risk of occurs primarily in tangential weathering for 33 months. In large artefacts in the preparation surfaces. This has been shown in parts of the exposed surface, the (Seltman 1995, Stehr et al 1998). an earlier investigation (Sandberg cell structure is exposed and it is Otherwise, no preparation of the 1998) and has been treated only possible to distinguish different samples was carried out. marginally in this investigation. The structural elements, which it is impossible to see in the unexposed planed wood surface, Figure 1a. Figure 1b shows how the uppermost layer of crushed fibres has cracked as a consequence of the weathering. Sunlight during simultaneous moistening and drying of the surface means that cracks develop and that the fibres which have been partly compressed in the planing operation rise. The surface is opened successively and fibre fragments are broken apart and eroded by Figure 1. (a) Planed radial wood surface of pine with crushed fibres, magnification: 213 x. The rain and wind. In Figure 1b, wood surface in the lower part of the picture has been ablated with UV-laser, which has removed most of the fibre layer which was compressed during the planing. (b) The fracture across the fibres as a corresponding surface after weathering for 33 months, 275 x magnification. consequence of tensile stresses is

134 also shown. The fracture region The crack extension almost always before the surrounding material has normally stretch 3-10 cell rows takes place in one of the first early eroded away down into the surface. wood cell rows and seldom more In the latewood, cracks seldom than two rows from the annual occur and, in those cases where Exposed radial surface ring border, (Figure 2b). cracks do appear, they are small, On the radial surfaces, cracks can In the earlywood, besides the only a few cell rows deep. be found in practically every deep cracks in the annual ring The earlywood erodes more annual ring border, (Figures 2a and border, relatively small and shallow rapidly than the latewood, which 2b). The cracks are usually a few cracks occur, (Figure 3). These probably depends on the fact that millimetres deep and extend cracks do not in general become the latewood has a higher density parallel to the annual ring border. deeper than about 10 cell rows than the earlywood. Figure 4 shows a latewood strip where the surrounding earlywood has been decomposed and removed by wind and weather. The thin and brittle strips of latewood are broken after some time. The radial cell walls of the earlywood of both pine and spruce contain a large number of bordered pits. These are degraded at an early stage. The same phenomenon has been observed by e.g. Miniutti (1967) and Borgin (1970, 1971). Figure 5 shows pits at an early Figure 2. Cracks at the annual ring border on a radial surface. (a) Crack opening at stage of degradation. The pit the surface of untreated spruce at 175 times magnification. (b) Crack a millimeter below the surface of impregnated pine at 250 x magnification. openings have been enlarged and through the pits, diagonally oriented cracks have developed. During the continuing decomposition, these cracks propagate through the cell wall. According to Turkulin et al (1997), these are a consequence of stresses which arise in contraction of the actual pit chamber. Torus also erodes at an early stage. Figure 5 Figure 5. Early stage of the degradation also shows that the wart layer is of a radial cell wall of impregnated and largely lacking and that cracks linseed oiled spruce at 1000 x have begun to develop in the cell magnification. Torus is destroyed, the pit Figure 3. Crack in the earlywood on a Figure 4. Latewood strips of untreated pine wall. This type of crack is found in openings have been enlarged and cracks radial surface of impregnated pine at in 265 x magnification, where the surroun- have arisen through the pits and in the 220 x magnification. ding earlywood has eroded. both the early- and latewood. The cell wall.

135 orientation of these cracks suggests that these follow the fibril orientation in the S2-layer of the cell wall. In the continuing degradation, the pit openings are further enlarged and the whole pit finally erodes away. The crack formation in the cell wall continues and degradation products are rinsed away until only a close-meshed network of fibril bundles remains, (Figure 6). The microfibrils are Figure 7. Exposed tangential surface of linseed oiled spruce in (a) 280 times thus shown to be the most stable magnification and (b) 445 times magnification. In the weathering, tangential surfaces part of the actual tracheid, and the develop (a) large deep cracks and small cracks in the cell wall, and (b) failure across the fibres. degradation continues with reduced adhesion between very few ring pores and earlywood layer and has reached microfibrils and between the diffe- consequently, the degradation the latewood. The crack rent cell layers. pattern for the tangential surfaces propagation in latewood in the becomes different. The tangential outermost exposed surface layer Exposed tangential surface surfaces develop more, but also showed a similar development. Figures 9a and 9b show dela- In contrast to the radial wall, the considerably larger cracks than the mination between fibres in early- tangential cell wall contains no or radial surfaces. Figure 7a shows a typical tangential surface, with on and latewood for an exposed tang- the one hand a large deep crack ential surface. The delamination is and on the other hand diagonal especially clear in the latewood. cracks in the cell walls, which is The cracks in the latewood extend also common in the radial cell into the middle lamella with a walls. Failure across the fibre depth of up to ten cell rows, occurs on the tangential surfaces, Figure 9a. In the earlywood, the (see Figure 7b). cracks are shallower and the On the tangential surfaces, delineation against adjacent cell cracks occur often in both the walls is not as sharp as in the early- and latewood. The cracks latewood, (Figure 9b). are initiated in the surface layer, in Figure 10 shows how the diffe- both the early and latewood, and rent layers in the cell wall are dela- propagate through the different minated during the degradation. annual ring layers. Figure 8 shows The different cell wall layers and a crack which has been initiated in the fact that the crack propagation Figure 6. Radial section of weathered spru- has a different orientation in the ce (linseed-oil-treated) in 465 x magnifica- the earlywood on an exposed tang- tion. The degradation has meant that only ential surface. The crack has different layers are evident. Figure a close-meshed network of fibril bundles 10 also shows a crack running remains from the cell wall. propagated through the whole

136 Figure 8. Exposed tangential surface of Figure 9. Delamination cracks in the middle lamella of (a) latewood and (b) earlywood Figure 10. Delamination and crack forma- linseed oiled pine in 235 x magnification of an exposed tangential spruce surface. tion in a tangential cell wall of impregnated with a crack in the exposed earlywood pine. Magnification 850 x. layer and with underlying latewood. along the length of the fibre which latewood. On radial surfaces, cracks has propagated through the whole occur primarily at the annual ring thickness of the cell wall. border, but to a certain extent also in the earlywood. The radial surface Conclusions develops a corrugated appearance Weathered pine and spruce show a because the early- and latewood degradation process which is decompose at different speeds. On strongly dependent on the annual an ultrastructural level, ring orientation in the exposed decomposition of the pores is the surface. At a macroscopic level, clearest difference between radial tangential surfaces develop long and tangential surfaces. The radial and deep cracks which open the cell wall has a large number of pits wood structure and facilitate which decompose at an early stage. continued decomposition. Radial In both radial and tangential surfaces, on the other hand, show surfaces, degradation of the cell only very small cracks. Sandberg wall takes place. Cracks arise which (1998) has analysed these follow the fibril orientation in the degradation processes. S2 cell wall layer. Delamination in On a microscopic level, it is also the middle lamella is especially possible to see clear differences in clear in the latewood on tangential degradation between radial and surfaces, but it also occurs in the tangential surfaces. Tangential earlywood. The microfibrils are the surfaces have more and deeper parts of the tracheid which are cracks than radial surfaces. The most resistant to weathering. cracks on the tangential surfaces occur frequently in both early and

137 Desai, R. L. 1968: Chemistry. Marcel Dekker Inc. Kind and magnitude of colour References Photodegradation of cellulosic New York difference on wood surfaces. Holz Arnold, M.; Lemaster, R. L.; materials - A review of literature. Kalnins, M. A.; Knaebe, M. T. als Roh- und Werkstoff 20(8), pp. Dost, W. A. 1992: Surface Pulp Pap. Mag. Can. 69, pp. 53-61 1992: Wettability of weathered 285-291 characterization of weathered Feist, W. C. 1982: Structural use wood. Journal of Adhesion Science Sell, J.; Leukens, U. 1971: wood using a laser scanning sys- of wood in adverse environments. and Technology 6(12), pp. 1325- Investigations on weathered wood tem. Wood and Fiber Science pp. 156-178, Van Nostrand Rein- 1330 surface. Part II: weathering 24(3), pp. 287-293 hold Co, New York Kalnins, M. A.; Feist, W. C. phenomena on unprotected wood Borgin, K. 1970: The use of the Feist, W. C., Mraz, E. A. 1978: 1993: Increase in wettability on species. Holz als Roh- und scanning electron microscope for Comparsion of outdoor and wood with weathering. Forest Werkstoff 29(1), pp. 23-31 the study of weathered wood. accelerated weathering of Products Journal 43(2), pp. 55-57 Seltman, J. 1995: Freilegen der Journal of Microscopy 92(1), pp. unprotected softwoods. Forest Kenaga, D. L.; Cowling, E. B. Holzstrukturdurch UV- 47-55 Products Journal 28(3), pp. 38-43 1959: Effect of gamma radiation on Bestrahlung. Holz als Roh- und Borgin, K. 1971: The mechanism Feist, W. C., Hon, D. N.-S. Ponderosa pine: Hygroscopy, Werkstoff 53(4): 225-228 of the breakdown of the structure 1984: Chemistry of weathering and swelling, and decay susceptibility. Stehr, M.; Seltman, J.; Johansson, of wood due to environmental protection. Adv. Chem. Ser. 207, Forest Products Journal 9, pp. 112- I. 1998: UV laser ablation - An factors. J. inst. Wood Sci. 5(4), pp. pp. 401-457 116 improved method of sample prepa- 26-30 Fengel, D.; Wegener, G. 1984: Kringstad, K. 1969: Degradation ration for microscopy. Borgin, K.; Paramewaran, N.; Wood - chemistry, ultrastructure, of wood and high-yield pulp by Holzforschung 52(1): 1-6 Liese, W. 1975: The effect of aging reactions. Walter de Gruyter, Ber- light - A survey of the present state Turkulin, H.; Sell, J. 1997: on the ultrastructure of wood. lin of knowledge. Tappi 52, pp. 1070- Structural and fractographic study Wood Science and Technology 9, Hon, D. N.-S. 1983: Weathering 1074 on weathered wood. An pp. 87-98 reactions and protection of wood Miniutti, V. P. 1967: application of FE SEM microscopy Browne, F. L.; Simonson, H. C. surfaces. Journal of Applied Poly- Microscopic observations of to the ”Thin strip” method. 1957: The penetration of light into mer Science: Applied Polymer ultraviolet irradiated and EMPA, Bericht 115/36 wood. Forest Products Journal Symposium 37, pp. 845-864 weathered softwood surfaces and 7(10), pp. 303-314 Hon, D. N.-S.; Ifju, G.; Feist, W. clear coatings. US Forest Service Derbyshire, H.; Miller, E. R, C. 1980: Characterization of free Research Paper FPL 74 1981: The photodegradation of radicals in wood. Wood and Fiber Raczkowski, J. 1980: Seasonal wood during solar irradiation. Part 12, pp. 121-130 effects on the atmospheric I. Effects on the structural integrity Hon, D. N.-S.; Feist, W. C. corrosion of spruce micro-sections. on thin wood strips. Holz als Roh- 1981: Free radicals formation in Holz als Roh- und Werkstoff 38, und Werkstoff 39, pp. 341-350 wood: The role of water. Wood pp. 231-234 Derbyshire, H.; Miller, E. R,; Science 14, pp. 41-48 Sandberg, D. 1998: Weathering Turkulin, H. 1995: Investigations Hon, D. N.-S.; Glasser, W. 1979: of radial and tangential wood into the photodegradation of wood On possible chromophoric surfaces of pine and spruce. 1. using microtensile testing. Part 1: structuresin wood and pulps - A Crack formation and colour The application of microtensile survey of the present state of changes in exposed surfaces. In testing to measurement of knowledge. Polym. Plast. Technol. press Holzforschung photodegradation rates. Holz als Eng. 12, pp. 159-179 Sandermann, W.; Schlumbom, F. Roh- und Werkstoff 53, pp. 339- Hon, D. N.-S.; Shiraishi, N. 1962: On the effect of filtered 345 1991: Wood and Cellulosic ultraviolet light on wood. Part II:

138 Material damage due to electron beam during Chapter creep test in environmental scanning electron microscopy (ESEM) VII G. Kifetew and D. Sandberg

Summary 1. Introduction of vapour or liquid. Thus, compared Wood is a complex anisotropic, to the conventional SEM, drying This study describes the heterogeneous and hygroscopic and coating a specimen can be development of cell wall damage structural material. Due to its wide eluded and therefore, allows (creation of checks) across or in structural application, the long investigation in the chamber at low the vicinity of pits during testing term behaviour of wood has been vacuum. Taking advantage of the microtomed samples in the a research area during the last development, Mott et al (1996) Environmental Scanning Electron decades. However, the relationship and Shaler et al. (1997) have tested Microscopy (ESEM). Four between the long term behaviour individual fiber in tension using microtomed sapwood samples of of wood under varying load and ESEM. Their observation suggested normal grown spruce (Picea abies moisture conditions and its that tension failure was related to Karst.) were prepared in green anatomical structure at a cell wall pit. Accordingly, in this study the condition. One sample was level has not yet been clearly authors have tested several investigated in an unloaded understood. Hence, the microtomed green sapwood samples condition, while three specimens conventional Scanning Electron of normal grown spruce (Picea were tested under constant tension Microscopy (SEM) and ESEM are abies) in the ESEM to relate the load and at different moisture some of the convenient tools that surface deformation of wood under level. Regions of the moisture can be applied for such constant tension load and varying cycled samples that had been investigation. moisture conditions to its exposed to electron beam during SEM technique has been utilized anatomical structure. Kifetew image acquisition showed damages in describing the failure (1996) have utilized a regular grid that run through pits and their morphology of wood under diffe- pattern for measuring deformation surroundings. Specimen loaded in rent loading conditions. The effect field on wood during drying. green condition and dried in the of low moisture content in a speci- Similarly, in this study an chamber without beam exposure men and the high vacuum of the earlywood zone with several pitted for two hours except during SEM on the failure behaviour of cells was selected to identify the imaging did not show any wood have been studied by identical points to obtain the sur- noticeable cell wall damage. The Kyanka (1976) and others (see face deformation. Further, during result indicated that the electron Hoffmeyer and Hanna 1989). In the process of utilizing ESEM in beam might be the major source of the past, Côte and Hanna (1983), determining the surface deforma- damage initiation. Therefore, it is Zink et al (1994) and Kifetew et. al tion of wood, damages have been essential to take care on the (1997) have used SEM technique to observed at or in the vicinity of circumstance of the test and in characterize wood fracture pits. However, most SEM fracture explaining the observations made surfaces. Hoffmeyer and Hanna surface studies have demonstrated in ESEM studies. (1989) have utilzed the SEM to the resistance of pits to a transwall study the influence of electron failure i.e. failure that runs all the beam on the failure morphology of way through a pit. A literature spruce (Picea abies). review of the subject has been Recent improvements have presented by Kifetew (1996). permitted ESEM to be a suitable Nevertheless, damaged pits have technique in investigating biological been detected by Turkulin and Sell tissues that contain water in form (1997) in SEM studies on Scots

141 pine and Norway spruce samples selected as region A and the initial rent moisture conditions are exposed to natural and artificial image was recorded prior to included, for each condition weathering. Feist and Hon (1983) loading and changing the moisture several specimens were tested and have observed a diagonal state. This was done in order to a number of micrographs were microchecks passing through the record any damage that might have recorded. bordered pits after 1000 h UV been introduced onto the specimen The experimental procedures irradiation. Jenkins and Donald surface during sample preparation for each sample can be (1997) have investigated the and air-drying. Another region of summarized as follows: sensitivity of cellulosic fibers to the interest, B, was also selected as a a) Unloaded and moisture cycled electron. They think that damage standstill position during the drying green sample: Region A was selected mechanisms are enhanced by the and wetting periods. At this posi- and an initial image was acquired contnual presence of water vapour tion, the gun alignment which in the green condition. To protect in the ESEM chamber. Therefore, normally has x,y-"coordinates of region A from long-term beam the principal objective of this study (0,0) was changed to another x,y- exposure, a new region of interest has been concentrated in studying co-ordinate position in order to B was selected as a standstill posi- the influence of electron beam, avoid a direct beam exposure of tion. The specimen was dried for moisture variation and loading this region. Other regions were 30 min and the second image of condition on the cell wall structure chosen as comparison sites in case region A was recorded. The of wood. damage was detected in region A sample was then exposed to two during the test procedure. moisture cycles of 10 min wetting 2. Material and methods The accelerating voltage during followed by 30 min of drying. Four sapwood samples of normal the tests was 15kV. Drying and During the wetting/drying grown spruce (Picea abies Karst.) wetting of samples was carried out procedure, the electron gun were cut in the green condition at about 666 Pa, 30oC and 1333 Pa, alignment was directed away from using a slide microtom along the 10oC respectively. In most cases, region B. Images of region A in the fiber direction. The cross-sectional images were recorded at a dry state were recorded after each dimensions were 0.2 mm by 10 magnification of 300X, but some wetting/drying cycle. Additional mm, and the length was about 45 images were also acquired at high sites were also selected and images mm. One test sample was air-dried magnifications. Failure loads in were recorded for comparison for about two days and the tension parallel to grain of two purposes. remaining three specimens were green and two air-dried samples b) Tensile loading and moisture preserved in the green state prior were determined in order to esti- cycling a green sample: An initial to tensile loading parallel to the mate the constant load level that image of region A was first grain. During the test, all the could be exerted on the sample. recorded and the sample was samples were mounted on a Thereafter, each sampel was loaded. After loading, the second microtensile stage that fits inside loaded at a rate of 0,1 mm/sec to a image of the region was acquired. the ESEM. leavel of 15N. One specimen was A new region B was selected as a Several regions, referred to as investigated in the unloaded standstill position and the sample A, B, C etc were selected for condition. was dried for 30 min and the dried image acquisition. On each sample, In the report, although only image of region A was recorded. a distinguishable earlywood zone four samples, i.e. one control The sample was then exposed to with several pitted cells was sample and three samples at diffe- two 10 min wetting periods

142 followed by 30 min drying cycles. d) Tensile loading and drying a A. Micrographs of other regions are Images of position A in the dry green sample: Region A was selected considered when a relative state were recorded after each and an initial image was recorded, explanation becomes essential due wetting and drying cycle. the sample was loaded and the to damage in region A. Additional regions of interest were second image acquired, and the selected for recording images in sample was then dried for two Unloaded and moisture order to identify changes. hours. During the drying period, cycled green sample c) Tension loaded and moisture the beam was completely shut off. During recording of the initial cycled air-dried sample: The initial After two hours of drying, images green state image, region A was image of region A was recorded, of region A were recorded. exposed to the electron beam for the sample was then loaded and a Additional regions were also 2.1 min, but the region showed no second image acquired, then it was selected for image acquisition for damage. Nor did the second image dried for 30 min in the chamber. descriptive purpose. recorded after the sample had The procedure that followed after been dried for a period of 30 min drying the sample for 30 min i.e. 3. Observations and exposed to the electron beam the selection of other regions, the Although, for each sample, images for a total of 5.1 min (Figure 1a) wetting and drying cycles and the were recorded at several regions, reveals any noticeable damage. acquisition of images were the the analysis presented in this study Damage started to appear on the same as for sample b). was concentrated only on region image (Figure 1b) which was made

a b

Figure 1. Micrographs of region A a) after 5.1 min and b) after 8.25 min of beam exposure

143 a b

Figure 2. Micrographs of region A after a) 4.0 and b)7.0 min of beam exposure.

a b

Figure 3. Micrographs after of region A a) 5.0 min and b) 8.0 min of beam exposure.

144 after the first wetting/drying cycle damage. An image that revealed wetting/drying cycle and a total of and a 8.25 min beam exposure. some damaged pits and its 6.5 min of beam exposure showed After two wetting/drying cycles surroundings was recorded after some slightly damaged areas near and a total of 15.55 min of beam the first wetting/drying procedure the pits. Damaged areas can exposure, region A exhibited more and after beam exposure of the clearly be observed on images damaged areas. region for a total of 7.0 min (Figure recorded after the second wetting/ 2b). drying cycle (Figure 3b) and 8.0 Tensile loaded and min of beam exposure. moisture cycled green Tension-loaded and sample moisture cycled air-dried Tensile loaded and dried While the initial and the second sample green sample (Figure 2a) images were recorded, The first undamaged image of re- To acquire the reference image, 4.5 region A was beam exposed for 4.0 gion A was recorded after 3.5 min min of beam exposure was required min. However, neither image of beam exposure prior to loading and the region showed no damage. revealed any noticeable damage. and additional drying. The second The specimen was then dried for Furthermore, the image recorded image was made after loading and two hours in the chamber while after the sample had been dried for exposure of the region for a total of the electron beam was completely 30 min, by exposing the region to 5.0 min (Figure 3a). These images shut-off i.e. 0 kV. Images were an electron beam for a total of 5.5 showed no noticeable damage, but acquired to detect any damage or min, still did not exhibit any the image recorded after the first structural changes on the surface of

Figure 4. Micrograph after two hours of drying and 9.5 min Figure 5. Micrograph of a comparison site. of beam exposure. 145 the region of interest, but no in the form of vapour or liquid. 5. Conclusion noticeable structural change or Therefore, specimen drying or The aim of this investigation was to damaged surface could be detected coating can be avoided and this study the effect of electron beam, after the region of interest A had makes possible the observation of constant load and repeated drying been exposed for a total of 9.5 min a sample in the chamber at low and wetting on cell wall damage. (Figure 4). vacuum. However, to acquire an The micrographs of both loaded image, sample exposure to the and unloaded samples exposed to Comparison sites electron beam can only be reduced moisture cycles and an electron Before any conclusions were and cannot be completely avoided. beam have revealed damage that in drawn about the structural change The micrographs presented in most cases runs through a pit and due to the development of damage, this study showed a clear its surroundings. However, a green images of some new and randomly difference in the occurrence of sample tested under constant load chosen regions on each sample damage depending on whether or without electron beam exposure were recorded. These regions were not specimens have been exposed during two hours of drying except considered to be comparison sites to electron beam for a longer time during the period when images for the changes noticed on the i.e. a minimum of 6.5 min and a were acquired, showed no surface of region A on each maximum of 8.25 min and have noticeable damage. These results sample. The maximum time been wetted and dried. All the indicate that the electron beam required to expose a region to three specimens that were tested may be a major cause of damage electron beam during image with or without a constant load initiation. Therefore, care should recording was 2 min. As shown in have experienced two wetting/ be taken when explaining the Figure 5, none of the images of the drying cycles at different and observations made in ESEM newly selected regions revealed any varying pressure and temperature studies. noticeably damaged area. levels. These samples have shown distinct pit damage and its surroundings. However, other sites 4. Discussion used for comparative purpose The reason for choosing region B showed no sign of cell wall was to use the site as a standstill damage. position during the drying and This investigation has disclosed wetting periods in order to protect the development of damage during region A from long-term beam attempts to study the creep defor- exposure. This was done because mation of wood and its anatomical we were not sure whether the structure relationship in the change in electron gun alignment ESEM. Therefore, the study at region A could achieve the suggests that precautions be taken expected goal and prevent the in identifying the weakest links and region from beam exposure. in labelling the sources of failure It is recognized that recent initiation in wood cell wall improvements and the use of structure under direct ESEM ESEM have made it possible to studies. investigate wood fibers with water

146 References CÔTE, W. A., AND R. B HANNA. KYANKA, G. 1976: Fracture 1983. Ultrastructural characteristics behaviour of single fibers and of wood fracture surfaces. Wood paper sheets in the scanning and Fiber Science 15(2), 135-163. electron microscope. Proc. 2nd Int. FEIST, W. C., AND D. N.-S. HON. Cong. On Mech. Behaviour of 1983. Chemistry of weathering and Materials. 1354-1357. protection. Pages 401-451 in Roger MOTT, L.; S.M. SHALER, L.H. Rowelleds. The chemistry of solid GROOM, AND B.H. LIANG. 1995. The wood. The 185th meeting of the tensile testing of individual wood American Chemical Society, Seattle, fibers using environmental scan- Washington, March 20-25. ning electron microscopy and vi- HOFFMEYER, P., AND R. B. deo image analysis. TAPPI HANNA. 1989. Electron beam 78(5):143-148. damage during testing of wood in SHALER, S.M., L.H. GROOM, AND SEM. Wood Sci. Technol. 23: 211- L. MOTT. 1996. Microscopic analy- 214. sis of wood fibers using JENKINS, L. M., AND A. M. DO- environmental scanning electron NALD. 1997. Use of the microscopy and confocal Environmental Scanning Electron microscopy. In: Proc. Woodfiber- Microscope for the Observation of Plastic Composites Symp. ]Forest the Swelling Behaviour of Prod. Soc., Madison, Wisconsin pp. Cellulosic Fiber. Scanning 19 (2), 25-32. 92-97. TURKULIN, H.; J. SELL. 1989. KIFETEW, G. 1996. Some aspects Structural and fractographic study on the deformation behaviour of on weathered wood. An wood in relation to its structure. application of FE ESM microscopy Doctoralthesis, KTH, Stockholm, to the ”Thin strip” method. Sweden, TRITA-TRÄ R-96-19. Dübendorf, EMPA Bericht 115/36. KIFETEW, G. 1996. Application of ZINK, A. G., P. J. PELLICANE, AND the deformation field measurement C. E. SHULER. 1994. Ultrastructural method to wood during drying. . analysis of softwood fracture Wood Sci. Technol. 30: 455-462. surfaces. Wood Sci. Technol. 28: KIFETEW, G., F. THUVANDER, L. 329-338. 15(2), 135-163. BERGLUND, AND H. LINDBERG. 1997. The effect of drying on wood fracture surface from specimen loaded in wet condition. Wood Sci. Technol. 32:83-94

147 148 The influence of pith and juvenile wood on Chapter proportion of cracks in sawn timber when kiln dried and exposed to wetting cycles VIII D. Sandberg

first, a 30-minute wetting phase in Sweden is concerned, the major Subject a water bath and second, a parts of the juvenile wood have The presence of macroscopic succeeding drying phase in the been removed if timber is sawn cracks has been determined after same climate as in the earlier further away from the pith than 30 cycles of drying and wetting of described conditioning period. millimeters. The sawing pattern has Scots pine (Pinus silvestris L) and After the last wetting cycle the been chosen in such a way that it Norway spruce (Picea abies Karst). total length of cracks for each is possible to find boards in groups board was determined. The boards B and C with an annual ring Materials and methods were divided into three groups: pattern that is horizontal and the Six logs of Scots pine and seven A. Boards with the pith whole range of annual ring pat- logs of Norway spruce from the enclosed in the cross section in all terns up to the ones that are north of Sweden have been sawn of, or in parts of the length. vertical (Sandberg 1995). into 45 and 70 boards respectively, B. Boards sawn with a distance less all with the cross section dimension than 30 millimeters between the 50 by 100 millimeters. The boards pith and any part of the board. were kiln dried with a conventional Boards from group A are not drying programme to a moisture included in this group. content of 18 ± 3 %. After two C. Boards sawn with a distance weeks conditioning at a tempera- greater than 30 millimeters between ture of 21 ± 1° C and 39 ± 5 % the pith to any part of the board. RH, the total length of cracks in The distance 30 millimeters has every board was determined. been chosen as a borderline Thereafter, the boards were expos- between group B and C based on ed to three cycles lasting 20 days the anticipation that, as far as slow and nights, each cycle including, grown timber from northern

Figure 1. The variation of the average moisture content in the timber during the test.

151 Results other things, implies that the radial clearly related to the distance from Figure 1 shows variation of the shrinkage is roughly 50 % of the the pith. In addition, the propor- average moisture content of the tangential shrinkage. When the tion of cracks can be related to the timber during the test. The charts pith is enclosed in the cross section pith cracks, which are present in Figures 2a and 2b show the of the timber there is a demand before the log has been sawn or relative length of the cracks before that the timber has to be the reduction of the radius of and after the wetting cycles. The plasticized during the drying period curvature of the annual rings close relative length of cracks is here in order to make the shrinking to the pith. This reduction in- understood by the ratio between stress relax without causing cracks. creases the stress level during the total length of the cracks and The fairly great relative length drying as a result of the shrinking the length of the timber. of cracks in group B compared to anisotropy, which in turn results in As can be seen in Figure 2, the the ones in group C cannot be increased cracking. It is also relative length of cracks in the explained only by referring to the possible that the properties of the timber is being reduced by the anisotropic moisture movement juvenile wood affect the cracking distance from the pith for both found in wood. The similarity behaviour. spruce and pine. The wetting between group B and C, conside- cycles, however, resulted in an ring the dimensions and the annual References increased length of the cracks for ring patterns in the cross section, Sandberg, D. 1995: Vertical all the groups. This increase is also indicates that the amount of annual rings in pine (Pinus silvestris slightly bigger for pine than for cracks is not only dependent on L) and spruce (Picea abies Karst). spruce. The high relative length of drying factors. Royal Institute of Technology, cracks in group A is a This study, however, Stockholm, Wood Technology and consequence due to the shrinkage demonstrates that the proportion Processing, Report TRITA-TRÄ anisotropy of wood which, among of cracks in timber is a factor R-95-13.

Figure 2. The relative length of cracks regarding pine (a) and spruce (b) after drying and after the following cycles of wetting and drying. The categories A, B and C represent different distances between the sawn timber and the pith.

152 The influence of annual ring orientation on Chapter strength and dimensional changes during moisture variation in finger-joints IX D. Sandberg

The joins were glued using a moisture content of 8 %. Thereaf- rences in failure strength between Subject PVAc-glue (Bostic 742, harder 743) ter the boards were loaded to fai- the two groups, i.e. finger-joined This investigation elucidates the and the end grain was sealed with lure in a 4-point bending test. boards with the same annual ring significance of uniform annual ring silicon. The boards were conditio- orientation contra finger-joined orientation on the dimensional ned to a moisture content of 8%. Results boards with both vertical and hori- changes and strength of finger The length:width ratio of the Figure 2 shows the mean swelling zontal annual rings. joined boards. finger-join was 10:3.5. The cross- values of the surface of the boards sectional dimensions were measu- as a function of the distance from Materials and methods red on both sides of the finger- the finger-join. Since the tangential 30 radially sawn Scots pine (Pinus joins at intervals of 5 mm over a shrinkage and swelling due to silvestris L) boards free from knots distance of 50 mm and then at a moisture movement was about and visable defects with a cross distance of 75 mm and 100 mm. twice as large as the radial section of 45 by 45 mm and 900 The boards were then soaked in shrinkage, non-uniform swelling mm length, were cut into two water for 7 days, which resulted in occurs in the finger-join with cros- equal lengths. Each pair was a moisture content above fiber sed annual rings, i.e. in the finger finger-joined in two ways. saturation at a depth of at least 10 joined boards with both vertical Half the boards, i.e 15 pairs, mm from the surfaces of the and horizontal annual rings, were joined using the same annual boards. After wetting, the cross (Figure 2b). In finger-joins with a ring orientation in both pieces. In sectional dimensions were uniform annual ring orientation the the remaining 15 pairs, one of the measured again as described swelling was uniform as shown in boards was turned 90 degrees. above. The boards were then dried Figure 2a. This means a board with vertical using a conventional drying sche- As a result of non-uniform swel- annual rings was joined to a board dule (drying temperature 50° C) ling and shrinkage in the join when with horizontal annual rings, for 10 days followed by a condi- boards with different annual ring (Figure 1). tioning period of 3 weeks to a orientations are joined, stresses will develop in the join. These stress variations may, ultimately, lead to a weakened join. As for the appearance, a uniform annual ring orientation is preferable. If the joined boards are orientated in a such way that the radial surfaces are exposed and colourless glue is used, the join has a very homogeneous pattern. This implies that the width of the annual rings in the joined boards and the shade of colour of the wood are practically the same. The result of the strength test Figure 2. Mean values for the swelling in finger joined boards with the same annual ring orientation (a) and with a difference of 90 degrees in annual ring orientation did not show any significant diffe- between the joined boards (b).

155 Figure 1. Finger joined boards from the investigation. The upper board has the same annual ring orientation in the whole length. The lower board was joined by boards with vertical and horizontal annual rings.

156 Bending strength of I-beams with webs Chapter of wood-fibre board and flanges of star- sawn triangular profiles X D. Sandberg, M. Stehr

Subject Results References Table 1 shows the mean values for BFS 1993: Boverkets To determine some bending 23 beams loaded in bending to konstruktionsregler 94, Boverket strength parameters of I-beams failure. The results obtained are 1993:58, ISSN 1100-0856 (in Swe- with triangular flanges and wood- equivalent to an I-beam with a dish) fiber board webs. rectangular cross section flange Nordtest method 1987: Light- Material and methods having the same section modulus weight beams of wood, NT Build as the triangular flange. 327 Twenty three beams with wood- Compared to solid cross-section Sandberg, D. 1996: Radially sawn fiber board webs K-board K40 beams, an I-beam has the advan- timber. Star-sawing - a new (BFS 1993) and triangular flanges tage of providing the greatest mo- method for producing timber with of pine (Pinus silvestris L) prepared ment of inertia for the amount of vertical annual rings. Holz als Roh- by star-sawing (Sandberg 1996) material employed. Using an I- und Werkstoff 54(3): 145-151 were prepared and loaded in ben- beam with a triangular flange rat- ding to failure, (Figure 1). The her than a rectangular flange, it is flanges were considered to be a possible to achieve additional adv- structural lumber K24 or lower antages such as: (BFS 1993). The cross-sectional • A better material utilization, dimensions of the beams are given where a 7% small volume of timber in Figure 2. The flanges and webs is needed in a triangular flange to were glued using a resorcinol- give the same section modulus as a formaldehyde adhesive (Cascosinol rectangular flange. 1711, hardener 2520). • Since it is free from the un- The beams were loaded to fai- desirable juvenile wood, a beam lure using a four-point loading with a triangular flange can offer a technique (Nordtest-standard high density structure with a better 1987), and the failure load, modu- shape and dimensional stability lus of elasticity and ultimate mo- • The beam can offer a better ment to failure were determined. shape stability if the triangular During the experiment, the mean flange is glued to the web, since moisture content and the density the position of the annual ring of the flanges were 10.3 % and 417 orientation is maintained as shown 3 Kg/m respectively. in Figure 2.

Table 1. Strength values in bending for the beams in the test (mean values of 23 beams at 10.3 % moisture content)

Property Mean value Standard deviation Ultimate moment to failure (kNm) 11.2 3.3 Bending stiffness (kNm2) 392 66.9 Edge bending strength (MPa) 35.7 10.5 Modulus of elasticity (GPa) 10.9 1.7 Figure 2. Cross-section dimension for the Failure load (kN) 16.9 5.0 beam in the investigation. The length of the beams was 4.2 m.

159 Figure 1. Cross-section of a conventional light-weight beam (left) and a light-weight beam with triangular flanges (right).

160 Appendix

Sandberg, D.; Johansson, I. 1995: utilization of wood. TRITA-TRÄ Non-recoverable deformation in R-97-27 (in English) Appendix 1 Scots pine after bending and Sandberg, D. 1998: The influence The thesis is based on the one simultaneous variation in moisture of annual ring orientation on de- hand on a number of technical content. TRITA-TRÄ R-95-17 (in formation of clear specimens of reports from the Division of Wood Swedish) Scots pine in bending and Technology and Processing at the Carlsson, S.; Eskilander, S.; Sand- simultaneous variation in moisture KTH (KTH-Trä) which are largely berg, D. 1996: Evaluation of the content. TRITA-TRÄ R-98-32 (in written in Swedish, but with an deplacement method with water Swedish) English summary, and on the other for determination of dry wood Sandberg, D. 1998: Star-sawing - hand on articles published in density. TRITA-TRÄ R-96-21 (in Volume yield and quality from one scientific journals. The articles are Swedish) test sawing of pine. TRITA-TRÄ presented in chapter I - X, and the Holmberg, H.; Sandberg, D. 1996: R-98-33 (in Swedish) technical report are listed below: Volume yield and quality of timber Sandberg, D.; Holmberg, H. 1998: Nordfors, L.; Sandberg, D. 1993: when using Star-sawing technique. Distorsion and crack formation in Model for simulation of the volume TRITA-TRÄ R-96-22 (in Swedish) timber before and after drying - yield when sawing according to the Sandberg, D. 1996: Improved influence of the location in the log. new sawing pattern ”Star-sawing”. wooden windows. Annual ring TRITA-TRÄ R-98-34 (in Swedish) TRITA-TRT-1993-53 (in Swedish) orientation, cracks and heart wood Sandberg, D. 1998: Influence of Sandberg, D. 1994: Degradation - three important factors isostatic compression on the cell mechanisms of wood. 1. Photo- influencing the durability of timber structure. TRITA-TRÄ R-98-35 (in chemical degradation. A literature used in windows. TRITA-TRÄ R- Swedish) survey. TRITA-TRÄ R-94-10 (in 96-23 (in Swedish) Sandberg, D. 1998: Star-sawing- Swedish) Sandberg, D. 1997: Degradation The guiding-star in the Swedish Holmberg, H.; Sandberg, D. 1995: mechanism of wood. 2. Outdoor wood-working industry. TRITA- Frequency and character of knots exposed radial and tangential TRÄ R-98-37 (in Swedish) in triangular profiles of pine (Pinus surfaces of Scots pine and Norway silvestris L) and spruce (Picea abies spruce. TRITA-TRÄ R-97-24 (in Karst). TRITA-TRÄ R-95-12 (in Swedish) Swedish) Holmberg, H.; Sandberg, D. 1997: Sandberg, D. 1995: Vertical Volume yield in manufacturing annual rings in pine (Pinus silvestris gluelam board of Scots pine from L) and spruce (Picea abies Karst). star-sawn triangular profiles. TRITA-TRÄ R-95-13 (in Swedish) TRITA-TRÄ R-97-25 (in Swedish) Sandberg, D. 1995: Influence of Sandberg, D. 1997: Evulation of a annual ring orientation on crack conversion method for Star- formation and deformation in sawing. TRITA-TRÄ R-97-26 (in water soaked pine (Pinus silvestris L) Swedish) and spruce (Picea abies Karst) Sandberg, D.; Wålinder, M.; Wik- timber. TRITA-TRÄ R-95-16 (in lund, M. 1997: The concept of Swedish) Value Activation- a better

163 Appendix 2

Division of the work in the papers between authors Paper II: Sandberg initiated the work and mutually the authors performed the experiments, analysis, and wrote the paper. Paper IV: Sandberg initiated the work and mutually the authors performed the experiments and analysis. The paper was written by Sandberg. Paper VII: Sandberg initiated the work and mutually the authors performed the experiments and analysis. The paper was written by Kifetew. Paper X: Martin Wiklund initiated the work. Stehr performed the experiments and mutually the authors performed the analysis and wrote the paper.

164

Royal Institute of Technology - Wood Technology and Processing