RCD Feeding RCD discrimination

This includes types of overcurrent, Short Circuit and RCD protection
Post Reply
Energy
Posts: 8
Joined: Sun Apr 19, 2020 8:40 am
Answers: 0
Has thanked: 0
Been thanked: 0

RCD Feeding RCD discrimination

Post by Energy »

Hi,

RCD feeding in RCD - discrimination
Can we have a 30ma RCBO acting as an isolator feeding two other 30ma RCBOS(one for lighting circuit and one for PowerPoint) ?

Cheers
JamieP
Posts: 478
Joined: Tue Apr 14, 2020 11:08 am
Answers: 0
Has thanked: 92 times
Been thanked: 18 times

Re: RCD Feeding RCD discrimination

Post by JamieP »

You can, but what's the point?

You're correct you will have no discrimination so it'll be luck of the draw what trips and what doesn't so you'll be putting yourself in a situation where enough leakage fault on either is just as likely to take out both as it is to take out the one with the leakage

Infact it's almost more likely you'll trip the upstream device as it'll always have the combination of leakage (small or not) but will mean in total it'll always be more than what's just being detected by the one
AlecK
Posts: 912
Joined: Thu Apr 16, 2020 11:24 am
Answers: 5
Has thanked: 2 times
Been thanked: 351 times

Re: RCD Feeding RCD discrimination

Post by AlecK »

There is an underlying requirement to minimise inconvenience in the event of a fault [1.6.5 (a)] .

Personally I don't like the use of "minimise"; because minimum means as low as it is possible to get
- which would translate to only having one point per RCD.
However a fundamental concept of "3000" is that complying with Part 2 is deemed to meet all requirements of Part 1.
So we can assume that when it comes to RCDs we need do no more than what Part 2 requires;
and there's no specific requirement for discrimination between RCDs.

So nothing to prohibit configuration of one 30 mA RCD feeding 20 other 30 mA RCds.
But no advantage to it, and it's contrary to the principle of 1.6.5.

There can be no discrimination between RCDs of same rated residual current.
True there may be slight differences in the level of standing leakage, depending on exact configuration;
but we can't know what those levels are at any particular time.

Unlike mcbs, RCDs do not have a nice, predictable, trip curve that lets us calculate for discrimination.
The trip curves for mcbs are based on predictable thermal & magnetic effects; in effect we can know with reasonable accuracy how long tripping will take at various levels of load current or fault current.
RCDs are more like a logic gate: is the residual current enough to trip: yes / no.
That tripping point isn't specifically known, somewhere between 50% & 100% of RRC, generally around 80%
and the operating times are also not well enough known to allow us to predict how any particular pair of devices will react.

Even between RCDs of different RRC, eg 30 mA upstream of 10 mA (as for schools), discrimination can't be relied on.
At leakage currents between 10 mA & 30 mA, good chance of the 10 mA unit operating and the 30 mA not.
But we can't assume that the leakage will be at that sort of level ; especially where the RCD is intended for "additional protection" against shock
Consider a person touching an active conductor, and also being in contact with earth
Resistance of a body (hand - hand or hand - foot) is approx 1 000 ohms (assuming dry conditions and no penetration of skin).
Apply Ohm's Law : 230 V / 1000 ohms means the shock current will be around 230 mA;
so with both RCDs having the same required operating time (at their respective RRCs), and the current being several multiples of the RRC for both,
not possible to rely on the 10 mA RCD tripping first and the 30 mA remaining untripped.

Same logic applies for upstream RCDs of higher RRC, eg 100 mA or even 300 mA.
True there's a better chance of the actual residual current being enough to trip only the downstream device.
But we can't rely on that being the case; as the actual residual current depends on the actual circumstances.

The only way to get reliable discrimination between RCDs for "additional protection" and an upstream RCD is if the upstream RCD is Type S;
therefore taking longer to trip. And Type S doesn't provide adequate shock protection.
PawPatrol
Posts: 27
Joined: Thu Jun 04, 2020 8:42 pm
Answers: 0
Has thanked: 1 time
Been thanked: 1 time

Re: RCD Feeding RCD discrimination

Post by PawPatrol »

Energy wrote:
Wed Sep 08, 2021 8:44 pm
Hi,

RCD feeding in RCD - discrimination
Can we have a 30ma RCBO acting as an isolator feeding two other 30ma RCBOS(one for lighting circuit and one for PowerPoint) ?

Cheers
Take the type s out and throw them in the bin, they will likely never trip as the upstream one will trip first cutting the current flow and they will confuse others.

I agree with Alec, whats the point!

How you have described it is unlawful as you can't use a type s for compliance with the esr, that would be deemed unsafe. (Reading it again you haven't claimed any of them are for complince with the esr, if this is the case put as many in as you want! Its not un lawful)

If it was described as the upstream rcd was providing the personal protection and they type s was the isolator then thats not unlawful.
User avatar
DougP
Posts: 242
Joined: Sun Apr 12, 2020 10:11 pm
Answers: 3
Has thanked: 19 times
Been thanked: 31 times

Re: RCD Feeding RCD discrimination

Post by DougP »

If you could get a 30mA Type S, they will still comply as they trip at less than 300ms. And installing them upstream of a standard type A is one means to provide a form of discrimination with RCDs in series. But there's no requirement to do so.
AlecK
Posts: 912
Joined: Thu Apr 16, 2020 11:24 am
Answers: 5
Has thanked: 2 times
Been thanked: 351 times

Re: RCD Feeding RCD discrimination

Post by AlecK »

Type S are generally 100mA or more, and used mainly for property protection.
Whereas RCDs for protection against shock can't be more than 30 mA.

When NZ moves to allowing TT supply for some installations, we will need to deal with the lack of reliable discrimination between RCDs in series.
Such installations will have an RCD as (or associated with) the main switch;
necessary in order to provide fault protection in absence of a low-impedance fault current path via MEN link to distribution N .
The RCD will be in the order of 300 mA; but not Type S because can't take longer than 400 ms to operate on an earth fault.
And that's only 33% slower than the 300 ms max (@ RRC) for shock protection.

Then downstream of that will be all the usual RCDs for "additional protection" and also protection against mechanical damage.

As per previous post, shock current will be approx 230 mA.
So there's a chance only the local 30 mA RCD will trip - but again if the max operating time is the same (eg both Type A) any discrimination will be down to luck rather than something we can rely on

An earth fault in wiring / appliance will generate a far higher current, and almost certainly trip the 300 MA unit as well - blacking out the entire installation.
So accommodate normal levels of leakage, and to reduce disruption, some installations may have a bunch of these 'front-end" RCDs;
each supplying part of the installation.

A "front end" RCD can only provide fault protection downstream of itself; so to deal with earth faults upstream of the RCD will mean having to earth associated metalwork to the incoming PEN rather than to the installation earth bar.
Post Reply