I came to think that my scope is lying to me.

Settings:

AC input
vertical 10 mV / div
horizontal: 10 ms/div

There is a board powered by external PSUs.

If the board IS NOT connected to the PSUs, I do not see anything in the scope.

With the board connected to the PSUs, but NOT turned on, I get, consistently, in both chanels, a signal as in the picture, no mater what points I touch with the probes even IF BOTH TO GROUND.

Impossible to get a better picture. Sorry.

The signal, evidently periodical, appears as having a frequency of 112Hz.

Mains is 50 Hz thus nothing to do with it.

The ground clips on the probes are connected to the ground.

 

Could be a ground loop, can you post a diagram of your setup, along with the conection of the psu, scope and mains earth?

 

Looks like it may be PSU switching noise. Where are you looking?

 

Both are simple linears with diode bridge and caps.

 

Here is the setup.

My last findings noted there.

 

Had a similar problem with strange signals riding on what I wanted to look at. Turned out to be a nearby local AM broadcast station and something on the board under test was picking it up.

 

A linear supply with full wave rectifiers connected to a 50 Hz line will draw current spikes 100 time a second. There is nothing linear about the diode charging the cap: all the current comes in short bursts of high current as the input wave nears the top.

With no unit attached there are no bursts as the caps stay charged.

If this theory is correct you would see similar bursts if you replace the PCB with a simple resistor dummy load.

 

So the power supplies only have two-prong plugs on them?

What about the scope? Is it a two prong-plug and then you have a separate earth ground that you are taking to the power socket?

Your diagram shows an explicit ground connection as well as a ground connection through the scope's ground connector. Try removing the explicit ground and just have the single-point ground via the mains wiring.

You might try disconnect each of the four supply leads form your board one at a time (if the circuitry can tolerate that).

 

So the power supplies only have two-prong plugs on them?

Yes, 2 prongs.

What about the scope? Is it a two prong-plug and then you have a separate earth ground that you are taking to the power socket?

It is the (local) standard plug with 3 prongs. The biggest/longest to ground. Nothing is separate.

Your diagram shows an explicit ground connection as well as a ground connection through the scope's ground connector. Try removing the explicit ground and just have the single-point ground via the mains wiring.

That "explicit ground" of the common is in fact a symbol I used to show that IT IS common and nothing else. As I say, the sole thing to ground is the prong of the scope. Sorry for that.

You might try disconnect each of the four supply leads form your board one at a time (if the circuitry can tolerate that).

Did that already with no problem (because I took out previously, all ICs) and still, what allows that signal to show up is the connection of the ground to the board.

After reinstalling all ICs I turned on the board. The probes to the common, STILL show that signal.

When probing the different voltages I can see the signal happily riding everything. Nor more nor less what moved me to check all this in the beginning...

Turned off everything at home except the fridge.

As suggested in another forum, I also disconnected the earth of the scope. A real mess on the screen with strange artifacts wandering all over.

By now I run out off ideas. As an alternative, growing water lilies instead is not an option. No space around.

That frequency value, steady as it is, doesn't suggest anything?

 

A linear supply with full wave rectifiers connected to a 50 Hz line will draw current spikes 100 time a second. There is nothing linear about the diode charging the cap: all the current comes in short bursts of high current as the input wave nears the top.

With no unit attached there are no bursts as the caps stay charged.

If this theory is correct you would see similar bursts if you replace the PCB with a simple resistor dummy load.

I am willing to do any test suggested but am not sure I understand you.

The signal shows up just with the common connected and no 220V AC power applied.

 

Been wondering if would be of any use to check how good is the earth connection at the wall sockets (vocabulary could be wrong - I know).

Can you tell me how to do it.

I have a reverent fear of 220 V so I am always very cautious.

 

What does your scope show when you connect the ground clip to the tip of the probe?

 

Both connected in the air, show nothing.

Both connected to ground, the signal shown in my OP.

 

Both connected in the air, show nothing. Both connected to ground, the signal shown in my OP.

I am not sure we are on the same wave. Do you mean scope ground or circuit ground? Try simply shorting one of the probes to its ground.

Also do you use a 1x probe or a 10x probe? What brand/model is your scope and what is the lowest volt/div setting it has? Also check continuity between the probe ground and the chassis and mains ground at the connector.

 

Had a break to clear my mind.

I eliminated the DUT and one PSU.

The test was made on the remaining one, which is now powered all the time.

The ground lead croc AND the tip were both hooked to the 0V point - common of the circuit (or, if you like, the CT of the transformer).

a) No changes; the signal shown in the OP still shows up.

b) Connecting that point to earth (brought with a wire from the wall socket), in about one second the signal diminishes to become barely distinguishable (if at all).

All in all what I have now is about 1 mV of noise(?) or maybe less. Yes, I am at the bottom of what my scope can do.

c) Touching instead the BNC sockets or any part of the case with that wire, nothing happens and the signal stays unchanged.

 

If I'm following what you just said, it sounds like the scope chassis is not connected to the grounding wire. Try using a wire to connect the scope chassis to the wire going to earth and see what that does.

 

If I'm following what you just said, it sounds like the scope chassis is not connected to the grounding wire. Try using a wire to connect the scope chassis to the wire going to earth and see what that does.

I would first measure to see if there is any voltage differential between chassis and ground.

 

I would first measure to see if there is any voltage differential between chassis and ground.

Duh!

Good point.

Although, check with what? The scope? A voltmeter may or may not see it, especially if it is average responding. But worth a shot.

 

Duh!

Good point.

Although, check with what? The scope? A voltmeter may or may not see it, especially if it is average responding. But worth a shot.

A voltmeter at it's highest setting would probably be safe.

Hopefully Argentina is not as bad as other parts of the world but it many countries, you have no guarantee that ground is ground.

 

Try using a wire to connect the scope chassis to the wire going to earth and see what that does.

No change at all.

I measured wrt ground prong in the scope's socket, the following resistances:

BNC sockets: 0.5 Ohms.
Chassis: 0.4 / 0.5 Ohms.
Leader probes' ground leads: 1.2 ohms
Greenpar probe's ground lead: 0.8 Ohms.

BTW, I simplified things even more testing everything with probe's tip and grounding clip shorted, connected directly to a wire coming for mains' earth.

Measured AC with my multimeter, between chassis and earth in all ranges up to 750V. Zero, nada.

Spent the whole Sunday and still here...

Thanks for helping.