HUMAN POWER TECHNICAL JOURNAL OF THE IHPVA NUMBER 50, SPRING 2000
Summaries of articles in this issue; mast ...... 2 Contributions to Human Power ...... 2 Articles On the efficiency of bicycle chain drives James B. Spicer and others ...... 3 Offset rims reduce the amount of dish Vernon Forbes ...... 10 Rolling resistance of bicycle tyres John Lafford ...... 14 Note on John Lafford’s paper and spreadsheet Jim Papadopoulos ...... 15 Reply to Jim Papadopoulos John Lafford ...... 18 Technical notes Power requirements for laid-back recumbents Bert Hoge and Jeroen Schasfoort in HPV nieuws Report and comment by Dave Wilson ...... 18 My propeller theory E. Eugene Larrabee ...... 20 Reviews Feet on!: pedal-powered museum exhibit Michael Eliasohn ...... 21 Human power: the forgotten energy Arnfried Schmitz, reviewed by Dave Wilson...... 22 Editorial Cycle Vision, Ronald van Waveren ...... 23 Letters Number 50 Update, Anil Rajvanshi ...... 14 Supplement to velocar variations, Arnfried Schmitz. . . . . 22 Spring 2000 $5.50 Kudos, Chet Kyle ...... 23 On the efficiency of bicycle chain drives James B. Spicer,* Christopher J.K. Richardson, Michael J. Ehrlich and Johanna R. Bernstein HUMAN POWER The Johns Hopkins University, Baltimore, Maryland 21218 Number 50 Spring 2000 $5.50/IHPVA members, $4.00 Masahiko Fukuda and Masao Terada Shimano Inc., Product Engineering Division, Sakai Osaka 590-77 IN THIS ISSUE beautifully produced Netherlands HUMAN POWER counterpart to Human Power: HPV Bicycle chain transmissions nieuws. They showed, by testing a range ABSTRACT is the technical journal of the Jim Spicer and his associates at Johns of Dutch recumbent bikes, that the The efficiencies of bicycle drive International Human Powered Vehicle Hopkins have written a paper that will aerodynamic drag decreases as the angle Association change minds, and design directions, on trains have been studied to understand of reclining increases. Number 50, Spring 2000 HPV transmissions. To give just one energy-loss mechanisms in these sys- Editor example: the losses associated with small My propeller theory tems. An analytical study of frictional David Gordon Wilson sprockets on the rear wheel make “mid- Gene Larrabee, whom Human Power energy loss along with a series of named “Mr. Propeller” many years ago, 21 Winthrop Street drives” or countershaft gears suddenly summarizes his propeller theories and the experimental efficiency measurements Winchester, MA 01890-2851 USA attractive, and possibly hub gears too. developments that have resulted from of derailleur-type chain-drive systems [email protected] They also find that chain lubrication doesn’t seem to reduce losses: that will be them. He also pays tribute to those who under a range of power, speed and Associate editors even more controversial. (Your editor has inspired him and gave him the basic lubrication conditions are given to Toshio Kataoka, Japan high confidence in the data: he asked Chet theories on which he built. Was Isaac identify loss mechanisms. These mea- Newton the first to state that we all stand 1-7-2-818 Hiranomiya-Machi Kyle, IHPVA founder and one of the surements of mechanical efficiency are Hirano-ku, Osaka-shi, Japan 547-0046 foremost researchers in bicycle perfor- on the shoulders of giants? compared to infrared measurements [email protected] mance in the world, to look at them before Feet-on! review publication. He had done a proprietary Mike Eliasohn writes a delightful review performed during drive operation that Theodor Schmidt, Europe study on the same topic, and has produced of what sounds to have been an equally show the heating of drive components Ortbühlweg 44 broadly similar results.) delightful and truly interactive museum resulting from frictional losses. The CH-3612 Steffisburg, Switzerland Offset rims exhibit devoted to human power. [email protected] results of this study indicate that chain Vernon Forbes gives us another of his Human power: the forgotten energy tension and sprocket size primarily careful studies of spoked-wheel construc- Your editor reviews a small fascinating Philip Thiel, Watercraft determine chain-drive efficiency. 4720 - 7th Avenue, NE tion incorporating derailleur clusters or book by Arnfried Schmitz, already well- Seattle, WA 98105 USA brake disks that cause the wheel to be known in these pages, on the origins of “dished” (spoked asymmetrically). He HPVs in France, on the characters of the INTRODUCTION Production protagonists, and (very modestly) on his shows that the use of rims that have off- When this study was performed, it Fig. 1. Experimental schematic of test stand showing elements of the drive JS Design & JW Stephens center spoke holes brings about a consid- own part in the revival of interest in these was hoped that through identification assembly erable reduction in the difference in spoke wonderful vehicles. IHPVA of the loss mechanisms primarily tensions that otherwise makes highly Editorial from the Netherlands Paul MacCready, Honorary president responsible for limiting the efficiency tion (Tordion 1996). Since some of these tions (Juden 1997). Without detailed Chris Broome, USA, Chair dished wheels prone to spoke failure and Ronald van Waveren, chair of the occasionally to collapse. of bicycle chain drives, methods for are dynamic effects, the work presented analysis of Keller’s results, it appears Ben Wichers Schreuer, The Netherlands, Netherlands HPV association, writes a Vice-chair, Rolling resistance guest editorial about the astonishing improving efficiency could be realized here was conducted on chain drives that they do not wholly support the idea numbers of recumbent bicycles and HPVs Open, Secretary/treasurer John Lafford has tested, on equipment by eliminating or decreasing the vari- during operation. that the efficiency is governed by fric- of his own design and construction, a in the Netherlands, and in particular about ous losses. Unfortunately, as will be Hollingworth and Hills (1986a) per- tion. Work by Kidd, Loch and Reuben Publisher prodigious number of bicycle tires, mostly the annual celebration known as “Cycle shown, definitive identification of these formed a detailed theoretical and (1998) has attempted to quantify drive IHPVA of a size particularly suited to the front Vision.” Through delays in our publication PO Box 1307 wheels of recumbent bikes. He measured we are too late to encourage you to visit mechanisms has not been successful. experimental study of chain contact efficiency in relation to models of drive- San Luis Obispo, CA 93406-1307 USA rolling resistance and power loss over a this exciting event this year, but we hope Rather, the results provide information forces during link articulation in heavy- train performance based on chain con- Phone: +805-545-9003; [email protected] range of speeds and inflation pressures. that you will do so next year…. on the efficiency of chain drives and, at duty chain drives and used these tact forces. While static measurements Jim Papadopoulos wrote a commentary on Letters include kind words from Chet Human Power (ISSN 0898-6908) is pub- the same time, lead to conclusions results to calculate the theoretical effi- of these forces have agreed with mod- the results, and John Lafford responded, Kyle; comments by Anil Rajvanshi on the lished irregularly for the International eliminating those mechanisms that do ciency of chain drives assuming that els, efficiency-measurement results all in this issue. publication of his article on rickshaws in not dominate efficiency. We hope that frictional losses primarily affected effi- have not appeared in the literature Human Powered Vehicle Association, a Power requirements for the last issue (Human Power 49); and non-profit organization dedicated to pro- unfaired laid-back recumbents comments and corrections by Arnfried the results here contribute to the over- ciency (Hollingworth and Hills, 1986b). (Kidd et al. 1996). moting improvement, innovation and cre- Bert Hoge and Jeroen Schasfoort wrote a Schmitz on his article “Velocar variations”, all body of work on chain-drive effi- They found that the efficiency of the In this work, the efficiencies of bicy- ativity in the use of human power gener- short but valuable technical note in the also in the last issue. ciency and also to the on-going discus- chain drive should increase with the cle chain drives are investigated both ally, and especially in the design and sion of this topic (Cameron 1999; Wil- number of sprocket teeth on each of experimentally and theoretically to iso- development of human-powered vehicles. CONTRIBUTIONS TO HUMAN POWER son 1999). the driver and driven sprockets. Unfor- late factors associated with loss in Material in Human Power is copyrighted The editor and associate editors (you may choose with whom to correspond) welcome Even though design of chain drives is tunately, no experimental results were these systems. A computer-controlled by the IHPVA. Unless copyrighted also by contributions to Human Power. They should be of long-term technical interest. News and fairly well-understood (Vogwell 1994) given to verify their models. drive-train-testing system was designed the author(s), complete articles or repre- similar items should go to HPV News or to your local equivalent. Contributions should be the factors affecting efficiency have In the work by Keller (1983) measure- to measure the performance of the sentative excerpts may be published else- understandable by any English-speaker in any part of the world: units should be in S.I. been considered only in passing as part ments of efficiency were made for dif- chain, chain ring and rear cassette in a where if full credit is given prominently to (with local units optional), and the use of local expressions such as “two-by-fours” should the author(s) and the IHPVA. Individual be either avoided or explained. Ask the editor for the contributor’s guide (available in of the design process. Generally design ferent transmission systems (derailleur, derailleur-type system. This system was subscriptions and individual issues are paper, e-mail and pdf formats). Many contributions are sent out for review by specialists. factors that are considered include internally geared hub, fixed), using used to measure chain-drive perfor- available to non-IHPVA and non-HPVA Alas! We cannot pay for contributions. Contributions include papers, articles, technical chain length, load ratings, roller impact chains exhibiting various conditions of mance under a variety of operating members. notes, reviews and letters. We welcome all types of contributions from IHPVA members velocity, rotational forces, contact repair (new, used, non-lubricated), conditions. Assuming that frictional and from nonmembers. forces, chordal action and chain vibra- under a variety of power-transfer condi- forces degrade the overall efficiency of
2Number 50, Spring 2000 Human Power Human Power Number 50, Spring 2000 3 the system, simple analytical models the pin/bushing interface, ρ is the pin Sprocket tooth, These results indicate that configura- measured using a speed sensor. The images of drive components while the for the losses of chain drives have been radius, T0 is the free chain tension, ϕ is link roller and inner link bushing tion A should have 63% more power loss drive-shaft torque was measured using drive was in operation. Since frictional developed to estimate and identify the the pressure angle (for new chains A similar analysis for the losses at than C and that configuration B should a rotary transformer torque sensor. losses result in heating of the drive primary mechanisms of frictional loss. equal to [35°–120°/N] where N is the the tooth/roller/bushing interfaces can have 28% more power loss. For exam- The chain drive was configured to the components, the infrared camera was These models for drive losses have number of teeth on the sprocket), and be performed with the following ple, if a test of efficiency indicated a 5% geometry found on bicycles with the useful in identifying those components been used to interpret experimental α is the articulation angle (equal to results: power loss in configuration C, then con- distance between the front chain ring that had the highest temperature rises. results. Additionally, since losses due 360°/N). The subscripts on the angles (4) figuration A should suffer an 8.2% loss and rear cassette being adjusted by The primary component of this system
2 to friction ultimately are manifested as refer to the front chain ring, 1, and the N11ω and B should have a 6.4% loss. mounting the entire output-shaft assem- is the infrared camera that operated PTrfR2Total=∑µ20 i heating of the drive components, rear sprocket, 2. In deriving Eq. (1), it 2π i=1 bly on a translational platform. with an InSb planar array sensitive in π π π 2 + ψϕ −+ ϕ 22ψ −+ ψϕ µ infrared images of the operating chain was assumed that the pin and bushing cosii sin1n cos sin cot i DESCRIPTION OF Mounting the output drive components the 3–5 m range. Again, using NNNi ii drive were taken. have a neat fit (all surfaces were THE EXPERIMENTAL APPARATUS on this platform allowed for accurate LabView software to control the assumed to be in contact) and that the where µ2 is the coefficient of friction at To assess drive efficiency, a test adjustment of the cassette for distance camera and the image-acquisition- THEORY coefficient of friction was a constant, the bushing/roller interface, rRi is the stand was designed to measure the and offset from the front chain ring. board operations, thermal images of If it is assumed that friction between independent of the chain tension. inner radius of the roller, and ψ is the overall or average mechanical efficien- Additionally, the derailleur unit the drive were acquired and stored for contacting components performs work The rate at which this work is per- roller rotation angle (the angle through cy of the chain drive from the front (Shimano® Dura Ace®) could be analysis and display. For these opera- during drive operation, then power loss- formed represents the power loss which the roller executes no-slip chain ring to the rear transmission adjusted independently to satisfy rec- tions it was important to acquire es from the drive necessarily reduce resulting from friction, Pf . The average motion on the tooth). Since the pres- components. To assess the efficiency of ommended mounting conditions. The images of the drive so that the drive efficiency. The normal force producing power dissipated per link by this sure angle depends on tooth number, a the drive under various conditions, the output shaft torque was measured using component of interest occupied the friction is related to the chain tension in source is written using the period of simplified form for the variation of the power input to the front chain ring was a second rotary torque sensor. The same position in the infrared image the link during articulation and engage- chain revolution (expressed in terms of power loss with sprocket combination measured and was compared to the entire drive system was loaded using an from frame to frame. By acquiring ment. An analysis for this tension can the drive-sprocket angular frequency) is difficult to obtain from Eq. (4). power that was output by the rear electromagnetic brake mounted to the images in this manner, the heating of be found in the work by Tordion (1996) along with Eq. (1). The resulting Since both of these loss mechanisms drive sprocket. The ratio of the output output shaft. an individual component, such as a and in the work by Kidd et al. (1998). expression must be multiplied by the involved friction between elements of power to the input power was used to There are three signals that are single chain link, could be tracked Since the friction depends on chain ten- number of links in the chain to obtain the chain, it can be assumed that the quantify the overall efficiency of the recorded under computer control in a accurately as a function of time. sion, there are perhaps two major loca- the total power loss for this mecha- coefficients of friction are approxi- system. To determine the powers in the LabView®† programming environ- tions for loss in the drive that should be nism. The following result is obtained: mately equal such that a total power drive and driven shafts, the torques ment: input and output torques and EXPERIMENTAL PROCEDURE identified beforehand since the chain (2) loss can be written for the chain drive. transmitted by the shafts were mea- input rotation rate. The torque-sensor Four major areas for investigation were tension is large and is transferred from PNWff1111Total =ω/ 2 π This power loss is obtained by adding sured along with the rotation rate of differential outputs were amplified identified and were pursued. the chain to the sprocket at these loca- the losses given as follows: the drive shaft. Knowing the rotation using low-noise instrumentation ampli- Time variation. Measurements of tions. These include the surfaces Note that the functional dependence (5) rate of the drive shaft and the gear fiers. Since the torque-sensor outputs efficiency were made over extended ω π 2 + ()αϕ() between the inner link bushing and of Pf1Total on the articulation angle and, N11 122tanii / cot / ratio, the rotation rate of the driven were harmonically varying signals with run periods (2 hours) to determine PTfTotal= µ 0 ρϕ∑ sin i1n + π = − ()αϕ() chain pin and between the sprocket consequently, on the number of teeth 22i 1 122tanii / tan / shaft could be calculated. The efficien- amplitudes proportional to the torque, whether or not efficiency varied as a 2 tooth, link roller and inner link bushing. on the drive and driven sprockets is rnRij∑+(()αψcos ϕ j− sin ϕ j1 cos() αψ j+ ++ cot ϕ jj sin() αψ)} cy of the system was calculated using the signals were measured using lock- function of time during drive = Rather than derive in detail the func- not clear owing to the form of Eq. (1). j 1 the following formula: in techniques to improve the signal-to- operation. tional form for the losses, the results of However, if the number of teeth on the where µ is the common coefficient of (7) noise of recorded signal amplitudes. Configuration. Efficiency was mea- τω22 models will be presented to give the sprockets is large such that the articu- friction. It is noted that the total power %%Efficiency =×100 To relate these reader a feeling for the anticipated lation angles are small, tan(α/2) ≅α/2, loss per sprocket is reciprocally related τω11 signals to actual results of experiment. The interested then the expression can be simplified to the tooth number if the roller angle torque values, reader can find the full derivation else- such that the total power dissipated vanishes. Using this expression for the where τ1, ω1 are the torque and the each of the torque where (Spicer et al. 1999). becomes: total power loss where N1 = 52, three angular frequency, respectively, of the sensors was cali- Inner-link bushing and chain pin (3) separate configurations are examined drive shaft, and τ2, ω2 are the torque brated using Since the chain tension is large on π and are given as follows: and the angular frequency of the driven known static ≈+ωµρ 11 only one side of the drive, between the PNf 1111Total T0 Configuration A: N2 = 11 shaft. To gather the data required for loads. By 2 NN12 front chain ring and the rear sprocket, Configuration B: N2 = 15 the efficiency calculations, the test measuring the this loss has significant contributions A similar analysis can be carried out Configuration C: N2 = 21 stand was automated using computer torque-sensor only at two points. One of these is at for the effects of chain offset with the These configurations were chosen to control. The essential elements of the signals as a func- engagement on the front chain ring and result that the power lost as a result of reflect situations that were realized test stand are shown schematically in tion of applied the other is at departure on the rear offset has a form nearly identical to experimentally in this study. Assuming figure 1 (see page 3). torque, a calibra- sprocket. Adding the contributions that given for friction at the pin-bush- that the roller angle is ψ ≅ 5.6° and also The drive shaft was driven by a vari- tion curve was from these two points, the total work ing interface except that a factor of the assuming that the geometrical pre-fac- able-speed electric motor system con- obtained that resulting from friction can be offset angle appears in the expression tors for each of the losses are approxi- nected to the drive shaft. The drive- related signal level expressed as follows: for offset losses. Since this angle is mately equal yields the following shaft rotation rate from this system to the applied (1) small, the frictional effects of offset results: could be varied continuously under torque. Figure 2. Measured chain-drive efficiency as a function of time for 2 π 122+ tan()αϕii / / tan() / should be small compared with (6) manual control. However, once the Finally, an WTfi11= µρ 0∑ sin()ϕ 1 n three different drive configurations (52–11, 52–15 and 52–21) in = − ()αϕ() pin/bushing losses. Any effect would and desired rotation rate was set, the rate infrared-camera 2 i 1 122tanii / tan / PfTotalA ≅ PfTotalB ≅ the no-offset condition. Tests were run for two hours, but only the necessarily appear in the largest offset 163. 128. was not actively controlled and the system was used first 100 minutes are shown here. Note that efficiency is fairly con- PfTotalC PfTotalC where µ1 is the coefficient of friction at conditions. actual rotation of the drive shaft was to acquire thermal stant during the entire time period.
4Number 50, Spring 2000 Human Power Human Power Number 50, Spring 2000 5 sured as a function of gear ratio in Table 1. with measurements of efficiency being efficiencies at high chain tensions the previous trends for efficiency as a the temperature of the component, the (52–11 etc.) and the effects of offset First, comparing the results here made sequentially at 5–8 minute inter- exceed 100% by a small amount (a few function of configuration and as a func- pixel intensity is directly proportional were investigated to assess the with those in the long-duration tests, vals. All recorded values for efficiency percent in the worst case). tion of chain tension are still followed. to temperature. This method for data effects of drive configuration. the effect of chain offset can be esti- have a precision (due to short-term Clearly, these data indicate that the However, these results also indicate acquisition allows the temperature Power/rate. Efficiency variations with mated. These data were obtained with noise) of ±0.2% and a long-term preci- fundamental operation of the drive that the actual lubricant used has little input power and rotation rate were the 52–11 and 52–21 configurations in sion of ±0.3% under normal operating must be related to the chain tension effect on the overall performance of 52–21: 60 RPM, 150W measured to determine if load or the offset condition while those in the conditions. such that the efficiency increases with the drive under laboratory conditions rate-dependent effects were present. long-duration tests were taken with no Again, the efficiency results showed increasing tension even though the given the precision of the measure- Lubrication. The effects of lubrication offset. Comparing the data for 60 RPM that higher efficiencies are obtained for frictional losses should increase. This ment. In addition, the chain used for and de-lubrication on efficiency were 100 W tests shows that the offset low- those configurations that have larger experimentally measured dependence the lubrication study was fully quantified. ers the efficiency by, at most, 0.5% sprockets in combination. It was also of efficiency on chain tension can be degreased and was re-tested for effi- when measurement precision is consid- found that the efficiency decreases explained only in a limited sense using ciency. This degreasing operation con- EXPERIMENTAL ered. Additionally, if the efficiencies are with increasing rotation rate for con- the models for loss developed previ- sisted of a five-minute scrub with RESULTS AND ANALYSIS normalized by efficiencies measured in stant power input and that efficiency ously. For example, if the pressure kerosene followed by a cleaning with
Time variation of efficiency the 52–15 configuration (both sets of increases with increasing power for angle changes with tension, then the Castrol Degreaser. The measured effi- 4:00 minutes A new chain used for these tests was data were obtained with no offset), constant rotation rate. Both of these calculated losses will vary with tension ciency of the de-lubricated chain for cleaned using Simple Green™ Bike then it appears that the offset has a data trends can be related to the torque in a manner not considered previously. the 52–15 combination at 60 RPM and Cleaner/Degreaser and was lubricated negligible effect on efficiency. applied to the drive and ultimately to Measurements of link tension during 100 W was 90.3% and at 200 W was using Generation 4: White Lightning™ More importantly, in these mid-dura- the chain tension. Analyzing the effi- articulation are currently being 96.5%. These efficiencies are essentially self-cleaning lubricant. These tests tion tests, the efficiencies show a con- ciency as a function of tension shows pursued using noncontacting optical the same as those measured for the lasted 120 minutes at 60 RPM 100W for sistent dependence on rear sprocket that the efficiency increases with chain measurement techniques to investigate chain in the re-lubricated condition. each of the following chain-drive size where the larger the sprocket, the tension regardless of input power or these effects. Infrared measurements configurations: 52–11 no-offset, 52–15 higher the measured efficiency regard- rotation rate. Additionally, for a given Lubrication during chain-drive operation no-offset and 52–21 no offset. As is less of the selected power or rotation chain tension, the efficiency was found In this phase of the study, the chain Infrared images of the chain drive shown in figure 2, the efficiency for all rate. If the efficiency for 52–21 is 95.2% to be independent of drive rotation rate was thoroughly degreased/cleaned during operation aid the interpretation long-duration tests showed little-to-no (as is found for 50 RPM 100 W), then (between 40 and 80 RPM). The depen- using commercially available degreas- of the efficiency measurements that long-term efficiency variations during the previous modeling results predict a dence of efficiency on chain tension is ing agents (Castrol Wrench Force have been presented and also support 7:30 minutes the tests. The measured efficiencies difference in efficiency of 2.6% between shown in figure 3 where the drive effi- Degreaser™ and/or Simple Green™ aspects of the modeling of chain-drive for the three configurations are as the 52–21 and 52–11 combinations. ciency is shown as a function of the Bike Cleaner/Degreaser) and was relu- loss since frictional losses should heat follows: From the data in the table, the mea- reciprocal chain tension. bricated using one of three commer- the drive components. Simply put, the 52–11 no-offset: 91.4% sured difference between the 52–21 This graph shows that for the ten- cially available lubricants (Castrol chain components responsible for 52–15 no-offset: 93.2% and 52–11 combination is 2.7%. These sions investigated in this study (76.2 to Wrench Force Dry Lube™, Pedro’s Syn mechanical loss should heat the chain 52–21 no-offset: 95.0% results indicate that the primary mech- 305 N) that the reciprocal of the ten- Lube™ or Generation 4 White Light- and cause its temperature to rise. Since These values represent an average anism for chain-drive loss could be fric- sion is linearly related to the drive effi- ning™). the average heat deposition rate equals over the duration of the test. As a tion at the pin/bushing interface and ciency. For each of the linear fits to the The results in table 2 indicate that the average mechanical power loss, the result of these long-duration tests, friction at the bushing/roller interface. experimental data, the correlation coef- heating should have a functional subsequent tests were run for no more Power and rotation rate ficient exceeds 0.996. The extrapolated Table 2. Efficiencies for different drive dependence similar to that found for than 30 minutes to assess efficiency or To investigate the rotation rates and sprocket configura- frictional losses. efficiency variations. effects of power and tions (input power 100W) To acquire infrared images, the chain 11:00 minutes Configuration rotation rate on drive- 2.1 Castrol Dry Lube™ drive was set initially to a low rotation The next series of tests investigated train efficiency, more 40 RPM 60 RPM 80 RPM rate and the components were allowed the effect of chain configuration on extensive measure- 52–11 92.8 89.4 84.2 to reach a steady state in this mode of efficiency. These tests lasted 30 ments of efficiency 52–15 94.0 90.9 86.5 operation. The rotation rate and brake minutes each and were conducted with under varying powers 52–21 95.2 92.0 88.3 resistance were then increased rapidly the 52–15 combination in the no-offset and rotation rates to achieve the desired rate and power condition. Consequently, the 52–11 and were completed. The 2.2 Pedro’s Syn Lube™ settings (100 W 60 RPM, etc.). Infrared 52–21 combinations were tested in an chain was the same as 40 RPM 60 RPM 80 RPM image acquisition began prior to (or offset condition as would occur for a that used in the other 52–11 93.6 89.9 85.6 during) the time of power and rotation properly configured bicycle. The tests and the lubrica- 52–15 95.6 92.6 88.8 increase. Images were acquired at set results of these tests are summarized tion was not changed 52–21 95.3 92.6 89.0 time intervals (e.g., approximately 30 seconds) for the duration of the test 15:00 minutes Table 1. Drive efficiencies for different chain configurations 2.3 Generation 4 White Lightning™ such that actual image acquisition 50 RPM 60 RPM 70 RPM 60 RPM 60 RPM 40 RPM 60 RPM 80 RPM occurred synchronously to a particular 100 W 100 W 100W 150 W 175W 52–11 – – – link in the chain. Since pixel intensity 52–11 92.5 91.1 88.7 94.6 95.5 Figure 3: Variation of chain-drive efficiency with the reciprocal of 52–15 94.2 91.1 87.2 is proportional to the infrared energy 52–15 94.7 92.3 90.4 96.2 97.5 the chain tension. For this graph, chain tension has been calcu- Figure 4: Infrared images of chain drive dur- 52–21 – – – 52–21 95.2 93.8 92.0 97.4 98.2 lated using the measured torque values and the radius of the reaching the detector and the amount ing operation showing effects of frictional front chain ring. of infrared emission is proportional to heating.
6Number 50, Spring 2000 Human Power Human Power Number 50, Spring 2000 7 increases in various components to be in the 52–21 configuration. These Fig. 5 (97.2% for 150 W, 94.4% for 100 W tested using three different chain lubri- instruments by computer. National energy loss location in bicycle chain mapped as a function of time after results clearly show that the chain pin and 85.5% for 50 W) indicates that 4.2 cants under a variety of test configura- Instruments Corp. See drives. Submitted to the Journal of imposition of the heat load. rises in temperature more rapidly than W of power were lost at 150 W input; tions. No significant quantifiable effect http://www.ni.com/labview for ven- Mechanical Design. Representative results are shown in the other locations regardless of the 5.6 W at 100 W input and 7.3 W at 50 W of lubrication could be inferred from dor information. Tordion, G.V. 1996. Chains. Mechanical figure 4 where a series of four infrared input power and indicate that heating input. Obviously, for a lower power these tests. design handbook. H.A. Rothbart, ed. images from efficiency tests for the results from frictional losses near the loss, the temperature rise should be Infrared measurements of drive com- REFERENCES McGraw Hill, 25.1–25.12. 52–21 configuration are shown. At the pin. lower if the lost power is converted ponents indicate that frictional losses Cameron, A. 1999. Measuring drive- Vogwell, J. 1994. Chain drives. Rotary beginning of the test, the infrared The data for the chain and the pulley entirely to heat by frictional loss. It in the chain cause the chain tempera- train efficiency, Human Power power transmission design. images showed only variations of emis- tooth indicate that the component tem- would be expected that the tempera- ture to rise during operation. This 46:5–7. K. Hurst, ed. McGraw-Hill Europe, sivity since all components were in perature rises with the input power. ture rise would be lowest for the 150 W increase in temperature did not corre- Hollingworth, N.E., and D.A. Hills. 79–95. thermal equilibrium. From these For the pulley tooth, at 50 W input input test since the measured power late with measured efficiencies under 1986a. Forces in heavy-duty drive Wilson, D.G.1999. Transmission effi- images (taken at the times indicated) power, the maximum pixel value is loss is lowest for this case. various conditions of operation. chain during articulation. Instn ciency, Human Power 48:20–22. various qualitative observations are approximately 23 units; at 100 W, 35 Infrared measurements on lubricated Mech. Engrs, 200:C5, 367–374. ——— given as follows. units and at 150 W, 65 units. These DISCUSSION AND CONCLUSIONS and delubricated chain links showed Hollingworth, N.E., and D.A. Hills. *Corresponding author: 1. The chain pins rapidly heat on the results are in rough agreement with the Tests of efficiency for the derailleur- that the frictional heating did not 1986b. Theoretical efficiency of James B. Spicer, Associate Professor radial surfaces and reach near loss models presented previously type chain drive indicate that the over- depend on lubrication. cranked link chain drive. Instn Mech. Dept. of Materials Science & steady-state temperatures within 4–5 where the frictional losses are directly all efficiencies for the transfer of From the results of this study, it Engrs, 200:C5, 375–377. Engineering minutes. proportional to the input power. power from the front drive sprocket to appears that the efficiency of the bicy- Juden, C. 1997. Measurements of effi- The Johns Hopkins University, 2. At high chain tensions and low rota- Unfortunately, the power loss in the rear sprocket range from 80.9% to cle chain drive depends intimately on ciency of chain and shaft drives. 3400 N. Charles St., Room 102 tion rates, both the guide and the each of these cases is not proportional 98.6% depending on the conditions of the chain operation as it engages and http://www.soroos.net/hbs/chains/sn1 Maryland Hall, tension pulleys heat primarily from to the input power owing to the depen- drive operation. Primary factors affect- departs from the sprockets on the high- _2_2.html, (1997). Baltimore, MD 21218, USA the chain. dence of efficiency on chain tension. ing the efficiency include the sizes of tension part of the drive. Owing to the Keller, J. 1983. Der Wirkungsgrad im [email protected]; 410-516-8524 3. Steady-state temperature conditions Using measured values for efficiency the sprockets in the drive and the ten- high efficiencies measured under high Fahrradantrieb. Radmarkt, 12:71–73. The Hopkins group pursues work in for the chain drive are established under the conditions for the data in sion in the chain. chain tensions, friction can account for Kidd, M., N. Loch, and R. Reuben. 1996. pulsed and ultrafast (femtosecond) within approximately 15 min- It was found that larger sprockets only a few percent of the overall losses. Bicycle chain efficiency. The laser-based materials processing and utes after the beginning of a provide more efficient transfer of Most probably, mechanical losses that Engineering of Sport, S. Haake, ed. characterization. In contrast, the test. power while smaller sprockets proved are not converted to thermal energy in Rotterdam: Balkema, 217–220. expertise of the Shimano participants These observations are in gen- to be less efficient. Simple, frictional the drive account for the remainder of Kidd, M., N. Loch, and R. Reuben. leans towards the design and manu- eral agreement with the proposed loss models were developed that gave the measured loss. 1998. Experimental investigation of facture of cycling components. All of frictional-loss mechanisms. The sprocket-size loss variations that bicycle chain forces. Submitted to the authors express their gratitude to rapid heating at the radial pin sur- agreed with those variations measured NOTES Experimental Mechanics. the Baltimore Orioles for the use of faces is consistent with heating at experimentally. Typically, a 2–5% loss †LabVIEW®, National Instruments Spicer, J., M. Ehrlich, J. Bernstein, Camden Yards where many fruitful the pin-bushing interface. At low difference was measured between the Corporation is a computer software C. Richardson, M. Fukuda, and technical discussions about this work chain tensions, the frictional loss- 52–11 and the 52–21 sprocket combina- program useful for running scientific M. Terada. June 1999. Efficiency and were held. es should be relatively low and tions depending on the drive operating the temperature rise in the chain conditions. SUMMARY OF EQUATIONS should be low. At high chain ten- Experimental results indicated that sions and low rotation rates, the the efficiency of the chain drive varied 6 PfTotalA 2 frictional losses in the chain as a function of chain tension. It was π 122+ tan()αϕii / / tan() / ≅ 163. WTfi11= µρ 0∑ sin()ϕ 1 n PfTotalC should be relatively high produc- found that the efficiency varied linearly 1 2 i=1 122− tan()αϕii / tan() / ing large temperature rises in the with the reciprocal of the average and chain. The pulley-bearing losses chain tension with the highest efficien- =ω π PfTotalB should be relatively low. Even cies occurring at high chain tensions 2 PNWff1111Total / 2 ≅ 128. though the infrared images pro- and lowest at low chain tensions. For PfTotalC vide a wealth of qualitative infor- example, the highest efficiency mea- π ≈+ωµρ 11 τω22 mation, quantitative analysis of sured in the study, 98.6%, was mea- 3 PNf 1111Total T0 7 %%Efficiency =×100 NN12 τω chain-drive heating must be per- sured at a chain tension of 305 N and 2 11 formed using the temporal evolu- the lowest, 80.9%, at 76.2 N. ω 2 π π π tion of pixel intensity. It was found that chain-line offset =∑+µ N11 2 ψϕ −+ ϕ 22ψ −+ ψϕ 4 PTrfR2Total20 i cosii sin1n cos sin cot i In Fig. 5 the infrared pixel and chain lubrication have a negligible 2π i=1 NNNi ii intensity for positions on a chain effect on efficiency under laboratory pin, on a pulley tooth and on the conditions. Calculations of frictional 2 N11ω π 122+ tan()αϕii / cot() / body of the pulley (midway loss resulting from offset indicate that 5 PTfTotal= µ 0 ρϕ∑ sin i1n + π = − ()αϕ() between the bearing and the pul- this loss should be small compared to 22i 1 122tanii / tan / Figure 5: Variation of infrared emission with time ley teeth) are shown as a function during chain operation for different drive compo- those produced by other mechanisms. 2 of time for different input powers. nents. At time equal to 0 s, the chain drive was This was verified experimentally. Lubri- rnRi∑+(()αψ jcos ϕ j− sin ϕ j1 cos() αψ j+ ++ cot ϕ j sin() αψ j )} = These data were taken at 60 RPM placed under full power loading. cation effects on chain efficiency were j 1
8Number 50, Spring 2000 Human Power Human Power Number 50, Spring 2000 9 Offset rims reduce the amount of dish ever, this changes the chainline and mm and the average for an 8-sp hub was 135-mm-long axle (locknut-to-locknut) unless the bottom bracket is also 55.3 mm between the flanges, center-to- as the overall length. by Vernon Forbes widened it can result in difficult shift- center; Forbes 1998–99). A 135-mm axle ing and premature chain wear. Another length, locknut-to-locknut (see figure 1) RESULTS ABSTRACT hubs are laterally weaker than wheels distances are those from the plane of solution to the problem of the DS being was used at all times. Each wheel was The chief interest for us was the Offset rims reduce the amount a built with wide hubs. The idea they hit the rim centerline to the outside sur- at a higher tension than the NDS is to tested at two different levels of dish difference in tension between the DS rear wheel is dished. Two offset rims upon was a compromise; they spread faces of the hub flanges. Increased free- use a “dishless” hub. A “dishless” hub approximating the mean freewheel and the NDS. The mean spoke tension are tested and the results compared to the rear dropouts a little and narrowed wheel width increases the severity of must be a narrower hub. It has the NDS thicknesses of 7-sp (44.5 mm) and 8-sp for each side of the wheel was first a standard rim. Possible implications the rear hub a little so they could move dish. The newer eight-speed freewheels flange moved in closer so that it is the (48.0 mm) freewheels established previ- determined. Tension here is expressed are discussed. the whole hub over between the require more dish than did older five same distance from the rim center as ously (Ibid.). Dish was changed by mov- as nominal tensiometer readings, not dropouts to make room for the free- speeds. Increased dish results in DS the DS flange. However “dishless” hubs ing spacers from one side of the axle to as Newtons or poundals. To measure INTRODUCTION wheel. spokes being tighter. Cyclists are long that are wider are known to be laterally the other and redishing the wheel, e.g., dish the tension on the NDS (which “Dish” in a bicycle wheel is a mea- familiar with the increased tendency for stronger than narrow, dishless hubs. the axle was first spaced for a 7-sp spac- was presumably less tight) was sub- sure of its lack of symmetry. Usually spokes to break on the freewheel side Recently, offset rims (Ritchey, Bon- ing on a 135-mm axle. The wheel was tracted from the tension on the DS to this is given as a ratio of two lengths (Forbes 1998–99). One spoke manufac- trager) have been developed that have then dished using a Wheelsmith dishing determine the difference in tension. (defined later). Sometimes dish is also turer recommends 785-1079 N (176–243 the spoke holes offset closer to the gauge. The Ritchey OCR was the first to The accompanying figure (see figure 3) expressed as the ratio of, and as the lbf) spoke tension for the DS.1 This is NDS (see figure 2). Both sides of the be built up and tested. A rim transfer for shows the tension differences between difference of, the average spoke ten- well below the threshold for failure, wheel are affected equally: the DS all subsequent rims used the same hub the DS and the NDS for two levels of sions on the drive side and the non- established by Brandt as 2649 N (595 spokes are put at a steeper angle as the and spokes. dish corresponding to the mean drive side of the hub. An undished lbf) for lighter spokes and 3l39 N (706 NDS spokes are put at a shallower one Previously we had suggested that 7-speed and 8-speed spacing. The two wheel is one with the rim centered lbf) for heavier spokes (Brandt 1983). (see figure 1). This equalizes tension, wheel dish could be predicted from a levels of dish are expressed in mm and over the midpoint of the axle, not of Another spoke manufacturer specifies reducing the failure rate of spokes on hub’s measurements which we expres- measure the distance from the outside the hub. In a front wheel, the rim is which hub to use and recommends the the DS due to fatigue. The ability of sed as a ratio (DS/NDS). To measure edge of the locknut on the DS to the centered over both the axle and the DS be tightened to 1079 N (243 lbf) and these rims to reduce the amount of dish on wheels that are already built up flange center. Table 1 summarizes the hub. When viewed straight on the front the NDS be tightened to 883 N (199 lbf). dish is tested below. we are expressing dish as the differ- data in figure 3. For both figure 3 and wheel’s rim can be seen to fall in the (This has the DS 22% tighter than the ence in spoke tension between the table 1, freewheel width (F/W width) is middle of the front hub, to rest equally NDS.2) While this is a moderate amount DS and the NDS (DS−NDS). All ten- measured as the DS locknut-to-flange between the flanges. A front wheel that of dish perfectly adequate for most sion measurements were taken with a center in mm. is not centered on the axle cannot be applications it does not allow for differ- Wheelsmith spoke tensiometer. At centered in the frame or between the ences in the amount of dish required each level of dish the wheel was mea- DISCUSSION brake blocks. resulting from hubs with different free- sured at three levels of tension corre- Figure 3 plots dish as differences in In the rear wheel the rim is centered wheel widths, axle lengths and distance sponding, e.g., to tensiometer read- spoke tension between the drive side over the axle; it is not centered over between a hub’s flanges. These are ings of 75, 80 and 85 for the DS. A and the non-drive side (DS-NDS). The the hub. This is the awkward solution properties of the hubs and since it is copper template insured that ten- graph plots three lines; one for a Bon- to the problems posed by the develop- rare for two hubs to have the same mea- siometer readings were taken 30 mm trager Asym, a Ritchey OCR, and a ment of the derailleur. Before the surements no one can specify the ten- Figure 2. Offset rim: Bontrager Mustang from the rim edge. Each reading was standard non-offset rim. A front wheel, development of derailleurs both the sions on each side of the wheel unless (reproduced here with permission). taken twice. If front and rear wheels were centered s/he also specifies which hub to use. the two readings over the hubs. With the development of Shraner (1999) gives the following ten- METHODS did not agree a derailleurs came multiple-cog free- sion recommendations for wheels with Three 26x1.75 32-spoke rims were third reading was wheels which presented the problem of “normal rims”. He recommends tested, two offset and one non-offset, taken. If there was how to accommodate the additional 900–1000 N (202–225 lbf) of tension for for comparison. The offset rims were a any doubt about cogs. Bicycle designers could do one of “normal rims” in the front wheel and in Ritchey OCR (Off-Center Rim) and the reading on a three things. For one, they could the rear wheel he recommends Bontrager Mustang Asym (Asymmet- spoke the proce- spread the stays and widen the rear 600–700 N (135–157 lbf) for the NDS rical). The non-offset rim was a Matrix dure was repeated axle to make room for the freewheel and between 1000–1100 N (225–247 lbf) Singletrack. Each of these rims was until a true reading block. If they did this the bottom for the DS (this has the DS between built up 3X on a 32-spoke low-flange was taken. After bracket would have to be widened also Figure 1. Wheel cross-section for symmetric 57–66% tighter than the NDS; Schraner Suntour XC Pro hub with a 45-mm one level of dish to keep the chain in line with the free- rims (after Brandt) with modifications for 1999). However, he does not specify flange diameter, or spoke circle (mea- was tested at three wheel. Many consider wider bottom asymmetric rims. which hub must be used to achieve this sured center-to-center) using DT stain- levels of tension Figure 3. Tension differences (dish) for regular and offset rims. brackets harder to pedal. Their other range of ratios. less-steel 14g spokes, 264-mm long. The the axle was choice was to center the rim over a One effect a freewheel has is to Unequal spoke tension caused by Suntour hub measured 55 mm between respaced and the very narrow hub. They could move the increase spoke tension on the right, or rear-wheel dish has caused some manu- flanges, center-to-center (see figure 1). process repeated Table 1. Tension differences (DS-NDS) # F/W % reduction hub flange on the left side in by the drive side (DS) relative to the left, or facturers to experiment with previously The 55-mm distance was chosen for the remaining speeds width Reg OCR Asym OCR Asym same amount the freewheel block non-drive side (NDS). The ratio of rejected solutions. To make room for because it was midway between the dis- levels of dish. Each 7-sp 44.5 12.9 3.3 5.6 74% 57% moved the right side hub flange in. DS/NDS measures what is commonly the wider freewheels some manufactur- tances measured in a previous article. level of dish was 8-sp 48.5 22.6 12.0 14.8 47% 35% However, wheels built with narrow referred to as dish. In this ratio the two ers have used longer rear axles. How- (The average for a 7-sp hub was 58.4 established using a
10 Number 50, Spring 2000 Human Power Human Power Number 50, Spring 2000 11 or any wheel with no dish would plot a nearly 70 g heavier. This is in light of the CAUSES OF DISH the Ritchey front OCR is 24 mm wide, That said, any vertical loading will found that even differences as small as value at zero. The higher the number fact that while the Bontrager is eyelet- The utility of these rims may be the same width as the Rock OCR we result in the rim having increased DS 0.5 mm in axle spacing would affect the more severely a wheel is dished. ted the Ritchey is not. especially suited to wide rear hubs. tested which is no longer made. Unlike bending movement. At the very least tensiometer readings. Less dish is desirable. What is obvious The graph illustrates that for all free- Previously we noted how dish is the the Rock OCR, it does not appear to this might result in increased denting is that the non-offset rim has the great- wheel thicknesses, offset rims give a result of several variables (Ibid.). For have any internal reinforcing ribs and is (“flat spots”) of the DS half of the rim CONCLUSIONS est tension differences, for all levels of dramatic reduction in dish, as mea- any given freewheel both freewheel eyeletted. Phil Wood is listed as offering during cornering. AND RECOMMENDATIONS dish. Both the offset rims show a signif- sured as tension differences. Because thickness and hub width contribute to a front hub threaded to accept a disc Vertical loads are thought to loosen Offset rims were found to significant- icant reduction in dish. While both off- the Ritchey yields lower values pre- dish in different ways. brake in both a dished and undished the bottom most spokes making them ly reduce rear-wheel dish for both 7-sp set traces are below the standard rim sumably the spoke holes are more off- Consider the effects of both free- versions (Sutherland 1995). Looking at likely to become unscrewed. Another and 8-sp spacing. The rims have addi- trace, the Ritchey rim trace is below set. Although both offset rims reduced wheel thickness and hub width on dish. their dimensions it appears that the possible effect of bending-moment tional application in reducing front- the Bontrager rim trace. dish compared to the standard rim the Increasing the freewheel thickness undished hub has a narrower hub shaft asymmetry is that any vertical load will wheel dish induced by hub brakes. There exist a number of differences Ritchey had slightly lower tensiometer moves the entire hub over to make (flange-to-flange) than the dished hub. If bend the rim on the DS and the spokes between these two rims which may readings. Compared to the Bontrager room for it. This pushes the DS flange it is for a dished front wheel then using the DS might loosen more and their NOTES explain our findings. Bontrager adver- rim, the Ritchey readings were 2.3 in closer to the rim center as the NDS an offset rim would be an ideal solution. nipples be more likeley to unscrew. 1. Wheelsmith tensiometer: a profes- tises his rims as having 2.5 mm of lower for 7-sp spacing and 2.8 lower for flange is pushed further from the rim The Ritchey literature used to claim a 3- The wheel might be more likely to get sional wheelbuilding device for exact offset,3 while Ritchey ads used to 8-sp spacing rim. This, however, may center. By comparison, increasing the mm reduction in dish for their rear out of true. Cornering, on the other spoke-tension measurements [pam- claim a 3-mm reduction in dish; they be due to the Ritchey’s being a wider hub width affects dish by pushing the wheel. Dished models of Phil Wood hand, can subject the wheel to lateral phlet; 1989]. Menlo Park, CA: currently advertise a 50% reduction in and heavier rim. NDS flange further away from the rim, front-brake hubs threaded for a disk loads. Right-hand bicycle turns Wheelsmith Fabrications, Inc. The dish.4 One possible explanation is rim leaving he DS flange alone. Because brake are dished between 2–13 mm, increase the tension on the DS spokes pamphlet comes with the tensiome- width. Our measurements showed the DISH AS HUB MEASUREMENTS freewheel thickness affects both depending on the front axle length in “tension peaks” as the wheel rotates ter, although it does not give a NDS Bontrager Mustang was 21.8 mm wide Several claims are made for Ritchey flanges and hub width affects only one (locknut-to-locknut). Only two hubs had while it leans into a turn. Now if the tension recommendation. For more and the Ritchey Rock OCR was 24 mm OCR rims. One of them is that the flange we view freewheel thickness to 2mm dish and the other eight were rim on the DS were also to experience information, Wheelsmith Fabrica- wide. With a wider rim you can move Ritchey OCR rim will reduce dish from affect hub dish at twice the rate of hub 3 mm or more. The Sturmey-Archer some additional vertical bending in tions, Inc., 355l Haven Ave., Ste. R the spokes further over to the NDS. 6 mm to 3 mm using a Ritchey hub; a width. It is for this reason that we have BFC drum brake is dished 4 mm. We response to cornering then that would Menlo Park, CA 94025-l009 While the Bontrager rim is 2.2 mm 50% reduction. We previously found suggested that increasing hub width can think of no more appropriate a use tend to loosen the spokes. So the Telephone: 4l5-364-4930 narrower it has only 0.5 mm more dish that the dish on most hubs is consider- was not half as bad as increasing free- for an offset rim than for a front wheel “tension peaks” caused by lateral 2. Mastering the art of wheelbuilding. than a Ritchey. There is a weight ably more than 6 mm. We surveyed the wheel width. with a disc or drum brake. We have loading might be offset by any tension l996. DT Swiss Bike Technology USA penalty for a wider rim. We measured same 74 hubs we did in a previous The ability of an offset rim to attenu- examined the benefits of offset rims in losses as a result of vertical loading. and Campagnolo. Certification the Bontrager at 415 g (he advertises study (Forbes 1998–99). ate dish depends on the source of the reducing rear-wheel dish only. Brandt (1983) has suggested that seminar brochure. Extreme Skills 430 g) and the Ritchey at 484 g. Wider Table 2 summarizes the actual dish. Because offset rims affect spokes fatigue is the result of the interaction Seminars and Extreme Exposure rims are laterally stronger. In terms of average amount of dish expressed as on both flanges an offset rim would BENDING-MOMENT ASYMMETRY between peak and baseline tension in a Promotions. Grand Junction, CO: height the Ritchey advertises 12 mm distance measurements for 74 hubs reduce freewheel-induced tension dif- It is natural, however, to assume that tension cycle. If this were so then we LLC, p. 11. height and the Bontrager advertises with a 55-mm-wide hub on a 135-mm ferences at the same rate as the free- if form follows function we would might expect that an offset rim might 3. Bontrager components nineteen 13 mm height. Deeper rims are radially axle we previously examined. wheel causes them. By comparison, expect to be able to see the symmetry reduce the frequency of DS spoke ninety nine [promotional pamphlet], stronger. Both rims are made from We see from the table that actually the hub-width-induced tension differences and balance reflected in force-balanced failure in two ways. First, it would p. 11. See also www.bontrager.com. 6061-T6 aluminum alloy. Ritchey OCR would reduce dish from affect only one flange and offset rims structures. In offset rims the spoke equalize spoke tension thus decreasing 4. Ritchey web site: http://www.ritchey- Offset rims have been criticized by 9 mm to 6 mm on a typical 7-speed and attenuate dish by affecting both holes are drilled closer to one side. One the baseline tension in DS spokes. logic.com/products/components/n_ri some as being weaker because they from 17 mm to 14 mm on a typical 8- flanges. We would expect for offset cannot help but suspect for such a glar- Second, it would reduce the tension ms/home_rims.htm. lack ferrules, or “spoke sockets” that speed. These were reflected in lower rims to reduce hub-induced dish at ing asymmetry to somehow reflect an “peaks” experienced by the DS spokes join the upper and lower parts of a box- tension differences. Inspection of twice the rate caused by hub width. underlying unequal distribution of during right-hand turns due to lateral REFERENCES section rim so the spoke pulls on both table 1 reveals that when compared to Offset rims would be twice as effective force, or lack of balance. movement because of the rim’s Brandt, Jobst. 1983. The bicycle wheel, sections of the box-section rim. In a conventional rim, the Ritchey OCR in attenuating hub-induced dish as they Offset rims have what can be called increased complementary vertical rev. ed. Menlo Park, CA: Avocet, answer to this both rims come with rim yielded a 9.6 smaller difference in are in attenuating freewheel-induced bending-moment asymmetry. N.B.: we bending movement during cornering. p. 131. A book no wheelbuilder’s “internal reinforcements”. Figure 2 tensiometer readings for 7-sp spacing, dish. That said, it is reasonable for off- make the distinction between vertical, However, this is all highly specula- library is complete without. depicts a rib which joins the top and (a 74% reduction) and a 10.6 difference set rims to make for an increased wide- or up-and-down bending moment and tive. The possible effects of bending- Forbes, V. 1998–99. Predicting wheel bottom as did Wilderness Trail Bikes in tensiometer readings for 8-sp spread application of wider hubs, with lateral, or side-to-side bending moment. moment asymmetry in offset rims are strength from hubs. Human Power. (WTB) in its PowerBeam rim. Bontrager spacing, (a 47% reduction). all the benefits of increased lateral The longer DS half of the rim the more not known. Possible effects include 46:8–10. rims use one rib (“dual cavity” or dou- strength. bending-moment it has. The shorter the the increased likelihood of rim dents Schraner, G. 1999. The art of wheel- ble box) while Ritchey rims currently Offset rims may find an additional NDS half of the rim the less bending- on the DS as well as spokes building. Denver: Buopane Publi- use two (“triple box”). The application for offsetting moment it has so that a vertical load, unscrewing on the DS due to vertical cations, p. 44. Recent publication ability of internal reinforc- Table 2. Dish measured as distance difference (DS-NDS) front-wheel dish induced from, e.g., sitting on a bicycle, exerts movement and a decreased incidence from DT with many insider tips from ing ribs to act like a fer- by disk brakes. During the more vertical bending-moment to the of fatigue failure due to vertical move- the European professional racing Difference (DS-NDS) ruled rim is unclear. While preparation of this article DS than the NDS side of the rim. ment attenuating stress peaks caused circuit. # F/W Center Center Standard OCR Asym we cannot be sure if our speeds width to DS to NDS (DS-NDS) (DS-NDS) (DS-NDS) both Ritchey and One possible effect of this is that by lateral movement during cornering. Sutherland, H. 1995. Sutherland’s Ritchey OCR is triple box 7-sp 44.5 23 mm 32 mm 9 mm 6 mm 6.5 mm Bontrager each introduced every bump in the road would deform Despite our attempts to limit or con- handbook for bicycle mechanics, 6th or not it would help 8-sp 48.5 19 mm 36 mm 17 mm 14 mm 14.5 mm rims for use with a front the longer DS more than it did the trol for tension differences, measure- ed. Berkeley: Sutherland explain why the Ritchey is disk brake. Interestingly, shorter NDS, relative to the centerline. ment error cannot be ruled out. We Publications, pp. 11–25 ff.
12 Number 50, Spring 2000 Human Power Human Power Number 50, Spring 2000 13 ACKNOWLEDGMENTS Rolling resistance of bicycle tyres COMMENTARY ON THE TABLE 8. For additional rolling-resistance diameter tires”. The author thanks David Gordon 1. Some of the tyres are tested at pres- data, see reports in the following My questions relate primarily to accura- Wilson for his invaluable comments by John Lafford sures above the manufacturer‚s rec- British magazines: cy and meaningful precision of his data. during the preparation of this article, ommended pressure. This was done Cycling Plus, issue 62 (Feb. ’97) How repeatable is his test? The fact Helios Studios in Columbia, MO, John INTRODUCTION The power to propel tyres along the for scientific interest. This does not “Winter tyres”; Cycling Plus, issue 68 that an increasing series of inflation W. Stephens of Garden Grove, CA, and I have been a research and develop- road (Prr), is directly proportional to imply approval of operating tyres (Mid-summer ’97) “Road tyres”; pressures leads to a decreasing series Jean Seay of San Luis Obispo, CA, for ment engineer for many years, and the rolling-resistance coefficient (Crr), above the manufacturer‚s recom- Cycling Plus, issue 81 (Aug ’98) of rolling resistance coefficients sug- their assistance in the preparation of when I started recumbent racing about the weight of the vehicle plus rider mended pressure. “Time-trial tyres”; Total Bike, issue 6 gests that his final digit is meaningful this article. fourteen years ago, it was only natural (W), and also the speed of the vehicle 2. Most of the tyres are tested at (Oct ’97) “MTB tyres”; and BHPC (+0.0001). However it would be good to Figure 1 is after The Bicycle Wheel for me to want to obtain reliable data (V). several pressures to show the effect Newsletter, issue 58, “MTB tyres”. have a statement from John about how by Jobst Brandt and appears by his on cycling performance. The most Prr = Crr x W × V (times a constant for of pressure on rolling performance. 9. Thanks are due to the following for well his tests repeat, what averaging he courtesy and with his permission. The important factor is aerodynamics, and I users of ancient units –ed.) 3. The list includes the effect of grind- the loan of tyres for testing: Hilary employs, etc. How accurate is his test? author thanks him for his comments in addressed the measurement of aerody- The power absorbed is the same ing the tread down on some tyre Stone, Richard Grigsby, John Here my questions spring from a cou- reading an advance copy of this article. namic drag in the real world by doing whether the vehicle has two, three or types. This simulates the effect of Kingsbury, Michelin Tyres, Cambrian ple of disparate perspectives. Figure 2 is after the Bontrager coast-down tests. The procedure for four wheels (so long as they are per- tyre wear. This generally gives an Tyres, Dillglove (Nokian tyres). 1. How do his numbers compare Components homepage, this is described in “So you want to fectly aligned, as the author confirms improvement in rolling performance. ———— with others’? He gives “good 700C” http://www.bontrager.com, and build an HPV”‚ (p. 40 ff) a publication below –ed.). In the table, the power is In some cases it is not beneficial as John Lafford
14 Number 50, Spring 2000 Human Power Human Power Number 50, Spring 2000 15 Rolling resistance power absorbed Rolling resistance of tyres - test data Unfaired wgt = 185 lb/84 kg / Faired wgt = 200 Lb/91 kg © John Lafford, Nov 1999 Time lost @ Generally available recumbent tyres Test Prr Prr Prr Prr Prr Prr 25mph pres- Rolling 20mph 25mph 30mph 30mph 40mph 50mph over (Butyl inner tubes unless stated) Rolling resistance power absorbed sure res coef. 32kmh† 40kph 48kph 48kph 64kph 80kph 10 mi.‡ Unfaired wgt = 185 lb/84 kg / Faired wgt = 200 Lb/91 kg Tyre name Size psi* Crr test watts watts watts watts watts watts sec. Time lost @ Primo w/ground off tread, latex tube 37-349 120 0.0061 45 56 67 73 97 121 35 Test Prr Prr Prr Prr Prr Prr 25mph Primo w/ground off tread, latex tube 37-349 140 0.0058 42 53 64 69 92 115 29 pres- Rolling 20mph 25mph 30mph 30mph 40mph 50mph over Vredestein Monte Carlo (new) 37-406 90 0.0069 51 63 76 82 109 137 50 sure res coef. 32kmh† 40kph 48kph 48kph 64kph 80kph 10 mi.‡ Vredestein Monte Carlo (new) 37-406 100 0.0067 49 62 74 80 107 134 47 Tyre name Size psi* Crr test watts watts watts watts watts watts sec. Vredestein Monte Carlo (new) 37-406 120 0.0064 47 58 70 76 101 126 40 IRC Road Lite (new) 20" x 1 1/8" 100 0.0090 66 82 99 107 143 178 92 Hutchinson 600 x 28A 100 0.0059 43 54 65 70 94 117 31 IRC Road Lite (new) 20" x 1 1/8" 115 0.0089 65 82 98 106 141 176 90 Hutchinson 600 x 28A 120 0.0052 38 48 58 62 83 104 18 Conti Top Touring (new) 37-406 70 0.0092 68 84 101 110 146 183 96 Conti Grand Prix w/latex tube 28-406 100 0.0077 57 71 85 92 123 153 67 Conti Top Touring (new) 37-406 90 0.0079 58 73 87 94 126 157 71 Conti Grand Prix w/latex tube 28-406 125 0.0069 51 64 77 83 110 138 52 IRC Road Lite (old) 20" x 1 1/8" 100 0.0068 50 63 75 81 108 136 49 Conti Grand Prix w/latex tube 28-406 140 0.0067 49 61 74 80 106 133 47 IRC Road Lite (old) 20" x 1 1/8" 115 0.0064 47 59 71 77 102 128 41 Schwalbe Spezial City Jet (used) 32-406 100 0.0065 48 60 72 78 104 130 44 Michelin 28-440 70 0.0078 57 71 86 93 123 154 68 Schwalbe Spezial City Jet (used) 32-406 120 0.0063 46 58 70 75 100 126 39 Michelin 28-440 90 0.0071 52 65 78 84 112 141 54 Nokian City Runner (new) 40-406 72 0.0083 61 76 91 99 131 164 78 Tioga Comp Ramp (new) 20 x 1 7/8 85 0.0080 59 73 88 95 127 159 72 Nokian City Runner (new) 40-406 100 0.0078 58 72 86 93 124 156 69 Tioga Comp Ramp (new) 20 x 1 7/8 100 0.0079 58 73 87 95 126 158 71 Nokian City Runner (new) 40-406 120 0.0071 52 65 78 85 113 141 55 Hutchinson HP 25 25-451 115 0.0080 59 74 88 95 127 159 73 Nokian Mount & City (new) 47-406 50 0.0083 61 76 91 99 132 165 78 Hutchinson HP 25 25-451 90 0.0089 65 81 98 106 141 176 89 Nokian Mount & City (new) 47-406 80 0.0072 53 66 79 86 114 143 57 Hutchinson HP 25 25-451 100 0.0086 63 79 95 103 137 171 84 Nokian Mount & City (new) 47-406 100 0.0068 50 63 75 81 109 136 49 Conti Top Touring (new) 37-406 70 0.0090 66 83 99 108 143 179 93 Nokian Mount & City (new) 47-406 120 0.0063 47 58 70 76 101 126 40 Conti Top Touring (new) 37-406 90 0.0081 60 74 89 97 129 161 74 Nokian Mount & Ground (used) 47-406 45 0.0104 76 95 114 124 165 206 119 Conti Top Touring (new) 37-406 100 0.0080 59 74 89 96 128 160 73 Nokian Mount & Ground (used) 47-406 80 0.0100 73 92 110 119 159 198 112 Conti Top Touring (new) 37-406 115 0.0084 62 77 93 100 133 167 80 Nokian Mount & Ground (used) 47-406 100 0.0089 66 82 99 107 142 178 91 Michelin (450 x 28A) 28-390 100 0.0065 48 60 72 78 104 130 43 Schwalbe City Jet (new) 32-406 100 0.0087 64 80 96 104 139 174 87 Michelin (450 x 28A) 28-390 115 0.0064 47 59 71 77 102 128 42 Schwalbe City Jet (new) 32-406 120 0.0082 60 76 91 98 131 163 77 Conti Grand Prix (new) 28-406 100 0.0072 53 67 80 86 115 144 57 Schwalbe Marathon (new) 32-406 100 0.0097 72 89 107 116 155 193 107 Conti Grand Prix (new) 28-406 120 0.0067 49 61 74 80 106 133 47 Schwalbe Marathon (new) 32-406 115 0.0101 75 93 112 121 161 201 114 Conti Grand Prix (new) 28-406 140 0.0066 49 61 73 79 105 132 45 Schwalbe Marathon (new) 32-406 130 0.0100 73 92 110 119 159 198 112 Vee Rubber (new) 20 x 2.125 40 0.0114 84 105 126 136 181 226 139 Vredestein Monte Carlo (used) 37-406 90 0.0077 57 71 85 92 123 154 67 Vee Rubber (new) 20 x 2.125 60 0.0105 77 96 116 125 167 209 122 Vredestein Monte Carlo (used) 37-406 100 0.0069 51 64 76 82 110 138 51 Vee Rubber (new) 20 x 2.125 80 0.0097 72 89 107 116 155 193 107 Vredestein Monte Carlo (used) 37-406 120 0.0068 50 63 75 81 109 136 50 Primo 16 x 1 3/8 (new) 37-349 100 0.0078 57 72 86 93 124 155 69 Nokian Mount & City (new) 47-305 50 0.0135 100 124 149 161 215 269 182 Primo 16 x 1 3/8 (new) 37-349 110 0.0071 52 65 78 85 113 141 55 Nokian Mount & City (new) 47-305 80 0.0103 76 95 114 123 164 205 118 Primo 16 x 1 3/8 (new) 37-349 120 0.0070 52 65 78 84 112 140 54 Primo 20 x 1.5 (used) 37-451 85 0.0066 48 61 73 79 105 131 45 Tioga Comp Pool (new) 20 x 1.75 90 0.0074 54 68 81 88 117 146 60 Primo 20 x 1.5 (used) 37-451 100 0.0065 48 60 71 77 103 129 43 Tioga Comp Pool (new) 20 x 1.75 100 0.0071 52 65 78 84 113 141 54 Primo 20 x 1.5 (used) 37-451 120 0.0062 46 57 69 75 99 124 38 Tioga Comp Pool (new) 20 x 1.75 110 0.0069 51 63 76 82 110 137 51 IRC Roadlite 20 x 1 1/8 (used) 28-451 100 0.0068 50 63 76 82 109 136 50 Tioga Comp Pool (new) 20 x 1.75 120 0.0065 48 60 72 78 103 129 43 IRC Roadlite 20 x 1 1/8 (used) 28-451 120 0.0064 47 59 71 77 102 128 42 Kenda w. tread ground off 16 x 1.75 80 0.0068 50 63 76 82 109 136 50 Primo Comet 37-406 100 0.0074 54 68 82 88 118 147 61 Kenda w. tread ground off 16 x 1.75 90 0.0064 47 59 70 76 101 127 41 Primo Comet 37-406 120 0.0070 52 64 77 84 111 139 53 Kenda w. tread ground off 16 x 1.75 100 0.0066 48 60 72 78 104 130 44 Michelin (used) 600 x 28A 70 0.0081 59 74 89 96 128 160 74 Michelin Hilite S'compHD 650 x 20c 100 0.0062 45 57 68 74 98 123 36 Michelin (used) 600 x 28A 90 0.0073 54 68 81 88 117 146 60 Michelin Hilite S'compHD 650 x 20c 110 0.0058 43 53 64 69 92 115 29 Continental Grand Prix (used) 28-406 120 0.0082 60 76 91 98 131 163 77 Michelin Hilite S'compHD 650 x 20c 120 0.0055 40 50 60 65 87 109 22 Haro (used) 20 x 1.5 65 0.0073 54 68 81 88 117 146 60 Hutchinson HP 25 "600A x 24""" 100 0.0070 52 65 77 84 112 140 53 Haro (used) 20 x 1.5 85 0.0068 50 63 75 81 108 136 49 Hutchinson HP 25 "600A x 24""" 120 0.0068 50 63 76 82 109 136 50 Haro (used) 20 x 1.5 100 0.0063 47 58 70 76 101 126 40 Hutchinson HP 25 "600A x 24""" 140 0.0068 50 63 75 81 109 136 50 Haro (new) 20 x 1.5 65 0.0077 56 70 84 91 122 152 66 Primo V Monster (new) 20 x 1.75 65 0.0074 55 68 82 89 118 148 61 Haro (new) 20 x 1.5 85 0.0067 50 62 74 80 107 134 48 Primo V Monster (new) 20 x 1.75 80 0.0078 57 72 86 93 124 155 68 Nokia Mount and City (new) 47-406 50 0.0104 77 96 115 124 165 207 120 Michelin 600 x 28A 80 0.0075 55 69 83 89 119 149 62 Nokia Mount and City (new) 47-406 70 0.0087 64 80 96 104 139 174 87 Michelin 600 x 28A 100 0.0072 53 66 79 86 115 143 57 Nokia Mount and City (new) 47-406 90 0.0082 60 75 90 98 130 163 76 Primo w/ground off tread 37-349 85 0.0066 49 61 73 79 106 132 46 Nokia Mount and City (new) 47-406 100 0.0073 54 67 81 87 116 145 59 Primo w/ground off tread 37-349 100 0.0063 47 58 70 76 101 126 40 Vredestein S-Lick (new) 32-406 60 0.0155 114 142 171 185 246 308 220 Primo w/ground off tread 37-349 110 0.0064 47 59 70 76 102 127 41 Vredestein S-Lick (new) 32-406 90 0.0106 78 97 117 126 168 210 123 Primo w/ground off tread 37-349 120 0.0060 44 55 66 71 95 119 33 Vredestein S-Lick (new) 32-406 100 0.0098 72 90 108 117 155 194 107 Primo w/ground off tread 37-349 140 0.0060 44 55 66 71 95 119 33 Moulton Wolber line tread (new) 17 x 1 1/8 70 0.0092 68 85 102 110 147 184 97 RECOMMENDED 700C TYRES FOR REAR WHEELS for comparison. Very good performance at reasonable price. Moulton Wolber line tread (new) 17 x 1 1/8 100 0.0084 62 78 93 101 134 168 81 Nokia Roadie (used) 700 x 25c 100 0.0046 34 42 51 55 73 91 5 Moulton Wolber line tread (new) 17 x 1 1/8 120 0.0079 58 73 87 94 126 157 71 Michelin Axial Supercomp 23-622 110 0.0046 34 42 51 55 73 91 5 Primo w/ground off tread, latex tube 37-349 100 0.0066 49 61 73 79 106 132 46 Vredestein Fortezza Piste 700 x 23c 10 bar 0.0043 32 40 48 52 69 86 0 * 1 bar = 14.5 psi; † 1 m/s = 3.6 kmh; * 1 bar = 14.5 psi; † 1 m/s = 3.6 kmh; ‡ compared with the time using the tyre (at the same power input) with lowest Crr, the last entry in the table ‡ compared with the time using the tyre (at the same power input) with lowest Crr, the last entry in the table
16 Number 50, Spring 2000 Human Power Human Power Number 50, Spring 2000 17 • What was his load? “Representative” that gave the same rolling distance, TECHNICAL NOTES recumbents required less power to although only the faired M5 and the Figure2. Bicycles tested. probably means it was 45 kg, but it running in both directions. Then, to Power requirements propel them than did the “regular” regular bikes were common to both would be nice to know for sure. cope with the very slight undulations in for laid-back recumbents racing bike in this configuration. The tests. The LWB “Peer Gynt” with low None of this is meant as direct criti- the floor, the floor is marked out at 4' Report and comment by Dave Wilson, reduction appeared to be a function of bottom-bracket was found to have a cism of Lafford’s careful efforts, but intervals from the start point of the test with translation help from Jan how low the rider was (see photos). higher drag than that of a regular rather as an invitation to discuss some run. A stop watch is used and the time Limburg and Ellen Wilson. The lowest power of the unfaired racing bike with the rider in a full vitally interesting issues! recorded for the test rig to pass the This is an interpretation of the high recumbents was needed by Bram crouch. Because of the difference in —Jim Papadopoulos 4', 8', 12', 16'‚ 20'‚ 24', 28'‚ etc. markers. points of an article by Bert Hoge and Moens’ M5, which at 40 kmh took the rider positions on the racing bike, In the coast-down test, the retardation Jeroen Schasfoort in HPV nieuws no. 228 W, while the racing bicycle the principal interest in the results Reply to note by Jim Papadopoulos, should be completely uniform, and so 4, 1999, the magazine of the Nether- required 389 W input. The fully-faired shown here is in the differences among approximately in the order of his several of the 4' zones are averaged to lands NVHPV. Its topic is the use of an M5 required less than 130 W at the the recumbents, and in the accuracy of Optima Dolphin questions, give a mean average retardation. This SRM instrumented crank (giving same speed. (Table 1) actual power measurements taken on Tires tested. The number of tyres avoids any errors caused by cresting or torque) on a “regular” racing bike and These results can be compared with different bicycles with the same rider tested is approaching 400 and I have failing to crest a slight rise at the end of on five recumbents. All of the recum- the aerodynamic drag measured in the on the same circuit with similar tires. tested tyres of all sizes. The article was a test run. bents were of the very-laid-back “Tour” tests of stationary bikes in the These are very valuable data. Thank focused on small-diameter tyres as they Comparing my data variety, having seat-back angles with wind-tunnel and on bikes being ridden you, Bert Hoge, Jeroen Schasfoort and are of particular interest to HPV riders. with other sources the horizontal of down to 15 degrees on a velodrome, reviewed in Human the NVHPV! I included three good 700C tyres of There are some better 700C than (see photos). All of them had the Power, 12:4, spring 1997. There is —Dave Wilson good value. those I listed, but I did not include bottom-bracket considerably above the general qualitative agreement, Repeatability these as they are more expensive and lowest part of the seat: I believe that 1. If I run a tyre up the track and it less durable. There are many 700c tyres this is important because whirling legs runs 20' 3" (say), and then I repeat it a whole lot worse. normally give a high drag, and having Flevobike Fifty Fifty Table 1: Power required to propel bicycles straight away and follow the exact The Moulton data are for the line them in the “forward shadow” of the Expected same piece of the floor, then it also will tread‚ touring tyre, not the high-pres- body must reduce this drag. The Bicycle type or name Power, watts increase run 20' 3". If the direction is a bit off sure slick. authors write, “A smaller frontal 30 kmh 40 kmh in speed* then the distance will be slightly differ- I am not looking to repeat other surface gives less air resistance and Racing bike, touring position 181 389 0% ent due to the slight imperfections in people’s results. It would be relevant higher speed. It can be achieved by Optima Dolphin 161 336 6% the floor. only if I had the same tyres that they increasing the height of the bottom Flevobike 50-50 152 309 10% 2. If I were to repeat the test on that had used. bracket above the lowest part of the M5 20-20 131 265 17% tyre on another day, then I might not The test wheel/tyre is fitted to a tri- seat to about 270 mm (10.6"), and Baron Low Racer 128 251 20% have exactly the same pressure in the cycle. The two other wheels are per- reducing the seat height to about 250 Moens (M5) Low Racer 114 228 24% tyre by 1 or 2 psi, and the temperature fectly aligned, I know their rolling- mm (9.8").” (These are approximately *Relative to racing bike would probably be different and so resistance coefficient and they are only the relevant measurements of the M5 M5 20 20 there would be a slightly different lightly loaded. Even so, their rolling Low Racer.) All six bicycles were result. I have a simple test for this type coefficient and load and drag compo- ridden by one tester, Dries Baron, of repeatability, where I have a partic- nents are taken into account. weight 90 kg (198 lb.), height 1.86 m. ular Michelin tyre which I run from The weight on the test wheel is (6'1", almost a midget by Dutch stan- time to time. This has shown over a 66 lb. This is representative for three dards) wearing racing clothing, on a period of 18 months and 7 test sessions wheelers, which many HPVs are, but is 200-m-long velodrome track, presum- that the Crr value is repeatable at light for a two-wheeler. It is chosen for ably oval or circular. Ten circuits (2 km 0.0051 to ± 0.0001 and the power convenience of installation and the total) were made for each test point. absorbed at 25 mile/h to be 47 watts to number of times I have to pick it up All bicycles used Continental Grand- ± one watt. I have always rounded the and load it onto the trike test rig. Prix or IRC tires pumped to about 8- power-absorbed figure off to a whole Further, I have tested with a heavier bar pressure (116 psi). Two speeds Baron low racer number as it would be unreasonable to weight and got a very similar result. As were chosen: 30 and 40 kmh (18 and assume better accuracy than this. long as all my tests are done under the 25 mph), and the cadence was kept to Averaging same conditions then the results are about 88 rpm. The temperature was Jim Papadopoulos clearly appreci- properly comparative. Most of the about 15 C, 59 F. The measurement ates some of the practical problems in applied weight is directly on the test accuracy was reckoned to be ±2% (see running the tests. Yes, the floor does wheel. It cannot be exactly on it graph, figure 1). have very slight undulations in it. To because the tricycle would not then be The racing or “road” bike was ridden cater for this, in the area available, I stable. However, as the position is in the “touring” position, which I ran the tyre-test rig backwards and for- known, moments are taken, and the believe meant that the hands were on wards in many directions to find a line exact weight used in the computations. top of the handlebars, rather than the M5 low racer —John Lafford rider being in a full crouch. All of the —Photos and chart courtesy HPV nieuws; Figure 1. Power (watts) required vs. speed (kph). prepared for Human Power by JW Stephens
18 Number 50, Spring 2000 Human Power Human Power Number 50, Spring 2000 19 REVIEWS the ping-pong balls around came from “Feet On!” a furnace. “I had a blower...,” Caskey a pedal-powered museum exhibit said in explaining how he came up with By Michael Eliasohn the idea. “What are you going to do A bicycle is a relatively simple mech- with a blast of air? I wanted something anism. A frame holds two wheels. A so that as people walked in the door, chain runs from the chain-ring to the they could see a lot of action.” sprocket attached to the rear wheel. But the 300 ping-pong balls blown Turn the pedals and the bike moves. around in the clear plastic-walled
But how do you get those pedals to chamber did more than move around. Michael Eliasohn power a television or an organ, to light There were 294 white balls and six red 2) Leslie Gerschoffer (rear) and Marita English pedal what were exercise bikes, up light bulbs or create a vacuum, to ones. What were the odds of getting all which power a generator which powers the Figure 1. Blade-element velocities Figure 2. Blade radial-lift distributions blow 300 ping-pong balls around inside the red ones to land in the six “pock- television set they are watching. The TV a chamber, or a moving sculpture, ets,” which was half of an egg carton? could be “powered” by one or two people. My propeller theory the virtual slip velocity. Fortran code HELICE, written by which consists of twelve bicycle “...about the same as winning the lot- by E. Eugene Larrabee I calculated the radial bound circula- Susan Elso French at MIT. In the case wheels mounted on the wall? tery,” the accompanying sign stated. The mechanisms of the pedal- In 1978 I developed a useful form of tion distributions for minimum of windmills the displacement velocity That was the challenge facing Tom Caskey said he didn’t expect anyone powered exhibits for the most part propeller theory based on the work of induced loss by a process suggested by is against the wind direction and the Caskey, science exhibit designer at the would beat the odds during the life of were exposed, so visitors could see Hermann Glauert (1926 and 1938) and Glauert in 1938. The distributions are more curved portions of the blades are Southwestern Michigan (community) the exhibit, which taught visitors prob- how everything worked. “These are Sidney Goldstein (1929). It was functions of the advance ratio and the downwind. They were intended to College Museum near Dowagiac, MI. ability—and that it was unlikely they purposely made kid-understandable,” successfully applied to the propellers number of blades as shown in figure 2. leave a minimum hole in the air for a Caskey designed and, with some will ever win a lottery. said Caskey, whose museum job is of the Chrysalis and Gossamer Alba- They correspond to elliptic span given power output for the average assistance, made the exhibits for the The device that shot a spark across a part-time. tross human-powered airplanes in loading for a wing. wind speed of a “windfarm of many museum’s “Feet On! The Power of Ped- gap, to show that air is an insulator and The “Feet On!” exhibit was located 1979, and (in reverse) to windmills for Apparently these circulation distribu- windmills.” aling” exhibit, which ran from March 9 the absence of air isn’t, presented the in the part of the SMC Museum, to US Windpower, Inc. in 1980. tions are slightly in error, as suggested Since then Prof. Mark Drela has to June 12, 1999. He said he and other opposite problem. Caskey adapted an paraphrase from its flier, devoted to It is related to lifting-line theory as by Goldstein in 1929 and by my former developed his own XROTOR code museum staff members came up with old refrigerator compressor to act as a hands-on science and technological developed by Ludwig Prandtl and his student, Mark Drela, in 1982. In any which is a finite-element adaptation of the idea for the exhibit. vacuum pump, to remove the air from exhibits that investigate scientific prin- associates at Göttingen during World event they were good enough to form Goldstein’s 1929 paper. XROTOR was Among the challenges in creating it inside a clear-plastic dome. ciples and the technological world that War I. In it an induced velocity is devel- the basis of a Fortran code written by used to design propellers for the was a budget of less than $1,000. So The television could be powered by surrounds us. oped parallel to the blade lift direction Hyong Bang in 1978 to define the blade Monarch and Daedalus airplanes. many of the pieces, such as bicycles one or two people sitting on a couch. The sign at the entrance read: “This and perpendicular to the relative veloc- chord and pitch angles for the French’s HELICE code was rewritten in and exercise bikes, were purchased at The pedal mechanisms were two for- exhibition is an exploration of energy ity of the blade section with respect to Chrysalis and Gossamer Albatross air- Pascal as ELICA by Robert S. Grimes in a Goodwill store (which sells second- mer exercise cycles, with the chains transformation. The exhibits demon- the air mass as shown in figure 1. The planes so that they had not only mini- a form suitable for IBM PCs in 1982. hand goods) and at rummage sales. running to flywheels, and V-belts run- strate how your energy is converted flight (or axial) velocity, the rotational mum induced loss but also minimum Both Prof. Ernst Schoberl and I have Some items were donated. ning from there to a generator. into other forms with interesting out- velocity, and the induced velocity com- profile drag by choice of blade section used ELICA for many years personally. A lot of creativity was used. For To prevent one or two strong ped- comes. bine to produce the resultant velocity. and lift coefficient at the design point. I published my algorithms in 1980. instance, the organ parts were pur- alers from “overpowering” the televi- “You use chemical energy (namely The induced velocity is caused by lift They “were propellers of highest effi- I am told that AeroVironment uses a chased at Goodwill, but the exhibit also sion, Caskey hooked up an electromag- food and drink) to feed your muscles— on each blade section due to bound cir- ciency” in Glauert’s words. form of them to design propellers for used a metal trash can, garden hoses, a netic brake designed for trailers, which they are energy transformers. Your mus- culation according to the Kutta- At the relatively low advance ratios their airplanes including the Pathfinder, bellows from a previous museum acted on a disk on the generator shaft, cles allow you to move and give you the Joukowski Law. of these propellers they are character- which holds the altitude record for pro- exhibit and a bicycle pedal mechanism. making it impossible to “blow up” the ability to move different things.” Strangely enough, if the induced ized by narrow outer blade chords and peller-driven airplanes. The squirrel-cage blower that moves television. Making the exhibits pedal-powered velocity is small enough compared to wide inboard ones with strong twist, —E. Eugene Larrabee, The other two pedal-powered was a means to make them “hands-on,” the axial velocity it can be shown that having almost true geometric pitch, as professor emeritus, MIT exhibits were four light bulbs—the or more correctly, “feet-on.” “You’re the induced loss of the propeller is shown in figure 3. 1800 Knoxville Avenue harder one pedaled, the more bulbs lit really involved,” Caskey said. minimized if the virtual slip velocity is The same is true of the US Wind- Long Beach, CA 90815 USA up—and a sculpture consisting of The 59-year-old Caskey, whose back- radially constant, corresponding to a power windmills generated by a later November 1999 twelve bicycles wheels mounted on a ground includes product and graphic certain radial variation of the bound wall, linked by ropes and chains, so design, making dulcimers and building circulation. As Albert Betz, Prandtl’s that pedaling made the wheels rotate. a house, recently earned a master’s associate,wrote in 1923 (NACA TR Caskey said a challenge in creating degree in science education at Western 116), “The flow behind a propeller the pedal-powered exhibits was that Michigan University and wants to get a having the least loss of energy is as if they had to accommodate, in terms of doctorate in the same topic. the screw surfaces passed over by the Michael Eliasohn muscles, everyone from little kids to The Feet On! exhibits illustrates propeller were solidified into a solid 3) Marita English of Edwardsburg, MI, pedals football-player-sized men. “It’s got to be Caskey’s goal of making science learn- the lottery exhibit. A squirrel-cage blower figure and this were displaced back- blows air into the chamber, to put 300 ping- responsive for both.” able by being fun, not just by learning ward in the non-viscous fluid with a pong balls into motion. “Winning” required To accommodate various-sized rid- facts. “You can learn physics and sub- given small velocity.” The small putting the six red balls into the egg-carton ers, Caskey made super-long banana tlety and have fun ...,” he suggested, “or displacement velocity is exactly twice Figure 3. Shape of windmill blades produced by these methods. pockets. seats for some of the exhibits. you can think science is dumb stuff.”
20 Number 50, Spring 2000 Human Power Human Power Number 50, Spring 2000 21 Michael Eliasohn However, almostnothing abouthimis explains the“goat-herd”reference. is afarmerandraisesgoats, sothat felt way. We learnincidentallythathe delightfully casual,modest, yet deeply send me,explainsagreatdeal, ina hewaskindenoughto pages—which mystery man.Thissmallbook—128 1994). Allthismadehimsomethingofa changed for106years”(vol.11no.3 er’s) account“Whyyourbicyclehasn't historical authentic(almostaninsid- readers of Schmitz hadearlierbecomeknownto herd fromProvence….”Arnfried himself asbeingknownthe“goat- gurus acrossEurope.Hedescribed involved intheHPVmovementandits France andbecameenthusiastically German student,andlaterhesettledin building andasamechanicWest cal notestatesthatheworkedinship- recumbents hehadbuilt.Hisbiographi- which hebrieflydescribedaroundten fried Schmitz:“Velocar variations”,in 49, ledwithanintriguingarticlebyArn- 2Number50,Spring2000 22 T Schmitz,with 0 953617416)byArnfried (ISBN: energy Human Power:theforgotten 4) Tom Caskey, designer/builderofthe“Feet that newspaper. by theauthor, originallyappearedin and someofthephotographs,alltaken St. Joseph,MI.Portionsofthisarticle The Herald-Palladiumnewspaperin Dowagiac MI49047•USA Museum, 58900CherryGroveRoad, mail toSouthwesternMichiganCollege mail [email protected] ony Hadland accommodate various-sizedriders. long bananaseatenabledtheexhibitto a gap insideavacuum.The spark across On! exhibit,pedalstheexhibitthatshota The lastissueof To Michael Eliasohnisareporterfor m Caskeycanbecontactedbye- Human Power Human Power through his , no. on space-frame MoultonsandSturmey- fine booksonBritishportable bicycles, few wartsexposed. racing inEurope,aswell having a credit forpromotingbicycle andHPV comes alive:heisgivenagreatdealof acter ofthelateWolfgang Gronen others. Forinstance,thecomplexchar- Schmitz givesinsightfuldetailsofmany takes aminorrolebecauseArnfried and hisfamily’s HPVactivitiesoften the machines.)Butdetailsofhis some difficultypersuadinghimtoride himself andhissonJurgen.(Hehad bents, partlyforothersand describes howhetriedbuildingrecum- becomes partlyautobiographical,ashe potential. Fromthenonthebook Cycle in theFrenchbicyclingmagazine the racesandaboutthoseinU.S.A. or theMochets.ButSchmitzreadabout Schmitz, nooneknewanythingaboutit breaking Velocar of1933,but,states with aversionoftheFaurerecord- championships) andtraveledthere ulated bytheannualIHPVA speed Challenge atBrightonUKin1980(stim- Mochet heardabouttheAspoSpeed There arealsodetailsofhowGeorges cycles” [usinga“universaljoint”]. its “indirectsteeringforrecumbent in theinventors’GrandPrixLepinefor ‘VV’ [Velocar] wasawardedfirstprize between 1932and1940.”“Inthe They builtabout800[recumbentbikes] pedal-cars between1925and1944…. 1960. Theyconstructedsome6,000 Mochets builtmini-carsfrom1920to Human-Powered Vehicles… [The] and bikes,whatwewouldtodaycall were thenprofessionallybuildingcars late 1933orearly1934).“TheMochets cles regardlessoftype.”(Ibelievein hour recordforhuman-poweredvehi- Mochet sponsorsacupfortheabsolute believe thatwasin1913.)“Charles for streamlinedbikestakesplace.”(I “In Berlinthefirstinternationalrace details ofwhichIwasn’t fullyaware. anywhere. Herearesomeexamplesof recumbents thanIhavepreviouslyread streamline bicyclesandtoproduce of thehistoryearlyeffortsto These aredevotedtoafullerre-telling revealed inthefirstninechapters. T ony Hadland,whohaswritten very , andbecameexcitedbythe Le Arnfried Schmitz Arnfried Supplement to“Velocar variations”by LETTERS HPVs shouldreadthisbook. from theauthor. Everyenthusiastfor more, butI’msurethatwe’llhearagain thing aboutwhichwe’dliketoknow tense), anditisn’t preciseaboutevery- strangely inthe“historic”present is alivingdocument(writtenlittle book isn’t adryhistorybookbutrather rest oftheworld£14.95(airmail).The UK £12.95;Europe£13.95(airmail); try CV77DU,UK.InBritishpoundsitis Malvern Road,BalsallCommon,Coven- francs, orfromRosemaryHadland,39 84220 LiouxGordes,France:140 Arnfried Schmitz,QuartierGallas, Amazon.com orbydirectmailfrom lished thisbook.Itisobtainablefrom Archer gears,hasdesignedandpub- or running?Whatdoyouthink? with ourarmsaswedowhen walking use ourlegsaloneanddon’t balance want togostraight?Isitbecausewe posedly perfectmachinesiftheydon’t in snow. Whatiswrongwithoursup- wet tiresonadryroadorwhenweride wavy line,aswecanseewhenhave an uprightorarecumbent.We ride ina impossible whilepedalling,whetheron cle: ridinginadeadstraightlineis bicycles asIwasworkingonthearti- more obviousduringmyridingvarious an aspectofbicyclingthatbecame Henry Lemoine,inthepursuitrace. photographed passingachampion, far frombeingachampion.Hereheis cyclist ofthetime,buthewascertainly Faure, wasayoungwell-knowntrack Cycliste Internationale. recognised aslegalbytheUnion the verymomentthatarecumbentwas on February20,1934inParis.Thiswas illustrate theVelodrome d’Hiverevent published intheFrenchsportspressto Here itis,withourapologies.Itwas ly lostbetweenFranceandtheUSA. 49 (winter1999–2000)wasunfortunate- —A I wanttomakeanothercommenton The rideroftheVelocar, Francis A “key”pictureforthisarticleinHP rnfried Schmitz,QuartierGallas, Lioux, Gordes,F84220France, Reviewed byDaveWilson 24 March2000 Human Power Human Power called CycleVision. Itisonthetopicof tional recumbent-promotionalevent state-run events;andalargeinterna- association withanumberofsmaller winter; NVHPVannualmeetingsin the summerand“warm-updays”in presentations atfairs;competitionsin recumbent. strained byourallegiancetothe of HPVsingeneral.Andthisiscon- development andpromotionoftheuse objective shouldbetostimulatethe wants tobeinthelimelight,butits within themembership.TheNVHPV owners hasincreasedproportionately factories, thenumberofrecumbent commercial productthroughdiverse now beenmadeavailableasaserious example. Butsincetherecumbenthas and builders,followingtheAmerican tion madeupofrecumbentdesigners bers. Originallythiswasanorganiza- resent thelargestpercentageofmem- and recumbentownersridersrep- The NVHPVhasalmost1600members, 25,000 recumbentshereatthistime. estimated thattherearemorethan of recumbentsintheNetherlands.Itis totheincreasinguse bike-paths—and country isflat,windyandhasalotof recumbent-friendly infrastructure—our years. Perhapsthisisthankstoour has grownconsiderablyinthelastfew associations intheworld,NVHPV (the DutchHPVassociation). four years,chairmanoftheNVHPV father oftwogrownchildrenand,for am RonaldvanWaveren, 48yearsold, (translated byEllenWilson) Ronald vanWaveren EDITORIAL We In comparisonwithmanyotherHPV I’d liketointroducemyselfyou.I
Courtesy Arnfried Schmitz organize activitiessuchas Francis Faure passing Henry Lemoine inaUCI-sanctionedpursuitraceatthe passingHenry Francis Faure Ve lodrome d’Hiver in Paris, February 1934. d’HiverinParis,February lodrome to Lelystad.AbusforHarderwijkwill train fromSchipholAmsterdamairport er) andChallenge’s “Hurricane”. Bike”, Flevo’s “All-Weather” (Allewed- class criteriumssuchasThys’“Row- allowed. Newthisyeararethesingle- participated whenweatherconditions Britain, Belgium,FranceandHolland, among themthosefromGermany, km/h). Forthisattempt,foreignteams, break theworldhourrecord(over80 was madeinanearlierCycleVision to international frameworkarealeffort eminently suitableroadandinthis out foralldistanceraces.Onthispre- in USdollarsorEuros)willbegiven prizes totallingNLG10,000(over4500 time trialsandasix-hourrace.Cash the-hindmost” dragraces,one-hour criteriums, 200-msprints,“devil-take- cent accommodations.Onecanenjoy held onthe2700-mtest-trackwithadja- broadcast overloudspeakers. events, togetherwithcoolmusic,are parking lot.Announcementsofall take testridesonaspecialadjoining certain vehicle,neworused,onecan hand market.Ifoneisinterestedina strations, andasimultaneoussecond- there willbepresentationsanddemon- companies. Underthesame“roof” ucts ofDutchandforeignrecumbent can holdallthedisplaysofnewprod- on thisarea.Asingletentof1000m vehicles. Therewillbemanyactivities ment’s testinggroundsforhighway located inLelystad,onthegovern- 3rd and4th,2000.Itwillagainbe early inJune,ontheweekendofJune fourth successivetime,willbeheld your attention.CycleVision, forthe this lasteventthatI’dliketoaskfor Cycle Vision iseasilyaccessibleby International competitionswillbe ubr5,Srn 0023 Number 50,Spring2000 2 2000 m dren’s recumbenttrial/obstaclecourse; show; atoddlers’activityarea;chil- space, arecumbent-clothingstyle wonderful eventnextyear. that arecordnumberwillvisitthis to traveltherethisyear, butwehope has cometoolatetopersuadereaders W of thisissuemeansthatRonaldvan Editor’s note:Delaystothepublication ——— not affordtomiss. pening, whichyouasanenthusiastcan- ciation thathasbecomeanannualhap- is aninitiativebytheDutchHPVasso- that isatruebikerevival.CycleVision Cycle Vision 2000,asensationalfeast part ascompetitorsorspectatorsin from everypartoftheglobetotake ertheless inviteallHPVenthusiasts the reachofeverynon-European,nev- the Europeancontinentisnotwithin for moreinformation. NVHPV initiative).Seewww.ligfiets.net 2000 ConstructionContest(likewisean for alargedesigncompetition,theBike available recumbents;andtheawards cle coursewithalloftheNetherlands’ Chester R.Kyle IHPVA’sPraise from founder, In 1999ithad1000m visitors andmorethan100competitors. one-hundred volunteers,attracted3000 Vi only NLG7.50perday. In1998,Cycle B&Bs etc.Thepriceofadmissionis Oppertje.” Lelystadalsohashotels, overnight campingsitesat“The the eventforbothdaysthereare parking. Forthosewhowanttovisit the last3km.Thereisalsoadequate Cycle Vision shuttlebuswilltakeyou take youtoLelystadairport,anda Best Wishes, ulations. Keepupthegoodwork. photos, graphics,editing,etc. Congrat- and itisoneofthebestever—content, averen’s descriptionofCycleVision sion, anorganizationwithmorethan We I justgotmycopyof
Europeans, realizingthatatripto 2 of adultrecumbenttrial/obsta- Chet 2 of exposition Human Power —D ave Wilson , International Human Powered Vehicle Association
IHPVA PO Box 1307 San Luis Obispo, CA 93406 USA http://www.ihpva.org