In the schematic below I'm measuring the scaled-down AC voltage of a floating power source @120VAC (model is FC 205 from Adaptive Power). Both the op-amp (any standard op-amp) and ADC (ADS131E04) are running off Vss=-2.5V and Vdd = 2.5V from an isolated supply.

To address all safety concerns: instead of using a transformer on the AC voltage from the FC 205 to give the ADC and op-amp an isolated signal, I’m feeding the scaled down FC205 voltage in directly while powering the ADC and op amp off an isolated +/-2.5V supply. For communication I’m also using an isolated SPI module. So the isolation factor needed between the AC line and the end user is still there, just not shown in this schematic.

The ADC is missing all connections aside from the one analog input because they are not relevant to this question and have been removed.

My question here is: when measuring the voltage between IN1P and IN1N, does it matter that the Neutral Line from AC Source is not tied to the same ground as the ADC? I need the Neutral line of the AC Source to be separate from the ADC or Op-Amp's Ground Rail for other measurements I need to perform (the rest of that circuit is not relevant and not shown), but want to measure voltage directly without isolation (ADC and OP-Amp are powered off of isolated supplies, so there is no safety hazard).

I'm having a hard time figuring out what the voltage difference between the Neutral Line of the FC 205 and the -2.5V Vss supply would be, and if this could blow the ADC. (so looking for V_Neutral - Vss). Thanks for your help! enter image description here

  • \$\begingroup\$ This is highly ill-advised! Never connect a low-voltage circuit directly to AC-mains!. You should use a transformer instead of a voltage divider to both reduce the voltage and isolate the low-voltage circuit from the AC-mains. That said you really shouldn't be playing around with AC-mains when you don't have the necessary prerequisites to know how to do so safely, ie. when you need to ask a basic question like this one on EE.SE... Be safe don't be an idiot, for your own sake, you could easily get yourself killed.. \$\endgroup\$
    – user173292
    Commented Jan 31, 2021 at 1:06
  • \$\begingroup\$ Do not worry, this entire circuit is powered by a +/-2.5V isolated supply so the worst that could happen is the op amp and ADC blow. Human interaction is on the other side of the supply \$\endgroup\$
    – cdubs
    Commented Jan 31, 2021 at 1:09
  • \$\begingroup\$ I have an isolated 2.5V supply for the op amp/ADC instead of a transformer on the incoming voltage. Same factor of safety \$\endgroup\$
    – cdubs
    Commented Jan 31, 2021 at 1:14

2 Answers 2


If you need to keep Neutral separate from the local ground, then I think you'd be better of using a differential amplifier configuration to drive IN1P and tying IN1N to ground.
Something like this:


simulate this circuit – Schematic created using CircuitLab


Yes it does matter - if the common-mode voltage swing goes outside the isolated supply rails then bad things will happen - at best you won't get a meaningful conversion, at worst it could damage the ADC and op amp. Somehow or another you should either bias or just connect the neutral to the 0V if you want meaningful results. As has been pointed out, any circuit that uses AC mains is potentially hazardous, although sometimes there's no getting away from it (e.g. light dimmers). You could couple the AC live and neutral via some Y-rated capacitors and earth the 0V of your +/-2.5V supply, or use a transformer (this seems like the best idea IMO). Depending where you live, you may find that your neutral connection is in fact connected to ground at your fusebox. If this is the case then you should see almost no voltage between neutral and earth unless there's a fault condition, so you could find that you can earth the 2.5V supply and you'll see very little voltage on neutral. However, do put some protection (series resistor and zener/MOV to ground) between neutral and your ADC input.

  • \$\begingroup\$ Thank you for this answer! I know everyone’s go-to is a voltage transformer, but I’ve seen plenty of AC measurement devices (Keysight 34461A for example) that I think don’t use a transformer for measuring AC voltage. Figure the workaround was isolated power/communication instead \$\endgroup\$
    – cdubs
    Commented Jan 31, 2021 at 1:41
  • 1
    \$\begingroup\$ Sure, using a mains-coupled supply for the ADC and then an isolated digital signal is entirely valid, and arguably can give better accuracy, if that’s what you need. \$\endgroup\$
    – Frog
    Commented Jan 31, 2021 at 9:30
  • \$\begingroup\$ Awesome, yes that’s what I need. Thanks for all your help \$\endgroup\$
    – cdubs
    Commented Jan 31, 2021 at 15:53
  • \$\begingroup\$ Hmm @Frog if I were to leave the Neutral lead disconnected from the ADC ground, is there a way for determining what magnitude of common mode voltage I’d see in the ADC’s input signal? Or is that undetermined really \$\endgroup\$
    – cdubs
    Commented Jan 31, 2021 at 15:59
  • \$\begingroup\$ If the supply is floating then it’s pretty much undetermined, and likely to be noisy too. \$\endgroup\$
    – Frog
    Commented Feb 1, 2021 at 2:19

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.