Trimpot temperature coefficient and drift over time

[NaxFM]

Hi everyone, for a personal project I need to use a 25 turns cermet trimpot (the usual bourns ones) to fine tune an high precision circuit.
I never worked with trimpots for high accuracy, and I was wondering about their stability and drift over time.
The temperature coefficient of a (cheap) trimpot is terrible at the usual 100ppm/C, however the product will be ovenized and kept at a constant temperature.
I'm worried that even at a constant temperature the drift over time would be too much. Do you happen to have any figures or numbers about the stability of trimpots.
If I leave it on for a year at a constant temperature, how many ppm will I be off?
I can't find any document related to this topic, maybe you know something that could help me.

[mag_therm]

A look at the Bourns  pcb range shows that the long skinny through hole RT12 milspec type has among the best enr.
As you probably know, stability is often improved by putting trimpot in network with stable resistors above and below to minimize trim span and hence variations.

What drift can you tolerate?

[gbaddeley]

Can you design the circuit such that trimpot tempco has minimal effect? Eg  pad it out with lower tempco resistors?

[NaxFM]

What drift can you tolerate?

Let's say that I "can" tolerate any drift because I can pair the trimmer with precision resistors to reduce the effects of drift.
As in many cases, it's a matter of cost, I wanted to explore all the possibilities I had before going the route of an extremely reduced trim range which would give be best stability but also introduce the problem of not enough range.

[NaxFM]

Can you design the circuit such that trimpot tempco has minimal effect? Eg  pad it out with lower tempco resistors?

I'm already doing that, but the circuit should have a quite large trim range of 200-300mV, and I am already at the limit.

[SeanB]

Not possible to put in some padding resistors and shorting links, so you can have the trim range split up into 2 or 3 ranges, and then have a set of jumpers that you use to move the pot range into. Depending on the circuit you might even be able to do without the pot at all, and simply have a few preset ranges for offset adjustment. Otherwise if it is likely to be a rare job use the military approved method, which is to use 2 decade boxes, to select a set of fixed value resistors to use to set the offset.4 turret pins to solder them to on the PCB, and then select on test for the right resistors in the range.

[Zero999]

Can you design the circuit so the actual value is unimportant? The ratio of resistances from the wiper to either end should have virtually no temperature drift.

Please post a schematic.

[mag_therm]

A description of pot testing including enr and crv.

https://epci.eu/resistors-potentiometers-basic-principles/

[Terry Bites]

Can you show circuit and location of pot.
There may be a better solution.

[richard.cs]

A colleague of mine claims the multiturn ones are much less stable with temperature, based on testing with his previous employer. The datasheet accurately represents the stability of the resistive element, but thermal expansion of the mechanics apparently causes small jumps in resistance with temperature, the wiper physically moves. These jumps, so he tells me, are discrete steps, and have some hysteresis, so are hard to compensate for.

I have not tested it myself, but it certainly seems plausible, and as a result I tend to use coarse/fine pots rather than multiturns in precision designs.

[Terry Bites]

Pots will undermine precise analog circuit function. Resist the urge to tweak or the law of diminishing returns and frequent future twiddles will come to haunt you.

If you can, get rid of the pots and fix the offset, gain, drift and other compensations etc in code.