RCD testing anomaly in TT environment - i5 test at 180 trips downstream RCD

[djtip]

Hi All,


I was called in to follow up some nuisance tripping in a country cottage, which was 2 properties knocked into 1.


There are 2 CUs in the property, one at the East end of the property and one at the West end, there are still 2 earth electrodes.
Only one distributor supply now exists and this comes into the East end of the property next to CU-East.
The supply is split via some Henley blocks and one pair of tails feeds CU-East and the other goes into an enclosure
containing a 100A/100mA time delayed RCD from which the other CU-West is fed via a distribution circuit of 10mm T+E.


CU-East is a 10way protected by a 30mA RCD, there has never been any nusiance tripping on this half of the property.


CU-West is a 10way split CU protected by two 63A/30mA RCDs.


Having sussed the cause of the nuisance tripping I was in the process of running through my RCD tests to confirm compliance.


Zs at the origin and at CU-East are both good for a TT installation at 3 ohms.


The RCD test results were as expected except the last one.
In summary: 0.5I >1999 notrip / 1I @ 0degs 7.7msecs / 1I @ 180degs 18.2msecs / 5I @ 0degs 4.4msecs (last 3 all
tripped the local RCD) . . . but I5 @ 180degs 8.4msecs but tripped the 100A/100ma RCD at the supply end of the CU.


This overall result was the same for the other RCD in CU-East. All results as expected except the I5 test @ 180degs
tripping the 100A/100mA supply RCD not the local 63A/30mA RCD.

This happened repeatedly and was testing at the CU with all loads off.


I had already done R1 + R2 + Rn and insulation tests on the 10mm T+E feeding the CU-East and these were as expected.


But just to rule it out completely I re-installed one of the CU-East 63A/30mA RCDs next to the 100A/100mA time delayed RCD
via a short length (250mm) of 2.5mm T+E. I repeated the RCD test sequence and exactly the same occurred.


Can anyone suggest what might be going on here or suggest what steps to take next?


PS tester is Megger 1502/2.

 

[tony mc]

Have you tested the 100mA time delayed RCD.

It sounds like it is not seeing the 200ms time window on the 180 degree waveform IMO or possibly the 30 mA RCD is duff on that waveform!

Remember to test only on half and 1 X when testing the 100mA unit and select the RCD time delay function on your meter!

 

[djtip]

Have you tested the 100mA time delayed RCD.

It sounds like it is not seeing the 200ms time window on the 180 degree waveform IMO or possibly the 30 mA RCD is duff on that waveform!

Remember to test only on half and 1 X when testing the 100mA unit and select the RCD time delay function on your meter!

Thanks Tony,

Will try that when I pass the property on Friday, will update the post over the weekend.

Dave

 

[djtip]

Thanks Tony,

Will try that when I pass the property on Friday, will update the post over the weekend.

Dave

Have just checked the 100A/100mA RCD and it tests fine
at 0.5I no trip on either 0 degs or 180 degs
at 1I and 0 degs it trips in 18.7msecs (after 20 sec Megger test delay)
at 1I and 180 degs it trips in 28.6 msecs (after 20 sec Megger test delay)

So the only issue now is the minor anomaly in the testing of the downstream 30mA RCDs which fail their 5I test at 180 degs but pass it at 0 degs. 'Fail' in this sense is that they cause the S-Type 100mA RCD to trip in around 8 - 9 msecs.

However 5 * 30mA = 150 mA which is greater than 100mA so the S-type 100mA RCD has every right to trip in these conditions, but why only on the 180 deg cycle?

Dave


I guess

 

[tony mc]

The purpose of the S type is to let the downstream RCD to trip out during the time delay window and with 150 mA injected the 30 mA one should trip out first during that time window period !

I would put a main switch in place of the 100mA device for testing purposes ( bypass it on a temp basis ) and fully test the 30mA units!

I know that you have tested the main 100 mA unit but my money will be on it being faulty !

 

[Richard Burns]

Have just checked the 100A/100mA RCD and it tests fine
at 0.5I no trip on either 0 degs or 180 degs
at 1I and 0 degs it trips in 18.7msecs (after 20 sec Megger test delay)
at 1I and 180 degs it trips in 28.6 msecs (after 20 sec Megger test delay)

So the only issue now is the minor anomaly in the testing of the downstream 30mA RCDs which fail their 5I test at 180 degs but pass it at 0 degs. 'Fail' in this sense is that they cause the S-Type 100mA RCD to trip in around 8 - 9 msecs.

However 5 * 30mA = 150 mA which is greater than 100mA so the S-type 100mA RCD has every right to trip in these conditions, but why only on the 180 deg cycle?

Dave

A type S RCD (to BSEN61008) should trip in a period between 150ms to 500ms when a test current equal to the Idn is applied.
If you have the RCD tripping in 18.7 and 26.8ms then it is either not a time delayed RCD or the time delay is not functional.

If it is not a time delayed RCD then it is entirely possible for it to trip before a 30mA RCD if 150mA is applied to the RCDs as there will be no discrimination between the two RCDs, that is why the upstream RCD has to be time delayed to provide the discrimination.

 

[djtip]

Was clearly in 30mA testing mode and looking for sub 40msecs result. Will check that existing is marked up type-S and replace at earliest convenience. Should have spotted that myself - table 3A of the regs clearly shows this. Apologies must have had brain in neutral, which is not good.

 

[tony mc]

If its not a S type how did it not see the 150 mA on the 30mA test on 0 degrees waveform ?

 

[Richard Burns]

If its not a S type how did it not see the 150 mA on the 30mA test on 0 degrees waveform ?

30mA tripped first, pure chance which one will go but 30mA is at 5 x Idn and 100ma if only 1.5Idn.

 

[tony mc]

I was not thinking straight Richard !

 

[djtip]

If its not a S type how did it not see the 150 mA on the 30mA test on 0 degrees waveform ?

. . because the downstream 30mA RCDs tripped beforehand.

Just swapped the 100A RCD for a time delayed version and the installation is now working as designed with the upstream time delayed RCD not tripping at all when the downstream 30mA RCDs are put through the full testing cycle.

The replacement time delay 100A/100mA RCD has now been tested at 0degs with a trip time of 208ms and at 180degs with a trip time of 207msecs, which is within the 130 - 500msec range specified in table 3A.

so thanks for all your help.

 

[jaypp]

Nice to see a thread with a conclusion