Thanks for you replies. I have made quite experiments with oscilloscope but fortunately I hadn't blow up something yet and because I know all the danger with the grounding I ask to many things here. In same cases the the groung was not so clear how is arranged on the PCB I have used two identical probes and I 've done the differential math in the oscilloscope but I had noise on the measured signal because the ground clips are not connectioned. I have used differential probe but I would like to risk and measure with "ordinary" probes. Always there is something to new even with risk.
Okay, to learn abouts whats happening it absolutely makes sense to do the math in the scope. But if you (also) want to learn from other peoples experience, usually the quality of the results with a "analouge computer" in the probe circuitry is much better than whatever the scope can do digitally.
Personally I´d rather *only* use the differential probe, and if I want to measure against instrument ground I can also connect the "-"pole of the differential probe to instrument ground. It´s just THE big advantage of any differential probe, you can always select by yourself what part of the circuitry should be ground now. With the conventional probe you just don't have this choice.
Of course there are always many valid ways to many valid targets
