UK Change RCD type at CU to enable RCDs at socket outlets

[Chris-]

Hello All,
I am a homeowner (but not an electrician) and I am renovating a house that has an old CU containing MCBs and an MK LN5725 residual current-operated circuit breaker on the supply side (25A, 30mA trip current). The RCD is there because some of the circuits supply power to outbuildings.
In a few months’ time the old CU will be taken out and wiring upgraded to current wiring regulations.
In the meantime I would like to change some of the 13A sockets to a version with integral fast-tripping RCD so that an earth leakage fault would cut off an individual appliance rather than a whole circuit or the entire CU. Greenbrook’s Powerbreaker K22-WP10 looks like it might suit (RCD PowerBreaker Twin Switch Socket White 10mA Passive - https://www.greenbrook.co.uk/k22-wp10) but it is type A and should not be installed downstream of the MK which I am told is type AC.
Is there a good reason why I should not ask an electrician to change the MK type AC circuit breaker to a 30mA type A device and then fit some 13A sockets with integral 10mA RCD? Is there a better way achieve the objective, prior to replacing the CU?

 

[loz2754]

No not really. If you have 2 RCDs in series, both rated at 30mA, then there's no way to achieve the selectivity required. In other words, a fault at a socket is likely to trip the socket RCD as well as the one at the origin. Or it may trip the one at the origin only.
The same is true even if the socket RCD is rated as low as 10mA. This is because most earth faults will exceed the current required to trip an RCD by many times.

 

[pc1966]

Putting in RCD sockets is a reasonable approach if there is no RCD protection at all.

As above, if you have a normal (non-delay) RCD upstream of it then you won't get usable selectivity on any high-ish fault currents (anything above the highest of the two's trip threshold) as both will likely start tripping before the other has disconnected the load. This is not ideal, but not specifically unsafe either. In most cases just annoying that you need to find and reset the tripped device(s).

If your existing RCD works at all (I presume you have tried its test button?) then you might as well wait until the new CU is put in. If it is not working that is a different matter, more so if it feeds sockets or lights outside where contact with the Earth and a fault is relatively easy, and damp/wet conditions allow higher shock currents.

 

[Chris-]

Hi, thank you both for your comments.

So if I understand it properly, there would be no actual harm in putting a 10mA type A RCD in series with a 30mA one, but there would be little benefit in doing so because most earth fault conditions would trip both of them.

I thought it might work because the Greenbrook 10mA RCD has a “typical” trip response of 20ms and the spec for a current MK Sentry RCD is 300ms for 1xIAn and 40ms for 5xIAn. But the significant point is that an earth leakage current could reach far more than 30mA in less than 20ms, and then both RCDs would trip anyway.

The ‘proper’ solution, then, is not to have a single RCD supplying the whole CU – which is what we’re planning to do when the CU is upgraded.

The old MK does work – I’ve tested the test button and it tripped by itself after heavy rain not so long ago (earth fault found and rectified).

Thanks very much for your advice.

 

[Mike Johnson]

I would be tempted to change the RCD and MCB's too RCBO's they can be then retrofitted into your new board at your convenience later.

 

[newfutile]

those old mk rcds are very fast the figure you have quoted is the maximum allowed not the typical times,I have seen mk ones trip at 9 ms , the only real way to achieve any zoned trips are with rcbos .
you can’t really have an (RCD) Residual Current Device socket for each appliance unless the circuit does not require an (RCD) Residual Current Device ie:run in conduit and protected by only a mcb