Resistive Trimmer ratio accuracy

[moffy]

After searching the internet I haven't yet found definitive or rule of thumb info about how accurate the ratio of a trimmer is with temperature. The best I've seen is the Vishay AccuTrimmer document, but I am interested in more general trimmers. Does anyone have any guidelines or experience? It's not for anything specific, but something like if the trimmer has a bulk tempco of say 100/ppm/C then the ratio is stable to 10/ppm/C (I just made those figures up as an example). Thanks.

[matseng]

This is an interesting subject, I've searched for the same info as well without any joy.

Maybe someone with a 8.5 digit meter and an oven could run some tests?

[SArepairman]

typically its quoted at 30-60ppm for a decent cermet and 10ppm for a vishay metal foil trimmer.

the datasheets have a rheostat quote and a potentiometer quote, potentiometer = ratio (wiper), rheostat = overall (end to end).

http://www.vishaypg.com/docs/63054/1260.pdf
I think the wiper PPM pretty much dominates, so the 25ppm figure is used?
digital one is like 4ppm ratio, 30-100ppm rheostat.

you know, im curious about this, because, i have no idea where i read the numbers I gave you up top. I have some bourns 10 turn trimpots, tommorow I will wire one up in my heater box and give you a number.

[matseng]

I can't find any rheostat vs potentiometer data in the datasheet you linked to.

Temperature Coefficient of Resistance (TCR) 50 ? to 10 k? End-to-end : 10ppm/C
and
Temperature Coefficient of Resistance (TCR) 5 ?, 10 ? and 20 ? Through the wiper  : 20-25ppm/C

Neither of these figures describes the change of ratio, only the change of resistance end-to-end or end-to-wiper.

[SArepairman]

Yea, but when you are using the wiper, you get 25ppm pretty much, no matter how stable everything else is, since your two resistors are both end to wiper, you can say its 25ppm. I think you would need two wipers for it to cancel out.

I'm going to set this up by using a 6.5 digit meter and my styrofoam/halogen bulb tempco tester. I'm gonna measure voltages and use a external CC calibrator in order to get rid of the kelvin connection problems that would require me to redesign my box for testing a three leaded component, and I want to see how it handles a high current rather then relying on multimeter auto ranging.

If you do this experiment its critical to let it stabilize for like, 24 hours + prior to heating, because the mechanical strain effects absolutely dominate the drift. Don't bother doing this with a ebay trimpot, like a borang or boner. From what I read this wait time should not be required with a metal foil trimmer.

[matseng]

I've also got a 6.5, so I'll try to replicate your tests when I get back to my lab in a couple of weeks.

Please also test with a elcheapo ebay single turn pot as well, it might be interesting to see how it turns out.

But what is causing the increase of 15ppm when using the wiper?  Is it the metal in the wiper that have such a high tc?  If so I'd guess that it wouldn't affect the ratio tc too much.

[Kleinstein]

The problem is likely die contact of the wiper to the resistive part. This is not a perfect point contact, but more like a combination of several points. Over time and with temperature some of the local contacts can get better or worse.

The big trouble is, that this effect is not really reproducible - it can more or less randomly depend on the position of the wiper, age and many other effects. So it's more like a range of uncertainties, not a single fixed value that is easily measured. So measurement would need something like 100 individual measurements with something like 10 samples and 10 wiper settings each.

This is probably the reason one does not find much information in most data-sheets, even though this would be important.

I won't expect an easy rule of thumb, like 10 % of the end to end resistance TK. The effect will also depend on the physical size of the trimmer. This will be also a combination of a TK and some more hysteresis like and aging effects. So It's already difficult to describe it in numbers.

[DanielS]

One thing to keep in mind: it is a trimmer. It is intended to trim something like the last (0.(0))1% of a resistor's value. If you have 1000ppm drift and uncertainty on 1% of the total resistance, that 1000ppm only contributes 10ppm to the overall value.

[macboy]

Used as a rheostat, a trimmers usually have very bad tempco, since the resistive material is often carbon film which has a high tempco. Cermet are better and more stable over other conditions too, but have a limited life of ~100 adjustment cycles!. Metal film would be the best but horribly expensive.

Used as a pot however, the important thing is the ratio; you are using it as a voltage divider. Here, it doesn't matter if the overall tempco is 1000 ppm, as long as the tempco of the top part (one leg to wiper) is the same as the bottom part (wiper to other leg). If that is true, you will still get the same proportionate voltage output, even as the end to end resistance varies wildly. The tempco of the wiper may also be important if you are not feeding a high impedance circuit.

[Conrad Hoffman]

Agree, trimmers are evil, especially for long term reliability. I recently built a Geller "JCan" for resistor noise measurement and was shocked at how bad carbon trimmers can be, way worse than carbon resistors. If you're going to use a trimmer, pad it down so it covers the absolute minimum range required. Or, install posts and solder in a selected resistor!

[Marco]

Wirewound pots are supposed to be relatively low noise, but the resistance curve is a bit bumpy for obvious reasons.

[matseng]

I recently built a Geller "JCan" for resistor noise measurement and was shocked at how bad carbon trimmers can be, way worse than carbon resistors.

Did you use a real BF244A in your jcan? The 244 seems to be "no longer in stock" in most places...

[Conrad Hoffman]

Yes, got it from Geller along with a variety of other parts. There may be other JFETs that will work, though I haven't done a detailed search.

[moffy]

Thank you for all your feedback, very informative. I found a huge caveat for using a trimpot as a voltage divider with little load, it appears to need a certain minimum current to keep the wiper clean, below that the contact can go open circuit.
I agree with DiligentMinds that trimmers are generally undesirable. While working in the military aerospace industry, trimmers were forbidden. You had to use SOT(select on test) resistors. But there are certain precision applications like trimming voltage references where they are recommended.
Sorry for the delay in getting back I have been experiencing a: "Bad Gateway Error 504".

[Edwin G. Pettis]

@ Moffy,

I don't know where you found that 'caveat', but it is nonsense, having checked with some other engineers, none of us have ever found this to be true, the wiper contact is not current dependent.  I have pots in circuits that have not been tweaked in years and they are still working just fine.  As an example, the HP3456A is full of trimmers and they are quite stable after being untouched for years.  If they are a bit 'creaky' after being readjusted, just rotating the wiper back and forth a few times restores complete functioning.

A lot of trimmers are used in my primary resistance bridge, they have been stable and work just fine after decades.  I have not found any mention of such problems in the literature from trimmer/pot manufacturers including detailed handbooks on the subject.

In wire wound 'pots', it is the wiper that causes the higher TCR, it is a combination of the alloy it is made out of and the contact variation with the resistive element.  Reducing the current through the trimmer effectively reduces the TCR effect.  In other 'pots' with higher TCR resistive elements, the TCR of the wiper becomes less noticeable as it tends to be less than the overall TCR of the element, reducing the current in the wiper will still have the same effect though.

It is the selection of the correct pot for the application which is critical (when needed), it should be chosen for best resolution, minimal self heating, noise and TCR, it all depends on the circuit's requirements.

[moffy]

Thank you for that, you have reaffirmed my faith in trimmers somewhat, but the "caveat" is real. I have uploaded the relevant PDF from BI technologies. Read the section on page 103 "C: Correctly Using the Trimmer to Avoid Dry Circuit Risk" there it talks about a minimum current above the nano amp level being necessary. For a second reference "The Circuit Designers Companion" by Tim Williams writes on page 83 of some trimmers requiring 25uA to keep the wiper clean, but it is only some trimmers, others should be treated the exact opposite!Go figure! It's hard to know which is which.

[Edwin G. Pettis]

Hello Moffy,

I did some digging around and there has been updates to trimmer/pot data sheets in more recent years.  According to Bourns; Wirewound pots have lifetime of decades, even in dry circuits.  Cermets its more like 2 to 5 years in a dry circuit (< 25uA thru wiper) and some like even more >100uA.  And yes, for best stability and TC you want wirewound, with low current thru wiper for best TC.  Use true potentiometer mode, not rheostat mode when possible.  Bourns is the only suppler left that are certified mil-spec / aerospace suppliers for trimmers, and those are all generally WW for stability and low noise versions.  Now pots that expect to see a lot of use or need high resolution (like for servo etc.) will tend to be cermets.  WW pots like to be set and then stay there for a long, long time.

The required current through the cermetic wiper does explain some slightly odd behaviour I've seen in the past when re-tweaking a circuit after it had set for some time.  Some years back there was no mention of wiper current in the data sheets that I could find.

[moffy]

Thanks heaps for the clarification, that information is so hard to come by, it's worth gold. Wirewound for potentiometer operation.