SV: Re: [sdiy] Compensating PTC and NTC resistors

[René Schmitz]

Hi Karl and all,
>> I think the idea is that the change of the carbon resistor is so small
>> compared to the 3900TCR that it could be considered ~linear~
>>
>> So what you do is to make up a resistor where one portion of the total
>> resistance is the TCR, and the other fixed... so the overall tempco is
>> reduced. Its a ~mixture~ problem... maths left to the interested :^)
> 
> Yes, but still, where do one get the -600ppm resistor?
> The user cant compensate a +3900ppm with a +/-600ppm resistor! Nor are the ppm
> part fixed with temp, nor linear even manufacturers want to claim that. Most
> papers on the subject suggest xxxppm but they let the sign out! Elby papers for
> example does not suggest any sign nor do Rene's! So do the user use a carbon
> mass (hehe), carbon film, metal film, metal oxide, plastic, ceramic compound,
> or any of the myrriad of new resistor techs thats available today.

Thats not how it works, you can't just add and subtract tempcos like 
that. The idea is more that when you change the resistance by adding a 
series resistor (of assumed zero TC, don't bother with minute second 
order effects, and use MFR), the tempco of the whole deal becomes lower. 
The maths is basically a rule of three.

Cheers,
  René