EICR - Do I need to replace this fuseboard?

[Adam_151]

Thanks. That's what I thought - and why (by implication) I was asking why both boxes had been completed.

Under BS7671:2018+A1:2020 (as is valid until 27 sept) there are situations when both need to be completed, its not simple as the form makes out of being 1x for above 30mA, and 5x for 30mA and below.

RCDs for additional protection (which will be rated at 30mA and below) need to be tested at 5x
RCDs for fault protection need to be tested at 1x.

In some instances, you might be using the same device to serve both purposes, such as a dual split board on TT (without separate 100S incommer) or a C type rcbo on a TN system where you cant meet the Zs for the overcurrent device (ergh)

(Of course, for RCDs in excess of 30ma, they cant provide additional protection and testing at 5x has never been required)

But now they have changed it again, and we just test them all at 1x under BS7671:2018+A2:2022, I do wish they'd stop changing things!

 

[JohnW2]

Under BS7671:2018+A1:2020 (as is valid until 27 sept) there are situations when both need to be completed, its not simple as the form makes out of being 1x for above 30mA, and 5x for 30mA and below. .... RCDs for additional protection (which will be rated at 30mA and below) need to be tested at 5x
RCDs for fault protection need to be tested at 1x.

That's true.

I must say that I've never really understood the 5x requirement. I would say that one has to assume (because it's true) that, in terms of the current though a human being, 30 mA is enough to be potentially fatal, so it's the x1 test that primarily matters, and I very much doubt that any RCD which 'passed' the test at 1x would not also 'pass' at 5x. In fact, although my experience is limited, I'm not sure that I can personally recall an RCD which 'passed' tests at 1x which did not have a disconnection time (at 1x) which would not have satisfied the (≤40 ms) requirement at 5x - and it seems inconceivable that any RCD would have a longer disconnection time at 5x than at 1x! Maybe I'm missing something, but that's why I've always felt that the x5 tests are probably pretty redundant/irrelevant.

In some instances, you might be using the same device to serve both purposes, such as a dual split board on TT (without separate 100S incommer) or a C type rcbo on a TN system where you cant meet the Zs for the overcurrent device (ergh) .... (Of course, for RCDs in excess of 30ma, they cant provide additional protection and testing at 5x has never been required)

All true.

But now they have changed it again, and we just test them all at 1x under BS7671:2018+A2:2022, I do wish they'd stop changing things!

Such changes are, indeed, annoying but, as you will realise from what I've written above, my personal view (for what it's worth) is tht they have now changed it to something more 'sensible'!

In any event, I would say that, in terms of protraction against electric shock, the distinction between 'additional' and 'fault' protection (provided by RCDs) is really fairly artificial. As you may aware, over the years I have repeatedly tried hard (here and many other places) to canvas reports of occasions of which people have experienced (and survived!) electric shocks which caused an RCD to trip, and have only ever managed to identify a couple of such reports.

That being the case, I would say that the main value of RCDs in providing protection against electric shock derive from what you would presumably call 'fault' protection - i.e. clearing a fault before anyone has received a shock. In that regard, they are far more 'sensitive' than OPD-mediated ADS, since they will operate in response to L-CPC faults of far more than 'negligible impedance' (but still capable of resulting in potentially lethal shocks) - whereas (particularly if Zs is close to 'the limit') OPD-mediated ADS will only give the required disconnection times if the fault is close to being of 'negligible impedance'.

As a final 'in passing' comment, I wonder whether it is ever necessary (in a domestic situation) to have a 300 mA up-front RCD, since I would have thought that 100 mA would always be adequate (and 'better'). In my TT installation, although all final circuits are protected by 30mA RCDs/RCBOs, I have up-front time-delayed RCDs to protect long distribution circuits - but they are 100 mA ones.

Kind Regards, John

 

[flameport]

I wonder whether it is ever necessary (in a domestic situation) to have a 300 mA up-front RCD,

In most cases no, but it's certainly not impossible.
TT installation, submain to a distant consumer unit containing large numbers of individual circuits on RCBOs, and high leakage current on some of those circuits would be one example where it's necessary.
A 100mA RCD can only have 30mA of leakage current in normal use, once over that for even some of the time, it's a 300mA RCD.

The OPs installation claims to have a 300mA non-delayed RCD, which is exceptionally implausible.

 

[EFLImpudence]

TT installation, submain to a distant consumer unit containing large numbers of individual circuits on RCBOs, and high leakage current on some of those circuits would be one example where it's necessary.
A 100mA RCD can only have 30mA of leakage current in normal use, once over that for even some of the time, it's a 300mA RCD.

The OPs installation claims to have a 300mA non-delayed RCD, which is exceptionally implausible.

TN-C-S, cut-out other side of wall, until 2015 in a house near you:

 

[JohnW2]

In most cases no, but it's certainly not impossible. ... TT installation, submain to a distant consumer unit containing large numbers of individual circuits on RCBOs, and high leakage current on some of those circuits would be one example where it's necessary.

It goes without saying that very little is 'impossible'. In my TT installation, all but one of my CUs are on 'distant submains', but each CU only has a modest number of (RCD/RCBO-protected) circuits, probably none with "high leakage currents".

A 100mA RCD can only have 30mA of leakage current in normal use, once over that for even some of the time, it's a 300mA RCD.

As I keep saying, that 'rule of thumb' in relation to "leakage current in normal use" is pretty meaningless to a designer in the case of a sockets circuit, since he/she cannot do anything but guess what loads may be connected to the circuit, and what leakage currents that might result in. In any event, it's not an (electrical) safety issue - since, even if there is a high background leakage, the worst that happens is 'nuisance trips'.

The OPs installation claims to have a 300mA non-delayed RCD, which is exceptionally implausible.

I think that we are agreed about that. Again, it's not impossible, but I'm certainly rather doubtful.

You didn't comment on what I said about the (now 'gone') requirement for testing RCDs providing 'additional protection' at IΔn x 5. Do you think there was ever any 'need fir'/'/point in' that?

Kind Regards, John

 

[flameport]

The 5x is useful, as there are devices which take substantially longer than 40ms to trip at 1x. Older Hager RCDs, and the Memshield 2 RCBOs being likely culprits, both of which can have 1x trip times into the 100+ms range, but still trip under 40ms for 5x.

 

[JohnW2]

The 5x is useful, as there are devices which take substantially longer than 40ms to trip at 1x. Older Hager RCDs, and the Memshield 2 RCBOs being likely culprits, both of which can have 1x trip times into the 100+ms range, but still trip under 40ms for 5x.

Fair enough. As I said, in my (limited) experience, I don't recall any which tripped at 1x but took more than 40 ms to do so.

Mind you, if it were me who was receiving the 'additional protection' of an RCD, I don't think I would want it be one of the ones you mention which might take a lot more than 40 ms to trip at 30 mA

Kind Regards, John

 

[Adam_151]

The 5x is useful, as there are devices which take substantially longer than 40ms to trip at 1x. Older Hager RCDs, and the Memshield 2 RCBOs being likely culprits, both of which can have 1x trip times into the 100+ms range, but still trip under 40ms for 5x.

Interestingly enough, you'll probably find those Memshield 2 pods are type A RCDs, and this was back in the late 90's when very few were aware of the differnt types of RCDs.

I'm supprised that 2x RCD testing isn't a thing for time delay RCDs for TT fault protection, the table in BS61008 assures us that with 2x rated current, tripping time on a TD device will be 60-200ms, so it'll meet the 0.2 disconnection time, this table was in BS7671 but I beleive the latest admendment has deleted it. There was a slight rumour around when the 17th came out IIRC, but of course nothing came of it

 

[JohnW2]

I'm supprised that 2x RCD testing isn't a thing for time delay RCDs for TT fault protection, the table in BS61008 assures us that with 2x rated current, tripping time on a TD device will be 60-200ms, so it'll meet the 0.2 disconnection time, this table was in BS7671 but I beleive the latest admendment has deleted it.

Table 3A in BS7671:2018 did, indeed, indicate minimum and maximum trip times at 2x for time-delayed RCDs of 60 and 200 ms respectively - but as you say, that entire Table has been totally removed from Amendment 2, and I can find no reference to required RCD trip times anywhere else in Amd3. If people are going to continue to measure trip times, I wonder how they are now going to determine whether a device has passed or failed?

To require an RCD being used only for fault protection to be tested only at 1x IΔn does seem rather silly - since, as with ADS mediated via OPDs, the 'fault' is presumably deemed to be one of 'negligible impedance', in which case the residual current will probably be orders of magnitude greater than IΔn !

Kind Regards, John