BS1363 does not require double adapters to have a fuse. So with two unfused double adapters in a double socket you could potentially have four plugs with fuses adding up to a total rating of 56A.
Give or take your poor arithmetic (13 x 4 = 52), I agree - but that's a separate issue from the one I was talking about.
But more generally, outside of the specific exceptions for rings, is it acceptable to rely on overcurrent protection that is not part of the fixed installation to protect the fixed wiring?
Not in general. However, in the specific case of BS1363 sockets, the only way thay can be used is by plugging BS1363 plugs (each with a 13A max fuse) into them - so the downstream fusing is effectively part of the 'fixed installation'. Your observation about adapters still applies but, as I said, that's a different issue, and I agree it's difficult to understand why there is not a requirement for fuses in 2-outlet adaptors, just as there is for >2-outlet ones.
The whole system of BS1363 sockets and the circuits that supply them is ultimately built on the assumption that load will be spread around it's unlikely that two (or more) full-sized loads will be connected and used in the same location. BS1363 only requires double sockets to be tested at a total of 20A and BS7671 only requires cables in rings and spurs to be rated at 20A.
Agreed, although it's probably true that most households don't even have two 13A loads that they could plug into one double socket.
BS7671 certainly allows the cables of rings to have a minimum CCC of 20A, although it is not totally clear that that also applies to unfused spurs from a ring. Even if it does, I presume that the persistence of that in current versions of BS7671 implies that they share the view (per above) that it is very unlikely that loads totalling >20A will be plugged into a double socket ('for appreciable periods of time').
That IMO makes a double socket alone on a 32A circuit questionable design regardless of the cable size. At best it's a breaker that is larger than it needs to be for no good reason. At worst it's encouraging people looking for places they can get more power to overload the socket.
Other than for the 'adapter' issue, I don't see a problem, given that anything else plugged into a double socket will be providing, at worst, 26A of downstream fuse protection - so, provided that the CCC of the cable is ≥26A, there is no electrical problem.
In any event, if we were talking about Method C 2.5mm² T+E, then the difference between 27A and 32A is pretty 'trivial', particularly given the generous safety margins that undoubtedly exist in the CCC figures that we work with - don't forget that a 32A MCB will allow 46.4A to flow for around an hour, and that a hypothetical cable with a CCC of 32A would be deemed to be adequately protected by it.
Kind Regards, John
