Al,
I once struggled with these sorts of issues for years, especially in analyzing SMPS waveforms with live above ground issues and high vs low side driver issues.
And I tried all the tricks of course, isolating the test object's earth, isolating the scope earth, using coupling capacitors and coupling transformers, using the scope as a differential probe (by adding the two channels and inverting one and not using the probe earths) and finally using actual differential probes.
But in the end I abandoned all of those methods in favor of the only one that I found works all the time, without measurement errors and without mishaps or accidents:
This is the Tektronix 222PS (Power Scout) scope which has fantastic isolation on its probe inputs, you can connect the probe earth anywhere to any circuit under test with no problems.
These scopes are still readily available, they are yellow & easy to spot (not to be confused with the plain 222) .
The only thing is that the original probes had an oddball division ratio and have become very hard to get, but I worked out a small modification to use standard probes, which then means a X100 probe can be added further extending the high voltage capability.
If you do get a 222ps (which would forever solve your measurement issues in the high voltage department) here is the article on how to use standard probes with them;(the high voltage warnings in my article are to cover the fact that some people have a limited understanding of high voltage risks, but others will know where to set the limits for measured voltages)
http://worldphaco.com/uploads/USING_...CILLOSCOPE.pdf
Hugo.