- Joined
- 11 Jan 2004
- Messages
- 43,322
- Reaction score
- 2,758
- Country
What he said!!
Consumer unit replacement cost
- Thread starter pdhan
- Start date
-
- Tags
- consumer unit estimates quote
- Status
I've offered my suggestion of at least a partial answer!
I wonder to what extent this may be an 'age thing'?
In my younger years, my approach was very much to feel that anything 'new fangled' that had become possible as a result of technological advances had to be 'a good idea' and that I therefore 'must have it'.
I therefore went through a phase of having dimmers for many/most lights, PIRs and/or other sensors operating the lights in bathrooms and other rooms, all sorts of 'clever' variants on what one could achieve by way of (what would today be called) 'home automation back in the 60s/70s/80s etc. etc. However, a few decades on, much/most of that has been removed and I am back to only doing/having things if/when I perceive a need/advantage, not simply because they are (technologically) 'possible'!
Kind Regards, John
@securespark Did you perhaps miss this? If you can answer, I can then try to answer the question that you were ''pressing" me to answer ....
Kind Regards, John
Kind Regards, John
- Joined
- 11 Jan 2004
- Messages
- 43,322
- Reaction score
- 2,758
- Country
IME, the failures are either a no-trip situation, or a trip outside the required time-frame, be that too late or too early.
As I previously explained, if I experienced the first, I would operate the mechanism several times and invariably that would free the mechanism and a retest would be within spec. Ditto for late-tripping RCDs.
The half times test was the one I was thinking about when writing earlier.
As I previously explained, if I experienced the first, I would operate the mechanism several times and invariably that would free the mechanism and a retest would be within spec. Ditto for late-tripping RCDs.
The half times test was the one I was thinking about when writing earlier.
Do you mean that? I am not aware of any minimum trip time for an RCD (hence do not know what 'too early' means). Can you perhaps clarify?
Kind Regards, John
- Joined
- 11 Jan 2004
- Messages
- 43,322
- Reaction score
- 2,758
- Country
Sorry, don't know what I was thinking...
I meant the half times test. Because I had "times" in my head, maybe.
I meant the half times test. Because I had "times" in my head, maybe.
Fair enough, but what exactly is it you are now describing as the "half times test"?
Edit: per exchanges with Rocky below, I think I've now worked out what you meant - see my subsequent response to your earlier post!
Kind Regards, John
Last edited:
I would quess he means the 1/2 current test 15ma where it should not trip at all, sometimes in life it does trip when tested so i quess thats what he means by too "early"Fair enough, but what exactly is it you are now describing as the "half times test"?
Kind Regards, John
The testers we use at work sometimes trip Rcds purely when you touch the probes onto the Rcd, dont know why
That was indeed the nearest I could think of as to what he might have meant but, if that were what he intended, it would presumably mean that, by now talking of "half times test", he was again making the same mistake of which he said (in the same post!):
Edit: Ah, maybe I've 'got it', and this is perhaps what you had already realised
Maybe when he said "half times", he wasn't talking about 'time' (as in milliseconds) at all - but, rather about "half times IΔn". However, his then talking about "too early" tends to push my mind back into milliseconds, not mA, mode!I have to say that I would be very suspicious and nervous about any item of test equipment which behaved like that!
Kind Regards, John
Last edited:
Per above-discussion confusion (mainly on my part), I think I've now worked out what you meant by the "half times test" (and "tripping too early") - namely the situation in which the RCD tripped at or below 0.5 x IΔn (i.e. ≤15 mA for a 30 mA device) - is that correct?
If so, then, yes, if that happens when tested with the RCD 'in situ', then it's most likely due to L-E leakage due to connected loads, so the obvious thing to do is to re-test with no loads connected. If the test result is then OK (no longer trips at 15mA, but does at 30 mA), and if the speed of disconnection is also satisfactory, then I agree that can be regarded as a 'pass'.
As I thought you realised, I was talking about the situation in which, when tested 'in situ', it either failed to trip at 30 mA or did trip, but not quickly enough. That is the situation in which I would I would not be happy to say that the RCD had 'passed', even if it tested satisfactorily with loads disconnected (i.e. 'in isolation'). Do you perhaps agree with that?
OK, but (if you were successful) you would then end up with a device which tested satisfactorily 'in situ' (with loads connected), wouldn't you? If so, then I suppose that most people (probably including myself) would call that a 'pass'.
Even though I probably would call that a 'pass', my only hesitation is that I would perhaps be a little nervous that if the device had failed to work satisfactorily due to 'stiction' that might happen again, perhaps even soon after I 'walked out of the door' - so, if it were my installation, I would certainly feel inclined to test it fairly frequently after that.
That may sound ultra-cautious, but it brings to mind a fairly analogous experience of a friend of mine a few years ago. He suspected that the brakes on one of the wheels of his car was not working satisfactorily, so he took it to be checked, and testing confirmed that the brake on one wheel was not working at all. When he returned to collect the repaired car, he was not charged a lot and was told that the only problem was that the piston of the caliper had 'got stuck', and that as soon as they 'freed' it all worked normally. Not too long after that (days or weeks) the piston seemingly 'stuck' again, since when he tried to use the brakes 'in anger' he skidded all over the place, fortunately not resulting any harm. Examination of the vehicle confirmed that the piston had indeed 'stuck' again, so the whole caliper was replaced (as it probably should have been in the first place). There endeth the lesson
Returning to the general 'test in isolation' question, given that the simplest/easiest thing is to plug a tester into a socket and press the button (i.e. 'test in situ') doesn't it make sense to do that first and then be satisfied that it is a 'pass' if (as nearly always be the case) the RCD passed all the tests in that 'in situ state' (and only think about 'testing in isolation' if it fails the in-situ testing)? - or can you think of any (credible!) situations in which it would 'fail' when tested in isolation even though it had 'passed' when tested 'in situ'?
Kind Regards, John
OP here – this sounds like what happened when my electrician tested the RCD and gave it a "fail". Fairly sure he never tested it in isolation. I'd like a second opinion ideally but that would mean an extra invoice and the recommendation may still be the same: change the board. Any thoughts?
@JohnW2
I am Sure its a Kewtech KT63, its rare i use the Rcd test as i do mainly maintenance rather than install, but its calibrated yearly, last time it happened i recall was a large 36way board and similar to Securespark it passed the first tests but failed the 5 times test twice when tested at the socket, which as we do commercial stuff, was aprox 50 metres from the board.
due to this like securespark i opted to go test at the Rcbo, as usual like most pubs, the Db was a mess and jammed pack, it was unsafe to disconnect the load due to the wires buried deep at the rear.
Clipped on the Earth , prodded the N and the second the L was prodded. before the Button was operated, the Rcbo tripped.
I tried the tester on 3 other Rcbos in the same board and the same happened.
Baffles me why it does it, I wondered if it could be a voltage N to E causing it, so next time if it happens i may measure it.
At least 1 other engineer with similar tester has experienced it.
Maybe its wrong but like Securespark, due to the type of boards we work on, for safety reasons i often test at the socket first and if that fails, then do the Rcbo in isolation and if that passes then fine, in our case changing an rcbo can be troublesome and we should not do it on a Live Db, so can get quite involved with access etc, therefore ,
if it finally passes in isolation, im happy with that and if it fails when the yearly test is done a shutdown can be arranged to cover numerous issues
I am Sure its a Kewtech KT63, its rare i use the Rcd test as i do mainly maintenance rather than install, but its calibrated yearly, last time it happened i recall was a large 36way board and similar to Securespark it passed the first tests but failed the 5 times test twice when tested at the socket, which as we do commercial stuff, was aprox 50 metres from the board.
due to this like securespark i opted to go test at the Rcbo, as usual like most pubs, the Db was a mess and jammed pack, it was unsafe to disconnect the load due to the wires buried deep at the rear.
Clipped on the Earth , prodded the N and the second the L was prodded. before the Button was operated, the Rcbo tripped.
I tried the tester on 3 other Rcbos in the same board and the same happened.
Baffles me why it does it, I wondered if it could be a voltage N to E causing it, so next time if it happens i may measure it.
At least 1 other engineer with similar tester has experienced it.
Maybe its wrong but like Securespark, due to the type of boards we work on, for safety reasons i often test at the socket first and if that fails, then do the Rcbo in isolation and if that passes then fine, in our case changing an rcbo can be troublesome and we should not do it on a Live Db, so can get quite involved with access etc, therefore ,
if it finally passes in isolation, im happy with that and if it fails when the yearly test is done a shutdown can be arranged to cover numerous issues
I'm not doubting what you told us but am surprised (and a little concerned) and, like you, have to wonder about the explanation.
I had always assumed that the leads of a tester were (on 'ranges' like RCD testing, loop testing, IR etc.) essentially 'isolated' from anything unless/until one pressed the tester's 'test' button. However, from what you say this is not the case and that, even without the 'test button' pressed, there is a relatively low impedance between the L and E leads (or maybe an even lower one between L and N). This presumably means that one has to be careful (and should be warned to be careful) what one does with the test leads (i.e. allows them to touch) even before one presses the test button.
From what you say, I'm not sure whether this is a transient or permanent problem. If the latter, then it would presumably be impossible to do any RCD testing, since the RCD would trip every time one connected the leads. However, if it's 'transient' (only happens when one first connects the leads, but with the RCD 're-settable'), that's would be even more difficult to understand/explain!
As I've said/implied, for what it's worth that is what I would do, BUT ....
As I've said, my personally view is that this is crucially dependent upon the the nature of the 'fail' when tested at a socket. As I've said, if it 'fails' only because it trips with a test current less than 0.5 x IΔn (tripping time being OK), but 'passes' when the device is tested 'in isolation', then the 'on-load fail' is almost always going to be due to L-E 'leaks' due to connected loads - so, as you say, in most senses that is 'fine'. Having said that, the owner of the installation should presumably be 'warned' (maybe with some recommendations about loads, and distribution thereof), particularly if (when tested at a socket) it trips at well below 0.5 x IΔn - since there will then be a high risk of 'nuisance trips'.
However, IF the 'fail' when tested at a socket is because the RCD does not trip at IΔn, and/or if the trip time is too long, then I personally feel that it would be wrong, inappropriate and potentially dangerous to 'pass' it (say is was 'fine') even if it 'tested OK' in isolation. Do you disagree with that? Would you really be happy leaving an installation (and describing it as a 'pass') if the evidence you had (from your test) suggested that, in normal service, the RCD might well not trip if a leakage current (maybe through a person!) >30 mA were to arise?
In passing, I'm a bit confused by ...
Isn't it extremely unlikely that 'testing at the RCBO' will produce any different results from 'testing at the socket', if all the same loads remain connected to the circuit? Am I missing something?
Kind Regards, John
- Joined
- 11 Jan 2004
- Messages
- 43,322
- Reaction score
- 2,758
- Country
Not if it was tested in isolation....
I would not of thought the leads are totally isolated as from memory isn't there circuitry that indicates if the leads are connected wrong ie wrong polarity, or no earth prior to pushing the GO button.
As for the last bit, sorry to confuse you, a few times it has failed only one of the 5 x times tests, but then passed when tested at the Rcbo mainly with it "disconnected", which I tend to do if safe enough.
,
With a load still connected, which is rarer, I too cannot explain why it could be any different, but i am sure sometimes it has been, maybe its because the Rcd has become saturated through the previous tests, maybe due to the type of loads connected further down the line, often Electronics or Refridge
As for the last bit, sorry to confuse you, a few times it has failed only one of the 5 x times tests, but then passed when tested at the Rcbo mainly with it "disconnected", which I tend to do if safe enough.
,
With a load still connected, which is rarer, I too cannot explain why it could be any different, but i am sure sometimes it has been, maybe its because the Rcd has become saturated through the previous tests, maybe due to the type of loads connected further down the line, often Electronics or Refridge
- Status