measuring milivolts with EMI ghosts around

[MacIntoshCZ]

Hi,
just tried to scope my active PFC with scope, but it seems i am unable to see anything. When i turn pfc on and probes are not connected (rigol 1054z) ripple is low, when i move my hand to bnc connector ripple increase to range of hundreds of mVolts. Also if probe is connected to the osciloscope and shorted at the end that should go to pcb ripple is still high (range of 10-100mV).I got ripple instantly when i connect something to bnc connector. Even if i just stick wire to the midlle of that connector as  few cm antenna ripple goes crazy. When pfc is off i can see clearly trace near 0V level. I have no idea how should i block it. I would need to place it in hermetic isolated conductive ´ball or something like that.,

* i am measuring voltage on  current shunt resistance

[T3sl4co1l]

And you're measuring what?

It's a scope, get a trigger on it, show a screenshot.  Is it mains hum?  Is it a nearby radio station?  Is it switching noise?

I gather you have an active PFC project.  Is this powered from an isolation transformer?  Might the scope have been damaged from connecting it to this circuit if not?

Tim

[MacIntoshCZ]

I am ghostbuster measuring milivolts. Just kidding
I tried to measure rectified sine current after bridge on shunt resistor. It is switching noise. PFC is powered for isolation transformer. I doubt its capacitive coupling.

[MacIntoshCZ]

I see it as a big problem that the circuit is not in a grounded enclosure. My current project layout is disastrous.
Is there any way to measure in this environment? Maybe using some nearby filters and op amps at the same time.

[MacIntoshCZ]

ok maybe getting better probes would help, just tried to connect 5n6 cap for decoupling to bnc connector and now i get only 50hz sine wave when my hand gets closer enough to bnc.

[MacIntoshCZ]

Also i think runing osciloscope on battery coul significantly reduce noise coupling on probes

[CaptDon]

First off, any scope will show lots of noise with just a wire stuck into the center of its BNC input. Second, "The rectified output from a current shunt.....??" So most current shunts are 50mv max, how are you 'rectifying' that??? Diodes won't even conduct at that level. Third, 99.99% of the time you can't just connect the ground side of your scope to one side of a current shunt since without isolation you have now provided either a short or alternate ground path for the current you are trying to measure. At the very least you probably need an active differential probe solution. Your bench and layout are a nightmare. What sort of power factor correction are you designing? If you haven't sought out a differential solution to a measurement across a current shunt to reduce ground loop noise then maybe this isn't a great project for you. Wondering how you haven't killed the scope yet?

[Alex Eisenhut]

50mv max, how are you 'rectifying' that??? Diodes won't even conduct at that level.

As a Pedant Class I, I must point out that back diodes do. Then again, 50mV at a few KHz is not their wheelhouse.

[MacIntoshCZ]

Probably a translation error. I measure on a current shunt whose size is 25mR. Then 1A -> 25mV. As for my oscilloscope, I think he's still alive. The PFC circuit is galvanically separated by a variable transformer. I do not hide the fact that the measurement takes place in field conditions (war field). I tried to use twisted wires as best I could. The transformer is in a metal housing from an atx source that is grounded. The output from the probes is also a twisted wire and is as short as possible.

Why should my oscilloscope be destroyed? I'm not an idiot, maybe just a little.

thanks for help.

[Conrad Hoffman]

Measurement is an art and science. You won't go wrong by reading Jim Williams- https://www.analog.com/media/en/technical-documentation/application-notes/an47fa.pdf

When I do sensitive stuff I have an aluminum chassis with an insulator in the bottom. I drop the entire circuit in the box to shield it, paying lots of attention to where I ground the circuit, the box and the scope.