RCD tester design - sanity check

[fcb]

I'm building a precision RCD/GFCI tester for incorporation into a product.

In my infinite wisdom I've decided to do it with a pair of "resistor" simulators. There seems to be a lack of low-cost (or any cost!) high voltage PNP's these days.

There's a processor in the unit to compensate for changes in the line voltage and some extra circuitry as well as either a digital pot or a multiplying DAC.

Does this strategy make sense?

[fcb]

BTW: ^^This design works pretty well.^^

Now for the next part of this project - Earth Fault Loop Impedance tester.

So imagine you want to measure the impedance of the your wall socket and calculate the peak fault current. You are measuring the resistance from Line back through your fuse box back through the local utility step-down transformer to the earth and back to the earth pin on your socket.

The old way (and still the most accurate way is to momentarily (1 cycle) hang a low value resistor between Line and Earth (drawing perhaps 10-20A), and measure the dip in the mains voltage and calculate it using Ohms law.  This is still used for what is called 'Ze' testing, but cannot be done AFTER an RCD/RCBO/GFCI as it will fire the trip. Bridging out the trip is tiresome and removes protection (albeit temporarily), so not ideal.

So 'no-trip' EFLI was invented - this involves doing the same as above but using a much smaller current (perhaps 10-15mA).  This will typically take much longer and is prone to noise pickup etc...

Now 'no-trip' EFLI is often done in what is called 3-wire mode (a higher current, perhaps 6A, between Line and Neutral - and a low current (10-15mA) between Line and Earth (so as not to trip the RCD/GFCI) - it's more accurate than non-trip 2-wire testing but not as good as high current 2-wire testing.

And here lies my problem - I'm having a mental block on how it's done. Totally get the 2-wire methods, just not the 3-wire method. Any ideas?

[Rerouter]

For loop impedance, you already have a 3 wire circuit made for you for most systems, Active, and Neutral / Ground tied back at the fusebox,

For active / neutral, measure the ground to neutral voltage difference, if this is not almost 0, then you know you have other devices on the circuit which may interfere,

You would pulse a load on the active / neutral, measure the drop in voltage between them but also measure the difference between neutral and ground, this will allow you to calculate the impedance of the active and neutral wires independently (neutral is E-N difference / Test Load current, Active is the remainder of the change in voltage.)

As for measuring the ground circuit impedance, the easiest way would be plugging in to a ground on a different circuit, and using the neutral as a reference, but i understand this is rarely an option.

In reality i would do a chopper approach to remove X/Y filter capacitance leakage, you draw some power from A-N with an isolated supply, then pulse 15mA between neutral and ground, you know the impedance of the neutral from earlier, so after subtracting you have the ground impedance,

You would measure the voltage difference between the pulse being on and being off, this would remove any filter currents.

[fcb]

Thanks Rerouter - you cleared my mental block!

Now building a prototype.

[Jeroen3]

RCD testers are complicated to make. You have various types, characteristics and qualities. There are also digital configurable RCD's.
Most complicated are the cheap and digital ones.
The cheap ones don't clamp the trip coil, causing surges on your tester.
The digital ones are complicated with connection testing with an ohm meter, because they have active electronics. Considering you want to test their contacts as well.