Why is a too high zs reading on a circuit okay if it’s protected by an Rcd?

[Welsh_sparky_21]

Not long passed my test but was wondering why it’s okay for a circuit to be over the max zs if it’s protected by an RCD?

 

[Lucien Nunes]

The conventional 'max. Zs' is the highest impedance at which the fault current is sufficient for the MCB or fuse to clear an earth fault rapidly. An earth fault on an RCD-protected circuit should also trip the RCD, requiring only 30mA to do so instead of up to 5 times the circuit rating for a B-type MCB. Taking the maximum touch voltage as 50V, an impedance of 50 / 0.03 = 1.67kΩ is low enough to enable the RCD to clear the fault rapidly even though the OCPD cannot. Given the accepted lower reliability of RCDs vs. MCBs, opinion varies about the wisdom of this fall-back position, if the supply characteristics are good enough to allow the conventional maximum Zs to be obtained with a realistic choice of cable CSA.

 

[Sintra]

IMO a circuit should still be designed to be below the maximum permitted Zs for the over current device. However if this can not be achieved the maximum Zs can be increased by using an RCD. The acceptable touch voltage is 50V so to achieve this with a 30mA RCD in the circuit the Zs needs to be below 1667 ohms. The formula used is 50V / 0.03A = 1666.666r rounded up to 1667 ohms. The same formula can be used for a 100mA and 300mA RCD.

 

[andyb]

As above. Simply put, you shouldn't just use an rcd to cover up a high Zs, it's normally needed for a high Ze, normally a TT circuit. If you have a good Ze but not getting acceptable disconnection time, this will need investigating and rectifying.

 

[pc1966]

The regulations require that you can disconnect in (generally) under 0.4s on a fault to earth, but they do not specify how you achieve that.

I would agree with @Sintra that good design is always to meet the Zs value for over-current disconnection unless it is unreasonably to do so. Slightly crap design is not an excuse!

In a TT case you are almost certainly dependent on a RCD for disconnection as it is very unlikely you can achieve a low enough supply Ze by means of any feasible earth rod, but on a TN system you really should be looking at the RCD for "additional protection" against shock or rodent damage, not as the primary means of disconnecting on a hard fault since your Zs is too high.

 

[Deleted account]

Only if the Zs cannot be achieved is it ok for the rcd to be used to Achieve disconnection times.

good design will nearly always achieve acceptable Zs.

obviously on a TT system disconnection times are very unlikely to be achieved, that’s where the rcd comes in.

 

[Vortigern]

Simply put really it is not ok. The problem with relying on RCD is it does not really address faults between L-N. In certain situations that can be a problem. Before opting for such a means to say "cover" high Zs one must consider the implications of what would happen in a L-N fault. Could someone be harmed or a fire, in say a combustible environment?

 

[Lister1987]

When you consider that there are more deaths and injuries caused by fires (due to electrical faults) than electric shocks or electrocutions, it really hammers home the importance of good design.....even if the primary focus is on electric shock these days.

 

[telectrix]

Simply put really it is not ok. The problem with relying on RCD is it does not really address faults between L-N. In certain situations that can be a problem. Before opting for such a means to say "cover" high Zs one must consider the implications of what would happen in a L-N fault. Could someone be harmed or a fire, in say a combustible environment?

not really relevant as even with a high Zs, the L-N loop impedance is independentof the Zs and a fault L-N shoud still trip the OCPD.

 

[radiohead]

Where it can be useful to rely on an RCD is where it is desired to change an existing OCPD, say a type B to a type C or D to negate the effect of high inrush current where an inductive load is to be installed to an existing circuit.
Where the Zs slightly exceeds the max permitted for a C or D it is a waste of resources to rewire an otherwise perfectly good circuit where an RCD will meet the disconnection time.
I've seen RCD protected circuits coded 2 where a measured Zs slightly exceeds the max for the OCPD, resulting in unnecessary remedial work to a fully compliant circuit.
It's a bit like the old neutral in a switch argument, some people are inherently opposed to an RCD meeting disconnection time on a TN but will happily do so on a TT.
And as above, Zs is irrelevant to short circuit current.

 

[telectrix]

assuming of course that the high Zs is due solely to the impedance of the earthing conductors, not the phase conductors.

 

[Julie.]

Although use of a rcd to protect circuits where the ocpd does not do so is allowed under the regulations, it should always be as a last resort. The installation should be designed properly, unfortunately allowing it often means it is used to cover up faulty/poor design.

The biggest issue is that an rcd will only ever protect against an earth fault. If there is a phase-phase or phase-neutral fault you are back to the ocpd, which may, or may not provide suitable protection.

The earth loop impedance is usually slightly higher than the ph-n impedance because of the reduced cpc size, but not by much as the supply neutral and supply earth are often the same combined cable.

So if ze is 0.4ohm, and r1+r2 is .25 ohm (zs = .65 ohm) the ze equivalent (for ph-n faults) would still be around 0.4ohm, although r1+rn could be down to 0.2 ohm - that's still 0.6 ohm!!! - still bigger than the typical 0.55ohm limit for a 16a d or 32a c type mcb.

This example is likely to result in a trip time of ~3 seconds.

So is this acceptable?

Fault to earth ~40ms (rcd)
Fault to neutral ~3 secs (mcb)

 

[Welsh_sparky_21]

Thanks for all the replies, just trying to get my head around it, so my understanding that even with a zs (earth fault loop) reading that’s too high a live to neutral short circuit would still trip your MCB due to the size of the fault current created??

 

[Julie.]

Thanks for all the replies, just trying to get my head around it, so my understanding that even with a zs (earth fault loop) reading that’s too high a live to neutral short circuit would still trip your MCB due to the size of the fault current created??

Yes, it would still trip the mcb and NOT the rcd, but the time to trip may or may not be acceptable.

 

[pc1966]

my understanding that even with a zs (earth fault loop) reading that’s too high a live to neutral short circuit would still trip your MCB due to the size of the fault current created??

Yes, if the high Zs is due to the earthing arrangement, and not due to a stupidly long/thin cable.

Usually the L-N short will trip the MCB if the voltage drop requirements are being met, as if you have no more than 5% drop on the final circuit (and that is a small proportion of the total board current capacity) you are looking at a fault current of around 20 times the rated current.

But bad design know no limits! Though in reality it is usually "no design" that is the problem.

You could make up a pathological case of, say, a 50W LED flood lamp at the end of several hundred meters of 1mm cable. It meets the VD for its fixed load of around 0.22A, and with, say, a 10A MCB the 1mm cable is overload protected, but it might not trip on a L-N fault in any sort of reasonably time-scale.

 

[Julie.]

...
You could make up a pathological case of, say, a 50W LED flood lamp at the end of several hundred meters of 1mm cable. It meets the VD for its fixed load of around 0.22A, and with, say, a 10A MCB the 1mm cable is overload protected, but it might not trip on a L-N fault in any sort of reasonably time-scale.

Err

That's the stuff I come across sooooo.. often!

Pre covid, a lot of my work was hospitality rebuilds/renovations frequently in the highlands, almost all have long private approach roads with a few led lamps, but hundreds of metres away from the supply.

I would say over 90% have been wired in 1.5mm^2 swa because the load is so small, - even with the small loads the voltage drop is more than allowed, never mind meeting any trip times whatsoever

 

[pc1966]

And here was me making it up to show an example!

 

[Julie.]

And here was me making it up to show an example!

As you said "bad design knows no limits..."

Although use of the word "design" in these cases would be a new definition of the word!

Just no thought to what was being done at all.

I suspect the swa is the size it is because it's the smallest they could get.

(Two core, poorly installed to boot)

 

[pc1966]

I suspect the swa is the size it is because it's the smallest they could get.

(Two core, poorly installed to boot)

I think you meant to say "I suspect the swa is the size it is because it's the cheapest they could get."

 

[marcuswareham]

Interesting, A while ago I had conduit CPC fail on 2 circuits with 50-150ohm Zs, I pulled through a new CPC for one of them, and for the cooker ran a new bit of T&E as was close to CU.

Although I believe this is the best way to fix it as now has a nice low Zs and meets disconnection time for the MCBs, But in theory if the circuits were on an RCD or upgraded to be on an RCD then could have been left like this as the 150ohms is enough to meet disconnection due to the RCD.

If it was to later have an EICR though I cant imagine many people passing it without a code as the r1+r2 would be high