Purdue University Purdue e-Pubs
International Compressor Engineering Conference School of Mechanical Engineering
1974 Reliability Testing of Aluminum Magnet Wire Connections for Hermetic Motors J. L. Spears A. O. Smith Corporation
Follow this and additional works at: https://docs.lib.purdue.edu/icec
Spears, J. L., "Reliability Testing of Aluminum Magnet Wire Connections for Hermetic Motors" (1974). International Compressor Engineering Conference. Paper 94. https://docs.lib.purdue.edu/icec/94
This document has been made available through Purdue e-Pubs, a service of the Purdue University Libraries. Please contact [email protected] for additional information. Complete proceedings may be acquired in print and on CD-ROM directly from the Ray W. Herrick Laboratories at https://engineering.purdue.edu/ Herrick/Events/orderlit.html RELIABUITY TESTil!G OF ALUI1INU1'1 HAGNE'r WIRE CONNECTIONS FOR HERHETIC JJ[QTORS
Jerry 1. Spears, Senior Haterials Engineer A. o. Smith Corp., Electric Motor Div., Tipp City, 0., USA
1 IN'l'Jl.ODUCTION to the aluminum wire and system compati'oili ty ·was maintained. However, alwr,inwn wire even Hhen The heart of the hermetic compressor is pro the stator. perly stripped is not easily fused together, and As "•~ith the other components sealed into the her with copper metic wire in the connection, disslinilar envirom~ent, the stator is expected to pro melting points vide made heat fusing impossible. Be years of reliable service. The motor manu sides heat facturer, fusing, other methods of connecting however, does not look at the hermetic aluminum wire combinations stator as a unit but have been explored. rather as a selected group of For instance, sample connections materials and processes that will were made with provide the end such advanced techniques as capacitor user with the reliability and performance discharge desired. welding, laser beam fusion, and compression One of the processes that has years of proven cold field fusing. However, the feasibility of roliability is the internal stator connections these tech Nhere niques for a production line were soon apparent,and copper magnet wire is employed. In the past more feH years, practical approaches Here considered, such as, however, the hermetic motor manufactur compression ers have connectors, insulation piercing com been designing and producin~ an increas pression ing nlliQber connectors, and soldering. The balance of almQinum Hound hermetic stators. of this li.conomic considerations paper will present experimental life test notHithstanding, aluminUlll ing data on connection ma~net wire can offer inherent methods incorporating in physical advantages. sulation piercing compression For exampl:o, there is evidence of better connectors and sol kmce resis dering. to long term ther:nal aging when organic in sulations are applied to aluninum wire, as com p.cl.red vii th copper. It is corn.'llon experience that 3 Tli:ST CRITERIA AND SAMPLE PREPARATION al1J.minum Hire Hinds easier once proper tensions are adj-tlsted. Springback characteristics in alumi The development num allow for uniform of these reLtability tests did not coil conformation and more come about efficient use of available haphazardly. It was known from pre space. These advantages vious test experience, partially offset the lower conductivity with non-her.metic type values of Inotors, that four inherent characteristics alwninum wire. Historically, hovJever, the of alu primary minum were detrimental to a connecting drawback in using almninum magnet wire has been process. associated with These were: aluminum oxide formation, cold flow, the'development of reliable con coefficient necting techniques. The of thermal expansion and suscepti objective of this paper is bility to corrosion to discuss the state of when coupled with a disf.;i:::rlilar the art in reliability material. Since testing of aluminm~ the compressor environment mini ma,o;net Hire connections for mizes hermetic motors. the effects of oxygen and moisture, efforts The paper will also deal with the were types of almninum concentrated on evaluating connections from wire connections available for the standpoint the industry and a discussion of cold flow and thermal expansion. of their relative l'lli th these considerations Tneri ts. Comments and data resarding in mind, the follo 39 nections. All Connections Hade With Hochanical Clips b) Connection teclmiques repre senting proven methods, along ]; .s. 16 AWG aluminum magnet wire was not with large numbers of samples (3 samples) stripped. to statistically test new connection processes. c.s. 16 AWG aluminum magnet wire was (3 samples) chemically stripped. llo reduction E. Sample a) Mechanical connections pre in bare wire diameter. Preparation pared with manufacturers' prescribed tooling and pro H.S. 16 AWG alwninum magnet wire cedures. Fuse/braze con 5~6 stripped mechanically reducing nections prepared as per (3 sampl bare wire diameter 5%. established production pro cess specification. Sol~ H. s. Same as above, except bare wire dered connections were made 10% diameter reduced 10%. in accordance to solder man (3 samples) ufacturers' recon®endations. b) All samples were insulated M.s. Same as above, except ba~e wire l-Ji th heat-shrink polyester 15% diameter was reduced l5jb. film tubes. (3 samples) F. Test Criteria .Any condition, such as a radi This test was set up in a modified compressor cal change in resistance at the shell in order to introduce a hermetic envirorme,lt joint, overheating, or fluctua (R-22 & oil). The test was manually controlled ting millivolt drop readings with the follmving parameters: alerted the technician as to a possible connection failure. A. Constarttly balanced current - 30 amps. The test could be interrupted and individual connections B. Current off when slot winding temperature could be checked for deteriora reached 350° F. (Observed nonnal override of tion. 20° F.) It is important to realize from this point on that c. Cool dotm to 100° F. and repeat current on test results and recommended practices reflect on cycle. the materials and processes used by A. 0. Smith, and may not necessarily be true with other mater We found that the Lab ambient temperature had ials or Jnanufacturers. The evolution of our test some influence on the rate of rise, but typical 6.0° F. experience is described in the follm•ing test pro winding temperature rate of rise was 5.5' to grams. It is from these tests and supporting data per second. that we feel confident in the performance and re liability of the method chosen to connect aluminum Basically, we arrived at the following results wire in hermetic stators. after 1000 cycles: l. Connection temperatures ranked in order of ~ HOTORETTE TEST lowest to highest: This test was developed in order to closely simu A) 5% H.S. - lmvest operating temperature. late the conditions of an actual stator. A motor ette was wound with multiple turns of 16 A1rJG wire. B) Chemical stripped. At 15 intervals, the winding was broken and a con nection made. Each connection contained, in addi C) lO% N.s. tion to the 2 - 16 AWG aluminum wires, a 14 AWG stranded lead wire and a 20 A1"JG copper magnet lllire. D) N.s. Each of the 15 connections were thermocouple equipped. A thermocouple was also imbedded in the E) 15% H.s. - highest operating temperature. motorette slot to monitor winding temperature. The 15 connections were prepared as follows: Figure I illustrates the temperature pattern of the connections at various cycle intervals throughout the test. 40 sign foaturGs. FIG. I FIG. II .SERIES CONI!ECTIOlJS, ·300, HECHANICAL CONNECTOR I _ALUNINUM MAGNET WIRE .22._0, STRAlmED OP<. •. COPPER 200. •. 'llliRI1ETIC . ·.. .)_ l ·-· ·- _\[___ .~·CONNECTION TEMP • LEAD 'L-II]'IE ...... " --- ' :5% BARE WIP.E ,------\ .150 'DIA. REDUCTION COPPER L ____ MAGNET WIRE 290 490 6_00. BOO FIG. III _(a) Test Series #1: 20,000 thermal shock cycles in R-22 THERIW. & oil. SHOCK 2 min. heating at 40 amps; cool 8 min. & · CYCLES repeat. 16 AvlG al'l.llllinum wire, total connector CJ1A " 10230. 2. Connector Test Results Hesistance and millivolt drop readings were Description taken on retained samples and on the test sam Each ! Final MV Hax. Con- ples Group Failures 1 Drop Read- following the 1000 cycle test. The re 20 Samples nection sults Noted ings ( Avg. ) of these data also confirrned the opti 9 Serration Temperature Mized condition reached with 5% inax. bare dia Alwnimllll Wire F.eter reduction None 3.14 plus mechanical clip to achieve ~re-str:i.pped 235° F. a connection. 7 Serration Aluminum Wire None Pre-stripped 3.54 240° F. 5 AUTm;,TJ:C CURREIIT CYCLING TEST Soldered 65/35 Zn/Sn None The 2.66 I motorette test results were lirni ted in scope Alcoa 68 172° F. but provided Flux a direction to go with regards to ex ?"··serration panded testing processes. We felt the following Alumimllll hire I i teuts were essential 4 Samples 3.53 I in the expanded program: Hot Stripped Failed 242° F. I A. Constant current self heating. B. Tirne on/ti;ne off cycle adjustment. FIG. III (b) ITest Series #2: Essentially the same as Series r:;. Millivolt except #1, drop monitoring during the current 19 ArJG Al. wire; total connector GHA = on portion 6180; of the therraal stress cycle. current density adjusted to give same heat ~ng rat e. D. Continuous temperature monitoring. Description - 20 Samples/ Failures Final HV Hax. Con- ·"-• Use of control sWJ.oles (fused and brazed :Group Noted Readings nection copper equivalent)- in each test group. (Avg.) Temp. 9 Serr. Al. F. Increasing vJire Stl 'ipped the number of samples of each None 3.6 202° F~ type 9 of connection tested to arrive at a better Serr. Al. statistical result. 1·.'ire, No Strip None 3._95 218° F. 7 Serr. Al. 1.-Jire Strip_ped •o. Running parallel tests, R-22 as well as R-12 None 4.7 226° F. environments. 7 Serr. Al. · Wire, No StriQ ~ 4.57 Hith these factors in mind, the initial 65/35 Zn/Sn - group of Soldered SO connections were tested under autoMatic None 2.6 179° F. current Copper Wire cycling conditions. Figure II shows schematically the Fused & Brazed None series connections, i1s of this writing, four 2.8 191° F. (4) groups Copper Wire of connections have been evaluated under autOJT,atic constant coxrront cycling Banded & Fused None 2.5 199° conditions. A Barrel F. breakdoNn of the basic test parameters Con- and results nector, No are shown in Figures III(a), (b), (c), and For (d). Stripping Uone obvious reasons, tradename descriptions of 3.15 188° F. chanical me connections are deleted in favor of de- 41 for screening new materials and processes, It was for these reasons that several stators have been tested vii th the primary purpose of evaluating the FIG. III (c) I connection process. Although the test itsdf is Test Series #3: R-12 & oil. Test modified to pro-· quite rigorous, the individual connection is not duce higher peak temperatures. 19 AWG Al. wire; stressed nearly as much as in the cornponent test ' total C11A "' 6180. Test off at 10,019 cycles be described in the first part of this paper, Fip; cause of ext ·reme seven tv. •· ure IV sunrnar:i.zes the Ha.pid Reversal Test results Description - in regards to alwninurn-;wund stators and a partic 20 Samples/ Failures Final HV Hax. Con- ular connection process. Data is currently being Group Noted Readings nection : obtained on actual connection temperatures 1vhile (Avg.) Tenm. the rapid reversal tests are in progress. To do 7 Serr. Al. this, stator connections were therrnocouple equipped _l-LireJ_li.O Strio 9 7.1 31~° F. prior to varnish irrrprcgnation. Special fittings Soldered were used on the compressor shell to brine; up to '95 Zn/5 Al, None 4.9 270° F. 10 thermocouples out for direct temperature moni , Copper Wire toring. Temperatures being recorded in sorne ir:i Fused & Brazed None 3.9 265° F. tial tests Here dampened by the impingement of cold :-9 Serr, Al, liquid refrigerant returning from the condenser. JrJJ:.re Stripped 2 5.1 360° F. Hodification of internal compressor plur1bing pro 9 Serr. Al. vided for unifonn stator winding temperatures, -~ hfi1'~, --~~~ Strip 7 7.3 345° F, thereby enabling cornparative temperature studies Barrel Con- to be made on various connection processes. Fig nector, No ure V illustrates typical connection operating Stripping 6 6.4 310° F, ~ temperatures under rapid reversal conditions when FIG, III (d) refrigerant flm1 is diverted from the connection ! Test Series 1!4: t:ssentially the same as Series #3. end turns. ' Objective bei.n5 to detemine if l[V readinr:s can : be correlated to operating ternpEOrature. Data be loH is c\lru $000 cycles and test continues . Description - 20 Samples/ 5000 Cy- ' Group Except Failures cle NV Hax. Con- ; '\".Jhere !Joted Noted R(adings nection FIG. IV Avrz.• ) TelTIP. RAPID ?.EVERSAL STATOR TESTS viiTH 7 Serr. Al. 14-(Failed GOill.r;CTIIm TSCH:iJIQULS EVALUATED \dire, llo Strip Less Than 8.3 457° F, Stator Total Connection Description ·Size (HP) Reversals and Test ~esults ~$2. Cyc,) . -9-~-s err-:--AI :-- 3 796757 7 Serr, Al, wire stripped 3 403320 Same as above 1•1ire 1 No Strip 5 7.1 402° F. 9 Serr, Al. 3 727088 Same as above Hire Strioped 1 6.8 376° F. 3 690722 65/35 Zn/Sn soldered -- --~-.~· -~-- --~--- s-aiderecc·· ~------5 12207 Barrel connector, Al, wire not stripped, short life ..2~35 -~!1/Sn None 4.7 269° F. -~~- Copper ~-.'ire from compressor bearing Fused & 3razed None 3-5 272° F. failure ··fiarret Con- 5 503690 Barrel connector, Al. wire nectar, No not stripped Strip 1 6.4 334° F. 5 $01242 Same as above 13a.·r-:re·1 Con- 5 50h2ll 7 Serr. Al. 1-1ire stripped nectar, i'lo 5 814533 7 Serr. Al. wire not . Stri:e (Soecial None 6.2 303° F, stripped Cutting Tooth 4-(Failed 5 500597 Sa.Qe as above, 2 of 6 con- Fla"G Clip, !Jo Less Than 7-9 L,ll° F, nections overheated Strip. 4 500 Cyc.) 5 534976 7 Serr. Al, wire stripped s~~p_:t,es 5 583103 7 Serr. Al, wire not stripped ----~------··----· 5 514760 9 Serr. Al. wire stripped 5 601599 9 Serr. Al. wire not stripped 6 HAPJD HEVERBJ\.L TESTING OF ALUMINUM-WOUND 1 of 6 connections overheated JTE:CHANICALLY-CON1JECTED STATORS 5 577875 7 Serr. Al. wire stripped 40 252584 Barrel connector copper wire Added assurance of reliability for mechanically not stripped connected almninum-wound stators is being gained 15 167877 Barrel-connector copper wire through Rapid Reversal Testing. This accelerated not stripped test stresses a stator to its fullest, not only fran; the mechanical forces associated with rapid reversing, but also Hi th elevated i~inding tempera tunes and frequent cool-down periods. This test has gained wide acceptance as a recognized method 42 I Q, ~· Q) ,The reliability of a connection even (.'l::J 2! H T! at operational .ZE--4 •rl ~ temperatul'es as lm~ as 170° F. is questionable E-< H ;:2" ~ 8 SlWll1ARY 3. Klein, G. 1v., "Mechanical Connectors for Cop per and Aluminum Wires in Hermetic Motors, 11 From our testing, it becomes apparent that alumi lOth Electrical Insulation Conference, Sept. num magnet wire will have to be pre-stripped if 20-23, 1971. the conventional and commercially available 7 and 9 serration splice clips are used for mechanical 4. Petry, R. D., "Some Evaluations of Magnet Hire connections. The 9 serration clip appears to be in a Refrigerant Atmosphere," 9th Electrical the better of the t.~o, Early test results indi Insulation Conference, Sept. 8-11, 1969. cate that barrel connectors with special alloyed ,Perforated liners will perform well even on non 5. Spears, J. L., "Reliability Testing of Alurni stripped aluminum magnet wire. This connector num )l!agnet Wire Connections for Hennetic may be used pending the successful outcome of Hotors," llth Electrical Insulation Conference, additional tests. Sept. 30-0ct. 4, 1973. 43