Definition of high protective conductor currents

A 30mA 50Hz shock (the electricians classic shock) is not the same, physiologically, as [say] a 60kHz 30mA 'shock' leakage. The RFI leakage could be (depending on switching frequency design) anywhere between say 8kHz and 200kHz (or higher).

Then there is the "phasing effect" that it is capacitively coupled, and from the leak perspective, it's always in-phase with itself, but 90 degrees out of phase with the switching clock, and has no phase at DC, making discussions hard as each group uses it's own 'reference' concepts, while ignoring those [unappreciated] aspects from the other group.

It's real tricky stuff. (see https://tv.theiet.org/?videoid=1327, and my comment about 2/3rds the way through)

Are you planning to use a separate consumer unit fed independently via Henley blocks ?

I’m curious about your heat pump install setup. Would using SWA cable eliminate the need for RCDs in your case? Also, does your heat pumps have a dedicated earth connection point, similar to solar PV inverters, where you could attach a 10mm copper conductor to handle high protective conductor currents? 

In my setup, it's TN so Zs will be adequate, the supply HP cables won't be concealed in a wall or under a floor, or run underground, won't run through a bathroom, aren't supplying sockets or domestic lighting, and the HP although outside weighing in at over 100kg and being bolted down onto a concrete foundation (via large rubber feet) I would suggest wont count as "mobile". To my reckoning therefore I don't need an RCD (30mA or otherwise) from a BS 7671 perspective, regardless of the type of wiring system. If the cables were concealed or run underground I reckon SWA or similar might be needed.

Manufacturer's instructions are confused - the printed manual that came with the HP only asks for an RCD where the sit requires one, and diagrams show overcurrent protective devices, but no RCDs. Other (later?) on-line documentation, says "all" their HPs need a B-HP RCD, I've got an open enquiry with the manufacturer to try and clarify that point.

In my case I'm opting for 4mm² BS 8436 cable (partly because I rather like shielded cables, mostly because I have some left on a roll from previously) - run in flexible conduit outside to protect the sheath from UV, I've not spotted a dedicated stud (just a PE terminal on the main PCB) - but suspect I will be able to improvise something easily enough. I'm hoping to go down the 543.7.1.203 (ii) route - i.e. single 4mm with the BS 8436 cable providing equivalent protection to flexible conduit,

  - Andy,

Philip Oakley said:

and we need discrimination between real fault leakage and EMC/RFI leakage.

I think there could be some cross-over there. Normally we'd expect a 30mA RCD to provide a decent level of protection against shock in a wide variety of circumstances - including the single fault condition of a broken c.p.c. where "protective conductor current" becomes the same as shock current. If we've now got (non-standard) "30mA" RCDs that actually allow 150mA to flow without disconnecting (if not 50Hz components) it feels to me we might have lost something here.

   - Andy. 

To save others from spending time  looking for it that guidance is on-line here HPA-RCD-Guidance-March-2025.

Mike.

Hi Andy,

I’m curious about your heat pump install setup. Would using SWA cable eliminate the need for RCDs in your case? Also, does your heat pumps have a dedicated earth connection point, similar to solar PV inverters, where you could attach a 10mm copper conductor to handle high protective conductor currents? 

Have you seen the Heat Pump Association updated guidance from March 2025? Lot of useful information on RCDs, although doesn’t mention isolation transformers as an alternative to manage leakage.

Let's get back to the start. The question was about the technical spec of an RCD.

That spec (the special aspects) is not about the shock potential for the human, rather it's about all the side effects of the load and how it interacts with the RCD's detection apparatus, and the coordination between the two.

Things like needing to be a type B, and needing certain rejection characteristics. None of those characteristics, as best as I can see, are anything to do with effects on the human, and far more to do with a coordination between EMC/RFI requirements, filtering capabilities at the load and filtering within the RCD. The EMI/RFI filtering is the one that creates earth leakage that looks like residual current, and we need discrimination between real fault leakage and EMC/RFI leakage.

That's a bit sweeping - why do we worry about it tripping (or not) if the currents under discussion are not potentially a threat - put another way, if electric shock was not a thing, we would not need CPCs, let alone RCDs that tripped on earth currents,  at all. 
Mike.

The requirements are nothing to do with the 'shock', rather it's nuisance tripping and 'blinding' (failure to trip) of the circuit, that is being coordinated.

One of the big problems here is that as 'electricians' we are often unaware of the complexity of other competing regulations, especially the EMC regulations, which also require the measurement of (residual) electrical signals relative to the well connected mass of earth to ensure that the energy we supply doesn't re-emerge as RF radiation, relative to our common ground/earth.

Everybody thinks it's somebody else's problem, anybody could add extra costly components, nobody wanted to do it. 

And the RFI/EMC specs are just as susceptible to fudging as the electrical specs (e.g. see clock dither on microprocessors..) .