Testing A-type RCD at half times 100m/A ?

You may prefer this video describing some of the recent updates, it is a bit tongue in cheek, but seems to be more accurate.

About 3 and a half minutes in it reaches the RCD related stuff, and is good enough to be very clear, 30mA RCDs, of all types, are to be tested with a sinusoidal alternating current  fault of 30mA RMS, as per this screenshot.

On a different AMD2 surprise note

Amusingly and perhaps slightly nugatory,  the building I am part way through in my own garden will have foundation earthing anyway, as I had welded the studs onto the mesh before the AMD2 came out, and the floor is now poured, but it will still be interesting to see how it tests out.

Mike.

Splendid!

Had I known about foundation earthing when I had my garage extended, I (or at least my contractor) would have made suitable provision for it. Just because something is not required is no excuse not to have it, or at least make it simple to install at some future point.

Fortuitously, I bunged in a length of galvanised conduit for power and lighting in the pit, so that will have to do instead.

gkenyon said:

Chris Pearson said:

A 30 mA type AC RCD tested on the type AC setting at 1 x IΔn averaged a trip time of 33.2 ms. On the type A setting, it averaged 21.1 ms.

All this means, is that the RCD responds to different waveforms with different trip times.

I don't think that it goes so far. All that I can deduce is that either the waveform was different, which caused it to trip more quickly; the current was greater; or of course, a combination of the two.

gkenyon said:

On the Type A setting, it runs a waveform designed to test RCDs for the additional residual current characteristics (only) for Type A RCDs

Doubtless Graham is better informed than I about the inner workings of an electrician's tester (one could hardly be less well informed).

gkenyon said:

BS 7671 NEVER had a requirement to test with anything other than AC waveform, so there's nothing to communicate. It has been noticed that there is a lot of confusion in the industry regarding the functionality provided by MFTs, and BS 7671 requirements, and therefore this has been clarified in the recent GN3 ...

I am v. pleased to see the clarification because the proud new owner of Megger's latest would be fully entitled to follow their instructions.

The incipient obsolescence of type AC RCDs is all well and good, but if there is no requirement to test the additional capability of a type A, how do we know that it does what it says on the tin? It can only be an assumption, albeit no doubt a well-founded one.

Indeed - if it helps, then ignoring the safety electronics to cut the test off after 2 seconds, or earlier if the NE voltage moves too far, what is in an RCD tester on the 'AC' setting is little more than a suitable resistor between L and E. The electronic switch that puts it in circuit when you press the button comes on a few volts after the next zero crossing of the selected polarity, and starts a timer counting. The timer is stopped when the Live voltage has fallen below a certain fraction of the expected sinewave, as this indicates the contacts have opened. Some loads that hold the voltage up can fool the tester into thinking contacts are still closed after they have opened.

The A test is the same except now we emulate a resistor of slightly lower value with a diode in series so the RCD sees the rectified but unsmoothed DC waveform rather than a sinewave.

Actually most AC RCDs will respond to the A test waveform to some extent, but not necessarily at the right levels, and performance is not guaranteed.

The steady DC of a B test is far more likely to blind both AC and A types, as the DC detection is a very specific add-on internally.

This is entirely consistent with the results you report.

Mike.