# Changing the polarity of the transformer on DUT cause voltage offset on DC measurement with insulating amplifier

DUT (device under test) is powered by 230 VAC and for safety we use additional external 230 VAC - 230 VAC transformer (DUT also has its own transformer for buck converting). We are measuring DC voltage (see block diagram) on costumer DUT with ICT test device true insulating amplifier because of very unstable measurement.

DUT is motor driver and we are measuring voltage drop on the shunt resistor.

Isolation amplifier schematic diagram.

The problem is that we get -300 mV offset voltage in DC measurement when the DUT is directly connected to the main voltage. When we add additional external 230 VAC - 230 VAC transformer the voltage offset drop almost to 0 V, but if we switch polarity on the primary side then voltage offset come back again. If we switch polarity on the secondary side nothing changes. If we switch input polarity on DUT connected directly to the main voltage without additional external transformer the voltage offset is always present regardless if we switching input "polarity" 230 VAC. Basically we just turning the plug in the main socket.

At this measurement we are using several transformer (2 on the DUT side, one on the ICT test system side and DC-DC galvanic converter on the isolating amplifier). Max measurement voltage true isolation amplifier is 1.9 VDC so 300mV voltage offset has a large influence on the measurement.

Primary and secondary side of the isolating AMP power supply are galvanically isolated with DC-DC isolating converter. We are measuring DC voltage on DUT only differential and GND of the isolating amplifier and DUT are also separated. I try to ground all GND and try to connect GND together and there was no improvement.

What is the cause of voltage offset and why voltage offset is reduced ONLY with switching polarity on the primary side of the additional external 230 VAC - 230 VAC transformer?

• "We are measuring DC voltage on DUT only differential and GND of the isolating amplifier and DUT are also separated", well it can't be, since you do need a reference point. "I try to ground all GND and try to connect GND together and there was no improvement.", makes no sense. There is not enough information to answer the question. Upload the schematics of the isolating amp + DC/DC converter. Commented Nov 8, 2021 at 11:52
• I uploaded the schematic diagram of isolated amplifier. Commented Nov 8, 2021 at 12:01

This looks wrong:

It's also not clear whether GND1 is tied to the GND on the DUT. Your divider is very high impedance and any stray capacitance coupling to grounded case, humans etc. will cause issues. R3 and R4 are not helping.

• GND1 and GND of the DUT are separated because, we want to simulate multimeter. Commented Nov 8, 2021 at 13:32
• Now I change circuit and removed both 1M resistors and connect GND and DUT GND together. Now the measurement are stable. The problem was to big values of R3 and R4, thanks. There is still offset on the isolation amplifier but is fixed and not change whit transformator polarity. Commented Nov 8, 2021 at 13:37
• Multimeters float, the circuit has a limited tolerance for common mode range, so establishing a common is important. Commented Nov 8, 2021 at 13:45

The measuring board is wrong made.

• C1 is connected to GND2 where it should be connected to GND1.
• The input divider is wrong, use the image below. You can see that GND1 is connected to the DUT's GND.

• Not sure if the switching DC/DC converter 0505S without a LC lowpass filter is a good candidate for an analog device PSU.