# Facing problems measureing Input offset voltage of opamp

I setup the following circuit to measure the input offset voltage of an opamp; in this case an 741C opamp. The circuit is the same as in one of the EEVBlog video (https://www.youtube.com/watch?v=YeNpd-sXaHk) The datasheet (http://www.ti.com/lit/ds/symlink/lm741.pdf) states a typical value of 2mV and an maximum of 7mv. So if all goes correct, i should see an odd 200mV to 700mV in output. However I am reading -160mV there.

Now I have changed the values of R1 and R2 and here is my findings

  R1             R2            Voltage out             Calculated Vos
100k            1k              -166mV             -166mV/101 = -1.6mV
10k             1k              -16mV              -16mV/11 = -1.6mV
10k             100             -166mV             -1.6mV


It seams that -1.6mV is getting amplified at the output. Then is this the input offset voltage? Why is it negative and am I doing anything wrong?

PS:

• I have the circuit in breadboard.

• I tried with more than one 741 IC.

• I have placed 10uF caps at the opamp supply pins and ground. But the output gone even lower.

• I have tried the first circuit with LM358 opamp as well. The voltage out then is around -0.04mV or -0.44$\mu V$ of input offset voltage, which is absurdly low. (Its datasheet says the offset voltage should be around 6mV)
• Input offset voltages may be either positive or negative. Aug 3 '17 at 14:16
• I assume the -0.04 mV which is -0.44uV is a typo? -0.04 mV is 40 uV. Also, keep in mind that the datasheet states a typical and maximum value. It could just be that you are measuring a opamp with very good offset performance. Aug 3 '17 at 14:16
• (input offset V + offset current *Rin + input bias current * delta Rin ) * gain = output offset V. where delta Rin is difference of R for Vin+\- while bias current can be neglected for small R's only Aug 3 '17 at 14:54
• The offset voltage is a +/- range, within which you expect to find the actual offset. Perhaps your 358 just happens to be a very good one. Try warming it up and see what happens to the offset. Aug 3 '17 at 15:02
• @Neil_UK Ok thanks, I am just starting with opamp and did not know that the offset voltage can also be negative. But the 358 is strange. I will try with a different part. Aug 3 '17 at 17:00

Your circuit is not ideal. The relatively high 1K || 100K resistance the inverting input "sees" will lead to an additional term dependent on the bias current of that input. It would be better to use a much lower resistance (minimizing the effect of the input offset current) and to add a similar resistance to the non-inverting input (minimizing the effect of the two bias currents).

Here is a typical circuit. You should have bypass capacitors (10n to 100n ceramic is fine) near the op-amp power pins. Not really vital for very slow op-amps, but some day you may wish to test a fast op-amp. In such a case you may wish to also parallel R1 with a 10n NP0 capacitor to tame any oscillation. You may not be able to measure worst-case offset with low voltage supplies so you may have to lower the gain by lowering R1/R3 to 4.99K or increasing R2/R4 to 20 ohms. simulate this circuit – Schematic created using CircuitLab

The datasheet shows the typical/maximum (under specific conditions) magnitude of Vos. Vos can be of either polarity- it's just as likely to be positive as negative for most op-amps (the target value is zero).

It's (barely) possible the particular sample of LM358 you have happens to have very low offset voltage. Or maybe (perhaps more likely) you have it hooked up wrong- the pinout is different, of course, being a dual.

• I have also used a 10k and a 100E resistors as well with same results. See the table in my question. If I do your circuit with 10k and 100 (in my case} in parallel, then the output dips further to about -176mV. Do u think the values are completely wrong or such values are possible? Aug 3 '17 at 16:23
• @ArjobMukherjee Your numbers sound very plausible for the 741. Aug 3 '17 at 16:26
• I happened to think that the input offset can only be positive. If however it can be negative then my results are at par with the datasheet. As for the 358,I will try with another part. Aug 3 '17 at 16:36