Changing TCR with time

[analityk]

Hello.

Resistors have TCR. Ordinary carbon resistor have +-200ppm and more. But now i have looking for resistor, where i do not care about absolute value even in changing temperature, but i have to know how big change of TCR is in time. I want use it in thermostate, for polarisation of NTC. Nomen-oment for NTC resistor drift of TCR are specified.

Did i should use low TCR resistors (metal film) or ordinary carbon film resistor can fit in my ideas?

Sincerely.
Simon

[Teledog]

Check the spec sheets for the resistor(s) you intend to use?
Vishay & similar are available online, with derating factors, etc.

[Gyro]

Just use a metal film resistor. The change in TCR with time should be insignificant.

[Terry Bites]

Are you needing to linearise a thermistor. Is the TC of the linearising resistor that interests you?
The TC of the thermistor is very large compared with any fixed resistor. So the TC of the the fixed resistor has almost no effect.

[strawberry]

carbon film ~200ppm minimum
metal film =~ thick/thin film 
thin film is relatively better stability
bulk metal  0..10ppm max
resistor dimensions/construction affect long term stability

even some carbon film resistor go bad after ~40y

[Terry Bites]

Its worth mentioning that TC for ordinary resistors is sometimes specified as postive or negative or both. TC is a poor predictor of resistance change with T. There's no guarantee that the TC does change with T.
TC is not a parameter you want to use in a design- usually just the oposite.
Resistor ageing is unpredictable. But generally you get what you pay for.

[strawberry]

carbon film vs wirewound
heat them in oven to 100C deg.
what chemical compound or solid metal will change its properties