V vs. R of Polysilicon Silicide resistor

[maxporter]

Hi all,

Small question regarding the ‘R’ curve of polysilicon silicide resistor when sweeping the input voltage from negative to positive. As a result, I get a parabolic curve, negative or positive depending on N+ or P+. After a some searching this apparently what I can expect: there is a (V/L) squared term in the equation. And indeed, length changes the curve. But what puzzles me is that, depending on the length, at a certain negative and positive voltage the curve gets capped. See picture as an example. For small lengths, this in the operating voltage range (e.g. min length -> 300mV). Would this be a normal behavior? If so, what’s the cause and if not, why is it modelled?

Thanks & regards,
Max

 

[dick_freebird]

Try a pulsed-mode pull, you may be seeing self-heating and tempco (which may be a combination of quadratic and cubic).

Poly has a worse thermal impedance than bulk resistors.

 

[maxporter]

Hello Dick,

thanks for the suggestion. Tried it a couple of times, but so far I get the same outcome (or I have an incorrect setup). Perhaps important to mention, it’s a SOI process. Other poly-like resistors show similar behavior, however high-res N+ poly does not.

 

[maxporter]

Update: turns out it was indeed a self heating effect (found a global switch to turn it on/off).
Thanks & regards,
Max