PIR

[RMS]

Hi all,

Tested a property today and allot of the zs values exceed the recommended values as per BS7671 calculated at 80%.

The earthing arrangement was TNC-S and the measured ze was 0.36ohms. Which is 0.01 of an ohm above the recommended DNO requirement.

The supply was wired into a 60amp BS1361(fuse) isolator which supplied a 16mm² twin and earth cable that ran about 6metres to a split load consumer unit with 100A main switch for lights and RCD for sockets.

What do people think the best way would be to comply with BS7671?
would it be okay to change the 60amp fused isolator for a 100mA RCD with 63A MCB as running a larger earth to bring down impedance is out of the question. Also, would the RCD need to be time delayed to discriminate for the 30mA?

Any views appreciated.

 

[Adam_151]

The problem with using an RCD to allow higher Z's in TNC-S has the issue that a phase earth fault might collaspe the voltage at the orgin between phase an combined earth/neutral so that the RCD doesn't work

Quick calculation (230-50)/0.36 = 500A You'd need a fault current of 500A to drop enough volts on the supply that it falls below 50v and RCDs fail to opperate, now do you have any protective devices that 500A would not trip (type C over 50A, type D over 20A, [type B over 100A if they exist!])

If you wanted to use a breaker that fell fowl of the above condition then you'd have to make sure it would disconnect the fault its self (ie. adequate Zs)

Another point, when you are using RCD to disconnect earth faults where the Z's are high at the end of the circuit... check L-N loop is ok on disconnect...

Not sure what the 'regs' are about the above, but you have my 'common sense' look at it, what more do you need

And yes it would need to be time delayed if you want it to discriminate, I'd try to use a 80A or 100A RCD you never know when the supply might be upgraded (if and when... the MCB could be changed at relativly low cost)

 

[Spark123]

I am not keen on electronic RCDs for that reason, I think most single module RCBOs fall into that category where you need to ensure the Zs is low enough for the fault current to operate the MCB side of it. If you stick with the normal RCDs without the electronics I don't think the collapsing voltage is too much of an issue.
Where did you measure Ze?

 

[RMS]

Thanks for the replies,

I cant get my head around the point about the voltage collapsing at the origin due to a phase to earth fault. Never heard of this before?

 

[Spark123]

If you have a direct short L-E at the fuse board the voltage between L&E will try to be 0v, the 240v needs to be dissipated along the impedance of the supply cable between the supply transformer and the fault. There will be quite a few amps flowing, see your PFC reading!!!

 

[RMS]

I understand that the PFC is the total amount of current supplied by the supply transformer in the event of a short circuit fault between live conductors or live and earth(short circuit fault) but its just the point that the RCD would fail to work.

In the event of a live short to earth then the two toroidal coils in the rcd would no longer be balanced and cause the device to trip.

Is is that the neutral and earth are combined then the fault current would return via the neutral as live to earth fault in a TNC-S system becomes essentially live to neutral fault and therefore the RCD coils would be balanced failing to trip?

Sorry if i'm coming across a bit slow, its just my experience and understanding is still restricted by my age. Still so much to learn.

 

[Spark123]

Not really, you are mixing the nice standard torroidal RCDs with those horrible electronic ones which RCBOs use (urghh). The electronic RCDs require a voltage to drive the electronic components in order for them to function correctly, which is why for RCBOs (combined electrinic RCD and MCB) to be installed you need to ensure the Zs for the MCB side of them is low enough for in the event of a fault, the MCB to operate on the fault current magnetic setting. In effect, the voltage collapse in a direct short can shut down the RCDs circuitry.
Back onto the good old torroidal RCDs, these do not have this circuitry and do not suffer the same problem, they will still trip on a current imbalance. They do not require a voltage to operate.

 

[jondiy]

Peronally, I reckon that 16mm T&E is to blame for your Zs.

majority of 16mm T&E's have a 6mm CPC , some even have a 4mm CPC.

as Zs= Ze+ R1+R2 , then obviosuly a smaller CPC will have a higher resistance, hence a higher Zs reading.

Why can't you run a 16mm CPC from CU to the MET ??? Basically , I would either:-

1)
Fail the PIR and tell them that the CPC is inadequate for shock constraints, and tell them that due to regs Table 54?????, the size of CPC should be same as the live conductors , and leave it in their hands.

or

2)
Increase that CPC to 16mm or greater - re-test and ensure Zs are satisfied (and all other tests obviously) , and pass PIR as satisfactory.

TOUGH CHOICE Do you want the work or what?? Bear in mind with the all RCD option as you mentioned , nuisance tripping will cause total darkness at night - not a good option.

 

[RMS]

To be honest its the first time i've ever heard of electronic rcd's. It make sense know, i understand the logic that if the voltage collapses in the event of a short circuit then the rcd will fail to operate. Like you say i was getting confused with the toroidal type. Seems a bit strange as this was never discussed when i was at college.

Anyway, thanks for the info Spark123 and Adam

 

[RMS]

jondiy,

Your right about the earth size of 16mm² twin and earth. It was the first thing i noticed and seemed to be more like 4mm².

The best option is definitely to increase the size of the earth conductor to bring the zs values down but it was my boss that insisted that this would not be possible and installing the 100mA rcd for supplementary protection was the only way.

Also the rcd will be sited in the cupboard adjacent the entrance of the property which is an upstairs flat so like you say nuisance tripping is going to be a problem and fail to satisfy BS7671.

Would it be best to fit the rcd in the consumer unit for the flat as the main switch(time delayed) as long as the cable feeding the board from the origin satisfied disconnection times?

 

[Adam_151]

An interesting read, possibly? : http://www.theiet.org/Forums/forum/...5&threadid=8001&highlight_key=y&keyword1=rcbo

Onto the question, yes I would put RCD in flat as long as submain met disconnect times

 

[JohnD]

... they will still trip on a current imbalance. They do not require a voltage to operate.

Didn't understand that. How do you get a current with no voltage?

 

[Spark123]

The balancing toroid is driven by the current, the voltage across the input and output windings is probably only a few uV in normal operation. You are correct in if there was no voltage there would be no current, the current in this case being the full fault current of the given by the external loop impedance and the internal loop impedance up to and including the fault and the voltage Uoc of the supply transformer. I would probably have been better saying toroidal RCDs are not supply voltage dependent
Still prefer them over RCBOs anyhow
Didn't realise the electronic ones need 50V to work

 

[asc]

what part of BS7671 disconnection time is not met....is it 5s or 0.4s?
if 0.4 it's easy enough to fit rcd protection to the specific circuit(s) that require it. 0.36 Ze is not too bad!.......or am I missing sumfink?

 

[Spark123]

Where are you measuring Ze from RMS? Sub mains may have higher Zs than Ze readings.