When (where) is it safe to measure with oscilloscope

[czbaterka]

I ordered my first oscilloscope (Siglent SDS804X) and before it arrives, I would like to learn how to not blow it up. I originally wanted to buy Rigol DHO804 which is isolated, but you guys recommended me the first one so I am little worried

I already studied the basic ways how to not mess up the probing (from EEVblog #279 video), mainly that I must always make sure that oscilloscope probe's ground terminal is always at same ground potential as the thing I am measuring, or the thing I am measuring is floating.

But I am not sure about switching power supplies and dual laboratory power supplies. I would like to buy two lab power supplies like OWON SPS6051 (or similar using switching power supply), or some dual power supply like OWON ODP3063.

For the first example - Two single channel power supplies:
 - Can I connect (+) terminal of the first one to (-) of the second one to get e.g. +-15V symetrical power supply with ground at the middle point? Is that still considered as floating widget that I can probe whenever I want until I connect GND terminal to the middle point (if I can even do that)?

For the second example - One two channel power supply:
 - I know that those are surely made in way that I can make symetrical power supply and they will probably work as I expect right? What about the GND terminal? Can it be connected to the middle?


I am just a little confused by the whole linear vs switching power supplies as linear power supplies have transformer isolation and is almost always floating. But I am not sure about the switching power supplies in those laboratory power supplies as it is not in the description of the lab supplies (at least not for the cheap ones).

Same goes for the lab power supply modules like RIDEN RD6006 and similar connected to the "silver brick" switching power supply. Are those floating or not?

Last question I want to ask is about differential probing with multichannel scope. Can I always probe by two positive oscilloscope leads (from two channels), while not connecting ground leads at all and use math function to compute real waveforms same way as if I buy one of those expensive differential probes? Or there are also some limitations when using it like that?

[PGPG]

My first oscilloscope was one I have designed and made myself (in 1982). These were times when standard AC220V sockets in our home had only 2 contacts (there were only few sockets with grounding pin in kitchen and bathroom). So natural was for me that my oscilloscope have (I still have it) plug with only 2 pins and its case (so also BNC sockets shield) were simply not grounded.
All those time I was also using power supply I had made several years earlier (power supply was my first need so I started with doing it first). My supply has (even I have made it 50 years ago I still use it when need a power source at home (it is 0..40V, 100W) flat plug, so also is not grounded.
That's probably why your concerns are foreign to me and I've never thought about it too much.
Typical one voltage power supply has 3 banana sockets (-/GND/+) and a piece of metal connector that can be used to connect - or + with GND. All lab power supplies I have ever used had these metal not used so their outputs were simply floating and I had never any worries about my oscilloscopes.
I think it is correct solution to just have all supply sources at desk floating. In such situation I don't worry if I connect probe GND terminal to - or to + supply or whatever net of circuit I want. But after thinking a moment I have to say that I just have never get an idea to connect GND probe terminals of two probes to different points. It is kind of automatic habit that whenever I use two probes (in many cases you use only one) I start from connecting both probes GNDs to the same potential.
So my suggestion can be summarized to:
- use all lab supplies as floating,
- if you use two probes always connect their GND terminals together.

I don't like the idea of not connecting GND terminals at all and using only probe tips to make differential measurement. That way you not control the voltage at each input so I suppose it can happen that input amplifiers will be working just near one of their limits making them being not linear. But it is just my guess.

[timeandfrequency]

hello czbaterka,

As you watched Dave's video on that topic, you already have first class information.
If you need to probe on live circuits, today's safest method is using an isolated scope or differential probes + non isolated scope.

In the 20th century, the two probes method (non isolated oscillocope setup to A - B traces) or an isolation transformer were used. Both are much less safe that the ones stated above.

And install a 10 mA GFCI (Ground Fault Circuit Interrupter) to power your electronic bench, or at least the DUT (Device Under Test).
 

[tggzzz]

Welcome to the forum. Good to see someone thinking and questioning in advance

Have a look at the references given in https://entertaininghacks.wordpress.com/library-2/scope-probe-reference-material/ They give solid professional advice and the reasons for that advice.