What's in a GM1004 "ArcTrap"?

[LiteYear]

I've come into possession of a GAI-Tronics GM1004 ArcTrap. It's a two terminal device that looks like a big toroidal inductor and is designed for use in earth fault protection networks. I'm trying to characterise it and coming up with very confusing results. I've done a frequency sweep, an amplitude sweep and a few impulse tests. In all cases the behaviour of the device is not well described by any model of a non-ideal inductor I can come up with. There is obviously something more sophisticated going on here.

Is anyone familiar with these devices? It seems they were originally manufactured by National Mine Service Company (Femco), and the MSHA has some record of them.

Failing that, can anyone explain the step response seen in the attachment? This is what happens when a 33mF cap, charged to 3.2V, is discharged into the device while monitoring current into and voltage across the terminals. It's the clearest picture of the behaviour that I've been able to distil.

As you can see, there's an initial rush of current which then plateaus at about 180mA (can be as high as 300mA for higher initial voltages). This plateau continues for about 10ms (can be as high as 80ms) before ramping up and becoming very steep until the capacitor discharges. None of this is typical of an inductor. Can anyone explain either the design intent or a suitable electrical model?

 

[VEBill]

I'm only spending a few seconds on this, so this may be off-base...

But is it related to www.google.com/patents/US4340920 ?

 

[desrod]

Try the following link:

http://www.gai-tronics.com/literature/industrial/gtc070503gm8000.pdf

You can even call the guys there, I am sure they can help.

 

[LiteYear]

VE1BLL, that patent certainly makes use of the device, although in a non-standard way and without further enlightenment.

desrod, yes that's about all I've been able to find as far as documentation goes. Still no details on design intent or physical arrangement of the device itself. Calling is an option - I'm in Australia so will have to negotiate the international hurdle but that's surmountable. I just noticed that they even have a distributor here which could be a start. I was concerned that it was a legacy product and finding someone that knew anything about it could be very difficult, but now I see that at least it is still listed on their webpage. Still - I suspect they're not the OEMs, and may not have all the details either.

Through further digging I've been able to determine the purpose of the device is to prevent the 4kHz signal reaching the sentinel device via paths other than the cabled return (the parallel earths problem). It's also designed to saturate in fault conditions to prevent excessive earth fault voltages. The specifics of that behaviour were highly unpredictable in the testing I performed initially.

Last night I scaled up the tests and subjected the device to 240Vac series limited to 6A and the behaviour was a lot clearer. The dramatic saturation characteristic is very evident, though the inductive properties are still very difficult to explain. If I don't make further progress this morning a few phone calls might be in order.