Weird USB oscilloscope ground issue?

[m k]

Different grounds can have different levels on purpose.

180mA can be an overcurrent protected short.

[magic]

To the level of what he'd seen it's not normal.

Fair enough, I measured my own computers and found only up to 10mV between USB port grounds and PSU chassis under full load, so 250mV may be a little excessive. I also have a few mV difference between the chassis and earth on a nearby mains socket, indicating that I already have ground loops sending DC through mains earth, maybe through the monitor.

[Alex_Baker]

I have tested earth paths around my workbench, which is where I am using the scope. I have not tested the earth path of the actual outlets in the wall since that would need me to turn the power off, but other than that I have not found any problems.

Correct me if I am wrong, but I don't think this stray current issue actually poses any risk to me or my electronics, so I don't think it is worth chasing any more. If I had to guess, a mixture of long usb cables and a sub-optimal set of power strips feeding my PC and workbench make this seem much worse than it normally would be.

[wraper]

I have tested earth paths around my workbench, which is where I am using the scope. I have not tested the earth path of the actual outlets in the wall since that would need me to turn the power off, but other than that I have not found any problems.

Correct me if I am wrong, but I don't think this stray current issue actually poses any risk to me or my electronics, so I don't think it is worth chasing any more. If I had to guess, a mixture of long usb cables and a sub-optimal set of power strips feeding my PC and workbench make this seem much worse than it normally would be.

You don't need to shut down the power to check the earth connection. Yes it poses safety risk as some devices my be working without proper protective earth connection.

[Alex_Baker]

Yes it poses safety risk as some devices my be working without proper protective earth connection.

I have tested all the earthed devices in this room, I cant find anything wrong with any of them.

[wraper]

Yes it poses safety risk as some devices my be working without proper protective earth connection.

I have tested all the earthed devices in this room, I cant find anything wrong with any of them.

So what? Device can be 100% fine but dangerous at the same if there is an issue with mains wiring.

[Alex_Baker]

How about this, I know that there exists and earth connection between my workbench and my PC, which are on different wall outlets. I know one exists because I can create a current loop by connecting my scope ground to earth at my workbench, which must mean that current is flowing back to the PC through the outlet earth wiring. I don't exactly know the condition of the earth path, but it cant have too high a resistance or 100mV would not result in much current flow.

[m k]

Famous last words.

Here all deadly electric accidents have had two faults.
You seem to have one already.

[magic]

If there is abnormally high voltage between different "grounds" in the PC then the problem is likely in the PC.

If in doubt, disconnect everything from it and measure if the motherboard to case offset is still there. Theoretically, there is possibility that something else is broken and sends tons of DC through a ground loop passing through its PC connection cable, the PC and mains earth, but IMO the PC is the chief suspect.

[Alex_Baker]

Famous last words.

Here all deadly electric accidents have had two faults.
You seem to have one already.

I get the message, thanks

If there is abnormally high voltage between different "grounds" in the PC then the problem is likely in the PC.

If in doubt, disconnect everything from it and measure if the motherboard to case offset is still there. Theoretically, there is possibility that something else is broken and sends tons of DC through a ground loop passing through its PC connection cable, the PC and mains earth, but IMO the PC is the chief suspect.

I could not find anything that was more than 6mV.

[m k]

With earthed extension cord you can check what nearest GFCI has to say.

Insulated area doesn't need protective earth, assume that it is not there.

When connecting devices to earthed socket assume that metal parts are live.

[Alex_Baker]

I plugged the power strip under my desk into a bathroom outlet, which does not have a GFCI itself but is connected to one elsewhere. I then repeated my earth-to-earth and earth-to-pc GND current tests from the outlet under my workbench, and I could never get the GFCI to trip. I got about 4mA going from PC earth to workbench earth, and 10ma going from PC GND to workbench earth.

[magic]

I thought typical GFCI trip on AC not DC?

And they trip on mismatch between phase and neutral, not on protective earth current.

[m k]

I plugged the power strip under my desk into a bathroom outlet, which does not have a GFCI itself but is connected to one elsewhere. I then repeated my earth-to-earth and earth-to-pc GND current tests from the outlet under my workbench, and I could never get the GFCI to trip. I got about 4mA going from PC earth to workbench earth, and 10ma going from PC GND to workbench earth.

That is good.
Now you(and me) can be much more certain that no hazardous situation is present.
You can also assume that original ground current is between secondaries and possible leak is less than dangerous.

Comparing this latest situation and earlier connections you should be able to narrow the problem down, at least some.

[Alex_Baker]

I thought typical GFCI trip on AC not DC?

And they trip on mismatch between phase and neutral, not on protective earth current.

That is what I was wondering, this house was built in 1990, so not sure what type of GFCI is installed.

[shakalnokturn]

If neither the PC's case-PSU- ground continuity is suspect, it isn't the PC's PSU leaking massively and it isn't a DC voltage drop linked to the Picoscope power (as I understand that has been ruled-0ut ?).

Could it be a power device, likely inverter-based (solar power, aircon...) sharing the same line but further down that's getting away with putting a fair deal of DC leakage current through the PE wire on that line due to lack of (adequate) GFCI?

I'm not familiar with the American NEC, here in France the MCB's allow up to 500mA of ground fault before tripping.

[Alex_Baker]

If neither the PC's case-PSU- ground continuity is suspect, it isn't the PC's PSU leaking massively and it isn't a DC voltage drop linked to the Picoscope power (as I understand that has been ruled-0ut ?).

Could it be a power device, likely inverter-based (solar power, aircon...) sharing the same line but further down that's getting away with putting a fair deal of DC leakage current through the PE wire on that line due to lack of (adequate) GFCI?

I'm not familiar with the American NEC, here in France the MCB's allow up to 500mA of ground fault before tripping.

I will try to follow up on this at some point, I want to try and unplugging things in the vicinity until it is only the pc plugged in, perhaps in that entire circuit. There are no large power devices around me, only lots of little ones.

[m k]

You can also measure PE resistances.
Bathroom and workbench PE sockets are pretty close, desk socket seems to be something else.

Use the extension cord to bring wall sockets side by side.

Probe tip can have better connection than its side.
You may get lower shorted resistance by measuring few millis of conducting material.

If probe is too short you must tilt it to get a PE connection.
Then the connection can be less than optimal and if it is missing completely you may need to go back to previous situation to get a confirmation.
Keep also in mind that wall socket can be less than optimal, connection can be loose or broken.
So uncertain connection is not necessary your meter.

In case of uncertainty I would be "watta..." and open the socket.
Don't do that if you are not sure.

[Alex_Baker]

You can also measure PE resistances.
Bathroom and workbench PE sockets are pretty close, desk socket seems to be something else.

Use the extension cord to bring wall sockets side by side.

Probe tip can have better connection than its side.
You may get lower shorted resistance by measuring few millis of conducting material.

If probe is too short you must tilt it to get a PE connection.
Then the connection can be less than optimal and if it is missing completely you may need to go back to previous situation to get a confirmation.
Keep also in mind that wall socket can be less than optimal, connection can be loose or broken.
So uncertain connection is not necessary your meter.

In case of uncertainty I would be "watta..." and open the socket.
Don't do that if you are not sure.

The only problem with this is that there are small voltages between earth sockets, which makes it impossible for my multimeter to do a resistance reading. I know there shouldn't be voltages on the PE wire, but nonetheless there is.

[m k]

What is small?
Digital multimeter can create AC volts from thin air.
You can demonstrate that by moving hanging probes apart.

For checks you can also use line voltages.
But then you must be sure that your meter and leads are correctly protected.
If PE is connected one side measures a line voltage contra it.
I guess you can quite easily lose near 5% of line voltage for bad connections.

If that stray voltage is real you can also use it and create a distance map.
If there is a DC component you can also create a direction map.
But don't take that result granted, situation can change depending how surrounding machines are on and off.
It's also time consuming since you must move extension cord around and trust that its contact is equal enough.

One overall downside with wall sockets is that some are extremely difficult to reach.
And that doesn't even include exclusive ones, like fridge and washer.