|> For those interested in why I would do that, I will post that at a later |> time if it turns out to be practical to do with existing products. |>
| | They are designed for operation at "normal" line voltage, plus or minus a | reasonable tolerance. I have tested a few, and they quit working below | around 80 volts. The electronics continuously energize the solenoid with a | loud buzzing, and it no longer trips. I suspect that component failures | would be imminent at this condition.
That's similar to the behaviour I have found with GFCIs when subject to an RF field of sufficient strength at the right frequency (for example my 2m ham radio transmitter triggered it within 2 meters at 1 watts, 222 Mhz did not trigger it at all, and 70cm took 5 watts at 1 meter to trigger it). It actually did trip, but the buzzing indicated to me it continued to try to trip.
This suggests to me a design flaw in the GFCIs, unrelated to its inability to work at low voltages. After tripping and opening, normally there would be no more fault current. But false triggers, of which RF is only one possibility, can still cause this. Maybe the solenoid needs to be powered from the load side of the contacts so it ceases when it opens them.
I wonder what dangers could be present from these situations. Are the component failures you suspect are imminent in the GFCI itself? Could the solenoid burn up under this condition, either at lower voltage or at full voltage? Could that start a fire? Likely it wasn't designed for continuous tripping duty.
In addition to the RF field trigger source, I see a couple other scenarios where this can happen:
Utility distribution single phasing at the delta primary side of a substation or intertie transformer can result in half voltage on 2/3 of the single phase services downstream of it. I've seen this happen at least 3 times, with the worst case covering half of a town that wasn't cleared for over 3 hours. If a GFCI gets triggered in this situation (or worse, falsely triggers because of the low voltage), could it withstand 3 hours of punishment like that?
Because balanced power systems, where each load takes power from two hot lines at 180 degrees (such as 240 volts loads connected via NEMA 6-20) have much lower common noise introduction into the power lines, this can be the preferred wiring method in special applications like recording studios. But to get the best benefit you have to do it for every (all noise sources). And given most things run on 120 volts in the US, that means a modified system where the center tap is 60 volts into a 120 volt secondary, and it wired like the 240/120 volt system, but at 120/60 volts. Putting that in a circuit breaker panel would mean 2-pole breakers for everything, and the half-voltage scenario for GFCI breakers.
GFCI receptacles, depending on how they are wired with respect to the neutral and grounding wire, may work fine, or be worse.
AFCI would likely have problems similar to GFCI.