Waveform shapes different with AC and DC coupling?

[LiberaVeritas]

I have two wires going to my apartment intercom/buzzer. They seem to come from a transformer stepping down mains AC, with one having a DC offset. This is earth ground referenced:



When I measure their difference by probing one while the ground clip is on the other, I get this:



which seems to indicate a diode between the two lines.

But AC coupling the same signals shows this:



whereas I expected the same shape.

Is this just due to the AC coupling capacitor? The cutoff is at 10Hz on this scope.

[Marco]

The offset is supposed to be div limited, but the manual says it can go to -6 div ... so it shouldn't be an issue. Still would be interesting to see what happens if you decrease sensitivity.

PS. just to make sure, there is only one ground clip connected right?

[CosteC]

Can you show connections or draw them?
Photo maybe?

[LiberaVeritas]

The offset is supposed to be div limited, but the manual says it can go to -6 div ... so it shouldn't be an issue. Still would be interesting to see what happens if you decrease sensitivity.

PS. just to make sure, there is only one ground clip connected right?

I tried with lower sensitivity, both on the V/div as well as on the probe sensitivity. Same result.
In the first measurement, I had the ground clips of both probes on a known earth ground. In the other, I only used one probe with its one ground clip.

Can you show connections or draw them?
Photo maybe?

[Marco]

If I read the manual correctly, this scope is not isolated from electrical socket earth or USB ground when connected to either.

Is it being used as pure handheld or not?

[LiberaVeritas]

Ah, yep it's being used handheld. No USB or anything else plugged in.

[Marco]

Seems to be something screwy in the input circuitry then.

Do you have an adjustable power supply? You could try to use the function generator to generate a sine and then increase the negative offset with the power supply to see if/when it starts clipping.

[CosteC]

Original oscilloscope screen shots disappeared. EDIT 2024.07.08 18:40 CET : Pictures returned. Or maybe it was due to filtering at work.
Was PE from wall/mains socked/similar used for configuration #1? If scope was floating (no PE connected to scope) so, and if transformer measured was class 2 of insulation (totally insulated) then your measurement may be result of random leakages in system, and is generally non-sense as expected.

If transformer is double isolated (no connection to PE either) then measurement configuration #2 is correct one and shall provide correct result. DMM can verify it.

BTW what DMM says in both cases?

[LiberaVeritas]

Seems to be something screwy in the input circuitry then.

Do you have an adjustable power supply? You could try to use the function generator to generate a sine and then increase the negative offset with the power supply to see if/when it starts clipping.

I generated a 60hz sine and added a DC offset which I gradually increased to 20V. I didn't see any clipping, hmm...
Oh, and the clipping on the scope happens with a positive offset as well. With the wires reversed, it's just a reflection about the time axis.

Was PE from wall/mains socked/similar used for configuration #1?
...
BTW what DMM says in both cases?

For #1, ground/earth was from a mains wall socket, after I compared it with a few other sockets to make sure there was nothing funny with it.

With a DMM for #1, the offset signal (blue) measures 10.4V RMS and +15VDC. The yellow measures 14.28V RMS, -130mVDC.

For #2, I get 3.3V RMS and -14.4VDC.

Seems fine? This is with both a dedicated DMM the one built-in to the scope.