RCD failing 5x test at 0 but not 180

[pcboffinuk]

Testing a newly installed Crabtree RCD today, and it passed all tests except the 5x at 0 Deg (12ms at 180DEG).

It did trip at 0 Deg, but not within 40ms. (it is performing supplementary protection so test is required)

Apart from a duff RCD, am I missing anything obvious that could also cause this?

Is it rare to find a defective RCD or are they more common that I realise?

Gavin

 

[Spark123]

I take it there was no loading on any circuits at the time of the test?

 

[pcboffinuk]

Spark123,

so far as I know there wasn't, though its possible a pesky householder plugged something in upstairs while my back was turned.

Surely any load would tend to make the RCD more likely to trip rather than take longer though?

Gavin

 

[Spark123]

Something that introduced any sort of DC pulse into an AC RCD might have a locking effect.
Failing that, the only option you really have is to swop it.

 

[pcboffinuk]

That's interesting - are there likely candidates for doing that? It may be worth another visit and test before I replace it.

The failure was repeatable, as were the passes.

Gavin

 

[EdwardCurrent]

Hi,

I have had problems with Wylex RCD'S in the past few months, some not working at all, others not functioning correctly, they were all purchased from that screwfix catalogue by a local builder, because of the hassles of returning goods they had to get new ones form the wholesaler. .


Ed

 

[Spark123]

I'm thinking some switch mode supplies and dimmers can introduce bad waveforms to the supply, though thinking more about it you'd probably need a fair amount of them. It is worth trying it in isolation just to see if it passes. Do you have a spare you can try also?

edit:

Hi,

I have had problems with Wylex RCD'S in the past few months, some not working at all, others not functioning correctly, they were all purchased from that screwfix catalogue by a local builder, because of the hassles of returning goods they had to get new ones form the wholesaler. .


Ed

That's interesting, I think Wylex and Crabtree are part of the same group, as are Volex.

 

[pcboffinuk]

Doubly interesting, since the Crabtree CU came from....Screwfix.

I will test it in isolation, and bear in mind to ensure I keep a spare in stock for future reference.

Is there one solid brand that causes practically no problems? I have used Wylex and MK successfully before, but liked the design of the Crabtree busbar so gave one a try.

Gavin

 

[securespark]

Spark123,

so far as I know there wasn't, though its possible a pesky householder plugged something in upstairs while my back was turned.

You're testing it wrong!!!

You're supposed to isolate all circuitry from the outgoing side of the RCD & test directly from the outgoing terminals, thus eliminating any dodgy wiring or appliances from the equation.

As for it affecting testing, very strange faults can appear with RCD's when attached to faulty wiring or appliances.

I got a call from a manager asking me why one of his lads couldn't get the RCD to trip from a socket outlet, but when tested as above, it worked fine.

Turned out it was an RCD incomer & a fault on a lighting circuit meant the faukt current was returning down the neutral, not the cpc, so the RCD was not seeing an imbalance.

 

[pcboffinuk]

You're testing it wrong!!!

You're supposed to isolate all circuitry from the outgoing side of the RCD & test directly from the outgoing terminals, thus eliminating any dodgy wiring or appliances from the equation.

Hmm, so one should never use the plug and lead that every test instrument supplier supplies to test rcds? All my on site guide says is that it should be tested as near as practicable to its point of installation (the nearest socket in this case)

Admittedly it says the load supplied should be disconnected during the test, but I never took that to mean circuit cabling disconnected too...

In this case the wiring had been fully tested since it was a CU change, but thinking about it, if one of the appliances in the house causes the RCD to not function as intended, isn't it a good idea to know about it?

I will be checking and it may pass the test isolation, but is that not avoiding a potential problem?

Looks like my shiny new PAT tester may get an outing

Gavin

 

[GaryMo]

Also, on a schedule of test results (PIR, EIC) you'll find RCD operating times for each circuit. I was taught to use the plug in lead (unless there wasn't a socket outlet present) with my multi function tester directly at a socket outlet during my 2391 course.
If RCD testing was supposed to be done at origin then why would a request for times be by the side of each circuit? Unless there's a socket fitted with built in RCD protection.
As a rule I normally only do an RCD test (0.5x 1x (0 and 180) 5x (0 and 180)on one circuit of a split load consumer unit as there's no point testing the same RCD numerous times.

 

[securespark]

You talk to the manufacturers. They all recommend testing from source. If you don't, you often get misleading results. It's happened to me loads of times.

I go to houses where RCD's have "failed" only to test them at source to discover they are perfectly healthy: the reason for the "failure" is something in circuit that is amiss. I have assessed engineers who have tested a "faulty" RCD, then changed it, re-tested & got exactly the same results with the new one...

9 times out of 10, you can test from a socket outlet OK, but if you get a failure, don't write it off without testing in isolation.

 

[ricicle]

You talk to the manufacturers. They all recommend testing from source. If you don't, you often get misleading results. It's happened to me loads of times.

I go to houses where RCD's have "failed" only to test them at source to discover they are perfectly healthy: the reason for the "failure" is something in circuit that is amiss. I have assessed engineers who have tested a "faulty" RCD, then changed it, re-tested & got exactly the same results with the new one...

9 times out of 10, you can test from a socket outlet OK, but if you get a failure, don't write it off without testing in isolation.

Good advice, as ever, Monsieur Secure