B Type RCD's and electronic loads

[cdtsilva]

I have seen recent changes that require (most) EV chargers and solar systems to be fitted with a type B RCCD.

I work with power semiconductors and inverters for a living and have one of these in my workshop with a dedicated circuit.

The question arises on any downstream RCD's that may be affected. We have other appliances at home which in my humble opinion could suffer from these same faults: Inverter heat pump, induction hob, variable frequency motor appliances, etc.
How are these devices being protected by a type A, when they can leek high frequency components in case of a motor insulation failure?

Using RCBO type breakers is also not an option. as type B are rather bulky and require separate MCB protection, which a normal consumer unit would not be able to accomodate.

My last question, again on the topic of inverters is frequency. Type F are labeled up to 1KHz. But most inverter type devices work well above this frequency. I have seen type B+ from ABB which can accomodate up to 20KHz, but don't know if this is standard amongst manufacturers (I can't seem to find details on the upper limit for a normal type B).

As an electrician, what would you sugest to a customer in this situation?

 

[pc1966]

In reality some EV/PV demand a type B DC-sensing RCD. Some have built-in DC sensing/protection and only need a type A RCD for the supply protection, I have even worked on an ABB PV inverter that stated that type AC is adequate as no risk of a DC fault!

How serious the risk of DC induced desensitisation of an RCD is depend on a number of factors, obviously the level and likelihood of a DC fault, as well as how important the RCD is for safety. For example, in many cases in the UK you have a TN supply (low impedance earth) in which case many earth faults (AC side, or DC side short) will cause a lot of current to flow and so will disconnect on the over-current protection in any case.

If, however, you have TT earthing (more common in the EU, but seen a lot in rural cases in the UK) then you are normally dependant on the RCD for fault disconnection, and the reliability of that in the presence of DC is a far more serious factor.

The same goes for HF ripple and type F RCD, again if it is class I (earthed) appliance and on TN earthing most likely it will be disconnected on OCPD under fault conditions.

If you have a dedicated circuit for the inverter you are playing with and it is on an RCBO (i.e. not sharing an RCD with other circuits) and you are on TN earthing (supplier's low impedance connection, not earth rod) then your real risk is touching L & E on that unit and it being under faulted condition so it does not sense a current that is dangerous to you. Hopefully you take a lot more precautions and the RCD really is your last-resort for living!

If you work in an experimental manner with DC capable equipment you probably should have either a type B RCD just for that circuit, or to use an isolation transformer with some means to monitor and disconnect the supply if an earth fault is suspected (as many PV inverters do if the see the PV panels leaking to true Earth).

for some examples of how RCD behave here is one I prepared earlier:

RCD "blinding" by DC | on ElectriciansForums

Find RCD "blinding" by DC Advice and Help. How-to RCD "blinding" by DC in the Electrical Wiring, Theories and Regulations advice boards on ElectriciansForums.net | Free Electrical Advice - Electricians Forums: Electricians' Talk Forums. RCD "blinding" by DC on...

www.electriciansforums.net

 

[pc1966]

Just to add if you do need something special these folks have a great (but not cheap) range:

Doepke UK Ltd - RCDs and equipment for safe EV charging

Doepke UK Ltd provide innovative residual current protection equipment that meets the most demanding modern requirements.

www.doepke.co.uk

Typical electrical supplier offering type B (should also cover type F):

RCDs 2 Pole Type B | CEF

Generally speaking you won't get beyond type A RCBO (with B or C curve) anywhere that are certified to fit UK consumer units (CU). If you need anything different you have to cascade your choice of CU-compatible MCB (B/C/D curve) with a separate DIN enclosure and RCD of the type A/F/B sort to make up the protection you desire.

Dopeke do make RCBO with some of these options, but again you are looking at a stand-alone DIN box for it.

 

[cdtsilva]

Thanks for the repplies.

In the UK, with our single phase low impedance earth things are generally simple, but I also work on some European instalations where the use of a 3 phase supply complicates things a little bit.

Some facilities share a common neutral wire with multiple phases, it is not possible to have independent RCBO's at the consumer unit, therefore one relying in a 4P RCD protection to sum all currents in the system. Here my concern is that any DC or half wave within that network could potentionally blind the RCD serving circuits on the remaining phases, potentialy putting others at risk, or putting myself at risk from others.

I do have an habit, for my own lab setup, to bond earth and neutral before the RCD, which reassembles a TN-C-S system. On a lab setup I do this after an isolation transformer to avoid ground loops, followed by an RCD or RCBO after which PE and N run separate.
On my own lab, I do have this setup right from the consumer unit after the meter and before other RCD's/RCBO's, so that any leackage currents not dealt with by the RCD would hopefully cause the MCB to trip.

As it is not always posible to know the supply side - Is this setup always the safest?

 

[pc1966]

Thanks for the repplies.

In the UK, with our single phase low impedance earth things are generally simple, but I also work on some European instalations where the use of a 3 phase supply complicates things a little bit.

Some facilities share a common neutral wire with multiple phases, it is not possible to have independent RCBO's at the consumer unit, therefore one relying in a 4P RCD protection to sum all currents in the system. Here my concern is that any DC or half wave within that network could potentionally blind the RCD serving circuits on the remaining phases, potentialy putting others at risk, or putting myself at risk from others.

Using a 3-phase RCD to protect separate single phase circuits is not something anyone on here would recommend! While for TT systems it is not uncommon to have an RCD device as incomer (typically 100mA or 300mA delay/selective type), you would still use separate SP RCD/RCBO for "additional" shock protection on final circuits for a couple of reasons, such as:

• With 3P RCD you could have high leakage on 2 of the 3 phases, increasing the trip threshold on the 3rd
• A fault on any of the phases takes out the lot. Fine for 3P loads where you don't want potentially damaging "single phasing" on 2 out of 3, but not so good if you have lights and assorted appliances split over phases

I don't know enough about the specifics of different countries, but if you do find you are dependant on RCD for fault protection (typically TT system) and you have systems with a high chance of smooth DC (EV and PV being the obvious cases) then type B RCD really are your only safe choice.

I do have an habit, for my own lab setup, to bond earth and neutral before the RCD, which reassembles a TN-C-S system. On a lab setup I do this after an isolation transformer to avoid ground loops, followed by an RCD or RCBO after which PE and N run separate.
On my own lab, I do have this setup right from the consumer unit after the meter and before other RCD's/RCBO's, so that any leackage currents not dealt with by the RCD would hopefully cause the MCB to trip.

As it is not always posible to know the supply side - Is this setup always the safest?

If you plan on linking N & E for a test system with low impedance you must isolate it from the main supply as the ESQCR prohibits TN-C within installations, not to mention the overheating risk from high circulating currents (e.g. if you are on TN-S and there is a few volts at low impedance N-E).

Often isolated supplies would be used for labs for safety, as they limit the current that can flow during contact to very low levels. However, unless you monitor for a fault you might have one happen without realising it and then the 2nd fault make it dangerous. Say L-E shorts and nothing happens, but now N is at high voltage and able to kill you if you touch failed insulation, etc. That is why floating 'IT' systems are only for supervised use and usually with some means to detect an insulation fault.

Older approach on ships, etc, was two lamps, one each line to E. Normally they are both dull and each line at ~115V to E, but under fault conditions one goes bright. Here not for shock protection but to keep things going under single-fault situations, as loss of power at sea being a hazard in itself.

here is a more modern example:

Insulation monitoring device (IMD) for IT earthing system in medical environment - 230 V~

Insulation monitoring device (IMD) - for IT earthing system in medical environmentUsed to permanently monitor the insulation level of the IT earthing system, in compliance with installation standard IEC 60364-7-710 (Group 2 hospitals)

www.legrand.com

 

[cdtsilva]

Thanks for the feedback.

For the circulating currents, I assume you mean because of the wire voltage drop from the substation, with the neutral rising a few volts above the real ground and causing some currents...

I guess at stake is how functional RCD's really are with the mixed variety of electronic loads present in the household these days. Protecting individual circuits with a type B RCD is not really feasible and I wonder if a type A, on the presense of a high impedance earth will provide sufficient protection to non linear electronic loads.

It seems the most safe, would then be to convert to a TNCS type supply, monitoring the current flowing to earth?

I remember years ago, there were some relays that measured voltage across the supply neutral and the Earth. They would trip if an excessive voltage developed between them. I can immagine most of these systems are now deprecated, but wondering what regulations say about them? Hypotetically, they could add a second level of protection to an exiting RCD, as (theoretically) the coil would be just as sensitive to AC or DC?

It seems very similar to your description of how it is done on the boats, ecept ther eit is used as a visual indicator of fault. Do they also employ a center tapped earth, instead of referencing one of the conductors?

Another aproach would be a type B with a delay. This would give time for any downstream type A RCBO's to act before shutting the main supply. Is there such a device?

 

[pc1966]

Thanks for the feedback.

For the circulating currents, I assume you mean because of the wire voltage drop from the substation, with the neutral rising a few volts above the real ground and causing some currents...

Exactly.

I guess at stake is how functional RCD's really are with the mixed variety of electronic loads present in the household these days. Protecting individual circuits with a type B RCD is not really feasible and I wonder if a type A, on the presense of a high impedance earth will provide sufficient protection to non linear electronic loads.

In most cases type A is fine. Also you have to remember that practically nothing will deliberately leak DC, that would fail the usual insulation tests, etc. So the RCD risk is during a period of a DC fault and someone can make contact with L & E.

It seems the most safe, would then be to convert to a TNCS type supply, monitoring the current flowing to earth?

That is dependant on that the DNO (supplier) is willing to do. Not all supplies can be safely made TN-C-S as the integrity of the PEN is critical to safety, given any added rods, etc, will not do much to divert open-PEN currents.

I remember years ago, there were some relays that measured voltage across the supply neutral and the Earth. They would trip if an excessive voltage developed between them. I can immagine most of these systems are now deprecated, but wondering what regulations say about them? Hypotetically, they could add a second level of protection to an exiting RCD, as (theoretically) the coil would be just as sensitive to AC or DC?

You are probably thinking of the VOELCB which have been depreciated for decades. Basically the fail if you have a good Earth on the CPC somewhere, and also fail to protect against outdoor shock L - true Earth.

It seems very similar to your description of how it is done on the boats, ecept ther eit is used as a visual indicator of fault. Do they also employ a center tapped earth, instead of referencing one of the conductors?

The supply is floating. Neutral is defined by the fact is it the Earth-referenced live conductor, so the have two line conductors in effect.

Another aproach would be a type B with a delay. This would give time for any downstream type A RCBO's to act before shutting the main supply. Is there such a device?

Yes, Doepek (and a few others like Schneider) make them. But at a price!

 

[pc1966]

Here is one example:

63A 4 Pole 300mA Type B Selective Time Delay DC Sensitive RCD

VISI-SAFE for safe operation and maintenance work

www.cef.co.uk

But at £1,029.74 Including VAT you might want to consider your design carefully...

 

[cdtsilva]

I guess for the VOELCB, the intent is to monitor for an earth or neutral fault and disconect power. This would need to be combined with an RCD to protect people. Probably the reason they are now obsolete. But I think they still serve a safety purpose on specific cases

The problem with ground loops, due to parallel earthed points along the supply (although less severe), could also affect the operation of an RCD on some specific scenarios.

On that subject, it is surprising that most RCD units (at least in Europe) are powered from the supply they are protecting. Effectivelly, in the event of a neutral failure, the RCD becomes useless and the neutral wire can potentionally become live through the resistance of conected loads. I guess this is only exacerbated the smarter the RCD is (electronic circuitry)

Some RCBO's have a separate earth connector, explicitly for this function, but not all.

Guess, at the price of type B RCD's in general - The better bet will be a toroidal transformer and associated inrush control mechanism, as used on hospitals.

Makes me think now... How severe would capacitive coupling be on these? I don't recall seeing the core being grounded on such transformers...

 

[pc1966]

I guess for the VOELCB, the intent is to monitor for an earth or neutral fault and disconect power. This would need to be combined with an RCD to protect people. Probably the reason they are now obsolete. But I think they still serve a safety purpose on specific cases

The closest modern equivalent are the EV chargers that use a rod to detect open-PEN faults on TN-C-S supplies.

On that subject, it is surprising that most RCD units (at least in Europe) are powered from the supply they are protecting. Effectivelly, in the event of a neutral failure, the RCD becomes useless and the neutral wire can potentionally become live through the resistance of conected loads. I guess this is only exacerbated the smarter the RCD is (electronic circuitry)

Most RCD will trip on the current even if supply N is floating. Maybe not quite as fast, and probably not the DC side of type B ones.

Guess, at the price of type B RCD's in general - The better bet will be a toroidal transformer and associated inrush control mechanism, as used on hospitals.

Type B detecting DC is much harder than a simple differential transformer.

Typically the DC side sensing is separate from the type A AC side, and uses either a hall effect sensor or an active flux-gate style of magnetic device that can detect "constant" DC flux from any imbalance.