RCBO tripping mystery

To reinforce other comments here:

1) Mere excessive length should be no issue unless inductive effects are messing with a hypersensitive RCD.

The neutral lead (and the cream reference earth if your device has one) of an RCBO *should* be cut to an appropriate length (Schneider instructs you to do so, possibly other manufacturers, I haven't read all the mfr installation guides in a while) which, depending on circumstances, may be long enough to reach the furthest part of the DB (allowing the device to be relocated when it is re-tasked in future alterations) but at any rate *should not* have 3 or 5 loops of slack cable-tied and stuffed behind the other wires to make inductor-like interference in the other circuits.  (I've never understood why Schneider make their neutrals big 30-loop pigtails making for a quite impressive air-core inductor coil, and then expect you to straighten it and tell you to cut it to length!)  *BUT* having cut it, you should crimp on a ferrule, like the mfr did, to protect the very fine Extra-Flexible strands of the Class 6 conductor from the force of the terminal screw.  MEM even instruct you to that, but most mfrs don't because (and here's a notable difference in professional practise) in much of Europe it is fully expected (almost compulsory) that a professional sparky will ferrule any stranded conductor before termination (except when using a cage terminal designed for that conductor size, rather than a screw terminal).  Over here, ferrules are still highly uncommon outside of panel-makers using small size wires in larger terminals, which is fine when a professional sparks habitually twists up the strands (and doubles it over in larger terminals) to spread the load caused by a screw terminal.  But when can’t-be-arsed amateurs leave the strands 'parallel' (with only the slight construction lay) and doesn't think to bend over smaller conductors, then the force of the screw terminal will damage 2 or 3 of the precious 7 strands of a normal stranded conductor, and a great many fine extra-flexible strands of a tri-rated cable, which may well break at that time or will likely fracture as soon as it needs to be released for testing, alteration, etc.  The effective reduction in cross-sectional area can often create hazardously hot terminations making insulation creep back and softening the remaining strands.

(Sorry for that off-topic rant, as an inspector I just triggered one of my own hot-buttons…)

2) Yes, you need proper test results from an installation (multi-function) tester to determine whether the RCBO is at fault (if the RCD is over-sensitive or reacting badly to DC current from digital circuitry) or the circuit conductors are leaking to earth (e.g. damp old cables of excessive length).  So if you’re not a sparky and have no proper tester, you need to get in a professional!  (In this instance I assume Mr Sakho is such a person, so my first advice him would be to do a ramp test on each RCBO to ensure it doesn't trip sooner than mfr specs, for instance iirc SquareD once told me their 30mA RCCB/RCBO must trip at >18mA.)

Otherwise it will be pure guesswork, although you can use trial-and-error logical experimentation to determine if one or more current-using devices or switching points are at fault.  For instance, cheap poorly made LED lamps may have controller faults that upset an RCBO, so swapping them out and trying them each in a single fitting using a single light switch may tell you something, or switching on different combinations and counting the number of lamps that cause a problem (combined leakage) or trying to turn on a number simultaneously (LEDs are notorious for their in-rush current) which may not be easy.  The alarm panel causing a trip is more worrying and begs for a competent person to disassemble it to closely inspect all joints and components.  But again, without test results you may be completely wasting your time examining the box when the fault is rodent damage to the circuit cable

To reinforce other comments here:

1) Mere excessive length should be no issue unless inductive effects are messing with a hypersensitive RCD.

The neutral lead (and the cream reference earth if your device has one) of an RCBO *should* be cut to an appropriate length (Schneider instructs you to do so, possibly other manufacturers, I haven't read all the mfr installation guides in a while) which, depending on circumstances, may be long enough to reach the furthest part of the DB (allowing the device to be relocated when it is re-tasked in future alterations) but at any rate *should not* have 3 or 5 loops of slack cable-tied and stuffed behind the other wires to make inductor-like interference in the other circuits.  (I've never understood why Schneider make their neutrals big 30-loop pigtails making for a quite impressive air-core inductor coil, and then expect you to straighten it and tell you to cut it to length!)  *BUT* having cut it, you should crimp on a ferrule, like the mfr did, to protect the very fine Extra-Flexible strands of the Class 6 conductor from the force of the terminal screw.  MEM even instruct you to that, but most mfrs don't because (and here's a notable difference in professional practise) in much of Europe it is fully expected (almost compulsory) that a professional sparky will ferrule any stranded conductor before termination (except when using a cage terminal designed for that conductor size, rather than a screw terminal).  Over here, ferrules are still highly uncommon outside of panel-makers using small size wires in larger terminals, which is fine when a professional sparks habitually twists up the strands (and doubles it over in larger terminals) to spread the load caused by a screw terminal.  But when can’t-be-arsed amateurs leave the strands 'parallel' (with only the slight construction lay) and doesn't think to bend over smaller conductors, then the force of the screw terminal will damage 2 or 3 of the precious 7 strands of a normal stranded conductor, and a great many fine extra-flexible strands of a tri-rated cable, which may well break at that time or will likely fracture as soon as it needs to be released for testing, alteration, etc.  The effective reduction in cross-sectional area can often create hazardously hot terminations making insulation creep back and softening the remaining strands.

(Sorry for that off-topic rant, as an inspector I just triggered one of my own hot-buttons…)

2) Yes, you need proper test results from an installation (multi-function) tester to determine whether the RCBO is at fault (if the RCD is over-sensitive or reacting badly to DC current from digital circuitry) or the circuit conductors are leaking to earth (e.g. damp old cables of excessive length).  So if you’re not a sparky and have no proper tester, you need to get in a professional!  (In this instance I assume Mr Sakho is such a person, so my first advice him would be to do a ramp test on each RCBO to ensure it doesn't trip sooner than mfr specs, for instance iirc SquareD once told me their 30mA RCCB/RCBO must trip at >18mA.)

Otherwise it will be pure guesswork, although you can use trial-and-error logical experimentation to determine if one or more current-using devices or switching points are at fault.  For instance, cheap poorly made LED lamps may have controller faults that upset an RCBO, so swapping them out and trying them each in a single fitting using a single light switch may tell you something, or switching on different combinations and counting the number of lamps that cause a problem (combined leakage) or trying to turn on a number simultaneously (LEDs are notorious for their in-rush current) which may not be easy.  The alarm panel causing a trip is more worrying and begs for a competent person to disassemble it to closely inspect all joints and components.  But again, without test results you may be completely wasting your time examining the box when the fault is rodent damage to the circuit cable