How to test RCD when RCD is RCD protected?

[pc1966]

Can someone explain the joke to me.

Don't worry, it is not that funny and probably people have been shot for better jokes in the past.

 

[GBDamo]

The A/C units we got specified 30mA protection, even though fed in SWA, etc. They may not actually need it, but only sane option was to follow the manufacturer's instructions.

Not sure if it stems from EU markets where TT earthing is the norm?

Somewhere on here I quoted a response I got from Daikin who specified type B RCD protection.

There response was something along the lines of "yeah, that's standard wording o. All quotes but we'll accept installation to local code"

Which was a relief as a three phase type B RCD is upwards of £600, and I needed eight of them...

 

[GBDamo]

Can someone explain the joke to me.

It's not so much of a joke as a measure of the knowledge of some trades that dable in electrics.

"One RCD good, Two double plus good" no?

 

[Calebp43]

As above, although you have to ask about the nature of the circuit design and apparent lack of discrimination if this is so. Very few reasons to protect a standard sub-mains feed with RCD, even less if it's not a cascading system of design with variations on trip time.

Been thinking a lot about your comment. Shouldn't the cable that's being fed from the main board to the sub main be RCD protected?

 

[GBDamo]

Been thinking a lot about your comment. Shouldn't the cable that's being fed from the main board to the sub main be RCD protected?

If its necessary, yes. If not, no?

 

[Rockingit]

Been thinking a lot about your comment. Shouldn't the cable that's being fed from the main board to the sub main be RCD protected?

In a commercial setting with cable methods in trunking or similar I can’t instantly think of any (although some may exist) - in domestic there may be a case if you have for example an extension sub board fed with a buried cable. However, you either protect the entirety of the sub board with an upstream RCD (which I’d suggest would be poor design) or you cascade the trip times - an S type (time delay) upstream and standard A types downstream.

 

[pc1966]

In a commercial setting with cable methods in trunking or similar I can’t instantly think of any (although some may exist)

Usual reasons are TT supply or high current sub-mains where you can't achieve disconnection on L-E faults due to Ze and/or affordable end of cable Zs being too high for the OCPD.

But typically in that case it would be a higher current delay RCD to be selective with final circuit protection, and once you have an RCD feed you get in to issues of trouble with N-E faults if not neutral-breaking RCBO in use, etc.

Most likely legitimate reason for two 30mA in series is an RCD socket fed from a typical domestic supply. You could argue such a socket is not necessary, but you might want one just for high risk areas where you don't quite trust one RCD's reliability.

 

[swaRRR]

Look carefully at what each device is.

Is it the type AC RCD protecting the type A breakers?

 

[mainline]

Is it the type AC RCD protecting the type A breakers?

Hint, they have a test button

 

[swaRRR]

Hint, they have a test button

Oh ffs how did i not see those!

 

[mainline]

Oh ffs how did i not see those!

Don't concentrate on the finger or you will miss all that heavenly glory.

 

[GBDamo]

Oh ffs how did i not see those!

Don't worry, I spent about half an hour the other day staring a picture of an FCU with the neon on before I realised the switch was off.

 

[Rockingit]

Usual reasons are TT supply or high current sub-mains where you can't achieve disconnection on L-E faults due to Ze and/or affordable end of cable Zs being too high for the OCPD.

But typically in that case it would be a higher current delay RCD to be selective with final circuit protection, and once you have an RCD feed you get in to issues of trouble with N-E faults if not neutral-breaking RCBO in use, etc.

Most likely legitimate reason for two 30mA in series is an RCD socket fed from a typical domestic supply. You could argue such a socket is not necessary, but you might want one just for high risk areas where you don't quite trust one RCD's reliability.

You'd use a time-delay or adjustable time curve device, though. Having a higher current let-through makes no difference under an earth fault situation - you're in the hands of the gods as to which device will react the fastest. Hence my earlier post above where when we put some maths into the discussion it all becomes obvious - doesn't matter how many RCD's there are in a series circuit, they ALL of them will sense the same fault, so the only way to achieve proper cascaded protection under a fault condition is to graduate the times. Typically on the systems I work with I'll set primary upstream RCD/RCM devices to around 1A, 1S running through 500mA, 1S > 300mA 100mS > 100mA 100mS > standard 30mA RCBO (depends on what the system is, though, some final ccts can be 300mA on their own... )

 

[timhoward]

Been thinking a lot about your comment. Shouldn't the cable that's being fed from the main board to the sub main be RCD protected?

Basically you'd only do it if you had to, as it introduces lots of complexity!
Impact protection (section 522). Check
Fault protection (section 41). Check
Keep life simple and save RCD's for the final circuits.

 

[Rockingit]

As a case in point about RCD’s in general… I’ve literally just clamped ONE of my many CPC’s on the show I’m currently looking after….. 2.5A and as a guess there’s probably about 30A total at the energy train (generator TN-S). Management is crucial!!

 

[Calebp43]

In a commercial setting with cable methods in trunking or similar I can’t instantly think of any (although some may exist) - in domestic there may be a case if you have for example an extension sub board fed with a buried cable. However, you either protect the entirety of the sub board with an upstream RCD (which I’d suggest would be poor design) or you cascade the trip times - an S type (time delay) upstream and standard A types downstream.

Okay I see. Putting a time delay RCD upstream and another RCD in the sub main board just to put in a garage supply with a few sockets and 2 lights seems kinda nutty.
But regs are regs I suppose!
Okay how about this scenario - you put a sub main into your extension to supply a newly built extension that has equipment with a fairly decent load. You supply it with 10mm flat T+E and that cable is running through the attic, contained in trunking up the walls. Does this supply cable need to be RCD protected? Or could you just put it in on an MCB and then put an RCD in the sub main?

 

[timhoward]

Okay how about this scenario - you put a sub main into your extension to supply a newly built extension that has equipment with a fairly decent load. You supply it with 10mm flat T+E and that cable is running through the attic, contained in trunking up the walls. Does this supply cable need to be RCD protected

If the wall parts were > 50mm deep OR the trunking was metal and earthed, and the attic parts were entirely visible and couldn't be construed to be "buried" then I'd say No, otherwise Yes.
The construction of the wall also comes into it, if metal then Yes.

The most likely answer in the real world is that RCD protection would be required for that scenario and hence choosing a different cable for the job might be wise.

 

[Calebp43]

So

If the wall parts were > 50mm deep OR the trunking was metal and earthed, and the attic parts were entirely visible and couldn't be construed to be "buried" then I'd say No, otherwise Yes.
The construction of the wall also comes into it, if metal then Yes.

The most likely answer in the real world is that RCD protection would be required for that scenario and hence choosing a different cable for the job might be wise.

So what cable would you say would be a good idea to stop the need of RCD protection? Maybe SY?
It just feels a bit much to have to put a new time delay RCD into the main consumer unit. Might even have to change the board if there isn't enough ways!

 

[pc1966]

So what cable would you say would be a good idea to stop the need of RCD protection? Maybe SY?

SY (and CY) is not to any BSI standard so its use is frowned upon for fixed wiring. Even for moving machinery there are limited use-cases as often it is not UV tolerant for outside use, etc.

The most obvious choice is SWA as it is protected both mechanically by the armour, and electrically against cable-penetration since the armour is earthed.

The other choice that is often favoured on farms, etc, is Flexishield cable, or similar ones to BS 8436 (UK) or IS 273 (ROI). It has a foil cover that is rated to safely disconnect faults to certain MCB levels so avoids the shock-protection issue. It is relatively easy to work with and fairly tolerant of rodents, in fact they even make a cat urine flavoured version for high risk areas:

Flexishield R How It Works

Flexishield R How It Works

www.flexcables.co.uk

It just feels a bit much to have to put a new time delay RCD into the main consumer unit. Might even have to change the board if there isn't enough ways!

There is also a point where poor existing design in terms of selectivity is just not worth fixing. Here a garage fault will probably trip both RCD which is inconvenient, but maybe tolerable if that is not likely to introduce other risks. An obvious example is a garage workshop being plunged in to darkness with sharp tools, etc, all around. There you might decide to put in emergency lighting as a better solution as even an external power cut becomes less risky.