then i think we can all agree that if we had a device called a "neutral grounding resistor monitor" connected up, which would then trip the breaker if there was a connection between earth and neutral while leaving 'neutral' floating and not bonded to earth, we would have the safest system going?
I believe that takes this full circle. Here, we call that a GFI/GFCI/RCD/PITA
I encountered this issue with the truck and it's umbilical cord connection, and wrestled back and forth with what was the best way to provision for every scenario I could think of... the answer? There wasn't one. Not that covered them all at least.
In the end, I went with looking at the truck as an "appliance" that has most of its threat potential inside of it, and gave it its own internal N-E bond, complete with a set of downstream GFIs. Once someone is inside, there's no more threat of getting zapped than in a normal everyday dwelling wired for AC.
The caveat to this is that because of the tires, it is for all intents and purposes inherently isolated from earth, as the "single appliance". The problem then became how to deal with bonding the frame to my upstream source's ground to keep it at earth potential, while keeping an upstream GFI happy under no-fault conditions. Put simply, it wasn't possible to do *both* internal and external safety protection to my satisfaction at the same time.
If the upstream connection is to a non-GFI protected source, the earth bond is more important than ever, because if I'm parked on grass or the asphalt is wet, there's a much higher chance that anyone with one foot on the ground and the other on the rear bumper will be part of a lower resistance path at both, than if it is parked on pavement and everything is dry.
Having the internal N-E bond however, this means that the normal current return path would share with the earth lead.
If there *IS* an upstream GFI (a scenario I only envisioned as likely to increase in probability of encountering), this results in an immediate trip of the GFI with nothing more than the *capacitive reactance alone* of the internal wiring connected as a "load".
As a side note, this was a bit baffling until I realized (and subsequently confirmed) that this was what was happening... the fluke was showing completely open resistance measurements, but that is done with DC. AC had snuck parasitic capacitance in under my nose. This however is ultimately moot, because even without it, the moment a real load gets applied, the same result happens.
What I ended up doing was placing a "jumper" in that could leave the truck frame and umbilical disconnect/breaker box bonded to the umbilical's earth, still leaving the internal bond in place to provide the protection on the inside. This of course then ultimately means that an open (or other) neutral fault between the truck frame and upstream source's N-E bond relies on the upstream GFI to prevent shock hazard conditions. Again, I weighed this against the idea that it's then essentially the equivalent of grabbing a toaster chassis in an up-to-code kitchen in terms of threat. GFIs fail too, after all, but that failure would need to be combined with the open neutral or other fault to become a problem.
Whether or not to have the jumper in place of course requires knowledge of the upstream system I'm connecting to... which leads me to a final point...
The thing I keep seeing here, and it's either overlooked, ignored, or rationalized away, is something that was ingrained in me early on: "Treat every wire as if it can kill you". You can rationalize and argue against the idea that its possible all day long, but get into an argument with physics, and (I really hope) I don't need to justify my reasoning for betting on my choice as to which one will win, right?
As for isolation transformers: Under the wrong circumstances, they *will NOT* save your neck. They're designed in general to isolate power and ground leakage to levels that *equipment* can safely tolerate, not people. Don't fool yourself into thinking otherwise. If you don't believe me, connect one with only the hot side connected to the appropriate primary terminal, and leave the neutral and ground disconnected... then check how much AC current is available between one of the secondary leads and earth...
You no longer have a transformer... you have a *capacitor*, a non-trivial one at that, which passes AC rather freely. Still think it'll save you?
</rant> </2cents>
Steve