Can an isolation amplifier be used to block common mode noises and capacitive coupling?

Question

A single-ended earth grounded data acquisition board have some channels fixed inputs as transducers. But when I connected extra channels as any of force transducer amplifiers, this causes standard deviation increase and some 50Hz noise in fixed channels and in the force transducer themselves as well. I already checked couldn't find any ground loop since the sources' DC grounds are not earth grounded.

But when I disconnect the force transducers the noise disappears and standard deviation decreases dramatically which is good. I notice that the force transducers have 78V common mode voltage which is a ghost voltage due to parasitic capacitance.

Im trapped because if I earth the DC supply of the force transducers I get rid of the common mode voltage but then this creates ground loops. But even I hook up the force transducer channel the noise comes back even though its earth is not grounded.

I was thinking to use an isolation amplifier at the end of each force transducer channel but that would require a lot of amplifiers.´

The force amplifier system with many cables shares the same ground and possibly corrupting the ground when hooked up to the daq because it is single ended(?)

Can an isolation amplifier be used to block common mode noises and capacitive coupling from the DC supply?

Answer 1

Yes, but they are not cheap. Try looking up AD202 and the AD210 analog isolators from Analog Devices. The source could be Digi-key or Mouser. Both are available as a 1" wide x 2" long module, or a vertical 2" long by 1/2" wide package.

They have an isolated 15 volt power pin and ground, which powers both the input and output sides.

You have 1,500 VAC isolation from the supply and inputs/outputs.

For creapage isolation you can slot-cut the PC board just inside the input pins, which also supply +/- 7 volts to power an op-amp if needed.

Answer 2

This might work. Try connecting capacitors from each line to earth/ground. Hopefully, if the irritating common mode voltage has arisen due to high impedance coupling, the capacitors should reduce this interference. Start with 10 nF per line and see what happens. If there is a 10% or greater reduction in CM voltage then it’s likely that it is higher-than-regular-power-AC frequency and, adding maybe 100 nF per line might solve the problem and save the hassle of an isolation amplifier.

If 10 nF doesn’t stop much CM voltage then it’s likely to be a power AC frequency and would require maybe 1 uF or more to have a significant impact but, the problem with this type of fix, is that you might make the transducer output too filtered and lose information. However if your signal bandwidth is low, you may be ok with capacitors up to 10 uF.