On 06/08/2013 09:58, Lobster wrote:
Currently being driven nuts by the above... I have a split load CU. When
the RCD trips, it's only that which goes; ie none of the MCBs trips at
the same time which means there's not much for me to go on.

Question 1: if, for diagnostic purposes, I switch off each circuit in
turn by manually flipping its MCB, does that actually isolate the
circuits for the purpose of elimination? Or would I need to open the CU
and physically disconnect all the cables from each circuit in turn,
completely?

MCB should do it but the circuit will no longer be live so you need to
put the freezer on full bore for a while first (which might actually
demonstrate conclusively that it is the root cause).

It is fastest if you drop out half at a time:

11110000
11001100
10101010

It could be ambiguous if no trip occurs. You may have to repeat with the
other handed pattern so that you have three patterns that all cause a
failure. Then depending on which pattern(s) trip out try just the
appropriate failing circuit on its own.

Question 2: what does the fact that no MCBs are tripping; and the trips
occur at apparently random intervals (4-5 times in the last day, with
interval of up to ~10 hours so far, and including an overnight trip; and
no problem with reinstating the RCD immediately) tell me about the
nature of the fault, if anything?

Thanks for any pointers!

My money would probably be on cooker, fridge, freezer or similar.
Transient earth leakage fault as a motor starts or something like that.

--
Regards,
Martin Brown