Edwin, you seem to know quite a bit about resistors. My company (RCD Components Inc) has been producing resistors (all kinds), as well as a wide range of passives for nearly 40 years. Foil resistors have their place but precision wirewounds (when properly designed) offer the best overall performance, overload capability, power coefficent, ESD, etc. This is the reason that esentially all Resistance standards utilized at various test labs around the world feature WW construction, not foil (RCD makes many of these Resistance Standards). Anyway, I have some questions about your engr company, capabilities, etc., so send me an email (mikea@rcdcomponents.com) Rgds, Mike Arcidy, RCD Components Inc, Engr Mgr

After viewing the Vishay videos again I've found more discrepancies that I will point out. Note, I am not saying that the Vishay foil resistor is a bad part particularly, I am saying that the test methods and resulting claims are rather questionable.

You'll notice that the displayed TCR/time graph has rather large 50 PPM /'C vertical divisions, which is fine for showing a higher TCR resistor, but the resolution stinks for showing any real variations in low TCRs. Very skewed! The claim of tying TCR to reliability is totally specious. The TCR of a resistor is the intrinsic characteristic of the alloy used in making it, while the TCR can change a bit during the manufacture of a resistor, it tends to be quite stable over the life of a resistor if operated within normal parameters. A resistor with a TCR of 500 PPM/'C can be just as reliable as a 5 PPM/'C TCR, TCR has absolutely nothing to do with the reliability of the resistor over life! How the resistor is constructed has everything to do with it, not TCR!

In the foil TCR trace, as fuzzy as it is, it would appear that there is a small 'bump' in the trace after spraying but the trace returns to near zero afterwards. Since the thin film trace shows that the temperature of the resistors is stable beyond the end of the graph, why does the foil trace return to near zero? This is not correct, the trace should change to the new temperature resistance value and stay there, not return to zero!

I find it quite curious that the so-called accelerated life tests of a 5 second, 7x overload is supposed to indicate the behavior of the resistor in the future. This does not comply with any life tests used on precision WW resistors, which range from hundreds to thousands of hours at maximum 125'C ratings. Thermal shock cycling (MIL-STD-202) from -65 'C to +125 'C would be a better indicator of a resistor's stability over time.

Another interesting discrepancy you'll notice in the first test, the TCR goes negative by about -50 PPM before shooting off positive at 1:10 into the video but watch the screen when he starts the second test......at 1:25, the negative loop drops to over -150 PPM and the time scale has changed from the 1:10 display! The loop started BEFORE the second time division. In the second test, the loop doesn't start until after the second division and is of much larger deviation and duration! What is going on here? Obviously something has changed, the time scale is still the same as the first test but the resistor traces have significantly changed. Compare 2:08 to 2:15 in the video, the traces have changed again between these two times! The first trace (at 2:15) now has the small loop back but look at the second (negative) trace, it not only doesn't go below the -500 PPM line but returns to about -400 PPM line, unlike the first one which stayed at -500 PPM (AT 2:08)! More bugs in the soup!

I also find the focus of the camera to be quite interesting. In the TCR test, it started fairly well focused but got progressively more out of focus with each following test until the foil test was nearly unreadable. In the static discharge test, the camera remained very sharply focused.

One other point I noticed in the life test, one resistor was 100.2 and the other was 100.4. The parallel value should have been 50.15, but the reading they showed was 50.5, a rather large error if you ask me! No indication of how they were doing the readings except that it was sloppy. With a video, a digital resistance bridge could have been used and the readings visible to the viewer. Here there was nothing, just a before and after with poor resolution and questionable results.

The static discharge video looks good and sounds good. But here again, was this test set up in such a way that the foil resistor was assured of surviving it? Notice the lack of resolution in the readings again. Just how long was that pulse? The time integral here is very important and you'll also notice when he was supposedly zapping the foil multiple times, the resistor wasn't being shown and the time interval was notably long in between each supposed pulse. There is no note of the pulse characteristics and no indication, beyond the first test with a visible flash, that further pulses were being applied, no monitoring of consequence. I have never seen a film/foil resistor that did not change value after exceeding its maximum specs in any manner to some degree. In fact, operating one near its maximum specs causes drifting of value over time as evident in these resistor's own data sheets. Multiple 10kV, 1 amp discharges through a film/foil resistor is guaranteed to cause some shift in value. Their own data sheets prove that, and I would suspect that the shift would be larger than the long term shifts in the data sheet.

Static discharge has never been a particular issue with wire wound precisions, only films! Their flimsy contention that Vishay foils are the only resistor technology for tough applications is just pure hot air! It appears that their foil resistor has some good characteristics, but their shabby attempts at making them look even better just makes it all the less believable. If they had done some really true comparisons with accurate measurements, that would have gone much further than all the tricks they used in these videos. We all know that components are not perfect, there are always some small flaws. To show these small flaws would be much more credible than to do limited resolution and questionable measurements and claim perfection.

To a fair degree, these videos prove very little. There is no visible monitoring of the test parameters, nothing to show what is actually taking place only that you accept at face value what Vishay wants you to believe. The same type of tests can be done with proof of measurements, to prove the claims, why doesn't Vishay?

Best regards,

Edwin G. Pettis

Pettis Engineering

Grand Junction, CO

Regarding the two Vishay videos presented in your e-mail 'bag', I have some comments to present on the subject. Having looked over the two videos several times, I find that they are fundamentally skewed, particularly the TCR 'stability' test. Sorry I have not written sooner, I'm busy with another project with Robert Pease's collaboration, and it is getting more involved as time goes by. Trying to catch up on my reading some.

First, Vishay manufacturers all three types of film resistors used in these tests. I am genuinely not surprised at the lack of any competing resistor types, particularly precision wire wound resistors. An overload of only seven times rated power for five seconds is rather on the low side for an overload, certainly for wire wounds. I congratulate Vishay on being able to at least withstand some overloading without gross characteristic change. Since the resolution of the reading was limited to a mere 0.1%, this could easily hide significant changes of tighter tolerances. However, such overloading in the past has always caused some permanent, significant shift in characteristics. This test did not have sufficient resolution to indicate just how good the stability actually was. I believe that it was intentionally designed to do just that.

An examination of their data sheet indicates a significant shift in value over a relatively short period of time at rated power. An overload of seven times rated power is guaranteed to produce bigger shifts. This fact was glossed over in the video, indicating the resistor had no significant shift. The video did not specify if the overload was the higher or lower power rating since these parts come with two ratings (70'C or 125'C?). I suspect it was the lower rating that was used. This is a significant point in my opinion.

The TCR video was obviously skewed by the selection of 'typical' higher TCRs of the other film resistors. These all had TCRs exceeding 10ppm/'C. I also note that only cooling was used as the test of choice, partly for convenience, of course, but most film resistors tend to have lower TCRs below 0'C than above because the ceramic substrate on which they are made contracts at a slower rate than the resistor element does at elevated temperatures. Above 0'C, the differential expanding coefficients of the materials involved causes more stress and, thus, higher TCRs. That is not to say that Vishay has not ironed out their TCR bumps over a limited temperature range. Again, I note the lack of any competing resistor types in this test.

Now, if you want to see some really good stability tests, how about thermal shock cycles, MIL_STD-202. I have precision WW resistors that make films look pretty sorry. How about fifty resistors, run through thirty cycles (six times the normal five cycles), 1500 component hours, with zero failures! Further more, the average shift in resistance was only 12 PPM, with a maximum of 15 PPM and a minimum of 8 PPM. This is over -65'C to +125'C temperature shifts. Note that these shifts are significantly better than the load life stability (a mere 2000 hours at 70'C), shelf life stability, high temperature exposure or moisture resistance of films. These resistors are capable of running indefinitely at 5 times rated power (at 25'C ambient) with no detrimental effects. Their ratings are given at 125'C with absolutely no de-rating of any sort nor effect on characteristics. No, these are not hermetically sealed either, quite unnecessary and they are standard production line parts.

The bone I'm picking here with Vishay is that they often compare apples to oranges and work very hard at making their parts seem to be so much superior to other resistor types when that is not necessarily true. Films have some very good characteristics, particularly with parasitics (when they are significant to the operation of a circuit) but they do not own the resistor world when it comes to stability, reliability or precision. I also point out that the MIL- specs concerning film resistors and wire wounds are different so that a direct comparison is not entirely possible.

Best regards,

Edwin G. Pettis

Pettis Engineering

Grand Junction, CO