An RCD tripping problem.

Originally:
o Ring-Main of adjacent twin double-sockets
---- ring enters skt #1, leaves to enter adjacent skt #2, ring exits skt #2
---- ring is 2.5T&E thro surface pattress thro skirting board into floor void
o Ring-Main extended by a single-spur of 3 sockets running 6ft, 2.5T&E
---- spur took feed of socket #1 terminals

Now:
o Single-Spur removed entirely
o Old Ashley wobble-switch twin double-sockets replaced - skt #1 Newey, skt #2 MK

Problem:
o Plug any appliance into twin double-socket #2
---- socket & appliance off - fine
---- socket & appliance on -- fine
o Plug any appliance into twin double-socket #1
---- socket & appliance off - fine
---- socket & appliance on -- RCD trips

Remove Ring-Main L (MCB off) & N still connected
o Plug any appliance into twin double-socket #1
---- socket & appliance off - fine
---- socket & appliance on -- fine (no RCD trip)

o No nuisance tripping in 20yrs.
o No MCB tripping.
o Appliances were 2 vacuum cleaners & 1160-epson-printer, lightbulb.

Suspicion:
o Twin pattress boxes are mounted via overly large screws to skirting board
---- risk of screw point-pressure on insulation, but no red-insulation shows white
---- when L-isolated = no-RCD-trip suggests no N-E earth leakage
o MK socket itself faulty, sealed new item from B&Q
---- does not seem possible
o MK L-N polarity reversed
---- possible, would allow skt #2 to function fine re continuity
---- but any appliances on skt #1 on powering up would trip the RCD?

Any other thoughts?

I'm tempted to redo the two pattress boxes with MK PVC surface conduit boxes
as they look visually identical but don't present any nasty brittle edges. I
can also examine the outer grey sheath tails better, although it looks fine.
I do note MK socket terminal clamps don't seem to grip bare ends that well.

Other house details:
o Yes - whole-house-RCD, which is bad practice (no nuisance trip in 20yrs)
---- should be split-load with lighting & freezers off-RCD
---- but 1) some lighting is 2-core PVC T&E - passed insulation, just no E
---- and 2) one freezer is on a ring shared with hand-appliance - and garden ones
---- so we have 4 flush-mounted emergency-lights fitted
o All relevant cables are PVC T&E
o Other cables are butyl rubber T&E and pass tests (build 1952)
---- Butyl done by senior eng of powercompany & very overspec
---- Immersion -- 6mm T&E, just to be safe tails shortened & sealing sleeve
---- 12"-initial-from-CU-to-1st-skt-on-ground-ring -- 4/6mm T&E replaceable anytime
---- 12ft-to-dryer -- 4/6mm T&E, pain to replace, passed, so told to ignore it

Proper solution is RCBOs ($) or split-load twin-RCD?
o Unsure what to do for a lighting circuit which is 1.5mm PVC-T & not T&E
o No metal fittings, disconnect time not a problem - but dumb fingers are
o Or simply just put the lighting on a normal split-load-with-non-RCD?

Only thing changed before/after was that socket to remove spur take-off
from its terminals - at least re targettable same-location & cause-effect.