# Op Amp circuit temperature drift compensation

#### spitsnsparkin

Joined Jun 19, 2011
29
Hi All,
I’m having trouble locating a reference paper/article about correcting the temperature drift of a sensor & op amp circuit over temperature extremes.
I’d like to assume that the circuit output is linear and characterized at 25C. But measurements of known inputs at the extremes, 0C and 50C, are not the same as 25C.
Given the measured points:
0% FS at 0C
100% FS at 0C
0% FS at +50C
100% FS at +50C
Is there an article/paper that describes a method of how to use this information to implement the correction?
Thanks in advance for you help.

#### Ian0

Joined Aug 7, 2020
7,441
There might be - but you haven't said what type of sensor it is.

#### spitsnsparkin

Joined Jun 19, 2011
29
Hi Ian0,
My question is generic. However, assume a load cell connected to an IN-AMP, which provides gain & offset to scale the output to the internal A/D of an MCU. The MCU’s VCC is the excitation for the load cell and VREF for the A/D. (I have implemented this in the past.) Likewise, a linear temperature sensor is also read by the same A/D.
In the application, the output from the load cell and IN-AMP circuit is found to be temperature dependent, which produces four different calibration points:
0%FS at 0C
100%FS at 50C
0%FS at 0C
100%FS at 50C
Given: 0%FS at 0C <> 0%FS at 50C and 100%FS at 0C <> 100%FS at 50C.
This results in two 2-point calibration equations at two different temperatures which produces two different values of gain and offset:
O=M1*X + B1 (at 0C)
O=M2*X + B2 (at 50C)
Is there a reference paper/article that describes a compensation method where the temperature (T) will scale the value of Gain (M1 & M2) and Offset (B1 & B2) to compensate for the circuits temperature drift?

#### spitsnsparkin

Joined Jun 19, 2011
29
I’m sorry, I goofed up the calibration points:
0%FS at 0C
100%FS at 0C
0%FS at 50C
100%FS at 50C

#### ericgibbs

Joined Jan 29, 2010
17,416
The MCU’s VCC is the excitation for the load cell and VREF for the A/D.
Hi spits,
The ADC in an MCU is ratio metric, in that if the load cell, the MCU and Vref are the same voltage source.
ie: Vcc, any change in Vcc will affect the Load cell and the Vref and the ADC will track any changes.

Try the temperature test again using the above set-up.
E

#### spitsnsparkin

Joined Jun 19, 2011
29
Hi ericgibbs,
I agree about the ratiometric performance of the A/D, VREF, etc.
What I meant, is that the gain (and offset) of the load cell/IN-AMP changes with temperature. If it was found to be +200uV/degC. Or a 100g mass at 0C measures to 102g at 50C (+2% error). Is there a paper/article that describes the math needed to implement temperature compensation for the gain (and offset) drift in the sensor/IN-AMP combination?

#### ericgibbs

Joined Jan 29, 2010
17,416
Last edited: