# What is the best method to compensating offset at the output of a peak detector circuit?

Brief of Application:

The signal conditioning circuit developed is used to detect or capture the peak values of dynamic systems such as the impulse tools. In this application, we use an impulse tool to tighten or loosen bolts and nuts, hence when the torque is applied to tighten or loosen the bolts at a set torque, my system should detect the peak torque applied to tightening the screws.

The Figure above shows the schematic of the circuit developed in Multisim i.e. An Instrumentation Amplifier followed by a Peak Detector.

Description of the Problem

The output across the capacitor goes as input to the TI-ADC1790. The problem is, the output at the peak detector has a small negative offset of -4.7 mV because of which some ADC counts are lost for the initial readings, since the ADC starts counting from the positive voltage scale. Hence, while testing I noticed that, when I apply the torque, the values are acquired only from the start of the positive scale. Therefore I need to compensate the negative offset to the ADC input.

Two ways I feel this can be done is by using an adder circuit at the output to switch the peak detector output to a positive scale or just use a voltage follower circuit while providing certain gain to amplifying the negative offset to a positive scale.

I am posting this question here, expecting to get a more appropriate solutions to this. Therefore, any help, suggestions or advice regarding this would be great. Thank you.

• I think finding the source of the offset is the first task. What amplifier model did you use for the instrumentation amplifier parts? In general, it is better to use a monolithic instrumentation amplifier as these have been designed with key parameters in mind. – Peter Smith Sep 21 '15 at 7:43
• For a classic 3 amplifier instrumentation amplifier, R13 and R11 should be the same. In the configuration shown, there is more gain in the inverting path than the non-inverting, which would indeed give an offset. – Peter Smith Sep 21 '15 at 9:16
• The offset is caused due to offset from the load cell. How ever I compensated the load cell offset while making the inputs to the differential op-amp equal, but a small amount of offset still persists at the output of the peak detector. And yes, will make a note and check the R11 and R13 by making same. – PsychedGuy Sep 21 '15 at 10:42

According to my calculations, there is a gain in this amplifier of almost 3800. Your 4.7mV offset implies < 2uV of input offset voltage, which is really very good for any amplifier.

Instrumentation amplifiers can have offset voltages lower than this under ideal conditions but across temperature they will not be quite as good.

My question would be, why do you want to null this tiny offset when it refers to a tiny input offset?

See my comment that R11 and R13 should be the same; this is quite likely the cause of offset in the simulation (I may run it later when I have some time)

If you really need to offset this output, then I would use a precision reference and a follower into the final output amplifier.

Note that the tolerance build-up of your components in a real application will very probably dwarf this offset you have.

To overcome this, you may wish to use a precision instrumentation amplifier.

HTH

• I figured out the issue. The Calibration factor in the program dividing the ADC counts was way too large, due to this, the resolution was too low to display the actual value. Now I will extract the raw ADC counts and then check. And before that. I will make the R11 and R13 same and check once. And about the INA114, I will check that too, once this one works. Thanks. – PsychedGuy Sep 21 '15 at 10:40