RCBO test failures. What's going on?

[BlueToBits]

I have been testing an installation with lot of RCBOs recently and found that a significant proportion of them will pass at 30mA fail at 5X30mA.

SO FAR:

• Tests have been repeated several times over a year and the same ones fail at 150ms but pass at 30ms,
• Most other RCBS pass without problem.
• All are tested in situ and may have a connected load.
• Typical results are 19ms @ 30mA and >150ms @ 5X30mA.
• 6 out of 30 failed
• Boards are Crabtree Loadstar circa 1999 4 TPN DBs all serving 13A ring circuits (bars, meeting rooms, offices etc.).
• After removing the RCBS from the board, banging them around a bit and manually operated them 50+ times, they then tested OK (not perfect) when put back in the board.
• The same ones failed again this year's test (annual).
• On the bench, they all pass at 19/19ms or thereabouts.
• They are the active type with the earth tail.
• They are installed on different phases.
• Tests using an Amprobe 0100+ MFT.
• Bench tests (not site tests) are confirmed with a 10 year old Metrohm RCD tester (uncalibrated but give the same test values)

So what's going on?


I'm now thinking that some connected reactive load is sending back EMF current which is confusing the tester, possibly from another circuit in the building, but I have found this problem on circuits in other installations that a purely resistive. Why should other RCBOs pass if this was the case? I have no reason to doubt the tester as faulty. It was re-calibrated between the test periods.

Any ideas please? no matter how obtuse, because I'm now seriously considering zodiacle influences!
Has anyone else had RCBOs fail at 5xI but pass perfectly at I?

 

[DurhamSparky]

i had a 30ma RCD trip at 15ma when connected to a circuit!! but 30ma when disconnected!!

 

[PMS]

try doing the 5x test first before the 1x test

 

[telectrix]

i had a 30ma RCD trip at 15ma when connected to a circuit!! but 30ma when disconnected!!

might well have been some earth leakage on the circuit.

 

[BlueToBits]

Thanks. I find most 30mA RCDs trip at around 20mA. Could there be some leakage on that circuit?

 

[telectrix]

quite possible.

 

[malcolmsanford]

Just a couple of things on your OP you said that loads were connected this will affect the operation of the RCBO though in fairness I would expect it to alter tripping times not to cause a non trip, but each RCBOs do act in different ways.

I noticed you said the DB was 1999 are the RCBOs a retro fit and if they are how old are the actual RCBO. I ask for as you said you take them out knock them about a bit and they work, abiet not all and not perfect as you say. I just wonder if it is a case of them not having a quarterly test?

Funnily enough on a thread you just replied to about RCD tripping I atached a link to an ERA report that outlines a survey about failed RCD's and they found of the ones that failed, the action of doing the quaterly test would have highlighted the problem of failed RCDs and perhaps even having not tested them quarterly may have contributed to the failure of the RCD.

Clutching at straws springs to mind.

 

[telectrix]

t5hen again, how many RCDs are out there that have not been tested since they were instaled. are these potential , undiscovered failures?

 

[malcolmsanford]

t5hen again, how many RCDs are out there that have not been tested since they were instaled. are these potential , undiscovered failures?

Yes mate they are. We go on about manufacters instructions in reg 134.1.1 and RCD quarterly testing must be part of that.

You know yourself Tel we stick up nice notices about this on our board and tell the customer that it must be done and as soon as your out the door and before you drive away it is ignored for the best part.

I always wondered if a gas plumber told his customer, there is a little button on your new boiler, every 3 months you must push it to test it, otherwise it could mean if you don't the safety device might not operate and it could kill you. It will mean your boiler will go off and you may have to reset timers and things but it needs to be done, I think the customer would do it without fail.

 

[BlueToBits]

The RCBOs were new with the boards in 1999 as part of a full building re-fit.
We have conducted a PIR on the installation every year since so we have a lot of historical data to make comparisons with. The first ones started to get slower when they were 4 years old, and the first full failure was 3 years ago. It's has always been the same RCBOs. 2 on critical circuits 4 non-critical (1 unused circuit). I am told that the maintenance electrician tests them quarterly as recommended, however I suspect that this may be limited only to the non-critical circuits. In any event they all got a 'proper' test every year. The boards are all in differnt environments, Some dusty workshops some clean offices and corridors.

What I can't understand is why they should be significantly slower (they all do trip) when tested at 150mA than they do at 30mA