RCD's Failing

[DefyG]

Investigating for remedial works, 2 identical but separate installations (MK Dual RCD 4+4 consumer units with 60898/B MCB's)
CU 1 RCD1 x1 >300mS, x5 >40mS, Test Button - Doesn't trip.
RCD2 x1 24mS, x5 17mS, Test Button - Trips

CU 2 RCD1 x1 >300mS, x5 >40mS, Test Button - Doesn't trip.
RCD2 x1 23mS, x5 15mS, Test Button - Trips

Identical failings for RCD 1 on both consumer units yet RCD 2 tests fine.
All other tests are as expected with good R1+R2's and Zs's
RCD's tested at output terminals, with final circuits isolated.

Thought it was more than a coincidence that the Both RCD 1's fail in the same way until I tested them myself .... any further thoughts?

 

[UNG]

Have you ramp tested them, had quite a few RCD's that exceeded the 30mA trip current by as much as 20-30%

 

[newfutile]

I would check the RCD is wired correctly , also neutral - earth (cpc) shorts or low impedance faults can blind the RCD

 

[240volts]

I would check the RCD is wired correctly , also neutral - earth (cpc) shorts or low impedance faults can blind the RCD

Can you swap RCD 1 with RCD 2 and check if it is the circuit or the RCD.
I had a MK RCD fail after 3 years and confirmed using that method.
If both RCD's fail on that circuit then there is a fault on the circuit - Take all the neutrals out and test again, put one neutral back in one at a time to find which circuit is at fault - switching the live MCB off does not always work as could be neutral to earth fault.

 

[loz2754]

RCD 1 may in fact be 100mA and would fail a 30mA test.

 

[pc1966]

OP reports they don't trip on the self-test button, so it looks like they are faulty or there is something non-obvious blinding them (DC, etc).

Disconnecting the load completely and re-testing would be the obvious thing, but they might just both be from a dodgy batch.

 

[DefyG]

Double checked (again) and RCD1's on both boards fail RCD2's on both boards pass
As previously all outgoing circuits disconnected.
Internal links (by MK) checked ok.
Came to conclusion just a coincidence, so will replace both with new.

 

[UNG]

Double checked (again) and RCD1's on both boards fail RCD2's on both boards pass
As previously all outgoing circuits disconnected.
Internal links (by MK) checked ok.
Came to conclusion just a coincidence, so will replace both with new.

Have you ramp tested them?

 

[DefyG]

Have you ramp tested them?

No but the 5x at 150mA test is way above the 30% trip value you mention in previous post, and even if it trips at X value, its still not tripping in the required time.

 

[UNG]

No but the 5x at 150mA test is way above the 30% trip value you mention in previous post, and even if it trips at X value, its still not tripping in the required time.

The 30% I was referring to was a ramp test, a 30mA RCD tripping at 36 - 40mA so quite a bit above the specified 30mA and the normally expected tripping current of 24 - 27mA

IMO the ramp test gives a better indication of the problem if the calibration of the RCD is out at the 30mA level then it is probable that the 5x test isn't injecting enough current to meet the required disconnection time

 

[Mike Johnson]

They may not have failed at the same time, first one may have failed just after the last time it was tested the second one possibly recently failed and upon checking, both failed assumption that both failed recently.

 

[DefyG]

The 30% I was referring to was a ramp test, a 30mA RCD tripping at 36 - 40mA so quite a bit above the specified 30mA and the normal tripping current of 24 - 27mA

Agreed, but as said the 5x test for additional protection at 150mA is way above the 30 or even 40mA so would expect it to trip within 40mS which it doesn't.
A ramp test could take it well beyond 150mA to trip but then also a failure. RCD2 does trip in the same board so concluding it to be a unit failure.

 

[UNG]

I have had a number RCD's fail straight out of the box, on one job went back 5 - 6 months after installation and an RCD that had tested ok at installation failed

 

[UNG]

Agreed, but as said the 5x test for additional protection at 150mA is way above the 30 or even 40mA so would expect it to trip within 40mS which it doesn't.
A ramp test could take it well beyond 150mA to trip but then also a failure. RCD2 does trip in the same board so concluding it to be a unit failure.

Had exactly the same problem numerous times on 1x and 5x tests and the ramp test gave me the result I mention, an RCD that is only tripping at 35 - 40 mA would potentially need a 5x current of 175 - 200 mA to possibly meet the disconnection time, you came looking for an answer I have suggested a reason as to why all my be not what you assume it to be. Yes the books tell you how to test RCD's but there are times when you need to go beyond the books. You keep doing it the way you are it's not a problem with me and I'll check the RCD's I install and test with a ramp test for my peace of mind

 

[DefyG]

I'm not sure what a ramp test will prove when a RCD is not tripping during the prescribed time, except maybe it is still a failure if over and above the 30mA (or as much as 30% above that).
My point is it is still not tripping at 150mA, the 5x test so a ramp test would need to go beyond that to see if it trips but if beyond 150mA it is still a failure!
I can see benefits and have used a ramp test when a RCD is tripping prematurely to prove it is tripping before it should do at a much lower threshold.

 

[pc1966]

To me the simple fact the self-test button is not working is a reason to replace it, obviously after disconnecting both L & N load cables and testing again to verify there is not some external (installation) reason for it.