3 phase at work - 1 phase down - questions?

[AdrianUK]

As a commercial user I believe (or rather have been told, but not seen the evidence) that we are charged for neutral imbalance.

I think you've been misinformed. The DNO doesn't care about balancing of the phases .... a zero neutral current is pretty much impossible in an installation that contains anything other than just motors anyway.

On a commercial tariff the DNO will be concerned about power factor & you may well be penalised for having a poor power factor but that's not the same thing as neutral current.

 

[JohnW2]

Are they not thinking of '3-phase circuits', with 3-phase loads? In that case, one can't really talk about faults affected just one phase, or for it to be safe to disconnect just one phase in the case of a fault, can one?

Yes, in the case of a 3-phase load, it seems quite logical.

I don't think it's just a case of 3-phase loads - it's more an issue of 3-phase circuits (into which a 3-phase load could be 'plugged') - in that situation, I think it would be very unsafe not to disconnect all line conductors in the event of any fault on the circuit, wouldn't it?

However there are instances, because of budget/time where this ruling does not make sense, and can impede the integrity of the design.

As above, I think the only situation in which it makes sense (or is even safe) not to disconnect all three phases is when, as in my installation, the three phases are being used totally separately - and we have agreed that they can then simply be regarded as three separate single-phase circuits.

Kind Regards, John

 

[mfarrow]

The use of 3-phase equipment (i.e. a motor) would obviously be safer if three phases disconnected at once, however where the equipment it's feeding is only single phase by design, even if the circuit is 3-phase, I can't see any advantage to disconnecting all the lines if only one is faulty.

 

[JohnW2]

The use of 3-phase equipment (i.e. a motor) would obviously be safer if three phases disconnected at once, however where the equipment it's feeding is only single phase by design, even if the circuit is 3-phase, I can't see any advantage to disconnecting all the lines if only one is faulty.

I'm not sure I understand what situations you are thinking of. If one knows for sure that part of a 3-phase circuit is feeding a load which is 'single phase by design', that load must either be fixed-wired to one phase or else plugged into a single-phase socket which is wired to a single phase. In either case, the (single phase) circuit [either feeding the fixed-wired load or single-phase socket(s)] can surely have its own (SP or DP) protective device(s).

If there are also 3-phase loads run from the supply, it obviously gets more complicated, because they do need protective device(s) which disconnect all three phases - so the SP/DP devices protecting single-phase off-shoots from the 3-phase supply would (unless separated at the origin of the installation) have to discriminate from the 3-phase devices, which could be difficult/impossible.

Think of my house. It's currently three totally separate single-phase 'circuits', each with their own SP/DP devices. If I wanted to run a 3-phase load, I'd probably have to take a feed off the supply upstream of all my SP/DP devices, and provide the 3-phase circuit with it's own 3P or 4P protective devices.

Kind Regards, John

 

[mfarrow]

I'm not sure I understand what situations you are thinking of. If one knows for sure that part of a 3-phase circuit is feeding a load which is 'single phase by design', that load must either be fixed-wired to one phase or else plugged into a single-phase socket which is wired to a single phase.

The OPs example where he could have a socket circuit split over several phases, one or more double socket per phase balanced along the length of the office.

Also the lighting example, with single phase luminaires staggered in a row across three phases, with shared neutral.

In either case, the (single phase) circuit [either feeding the fixed-wired load or single-phase socket(s)] can surely have its own (SP or DP) protective device(s).

Yes if you were to run both 'live' cables out independantly for each phase and single-phase devices then you're left with 3 single-phase circuits, but if you use one neutral and 3x 1P devices occupying one 'way' then it becomes a 3-phase circuit.

 

[mattylad]

Cheers guys.

 

[JohnW2]

I'm not sure I understand what situations you are thinking of. If one knows for sure that part of a 3-phase circuit is feeding a load which is 'single phase by design', that load must either be fixed-wired to one phase or else plugged into a single-phase socket which is wired to a single phase.

The OPs example where he could have a socket circuit split over several phases, one or more double socket per phase balanced along the length of the office. Also the lighting example, with single phase luminaires staggered in a row across three phases, with shared neutral.

Are you (as per your comment below) simply referring to a situation in which single-phase circuits using different L phases are sharing a common neutral? Is that what you mean by "a sockets circuit spread across several phases" (although I'm not sure I would call that a single 'circuit')?

In either case, the (single phase) circuit [either feeding the fixed-wired load or single-phase socket(s)] can surely have its own (SP or DP) protective device(s).

Yes if you were to run both 'live' cables out independantly for each phase and single-phase devices then you're left with 3 single-phase circuits, but if you use one neutral and 3x 1P devices occupying one 'way' then it becomes a 3-phase circuit.

OK, I think I'm starting to understand what you're talking about. If you have a situation in which there are not, and you can be sure that there will never be (e.g. no 3-phase sockets) be any 3-phase loads, then it would be nice to have protective devices that only disconnected one L phase. However, the nature of RCDs is such that one obviously cannot do that unless one has separate neutrals for each 'single-phase circuit' - if the circuits share a common neutral, the only sort of RCD protection one can have would be from a 3P+N RCD, which would disconnect all line (and usually all live) conductors.

So, it seems to me that the 'problem'/inconvenience you are referring to is nothing to do with regulations (i.e. 531.2.1), and everything to do with how RCDs work. If, in an installation in which there is no possibility of a 3-phase load, you want operation of an RCD to take out only one phase, you surely have to have separate neutrals for each phase (as in my house)? ... or am I missing something?

Kind Regards, John

 

[mfarrow]

OK, I think I'm starting to understand what you're talking about. If you have a situation in which there are not, and you can be sure that there will never be (e.g. no 3-phase sockets) be any 3-phase loads, then it would be nice to have protective devices that only disconnected one L phase. However, the nature of RCDs is such that one obviously cannot do that unless one has separate neutrals for each 'single-phase circuit' - if the circuits share a common neutral, the only sort of RCD protection one can have would be from a 3P+N RCD, which would disconnect all line (and usually all live) conductors.

So, it seems to me that the 'problem'/inconvenience you are referring to is nothing to do with regulations (i.e. 531.2.1), and everything to do with how RCDs work. If, in an installation in which there is no possibility of a 3-phase load, you want operation of an RCD to take out only one phase, you surely have to have separate neutrals for each phase (as in my house)? ... or am I missing something?

No you're not. I probably threw you by discussing RCBOs in my initial response, which of course wouldn't work on 3-phase circuits with a single neutral. Allowing myself some time for a little but more thought, that makes sense to me now, and you've answered my questions below :

I'm still trying to decide why the disconnection of all lines is required for an RCD anyway, instead of just the phase affected. Perhaps it's in the lack of available distinction in a 3-phase RCD regarding which phase is leaking. Answers on a postcard?

So all that 531.2.1 does is re-affirm what the function the RCD should (and only could) undertake if it were a 3-phase device.

Going back to the lighting example, the use of an RCD across all 3 phases never got raised at the design/install; the politics were limited to the choice and interpretation of regulations surrounding the use of single-phase OPDs (leading to this discussion. However when the maintenance contractor commented that he'd observed we hadn't gone for RCBOs, or an RCD per way, I did think it was something we'd missed. Obviously at the time there wasn't much that could be done (the job was finished), so it's useful that we've had the discussion now!

Is it not theoretically possible though to have an RCD operate only 1 phase of a 3-phase circuit? i.e. look at the current through one live conductor and compare it to the sum of the other three? Then trip that faulty phase? Obviously this is a deviation from current design, but am I right, and if so does such a 'device' (not a package of parts from a supplier) actually exist?

 

[JohnW2]

...So, it seems to me that the 'problem'/inconvenience you are referring to is nothing to do with regulations (i.e. 531.2.1), and everything to do with how RCDs work. If, in an installation in which there is no possibility of a 3-phase load, you want operation of an RCD to take out only one phase, you surely have to have separate neutrals for each phase (as in my house)? ... or am I missing something?

No you're not. I probably threw you by discussing RCBOs in my initial response, which of course wouldn't work on 3-phase circuits with a single neutral. Allowing myself some time for a little but more thought, that makes sense to me now, and you've answered my questions below :

I'm still trying to decide why the disconnection of all lines is required for an RCD anyway, instead of just the phase affected. Perhaps it's in the lack of available distinction in a 3-phase RCD regarding which phase is leaking. Answers on a postcard?

So all that 531.2.1 does is re-affirm what the function the RCD should (and only could) undertake if it were a 3-phase device.

Quite so. As you say, 531.2.1 is really stating nothing much more than the inevitable/obvious in relation to a 3-phase device (and obviously means nothing in terms of a 1-phase device).

Going back to the lighting example, the use of an RCD across all 3 phases never got raised at the design/install; the politics were limited to the choice and interpretation of regulations surrounding the use of single-phase OPDs ... However when the maintenance contractor commented that he'd observed we hadn't gone for RCBOs, or an RCD per way, I did think it was something we'd missed. Obviously at the time there wasn't much that could be done (the job was finished), so it's useful that we've had the discussion now!

Glad you've found it helpful. You could, of course, have had one RCBO/RCD per phase, but only if you had installed separate neutral cables for each phase.

Is it not theoretically possible though to have an RCD operate only 1 phase of a 3-phase circuit? i.e. look at the current through one live conductor and compare it to the sum of the other three? Then trip that faulty phase? Obviously this is a deviation from current design, but am I right, and if so does such a 'device' (not a package of parts from a supplier) actually exist?

I need to think much more deeply about that, but my first reaction is to think that it probably wouldn't be possible, particularly given the phase differences between the three phases. You are postulating that, in the absence of faults, the vector sum of the currents in the other two Ls and the N would be equal to the current in the L of interest, and I'm far from sure that such would be the case. Even if it were, surely one problem would be that an L-E fault on either of the other two phases (or any N-E fault) would inevitably upset the balance (by reducing the N current in relation to the L currents) and cause the device to operate, disconnecting the phase of interest (which had no fault)? I'll do some more thinking!

Kind Regards, John

 

[plugwash]

Is it not theoretically possible though to have an RCD operate only 1 phase of a 3-phase circuit? i.e. look at the current through one live conductor and compare it to the sum of the other three?

No the math simply doesn't work. If your proposed calculation detects a difference for one phase it will detect the same difference (possiblly with different signs) for all phases.