I created the following circuit in MultiSim:

741 op-amo with pos and neg rails attached to +/-5V, inputs grounded and output showing 4.11V

It shows a typical 741 op-amp with voltage rails connected to +/-5V and both inputs grounded. With (for all intents and purposes) equal voltages at the inputs, I would expect the output voltage either to be 0V, or very close to it (perhaps ~800mV or less).

Instead the op-amp saturates at the positive rail (or at least maximum Vout). Why not the negative rail? Why not 0V?

Would connecting pins 1 or 5 have an effect?


  • 1
    \$\begingroup\$ Because opamps have "infinite" gain, which will take any microscopic difference to either of the rails. \$\endgroup\$
    – Eugene Sh.
    Mar 12, 2019 at 20:19
  • 2
    \$\begingroup\$ What you're seeing is called the input offset voltage. Simply put, the voltage the 741 actually amplifies is the difference between its inputs, plus its input offset voltage (and since we're talking about the 741 here, the input offset voltage can be quite high. The 741 is not a good op amp.) \$\endgroup\$
    – Hearth
    Mar 12, 2019 at 20:20
  • \$\begingroup\$ @Hearth Aha, thank you. I didn't consider Vos and therefore the op-amp is actually not comparing two identical voltages. \$\endgroup\$ Mar 12, 2019 at 20:23
  • \$\begingroup\$ Given the earlier comments, I think you need to have a look at the model you are using for the 741. Does it assume some value of offset voltage? What is the assumed value of the open loop gain? Without that information we can only guess at why you are getting the results you see. \$\endgroup\$ Mar 12, 2019 at 20:23
  • \$\begingroup\$ @ElliotAlderson The model is the default MultiSim 741 model whose contents is not decipherable by me. I think a sufficient explanation has been offered by Hearth, though. ... I have to ask if that's a reference to Mr. Robot or your real name? \$\endgroup\$ Mar 12, 2019 at 20:32

1 Answer 1


You're also not taking into account the input offset voltage. This is a factor present in all non-ideal op-amps, and can be modelled as such:

enter image description here

The 741 series of op-amp may require you to "null" the input offset voltage, so even if you apply an identical voltage to both terminals, without any feedback or nulling on the terminals, I'd expect it to rail in one direction or the other.

Remember, op-amps have very very high open-loop gain (106dB in this case or about 200,000). The worst case input offset voltage for this op-amp is 6mV, or even 7.5mV across the whole temperature range. This offset voltage will be amplified by the open-loop gain of the amplifier, yielding an output voltage of 1200V! Obviously the op-amp cannot output this, so it simply rails in one direction, or the other.

As a general rule, unless you're using them as comparators, op-amps aren't usually used in open-loop configurations. The high open-loop gain serves the purpose of linearizing the closed-loop gain through negative feedback (but this is probably above the scope of this answer).

image credit: https://www.allaboutcircuits.com/textbook/semiconductors/chpt-8/op-amp-practical-considerations/

  • 1
    \$\begingroup\$ Why do you say its input common mode voltage is out of range? The input CM voltage is zero. \$\endgroup\$
    – Hearth
    Mar 12, 2019 at 20:34
  • \$\begingroup\$ Whoops, I had my brain wired for single-supply circuits. Edited out :P \$\endgroup\$
    – Platytude
    Mar 12, 2019 at 20:37
  • \$\begingroup\$ Thanks - For some reason I was blind to any voltage manipulations actually occurring within the op-amp and only focused on the fact the inputs were "exactly" equal. \$\endgroup\$ Mar 12, 2019 at 20:46
  • \$\begingroup\$ You also say that the 741 may require offset nulling and you imply that this means it has an offset voltage (with the implication being that this is unusual), but it's worth pointing out that all op amps will have some offset voltage; the 741 just happens to offer offset null terminals. \$\endgroup\$
    – Hearth
    Mar 12, 2019 at 21:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.