FWIW...
Every place I've lived I set up my station with a local earth ground whenever possible to drain off RF noise. I would NEVER open the ground return to the panel! But to isolate it for RF I would (do) installed a heavy RF choke (air core) on the return. About 50uH of conductor rated for the same or better current than the grounding conductor. There's no violation of NEC with this. Back in the day several lightning protection systems required inductors to ground. I don't think they do that anymore.
I have used isolation transformers with floating neutrals as Don described but I haven't seen any advantage to this in practical terms, but then again, I've never had any hum/ground problems to speak of.
Also, the last time I visited the sub panel issue it was pretty clear that NEC recommends only ONE panel (the service entrance) have it's neutral strapped to the local ground. This includes outbuildings! Unless fed from a utility service drop the panel does not get strapped to neutral. I can have it's own earth ground however. In the event of a nearby lightning strike there will inevitably be a ground potential gradient between the two earth grounds. That differential will be seen across your equipment (ground to line/neutral) unless you provide some break-over protection.
Given the drop any transformer will exhibit I don't see what advantage you expect to gain. I have to agree with John JN in his comments in that regard. I am a fan of Constant Voltage Transformers for load regulation but that's another issue. The independent "RF" ground and choke isolated "safety" ground thing has always worked well for me.
Mark
I agree about what NEC seems to call for.
I'm trying to fathom the thinking behind it.
To me it seems counter intuitive.
----------------
Actually the potential gradient between the two ground points is
the issue!
Or a big issue anyhow.
With a long ground run, and an indirect hit (a direct hit, all bets are off on any system with a direct hit?) I'm saying that the vdrop along that ground
is going to be a big problem.
Otoh, with the (what was it called?) separately derived service (??) using a large iso transformer and a local ground, the only opportunity for a voltage differential is
if there is breakdown across the iso tranny. Then the opportunity for damage comes if that differential can not be dissipated or otherwise shunted to ground. I suppose if the breakdown voltage of the iso transformer is exceeded at a sufficiently high current then damage could occur - but imo it would be worse with a solid connection running 100ft or so back to the panel. Especially if the inception of the "hit" is somewhere at the far away end of the 100ft run, since the self inductance of the wire itself is going to resist the pulse of the lightning! So, it will "look" for a better path to ground.
As far as line loss, I don't see that at all in practice.
I have a large iso tranny now in service, but running the ground back to the panel, and the 120vac taps out there are far stiffer than they were with the very same wiring split to 120-0-120 with no iso tranny. With the iso tranny running off 240vac and (for example) using just one single 120vac outlet one gets the benefit of having a 240vac line feeding it, stepped down AND the inductive "kick" of the large core. It works nicely and it kills common mode noise afaik.
Another big positive with the use of an iso tranny is the ability to kill transient pulses that often take out solid state microprocessors and cmos! A patent I read on a very excellent and advanced design method to protect gear from line transients made specific note that their system was approximately equivalent to using a large isolation transformer! The benefit they had was smaller size and lower cost! In my case, iso transformers can be gotten inexpensively (whenever they can) and the "large size" is not a big worry, so that works for me!!
That's my story and I'm stikin' too it!
And yes, I am still concerned about the ground differential possibility as well as the problems with a long run of ground wire.
Now what are you saying about breakover protection? Between what and what?? And by what means?
A potential problem with using an RFC in the safety ground return to the panel is that a hit out at the (shack) location will not have a low reactance path to ground - meaning it will seek a shorter path to ground, through your equipment! So, you still need a local ground (rod) at the shack???
_-_-bear