Hello Kamran, I have been faced with this problem many times in the (nearly 35) years that I have been involved with such measurements. In my experience, this is the major and problematic correlation factor (in the product safety domain). It is the one that stands out above all others, when faced with correlation difficulties (especially when the tests are conducted on different continents as they were in my situation :-)
Very slight changes in the immediate-area ambient temperatures and the electrical service where you are making the measurements can drastically affect the temperature readings as can the location of the thermocouples on the transformer under test (as you already noted). Consider the following possiblilites (not in any particular order) ; A) Be certain that there is no extraneous air-flow over the trafo (from a room-fan, people constantly walking by in the adjacent aisles, or room air-conditioning drafts. ( I believe ambient air-flow to be a major factor when you are faced with correlation difficulties) B) Be certain your input voltage is a clean sine-wave. Distortions in the input voltage waveshape will affect your data. (At one time we tried to use a low-distortion/low harmonics constant-voltage transformer to keep the line voltage constant during temperature tests - even that small amount of input distortion was sufficient to affect our results). C) If the trafo is in a switching-mode power supply, the currents being switched will likely be affecting the indicated temperatures. D) Why do you discount rise-of resistance ? I believe this technique integrates the temperature differences across a transformer and yields a more repeatable result (in my humble experience). E) One problem I've found in transformer construction is that air can become entrapped, by the impregnating materia, between the outer-wrap layers. This effects the temperatures measured on the outer surfaces (again, a thermocouple location problem). F) Another problem is the heating effect on the trafo of adjacent hot components (cold ones or sinks too, I suppose) when transformers are evaluated in an end-product. Be certain that you and your vendors (and your test agency) apply thermocouples in the same location/s. Consider submitting a transformer, with thermocouples attached by you, to your vendor/s and test agency/ies. This will, at the least, rule out thermocouple location (and its attachment) as a variable. G) Be certain that you are using high-grade (low error) thermocouples and calibrated equipment. Good Luck! Regards, Art Michael Int'l Product Safety News A.E. Michael, Editor P.O. Box 1561 Middletown CT 06457-8061 U.S.A. Phone : (860) 344-1651 Fax : (860) 346-9066 Email : i...@connix.com Website: http://www.safetylink.com ISSN : 1040-7529 ------------------------------------- On Tue, 24 Aug 1999, Kamran Mohajer wrote: > > Hello EMC-PSTCers, > > I wonder if anyone knows of the method of measuring temperature limits > on magnetic components. I happen to get involved in this and found that > my results are different than the vendors result by as much as 10-15 > degrees on measuring on a same magnetic component. Even applying the > thermocouple to different location on a coil seems to give you different > results. Is there a method that I should be following to measure > temperature with thermocouples methods, not change of resistance, on > magnetic parts such as transformers, coils, etc.? > > Thanks, > > > *********************************************************************************** > Kamran Mohajer > DSL Compliance Lead > Cisco Systems, Inc. > Phone(408)-525-6121 > Fax(408)527-0495 > kmoha...@cisco.com > *********************************************************************************** > > --------- > This message is coming from the emc-pstc discussion list. > To cancel your subscription, send mail to majord...@ieee.org > with the single line: "unsubscribe emc-pstc" (without the > quotes). For help, send mail to ed.pr...@cubic.com, > jim_bac...@monarch.com, ri...@sdd.hp.com, or > roger.volgst...@compaq.com (the list administrators). > > > --------- This message is coming from the emc-pstc discussion list. To cancel your subscription, send mail to majord...@ieee.org with the single line: "unsubscribe emc-pstc" (without the quotes). For help, send mail to ed.pr...@cubic.com, jim_bac...@monarch.com, ri...@sdd.hp.com, or roger.volgst...@compaq.com (the list administrators).