Old wife's tail, myths, and other stories we were told as an apprentice, before the days when BS 7671 started.

[ebee]

Can I just add to that Eric,
We should test without any of the circuit attached because we are purely testing the device itself and nothing else.
As soon as we connect a circuit then it tests the circuit itself and that could skew + or - the results, that would tell us how that that combination might result, not always a bad thing, in fact it was the way I used to do it.
In my mind, if I do a new install I think the connecting of circuits, sockets unplugged, lamps out of lampholders to be a good thing rather than a bad thing, usually.
But to do it correctly means the device only. I can see the logic

 

[securespark]

Yes, I was taught, and taught others, that the RCD should be tested in isolation.

I had a Wylex twin RCD board to install one morning and both RCDs were inop out of the box!

 

[JohnW2]

Can I just add to that Eric, ... We should test without any of the circuit attached because we are purely testing the device itself and nothing else.
As soon as we connect a circuit then it tests the circuit itself and that could skew + or - the results, that would tell us how that that combination might result, not always a bad thing, in fact it was the way I used to do it.

That makes sense,but it seems that what eric was suggesting (from his experience) was that connecting cables to the terminals of an RCD could 'warp its case' to an extent if impairing function - so,although such a problem would really be 'of the device itself', it would only appear when conductors were connected to its terminals.

Kind regards, John

 

[EFLImpudence]

Connect cables so they don't warp ???

 

[ebee]

Well, if we test the RCD(s) in isolation and we test when connected to a circuit with no loads and then compare then we have done all we can. If the circuits themselves are causing any warping we might see it.
What we can not control is what is plugged in, fridges and freezers in multiples on one circuit certainly do make readings differ as I have observed but we can not prevent that really.
So 1/ prime - check that device and 2/ Second - if you like - check device and unloaded circuit

 

[JohnW2]

Connect cables so they don't warp ???

Sure, but I imagine that eric probably thought that he had done that - i.e. didn't realise that the connections were 'warping the device. With the substantial sized conductors normally connected to an RCD, it's not at all difficult for them to apply appreciable stress (hence also 'strain') to the device, although that might not be apparent at the time.

If that has 'unknowingly' happened,then one test of the device with cables disconnected will show it to be OK, but if testing then reveals a problem when the cables are re-connected,that's likely to result in a (perhaps quite time-consuming) hunt for some cause of the problem in the wiring of the circuit or connected items.

That's really just a comment in response to what eric wrote. I'm not sure what one 'can do about it'!

 

[JohnW2]

Well, if we test the RCD(s) in isolation and we test when connected to a circuit with no loads and then compare then we have done all we can. If the circuits themselves are causing any warping we might see it.

Indeed - and I think eric was probably suggesting the need for that testing "when connected to a circuit with no loads", whereas many people probably only test 'in isolation', and maybe also "connected to a circuit with loads', (or maybe only the latter).

In practice, it might well be most 'time-effective' to first test when the RCD was "connected to a circuit with loads", since if that reveals no problems, then no other testing would be required.

Kind Regards, John