Most electronic designers will be familair with designing op-amp circuits with a single supply rail (e.g. Vs+ = 5V, Vs- = 0V) or matched split supplies (e.g. Vs = +18V, Vs- = -18V).

I have a design where I need to handle both +24V inputs and -10V inputs. Of course, there are some off-the-shelf parts that can handle +/-40V supplies, but my choice is rather limited.

On the other hand, op-amps with 36V supply voltages are plentiful, and given my voltage span is +34V, a carefully chosen rail-to-rail input/output (RRIO) op-amp should be up to the task, if I can make the supplies +25V, -11V.

This would make the design rather unusual (to say the least). What do I need to pay particular attention to when designing for this unusual arrangement of supply rails? Or, am I simply better off going for the high-supply op-amp?

EDIT: Do I, for example, need to take into account the following:

  • A higher than normal offset voltage/bias current?
  • Problems with matching between bias current and/or offset voltage between channels?
  • Changes to distortion?
  • Any other likely specification changes?
  • 2
    \$\begingroup\$ Related [sub]question: electronics.stackexchange.com/questions/85683/… \$\endgroup\$ Jan 16, 2015 at 3:17
  • 1
    \$\begingroup\$ Handling those inputs is never a problem but what outputs need to be produced for those inputs? In other words step back from how you think a solution might be implemented and state what you want, black-box style. \$\endgroup\$
    – Andy aka
    Jan 16, 2015 at 9:51

4 Answers 4


It's fine as long as you don't go outside the common mode input range.

Remember that the op-amp thinks only in terms of V+ and V- applied. It doesn't care what you call ground.


There's no reason in general to be afraid of unequal supplies. You might have to calculate some resistor values for the regulators or whatever, but that's no big deal.

You do have to be concerned a bit about headroom- if you're using an inverting amplifier, for example, the output can only swing to (say) -9V reliably, so you can't have a gain of -1 and a 20V input. So, you might have to scale the inputs, so that you do not run out of headroom, again no big deal, so long as you allow for it.

I would suggest avoiding RRIO op-amps where possible. RRO and single supply op-amps are generally a better choice. In particular, high-voltage RRIO op-amps are seldom necessary or appropriate.

  • \$\begingroup\$ Curious: why "I would suggest avoiding RRIO op-amps where possible"? \$\endgroup\$
    – Damien
    Jan 16, 2015 at 0:29
  • 6
    \$\begingroup\$ @Damien More expensive, fewer sources, poorer performance- all things considered, and they tend to do somewhat strange things as you move through the common-mode range.. they generally have two front ends that work at different CM voltages, all to gain an extra volt or two of input range. Inputs are where the performance matters the most for precision, noise and sometimes distortion. RRO is not nearly so limiting. \$\endgroup\$ Jan 16, 2015 at 0:31

@Caveman almost has it. Unless you're dealing with a genuine precision design, there is nothing to worry about. If you're dealing with 10's of microvolts, then there are a couple of extra factors to consider.

One thing to take into account is that it will change what the "common mode" (with respect to what you call ground) with a corresponding impact due to CMRR.

Consider also that the change in supplies will affect the output due to PSRR.

For actual simulation results, take a look at this answer.


Do you need very high input impedance? If not, the inverting amplifier configuration can be used for inputs that go outside the power supply rails, since the voltage appears across the input resistor and the -ve input is a "virtual ground".

Then you can use a 5 V Power Supply. Depends on the application? What are you trying to do?

  • 1
    \$\begingroup\$ Using inputs outside the voltage rails by more than 0.7V is beyond the absolute maximum ratings of most op-amps with notable exceptions. As I understand it, the protection diodes to Vss will be activated. \$\endgroup\$
    – Damien
    Jan 16, 2015 at 22:27
  • \$\begingroup\$ True, but not when there is a resistor in between and the Inverting input is used - the feedback resistor keeps the input at the same voltage as the +ve input. I have sensed 500V from 3 phase AC mains with a 15 V supply very satisfactorily. \$\endgroup\$ Jan 17, 2015 at 11:03

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.