Correct selection of RCD with regards to EICRs

[pc1966]

I don't think there is any equipment in normal use that has "DC leakage" as such. That would imply a semiconductor between L & E which is not going to pass the likes of class-Y safety for a capacitor under impulse over-voltage, etc.

As mentioned above, this is only going to show under double-fault conditions, such as an open CPC on a washing machine with a DC drive and a motor insulation fault to earth.

EV chargers are a bit of an exception, as they have a big DC power source and are often in wet areas with the true Earth present.

PV install possibly as well, though the inverters tend to monitor the panel's insulation resistance and shut down otherwise. Also PV setups should be properly earthed by the fixed wiring, where as cars depend on plug/socket/flexible cables.

 

[Dartlec]

I don't think there is any equipment in normal use that has "DC leakage" as such. That would imply a semiconductor between L & E which is not going to pass the likes of class-Y safety for a capacitor under impulse over-voltage, etc.

As mentioned above, this is only going to show under double-fault conditions, such as an open CPC on a washing machine with a DC drive and a motor insulation fault to earth.

EV chargers are a bit of an exception, as they have a big DC power source and are often in wet areas with the true Earth present.

PV install possibly as well, though the inverters tend to monitor the panel's insulation resistance and shut down otherwise. Also PV setups should be properly earthed by the fixed wiring, where as cars depend on plug/socket/flexible cables.

The electrical design is slightly above my pay grade, but is there a way that washing machine manufacturers could protect their devices from such a fault happening? That would seem to be a better solution given the number of Type AC RCDs that are going to be around for 20 years still.

 

[pc1966]

The CPC is really the protection as such, it would either trip a DC-sensitive RCD, or more likely just take out the OCPD (supply MCB, or plug's fuse in the UK).

It is possible to reduce the risk by having the DC isolated, but that pushes up the cost. Some older members might remember the "live chassis" TVs and radios of years gone by then the metalwork was connected to neutral and the supply half-wave rectified for the HT feed (or not, for series connected value heaters). Thankfully those days are more or less gone, and the resulting need for work areas to have isolated supplies, etc, for safer testing.

But the move to DC operated motors is almost certainly going to be cheaper (and slightly more power-efficient) by rectifying the mains and then chopping the resulting DC directly to feed the motor. In those cases a motor insulation fault will cause a DC component, but very pulsed as 100Hz (most likely), hence the "type A" as being OK.

Things like USB power points theoretically have the same risk, but for a small high-frequency transformer it is cheap to achieve quite high isolation voltages. And they don't work with water normally!

True DC systems like PV panels or EV batteries don't have any ripple, so they can't be detected by a transformer arrangement and that really pushed up the cost for a "type B" RCD detection system.

 

[Dartlec]

The CPC is really the protection as such, it would either trip a DC-sensitive RCD, or more likely just take out the OCPD (supply MCB, or plug's fuse in the UK).

It is possible to reduce the risk by having the DC isolated, but that pushes up the cost. Some older members might remember the "live chassis" TVs and radios of years gone by then the metalwork was connected to neutral and the supply half-wave rectified for the HT feed (or not, for series connected value heaters). Thankfully those days are more or less gone, and the resulting need for work areas to have isolated supplies, etc, for safer testing.

But the move to DC operated motors is almost certainly going to be cheaper (and slightly more power-efficient) by rectifying the mains and then chopping the resulting DC directly to feed the motor. In those cases a motor insulation fault will cause a DC component, but very pulsed as 100Hz (most likely), hence the "type A" as being OK.

Things like USB power points theoretically have the same risk, but for a small high-frequency transformer it is cheap to achieve quite high isolation voltages. And they don't work with water normally!

True DC systems like PV panels or EV batteries don't have any ripple, so they can't be detected by a transformer arrangement and that really pushed up the cost for a "type B" RCD detection system.

Do Type A inherently cost more to make? or is it just an economy of scale thing?

Given that a lot of Consumer Units in new builds etc are not likely to be changed for many years, it would make sense for them to have some marketing pushing replacement Type A RCDs for their existing units - but most manufacturers seem to be very bad at supporting retrofit on anything other than their very current range.

 

[pc1966]

Do Type A inherently cost more to make? or is it just an economy of scale thing?

I can't see them costing more than a few percent.

The issue is with the current balance (differential) transformer saturation on high DC-component currents, that can be largely mitigated by choosing a larger core and/or lower permeability material. You do need a more sensitive detector then, but that is a design cost and not something recurring as high part cost.

So I think it is down to scale and/or off-loading old stock or old designs for UK market. I suspect type AC will dissepear in a couple of years as Type A becomes the norm everywhere (even if only 80% of the markets regulate for it).

 

[Dartlec]

I can't see them costing more than a few percent.

The issue is with the current balance (differential) transformer saturation on high DC-component currents, that can be largely mitigated by choosing a larger core and/or lower permeability material. You do need a more sensitive detector then, but that is a design cost and not something recurring as high part cost.

So I think it is down to scale and/or off-loading old stock or old designs for UK market. I suspect type AC will dissepear in a couple of years as Type A becomes the norm everywhere (even if only 80% of the markets regulate for it).

If there is no case in which an AC would be chosen in preference over an A, then it would make sense for everyone to switch over and AC to stop existing.

Since they've been the standard in Germany for a while I believe you'd think that someone like Schneider might have an advantage - but then perhaps they are using us to clear out their old stock - pretty sure their Screwfix dual RCD boards are still type AC.

But how much DC could a washing machine leak? Type A covers up to 6ma so what if you have 3 washing machines in a large house - when do you need to start splitting them.

Perhaps we should just ban washing machines and go back to tubs and mangles...

 

[pc1966]

I don't know for sure why they make both type AC and type A RCDs, as Hager do the same. Sadly the market will, in the majority of cases, go for the cheapest option even if it is only a couple of percent so I suspect that is the underlying reason. But if the IET regs move to marginalise type AC it might just kill them off.

The thing about the "DC leakage" is it not likely to be cumulative like AC is. With most appliances having some form of noise filter they (along with the cables) always have some AC imbalance, even if only a fraction of a mA.

However, I seriously doubt that any normal bit of machinery will have a DC leak as part of "normal" operations, and if they have an insulation fault then is most cases the CPC diverts it and quickly it descends in to an over-current situation taking out the OCPD anyway.

So really it is a loss of CPC integrity and a DC fault combined that might lead to a person becoming the path for the AC+DC leak and that is when you rely on the additional RCD protection for safety. So for anyone worries I would be a bit less concerned about the type A/AC debate and more strict on verifying the socket used has a good CPC connection and the appliance is properly earthed (as PAT testing ought to verify).

Here the ring final has the advantage of double paths for CPC, but not all kitchens use rings. Also you might just have the CPC fault at the outlet socket used. Socket testers are not good in this respect, they will show an open CPC but only a few of the more expensive types show a high Z one. Also few private households will do PAT testing, though it is something that most electricians could offer.

I guess TT installs might have a bit higher risk as the rod's Ra may not be able to take out the OCPD, but again with proper bonding of utility pipes it is unlikely that a high CPC potential to true Earth would present much of a risk within the house.

 

[richy3333]

All of this has been discussed at length by e5 if anyone wants to look it up. Hager stated in their instagram cast this week that all their split load boards have type A rcds fitted. We’ve been fitting type A RCBO’s for quite a while now. The cost difference to us on Hager or Fusebox is only a couple of pounds per device.

 

[Dartlec]

All of this has been discussed at length by e5 if anyone wants to look it up. Hager stated in their instagram cast this week that all their split load boards have type A rcds fitted. We’ve been fitting type A RCBO’s for quite a while now. The cost difference to us on Hager or Fusebox is only a couple of pounds per device.

Yep, I watch all the E5 stuff and it's very informative. Their EICR coding ones were some of the best 'training' I've seen on the subject.

However, if the issue is as significant as they suggest, it seems more sensible to put pressure on the manufacturers who sell to screwfix etc, to stop pushing out their cheap boards with Type AC in, rather than to expect inspectors to start coding things that the landlord/tenant/homeowner won't understand.

One thing that they (E5) do seem to have been pushing recently is that adding RCDs is not (or should not be) a solution to poor practises elsewhere - within the fixed wiring and within appliances.

A self-testing RCD that doesn't require reliance on the end user to press 6 monthly (or monthly as some manufacturers seem to recommend) would probably add more to safety. I've personally run into a couple of RCDs just in the last month that weren't tripping on test button until 'unstuck' by flicking on and off several times.

 

[GBDamo]

All of this has been discussed at length by e5 if anyone wants to look it up. Hager stated in their instagram cast this week that all their split load boards have type A rcds fitted. We’ve been fitting type A RCBO’s for quite a while now. The cost difference to us on Hager or Fusebox is only a couple of pounds per device.

I was led down this rabbit hole by Sparkyninja but have since watched the E5 stuff, all very informative.

And

 

[richy3333]

@Dartlec we never buy CU’s or DB‘s from the likes of Screwfix. They have very poor choice and sell to a price. The whole point about screwedfix and tooledupstation is selling at a price point first and specification second.

 

[Dartlec]

@Dartlec we never buy CU’s or DB‘s from the likes of Screwfix. They have very poor choice and sell to a price. The whole point about screwedfix and tooledupstation is selling at a price point first and specification second.

Of course - though it's that much harder to explain to some customers why you are quoting more when they can say - well Screwfix sell it so it must be safe...

 

[Lister1987]

I was led down this rabbit hole by Sparkyninja but have since watched the E5 stuff, all very informative.

And

This is why in my recent posts I'm going on about checking for DC leakage just as a part of what you do, whether it's a complete fullscope analyse of every item on every circuit or just part of fault finding

 

[newfutile]

If the inverter has no isolating transformer or other form of simple separation I assume the worst and code a C2 as such ;”Solar power on type AC RCD reg: 712.411.3.2.1.2”