Adiabatic Lecture.

ebee said:

Graham, would be be happy if we thought of it in this way? an RCD on it`s own may be used for protection against earth faults.

However the RCD itself must be protected by a suitable overcurrent protective device.

Net result in actuallity , an RCD must be used in combination with an OPD

(an RCBO might be considered as one way of achieving this)

Well, not quite,  Just to make it very clear, I'd go for:

An RCD on it`s own may be used to achieve disconnection times for ADS to provide protection against electric shock resulting from earth faults.

However the RCD itself must be protected by a suitable overcurrent protective device, as RCDs are not capable of providing protection against overcurrent.

Net result in actuallity , an RCD must be used in combination with an OPD

I think the best thought experiment you could use to help understand why this is necessary is as follows. If the earth fault (in a TN system usually, but as per earlier examples, could be a TT installation also) is at the end of the circuit closest to the CU, the prospective fault current may be high, and the OCPD may have to operate first, to protect the RCD from overcurrent. On a long run, an earth fault at the end of the circuit furthest from the CU, the prospective fault current is at its lowest (ignoring parallel paths), and at this point the RCD is required to provide the disconnection times required by Chapter 41.

Dad my head hurts. Is this guy any good?

Nice video ... I think he has made one fundamental error though - he's only considered faults at the very end of the circuit.

Even with Ze at 0.35Ω a fault nearer the start of the circuit would give rise to a considerably larger current (say around 657A rather than the 151.3A he considered) - so using the same 0.1s disconnection time, the S=√(I²t)/k would yield around 1.8mm² (not so reassurring for a 1.5mm² c.p.c.)

There's also the issue that assuming 0.35Ω for Ze isn't really considering the worst case, neither is the 1.2 factor for the conductor resistance which presumes that the conductors will have been heated to 70ºC prior to the fault - which they may well not have been in practice. On the plus side his assumption of a 0.1s disconnection time is rather unrealistic the other way (even if BS 7671 fails to provide any better information),

There is a tradition in many textbooks of considering only faults at the far end - and that was justifiable when the protective device is a fuse since they tend to have increasing energy let-through with reducing fault currents so far end faults were the worst case for both disconnection times (ADS) and conductor protection. I suspect a lot of people have taken the old explanations and just substituted fuses with MCBs without realizing that MCB charactreristics are quite different.

On the RCD element of an RCBO, we can only assume (for larger earth fault currents) that it'll disconnect within 40ms (as the standard specifies - it may be faster under particular circumstances under tests that we do, but that's not a guarantee that it'll always be that fast under all allowable conditions). Even 40ms might not be quick enough for conductor protection if the fault currents are high - a 1.0mm² would only be good up to about 575A and a 1.5mm² up to 862.5A (presuming k=115).

   - Andy.

That's an interesting picture. The N bar looks to be well insulated from the case, but the L bar looks suspiciously like a re-purposed PE bar so I wonder how well it's insulated from the metallic encosure. (It might be nice to see some shielding on it was well, but probably not a requirement).

   - Andy.

Freshly taken yesterdayAJ. Imagine working on a Saturday at my time of life! Periodic inspection on a church hall. Thankfully nowadays I always have a younger chap with me to do the frustrating stuff like remove covers of this adapted SquareD board. He also removed a Perspex screen that covered the re-purposed earth bar. That could easily have caught out even an old hand like me. 
I need to dig a little deeper to get a grip on the Im and IdeltaM issue. In this case, Im is exceeded by at least twice the stated value on the RCD but I am unsure what that means practically. The DNO fuse is the upstream device which will clear a 2KA fault in way less than 0.1s but quite how that relates to the stated Im value, I am not clear.