314.1 , 314.2. No it doesnt mean each circuit must be controlled by its own RCD, account must be taken of the install and its use, if required, a split board can still be used, even an upfront single RCD could be compliant.

314.1 (i) avoid danger and minimise inconvenience, - having all socket outlets going off due to an aplliance fault may be no more inconvenient than getting up, unplugging the errant device, then resetting the RCD. If, however there was a dialysis machine, or, indeed, an important internet server on those same bank of circuits controlled by one RCD, then it could be rather more of an inconvenience.

It all boils down to the design for the intended User. Offering a Consumer Unit full of RCBOs could also be a bad design,when replacing an old fuse box,  in that it would work, technically, as required, but the cost to supply and fit may be more than the Customer can afford, so they will not take up the offer of fitting it, whereas fitting a 2 RCD consumer unit, at £100 less, could be affordable to them, and wold make their use of the installation far safer, with a little inconvenience if ever there was a fault.

So, basically, good design for the intended use is what is required, and no, a full RCBO consumer unit is not required in every installation.

Start with the definition of circuit in Part 2 of BS 7671.

Circuit. An assembly of electrical equipment supplied from the same origin and protected against overcurrent by the same protective device(s).

We then move on to what the RCD is doing (providing overcurrent protection, fault protection, or additional protection, and discuss which of those is necessary for individual circuits vs "circuits" in common ... and then consider selectivity. Is selectivity necessary for fault protection, overcurrent protection, and additional protection?

If the RCD is only providing additional protection, and it is acceptable in those circumstances for it to disconnect some circuits rather than just a single circuit, is that OK?

So is the busbar and the cable from the neutral bar back to the RCD a circuit? The origin of the socket, cooker, etc, circuits would be the output terminal of the MCB and the neutral bar.

Chris Pearson said:

So is the busbar and the cable from the neutral bar back to the RCD a circuit?

In BS 7671, a circuit has an OCPD by definition. The OCPDs could be at the "remote" end of the circuit ... so the outgoing OCPDs for each "busbar" being a possibility ... if you agree that, in totality, they protect that busbar against overcurrent? If not, the DNO cutout perhaps?

Certainly the devices would have to be in a circuit (distribution circuit or final circuit) somehow, because they are required to be protected against overcurrent (following manufacturer's instructions).

gkenyon said:

The OCPDs could be at the "remote" end of the circuit ... so the outgoing OCPDs for each "busbar" being a possibility ... if you agree that, in totality, they protect that busbar against overcurrent? If not, the DNO cutout perhaps?

I see what you mean. So the circuit in the CU is from the main switch via the busbar to the outgoing terminal of the MCB. We must include the MCB in the circuit as the OCPD.

Back to the OP and Alan's response: my preference is for an all RCBO board (save for any distribution circuits, or dare I say, to a PV array), but I am fortunate inasmuch as cost is not my primary concern.

314.4 sets out the requirement to connect each final circuit to a separate way in a distribution board. A split board provides that.

314.1 indent (i) sets out two of the reasons why installations should be divided into circuits, one of which is to minimize inconvenience in the event of a fault. The notion of "inconvenience" is a subjective one. It could be contended that the only person able to make that call would be the user of the installation. Without the input of the user, a designer could only speculate on what would cause "inconvenience". However, if it is established that inconvenience could be caused by the tripping of more than one circuit because of a fault, then the next step is ensure that the "inconvenience" is "minimized".

To "minimize" something is to reduce it to the lowest possible degree. If the user or designer considers multiple circuit disconnection in the event of fault would be an inconvenience then the designer must "minimize" that possibility.  A split board arrangement would not provide that and would therefore not meet 314.1 indent (i).

There is nothing in 314.1 about cost. It is a "shall" regulation.

Any one wishing to explore the meaning of "inconvenience" might like to ask folk in Manchester (today around 3pm) if it would be "inconvenient" if a fault in their kettle also disconnected their TV!

To "minimize" something is to reduce it to the lowest possible degree.

But not necessarily without regard to all other factors - e.g.. acceptable cost. That's why it's minimised rather than eliminated,

  - Andy.

Any one wishing to explore the meaning of "inconvenience" might like to ask folk in Manchester (today around 3pm) if it would be "inconvenient" if a fault in their kettle also disconnected their TV!

In which case, one final circuit for each socket? .. and only single socket outlets allowed.

Curiously before these modern fads of RCDs and circuit breakers, a 13A plug fuse discriminated very nicely with a ring 30A fuse, which in turn discriminated very nicely with a 60A+ DNO fuse - so selectivity even between appliances on the same circuit was pretty much guaranteed. Nowadays with a B32 RCBO tripping faster than a 13A fuse, it's a case of one out all out. That's progress for you....

   - Andy.

"The designer must "minimize" that ,,, nothing in 314.1 about cost. It is a "shall" regulation."

I take the point, and have RCBOs in a flat that we rent out, but an RCD here where I am near enough to reset it.

However, the counter argument is that one could raise similar assumptions about the  "inconvenience" at any price of not having a private substation, not having multiple incomer feeds, or a back up DRUPS, or even not having a small on-site nuclear facility in the grounds...

  Now at quite different price points, all of these make sense and are appropriate solutions, though the last probably only at govt agency level, as there are problems with the regulations around moving the fuel and the waste, (not the physics mind you) but the domestic installer world does not rewire a former council flat at all beyond a certain price point.

So regardless of the exact form of words in a non-binding British Standard,  there will actually be a limit.  It may still be that for some that limit is indeed  found with one RCD covering the whole installation or large chunks of it.  And there will be a good no, of folk who are not unduly worried about losing the footy who will be doing something else instead. 

Personally I'd be more worried about the loss of tea supply, and to that end I would  deploy the back-up kettle - and in a full street power cut, I have a further level of defence in the form of one that goes on the camping gas stove or in winter on the fire in living room

The political/ economic  meaning of minimize is not the same as the pure maths - what is really minimized is some sort of cost-effort- inconvenience blended quantity, if we like it or not,

Mike

lyledunn said:

To "minimize" something is to reduce it to the lowest possible degree. If the user or designer considers multiple circuit disconnection in the event of fault would be an inconvenience then the designer must "minimize" that possibility. 

If we take this is interpretation to its conclusion, one circuit per appliance or item of current-using equipment is necessary to reduce inconvenience to its lowest possible degree.

I therefore conclude this is interpretation is not what is meant (although only a court in the end could decide that).

lyledunn said:

A split board arrangement would not provide that and would therefore not meet 314.1 indent (i).

I therefore don't agree with this statement but it's subject to a court's decision.

Would custom and practice therefore provide an argument for an interpretation that did permit a split-load board interpretation of 314.1 (i) ?

I would add, though, that custom and practice is changing.

Whilst many existing installations operate satisfactorily with split-load board, it is not likely to be suitable when we start adding EV charging, induction hobs, heat pumps, etc.