0
\$\begingroup\$

I had a discussion with my colleges regarding Oscilloscope measurements and DUT (device under test) isolation. Specifically a device connected to a AC/DC wall-plug.

I think I am aware of the concepts and I have read How is using a transformer for isolation safer than directly connecting to the power grid?, Why do we need an isolation transformer to connect an oscilloscope? and Isolation transformer and scope when troubleshooting SMPS.

From my understanding, if you have the DUT behind an isolation transformer, one has to be potentially careful not to connected the scope GND to the DUT, because there may be a large potential difference between GND/PE (normally connected on the scope side) and the DUTs "GND reference".

A common way I often used to circumvent this problem is to use the “channel A minus B” pseudo-differential approach. Another would be to float the scope / disconnect PE which is obviously a safety risk but may be a acceptable approach when working with low voltages.

Back to the question:
A college asked why it seems that there is no need to consider the "isolation issue" when measuring on a DUT that is powered from a wall plug, since those use an isolation transformer as well.
My initial thought was that there may be a Y-cap between primary and GND that serves as kind of a potential reference. However thinking about that, one would want to connect that cap to neutral not to phase which can't be guaranteed.

If wall-plugs do not have some reference to the primary side, wouldn't it be even an issue to, say connect a USB cable to some wall-plug powered device, effectively connecting PE to "some" potential? Am I just missing something obvious here?

\$\endgroup\$
1
\$\begingroup\$

My initial thought was that there may be a Y-cap between primary and GND that serves as kind of a potential reference

That Y-rated capacitor is for spurious supression/EMI, not for providing a reference. Read more about the Y capacitor here

Fact is that isolated mains supplies which do not have a mains ground connection which is also connected to the low voltage output, are unreferenced. So essentially, if there is no connection between the ground at the output and mains earth then the supply is floating. Within certain limits of course, there always be some leakage and capacitive coupling.

Since that supply is "floating" you can set the ground reference yourself. In practice you could ground it to the oscilloscope's ground.

Compare this to using a lab supply, there the supplies are also floating but normal operation is to reference the negative supply rail (assuming a single supply voltage) to the ground plug on the lab supply. This plug will (should be) be mains earth connected.

wouldn't it be even an issue to, say connect a USB cable to some wall-plug powered device, effectively connecting PE to "some" potential?

That would achieve nothing if the wall plug is the "floating" type. If it has "some potential" it is useless to connect it to that as you do not know what it is. You better connect to a proper ground/mains earth which is a potential you do know (zero).

\$\endgroup\$
  • \$\begingroup\$ Yeah I think I am just confusing myself here. Those discussions where you think you are clear about something, then someone asks a "but why?!"-question and then somehow nothing makes sense any longer. But doesn't a Y-cap only make sense for EMI suppression if you have a plug with PE connecting N and L to? However, I am not implying that 2-prong wall-warts even have a Y-cap, I was just assuming it may be the case. \$\endgroup\$ – Rev1.0 Feb 9 '17 at 11:26
  • \$\begingroup\$ But doesn't a Y-cap only make sense... No, you are thinking that the Y-cap conducts spurious signals to ground or mains. That is not how it works. It actually prevents that ! The unwanted high-frequency signals are created by a fast switching transistor at the mains side of the transformer. Those RF signals also couple through the transformer to the output. Currents travel in loops. Without the Y-cap these signals do not have a short path back to the switching transistor so they emit as EMI signals at the low-voltage output ! That will make the supply fail EMI requirements. \$\endgroup\$ – Bimpelrekkie Feb 9 '17 at 12:28
  • \$\begingroup\$ To prevent this the Y-rated cap is used, it provides a short path, inside the supply, for these RF signals so that they stay inside the supply. The cap must be Y rated since one side is mains connected and the other side must be safe to touch for you as a user. If the cap fails it cannot short circuit as that would connect the output to mains and you would get ZAPPED when you touch that. Y-caps have thicker isolation and much less chance of shorting when they do fail. \$\endgroup\$ – Bimpelrekkie Feb 9 '17 at 12:30
  • \$\begingroup\$ Thanks for the info. I thought that the Y-Cap common mode suppression was meant to prevent switching noise "out of" the primary side. Because this is what a mains connected consumer device would be EMI tested against. \$\endgroup\$ – Rev1.0 Feb 9 '17 at 13:13
  • \$\begingroup\$ Actually this pretty much reflects my understanding of it. You say that its meant to suppress EMI at the low voltage output. But I suppose it does both since its all about the low impedance current loop at HF between primary and secondary. \$\endgroup\$ – Rev1.0 Feb 9 '17 at 13:20
0
\$\begingroup\$

... since those use an isolation transformer as well.

My initial thought was that there may be a Y-cap between primary and GND...

I'm not sure of what you're worried about, if the Y-cap is connected between the primary (L or N) and earth GND, since you're measuring voltages on the secondary of the transformer.

There is no connection between the secondary and the primary (Not taking the transformer's parasitic capacitance into consideration. This has to depend on the frequencies you're working with, I assumed they were low).

Unless your measurements are made on the primary side, there won't be any problems connecting the scope's GND to the DUT's.

The rest is on FakeMoustache's answer :P

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.