to Type A, as backfed DC current from appliances could saturate the RCD coil,

Not quite. A-types deal with pulsed d.c. - i.e. a.c. that's passed through a diode or similar before the fault (for single phase systems at least). A types are only guaranteed to deal with d.c. up to 6mA.  Real d.c. currents (e.g. from a battery or other d.c. power source) would require a type B if it's foreseeable it could escape onto the a.c. side. (Some 3-phase rectifier setups can produce something very much like d.c. with a bit of ripple on it too, so might require a B-type, but less of a worry for single phase applications (humm - thinks to self, what about single phase bridge rectifier with a decent smoothing capacitor?))

Now note, they recommend the Type B for PV, so that is another I will have to change.

Depends on the construction of the inverter - traditionally they had a transformer on the output that would reliably block any d.c. from getting onto the a.c. side - so A (or possibly even AC) types would be fine. More modern inverters tend to omit the bulky 50Hz transformer (for ongoing efficiency as well as initial cost reasons) but some still have internal means of blocking d.c. getting on to the a.c. circuit - so it depends on what the manufacturer says.

And If I go for a 'B', what does the Type F do differently?

F types are like A types but also deal with high frequencies (as you might get from a fault in a variable frequency drive), B types do everything an F type does, plus handle pure d.c..

Also there are other means of dealing with d.c. residual currents - a lot of EV charge points have residual d.c. detection devices (or are they residual d.c. protective devices?) that'll disconnect the circuit if the d.c. residual current hits 6mA - so allowing A-types to be used even though there's a lot of d.c. about.

   - Andy.

alanblaby said:

Personally, I think this is getting to be a real mess, that's why people are still putting in Type A's, when, if reading into it, a B or F is required, but, actually finding out what is required is so difficult.

Agreed - back to the £1,000 fuse box.

Could you not still fit type As in the fusebox and a free-standing type B or F by the heat pump?

Could you not still fit type As in the fusebox and a free-standing type B or F by the heat pump?

As I understand it, the trouble with B-types is that they tolerate rather than trip on small d.c. residual currents - e.g. a 30mA B type won't necessarily trip until the d.c. residual current reaches 60mA - so a downstream B type won't necessarily "protect" an upstream A type. A downstream RDC-DD (which trips at ≤6mA) might be a better approach.

   - Andy.

It is a mess, the RCD makers have scarecly caught up with type A, and the goal posts are on the move. And as you say, makers whose instructions are universal and 'translated from the foreign' are not up to speed either.

The only solution is to be very aware of the fault conditions that are credible in a given scenario, and those that are not.

An EV charger for example circulates a Dc between what is in effect a pilot signal wire and the CPC, so the length of CPC between car and charger carries a DC.

It is just about possible to imagine that damage to that section of cable may cause a fault current that returns via the supply NE bond through the neutral coil of any RCD between that link and the charge point, perhaps blinding those RCDs - and for this reason the advice is as it is. 

In equipment where there is no such additional "earth plus DC" path - and be careful, as some innocent looking installations with interlinked mains devices may in effect have this,  there is arguably a greatly reduced risk. 

Faults within power supplies that involve shorts between live and neutral or live and earth either via a diode or not, are likely not to persist for long,, as the diode will go bang and ADS will operate as normal. A diode between earth and neutral may do odd things, but at worst it will trigger a type A.

It is the sort of thing where some drawings and worked examples would be helpful.

Mike.

Faults within power supplies that involve shorts between live and neutral or live and earth either via a diode or not, are likely not to persist for long,, as the diode will go bang and ADS will operate as normal.

On TN systems for sure. On TT though where you might have a couple of hundred Ohms in the Earth fault loop, a reliance on the RCD for ADS and many more volts of touch voltage generated per amp of leakage or earth fault current, it might not be quite so reassuring. Hopefully bonding will limit the risks, indoors at least. Given that vast tracts of western and southern Europe are TT as the norm, I can see why a lot of the advise from international manufacturers takes a pessimistic view.

   - Andy.

Its going to be hard for electricians to research potential issues on a day to day basis and many companies look for excuses to always install the cheapest solution, even if it may not be safe . There are soany components going in to equipment to deal with emc etc there are are probably many potential failure points.

Also the only way to get more complex solutions down to a sensible price is to have everyone using g the same solution, which drives volume.

As an industry maybe we should go straight to an f type bi directional or simillar as a standard requirement. I don't know what's involved technically but suspect its much less than a afdd. Afdd prices have dropped quite a lot with modest volumes. 

Given the number of switched mode power supplies in use plus the 6mA limit of type A's I think just taking a big step would make good sense to provide a long term solution.

AJJewsbury said:

so a downstream B type won't necessarily "protect" an upstream A type.

There must still be a weakest link. If the A-type is rendered incapable, surely the B-type would trip.

There must be shed loads of VFDs out there: how are their circuits normally protected?

There must still be a weakest link. If the A-type is rendered incapable, surely the B-type would trip.

They don't seem to co-ordinate that nicely. Eg. the A type could be blinded by anything above 6mA d.c. yet the B type won't necessarily open until 60mA, so there's a bit of a gap in the middle. How likely it is that such a current would arise (either due to cumulative leakage or some fault) and persist is perhaps less clear. I could imagine one of those PE pilot arrangements that use various resistors for signalling (like for EV charge points) or a circuit intended to drive an LED (with say circa 20mA) shorting to earth but how likely is that in practice?

Similarly if you had several such dc-ish loads downstream of an A-type, would 6mA disconnection (by RDC-DD say) on each protect the A-type, or might it still be subject to several multiples of just less than 6mA?

On a tangent, I've read of some nasty happenings in installations near to railways with d.c. traction (the 650V/750V third rail stuff they tend to use still down south) - a simple N-PE fault can result in diverted traction N currents flowing backwards through a.c. installation RCDs - that sort of thing is going to be interesting to guard against.

There must be shed loads of VFDs out there: how are their circuits normally protected?

I suspect on TN with no RCD, in this part of the world at least. In TT land there's probably a risk they're still working on... not least because it's often not clear what the d.c. performance of "S-types" is.

   - Andy.

"I'm going to fit a Heat Pump"

"And, after reading the Hager explanation, I now also need a Type B bi-directional for any PV supplies"

Do either always require RCD protection? Which,  if any, Wiring Regulations in BS7671 say there is a fundamental requirement for these types of equipment to have RCD protection?

For cables supplying the HP at <50mm from the surface.

Agreed, If I could get a SWA through the house, no need for RCD protection, but then do we need to RCD the control wires? There will be at least 3 control wires going through the property, ELV, so presumably no need for RCD protection.