RCD tripping over 40ms on 5x

[Sidewinder]

I think a certain Mr @Sidewinder may be interested in your findings

Unfortunately Kerch, that ship has sailed.

A few people were trying to gather data to prevent RCD testing changes in the 18th, we failed to gather enough conclusive data, so the manufacturers got their way.

 

[Sidewinder]

Totally agree if it fails a low level current test on the circuit then how reliable is it going to be with a real fault

You are making an assumption that the fault current may be higher without any guarantee it will be. Anything above 30mA is sufficient to cause serious harm or death, the current situation as extolled by the IET for the testing of RCD's may as well tell us not to bother with RCD's as they may not operate when really needed

The only erroneous problem I have ever found with an RCD tripping at source and not tripping from a test on the circuit was the fact that on ramp test the RCD consistently tripped with a current exceeding it rated value. This was not an isolated case and some were brand new boards and some only a few months old that tested OK at install but not when doing other work some months later

Due to the blame game culture the UK has adopted I always ramp test any RCD I work on for piece of mind even though it is not a testing requirement to do so. The number of times I have found faulty RCD's is backing this approach up, I look at every job as a potential legal case these days if something god forbid happens then I know I have done everything I could do to mitigate any prosecution for negligence

The issue is that the testing required by BS 7671 is not to test the function of the RCD in circuit, it is intended to test the function of the RCD in isolation.

If the manufacturers have their way the only test we will be doing is the test button.

The testing in BS 7671 is copied from the product standard.

There are several issues with testing them in circuit which could result in false readings.  These are all down to circuit characteristics that can be identified by circuit investigations.

Do I test them in circuit, yes, but, if I see an issue or get a borderline reading, I will disconnect, test the RCD in isolation, then again in circuit.

This will give me the indication of the reasons for the original issue.

 

[Murdoch]

so the manufacturers got their way.

Which means?

 

[Sidewinder]

Which means?

The change in the RCD testing from the 17th to the 18th.

 

[Murdoch]

The change in the RCD testing from the 17th to the 18th.

I did wonder why this was even considered necessary.

It seems like change for change sake only - in an attempt to justify the 18th was necessary.

 

[Sidewinder]

The manufacturers of RCD's considered it necessary to reduce the number of what they consider to be false warranty claims due to devices that failed field testing under BS 7671.

So it was not, change for change's sake.

It was change to increase manufacturers profits.

There was nothing that the rest of JPEL/64 could or would do to go against the majority.

If greater evidence could have been gathered this would not have happened and the full suite of testing and recording of information would still be required.

However, a few people tried to gather evidence, and the response from the trade was somewhat pathetic, so the manufacturers won and got it reduced.

Their next aim is to remove it all together and rely on the device test button.

 

[Rob.]

The manufacturers of RCD's considered it necessary to reduce the number of what they consider to be false warranty claims due to devices that failed field testing under BS 7671.

However, you can still return the item under warranty if it still fails the tests?

Haven't looked over the RCD testing requirements from the 18th, what's the change?

 

[Sidewinder]

However, you can still return the item under warranty if it still fails the tests?

Haven't looked over the RCD testing requirements from the 18th, what's the change?

I don't see why you can't make a warranty claim if the device fails under 7671 testing, I would.

Push button test to 6 monthly, not quite testing but limitation of leakage current in designs to 30% of the RCD residual current rating, EFLI testing downstream of an RCD not essential if R1+R2 have been proven, for additional protection it is no longer mandated to do the x1 tests, apparently if the x5 test is fine then additional protection is ok.

It is still being recommended that all the tests are done x1 & x5 @ 0 & 180, but, there is only space on the form in the book to record the max time.

 

[Sharpend]

So does this mean we’ll have to buy a new 18th edition multifunction tester as our current ones have too many tests on them and it might confuse the novice??  :well-confused:

 

[roys]

The manufacturers of RCD's considered it necessary to reduce the number of what they consider to be false warranty claims due to devices that failed field testing under BS 7671.

So it was not, change for change's sake.

It was change to increase manufacturers profits.

There was nothing that the rest of JPEL/64 could or would do to go against the majority.

If greater evidence could have been gathered this would not have happened and the full suite of testing and recording of information would still be required.

However, a few people tried to gather evidence, and the response from the trade was somewhat pathetic, so the manufacturers won and got it reduced.

Their next aim is to remove it all together and rely on the device test button.

Well learn something every day, I didn’t realise that was the reason for change, thanks for posting that info Sidey.

 

[kerching]

It has been muted that for Amd 1 (and I have this from the man at the top)

if there is evidence of an old RCD box in the dustbin OR there is an 'A' in the day then NO testing is needed for anything....just drive past slowly and note down the house number

 

[Murdoch]

I don't see why you can't make a warranty claim if the device fails under 7671 testing, I would.

Push button test to 6 monthly, not quite testing but limitation of leakage current in designs to 30% of the RCD residual current rating, EFLI testing downstream of an RCD not essential if R1+R2 have been proven, for additional protection it is no longer mandated to do the x1 tests, apparently if the x5 test is fine then additional protection is ok.

It is still being recommended that all the tests are done x1 & x5 @ 0 & 180, but, there is only space on the form in the book to record the max time.

I can't remember the last time I had to replace a RCD under warranty ..................... maybe the manufacturers should make more reliable products!

 

[Sidewinder]

I can't remember the last time I had to replace a RCD under warranty ..................... maybe the manufacturers should make more reliable products!

Maybe not so much warranty as failure on initial commissioning, likely due to incompetent testing.

 

[Tony S]

So does this mean we’ll have to buy a new 18th edition multifunction tester as our current ones have too many tests on them and it might confuse the novice??  :well-confused:

So, the America syndrome again.

Turn on, press the button, if it doesn’t go bang run away. 

I thought we had more pride than this!

IEC RCD’s are in the main mechanical, NEC EFCI’s are electronic and admittedly 5x more sensitive. You’ll need a new MFT.

So Sharpie, dig deep, grin and bare the pain in you’re wallet 

 

[Sharpend]

I wouldn’t mind but I’ve only just updated my last fluke! 

 

[Sidewinder]

To be honest most of the RCD's in production now are electronic.

Your meters will be fine as long as they can do the AC, A, B & F types.

 

[Sharpend]

Thanks Sidey great information as ever. 

Appreciate your input.