AC voltage reading between ground and a two pin power supply

[manick]

Dear All,

I am not very experienced with the use of an oscilloscope and especially concerned about isolation. I watched Dave video and read some post here on isolation using a UPS.

That said I want to probe some points on a HP inkjet printer I have. The power supply supplies 22VDC via 2 pins of course. The power supply has two mains pins alone (american style outlet though that should not matter).

Using a multimeter before I connect the scope I am trying to determine if it is safe to ground the probe to the "ground" on the printer.

1) When I check the resistance from either of the output pins (DC side of the printer power supply) to either input pins (AC side of the printer power supply) its infinite (as expected)
2) When I put the meter into AC mode and check the potential between ground and either DC output pins I am getting about 57VAC.

Further confusing the matter I opened a similar printer power supply and there appears to be an isolation transformer just after the fuse. The output and input pins are not electrically connected but testing this open supply gives the same AC voltage.

Can anyone explain this to me? Seems like a short waiting to happen if I connect the scope.

[Jeroen3]

The power supply has two mains pins alone (american style outlet though that should not matter).

It's an isolated power supply. Wich means it will float at 0.5 * Vmains due to the filter capacitors inside if you do not connect mains earth.

All is normal. See the attachment for an equivalent circuit of a power supply. Although there days they will be switchmode. But the concept of isolation and filtering did not change.

I forgot to mention that the current is very low due to small capacitance. Any load will equalize the voltage with <1mA ac current between the two devices.
See brumby's post.

[manick]

Jeroen3 for some reason I am not allowed to see the picture you posted. I am getting an error when I try to click on it. EDIT - Now I can see the picture you attached. Weird but that was not possible till I posted.

It is a switch mode power supply. There is an initial isolation transformer that's 1 to 1. Powered I get 120V on either side across the terminals. Then there is the other transformer and a transistor I assume (did not look that up) on a heat sink.

My concern is the voltage I am seeing on the output reference to mains ground. Can I connect the mains ground on the scope to the DC negative off the power supply to probe the board?

[Brumby]

You should be fine.

The voltage you are reading is from capacitive coupling and the high impedance from the meter won't load it down much.

If you want a test rig to check, just get yourself an old fashioned household light globe - the one with a filament. Wire it up with two leads.  Connect one to an earth and then connect the other to the point of concern.  If the light globe lights up - even slightly - then you need to stop and check further.  If you don't see it light up at all, then get your meter and check the voltage across the globe.  It should be so low that you won't even be worried.

[manick]

Brumby thanks. I was considering putting a resistor and checking but I thought it wise to ask first instead of causing a mini explosion.

I will try this and see if there is any glow then probably try measuring the current on the 10A setting on my meter just to cross check.

[Brumby]

The advantage of using a mains rated light globe is that if you do have a mains path, it will survive happily.  (Just remember this as a possibility whenever you use it.)

The visible indication is good, too.

[manick]

Brumby I didn't have a 120V incandescent so I used a 12V car lamp. Figured the current was supposed to be negligible so I tried it.

The bulb did not glow at all. Tried the meter in series and could not get a current reading. Then I removed the bulb and shorted with the meter alone. 10A rating on the fuse. No current reading again so its safe as both replies indicated.

By the way with the bulb the voltage dropped to about 7V. Maybe its the resistance of the bulb that kept it from going to zero.

Thanks again.

[Brumby]

Everything you've said makes sense and is as expected - except this:

By the way with the bulb the voltage dropped to about 7V.

That seems a bit high - and, to my mind, isn't consistent with your other observations.  If this number is correct, there may be other causes like the meter leads picking up stray EMF.

For this exercise, the 12V globe was OK, since we already expected there to be no real current capability.  However, if you ever come up to a situation where you are really unsure, get yourself a 120V incandescent.  The last thing you want to do is put 120V through a 12V globe.

[Zero999]

For this exercise, the 12V globe was OK, since we already expected there to be no real current capability.  However, if you ever come up to a situation where you are really unsure, get yourself a 120V incandescent.  The last thing you want to do is put 120V through a 12V globe.

Unfortunately real mains voltage incandescent lamps are becoming harder to get hold of. In many countries governments have outlawed or put huge taxes on them, in a bit to increase efficiency and cut CO₂ emissions.

Fortunately halogen lamps are still often widely available and incandescents designed for use in ovens (where there is no alternative) are too but they're in smaller power ratings <60W.

[Brumby]

A low power globe is fine - just as long as it has no electronics and is rated for your local mains.

In fact, I would suggest a low powered one - 60W or less.  In case there is real current capability, I would think it a good idea to keep the current low.

[manick]

Trinidad is third world so I can still get incandescent in the stores. The test was being done in the wee hours of the morning so no where was open.

I checked it again just now and I am getting a jumping voltage with bulb. Fluctuates from practically 0 to 8V but never settles. Its a fluke meter but sometimes it acts a little funny.

With a direct short to ground I am getting 0V difference as expected. In this case its staying stable.

[Damianos]

You should be fine.

The voltage you are reading is from capacitive coupling and the high impedance from the meter won't load it down much.

If you want a test rig to check, just get yourself an old fashioned household light globe - the one with a filament. Wire it up with two leads.  Connect one to an earth and then connect the other to the point of concern.  If the light globe lights up - even slightly - then you need to stop and check further.  If you don't see it light up at all, then get your meter and check the voltage across the globe.  It should be so low that you won't even be worried.

Just a note:
To avoid the trap of a broken filament that does not glow, always double check with a reliable meter.

[Damianos]

...
I checked it again just now and I am getting a jumping voltage with bulb. Fluctuates from practically 0 to 8V but never settles. Its a fluke meter but sometimes it acts a little funny.
...

Something is wrong:
- the bulb is not a "pure" filament type (!)
- the instrument has a defect
- the instrument is in wrong mode (maybe DC)
- broken wire(s) of the probes
- ... something else?

[Brumby]

You should be fine.

The voltage you are reading is from capacitive coupling and the high impedance from the meter won't load it down much.

If you want a test rig to check, just get yourself an old fashioned household light globe - the one with a filament. Wire it up with two leads.  Connect one to an earth and then connect the other to the point of concern.  If the light globe lights up - even slightly - then you need to stop and check further.  If you don't see it light up at all, then get your meter and check the voltage across the globe.  It should be so low that you won't even be worried.

Just a note:
To avoid the trap of a broken filament that does not glow, always double check with a reliable meter.

That, actually, is an important safety tip.  Always check the filament before using it in this role!!