I have two high voltage nodes which I need to measure the differential voltage between. What I am thinking at the moment is the following...

1 - Use voltage dividers to probe the voltage at 1/1000, thus having a differential voltage of 0.500Vdc and 0.501Vdc.

2 - Use an operational amplifier with a gain of 1000 to obtain the original differential signal.

3 - Use the 1Vdc differential voltage with my analog to digital converter (this part is extra, the first two are where my roadblock lies.)

The problem I have with this approach however is the slight error in the resistors and other small components will affect the accuracy of the final reading. Even using resistors with 0.5% tolerance and coupled resistors (is this the proper name for them?) may not be enough.

Is there a more traditional way of obtaining the differential voltage between high voltage lines?

  • \$\begingroup\$ There are precision differential amplifiers with accurate DC offset capability which might fit the bill. Look at Teledyne Lecroy for example. \$\endgroup\$ – Adam Lawrence Sep 13 '18 at 17:21
  • \$\begingroup\$ I've used Caddock high voltage precision resistors in this type of application before. You might also look at optically isolated opamps, several semiconductor companies make them. Google or search your favorite online supplier; I just saw three on mouser. \$\endgroup\$ – AlmostDone Sep 13 '18 at 17:22
  • 1
    \$\begingroup\$ Is this a mass production product? Is there a reason not to use trimmers on your HV divider, then tweak the calibration? Also note that you need a HV divider, since many resistors aren't stable up at hundreds of volts. \$\endgroup\$ – wbeaty Sep 13 '18 at 17:58
  • \$\begingroup\$ Use a battery operated V to F at the high voltage, an optocoupler to bring the frequency down to your microcontroller, and you can count the frequency directly as an alternative to converting it back to a voltage which you then AtoD. Or, have a DPDT relay, which charges a capacitor to the difference between the two high voltage nodes, then switches the capacitor down to your ADC. Use an opto-isolated high-side current monitor IC (they do exist). Use a cascode driver at the high voltage nodes to source a current proportional to voltage difference into a resistor to ground. \$\endgroup\$ – Neil_UK Sep 13 '18 at 19:53
  • \$\begingroup\$ What is the average signal duty cycle? How does it look like exactly? The capacitive coupling solution may make sense, but you didn't list it. \$\endgroup\$ – dim Sep 13 '18 at 20:40

You can build an amplifier that lives at the 500V level and then transmit a current down to ground level.

This avoids the extreme error and drift issues of trying to use matched resistors. To get a relatively coarse 1% error from the output means you have to have +/-20ppm (+/- 0.002%) accuracy and stability which is expensive and may be hard to achieve with a DC divider even with trimming (resistors exposed to high voltage DC tend to drift).

For example:


simulate this circuit – Schematic created using CircuitLab

Another option is to use an isolated amplifier designed for current sensing in VFD motor control products. You will have to add a DC-DC converter most likely to power it, and you'll have to ensure that both parts are rated for continuous operation at 500VDC input-output isolation.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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