Lost neutral and surge protectors (TN-C-S)

[p=iv]

Would a surge protector (on an RCD) circuit protect from a lost neutral overvoltage? I know that isn't the aim of them (transient spikes only rather than sustained overvoltage) but I am thinking it may be possible (although very alarming if there isn't a thermal fuse in the surge protector). This is referring to a single phase TN-C-S installation.

If the neutral is lost, the line voltage will float. The surge protector's MOVs theoretically would start to conduct. That is of course making the assumption that the floating voltage goes above the clamping voltage of the MOVs. I would assume that if there was a persistent current flow through the MOVs (to ground) that the MOVs would burn out very quickly (or rather explode or catch fire unless they have a thermal fuse). If the MOVs are shunting the current to earth for over 30ms, I would assume the RCD would trip due to the L-N inbalance which now exists.

If that is correct, then hopefully the elevated voltages won't last for very long and the potential for damage to occur (from both the voltage and the now orange hot MOVs) be very small on every RCD protected circuit. The non-RCD circuits are pretty much stuffed.

On another tangent, it is well worth checking that any surge protector you buy does have a thermal fuse. Bigclive did a teardown of one and there was no thermal fuse. Given how MOVs behave when they fail, it is well worth checking if it's possible to open the case. It's completely irresponsible of manufacturers to skimp on thermal fuses in what is essentially a time-triggered incendiary bomb encased in flammable plastic. Given the influx of deathdaptors, exploding fake chargers, thin conductors in 10A cables and all the other dangerous tat out there, it's not surprising in the least. I wonder how many fires surge protectors (even those with thermal fuses) have caused. I wouldn't trust the plastic to be fire retardant either. If they skimp on a thermal fuse, the plastic is going to be the same.

 

[Lucien Nunes]

I assume you mean if an overvoltage occurs due to an open CNE on the 3-phase section of the network (as a fault in the single-phase supply or installation won't cause an overvoltage). In that case the answer is maybe. Obviously the N & E should track closely on a TN-C-S supply and if the L-N protection starts to conduct, then the L-E should do about the same time and trip the RCD. But the problem is that depending on connected load balance, open-neutral faults can produce slight overvoltages that won't push the surge protector into conduction at all, but can still damage loads. There are different clamping voltages of MOV used in commercial surge protectors, and at say 300V, all sorts of damage might occur as that voltage could be present indefinitely. I have certainly seen situations where a lost neutral resulted in equipment damage - e.g. professional monitors found to be smoking after 15 minutes use due to electrolytic capacitors breaking down - with no evidence of the surge protector going into conduction.

The only way to reliably guard against this condition, and it is rare, is an adjustable voltage sensing relay and a shunt trip coil on the RCD or contactor to isolate the loads.

 

[p=iv]

The open CNE is exactly what I was thinking of, say on a substation or transformer supplying a street and having a problem with a copper-robber or corrosion (or some accident that results in a lost neutral for the street).

But yeah, a surge protector giving "protection" would require the floating voltage to get above the clamping, which may never happen. Until then (or if) there'll be a fair amount of damage done by the excess voltage.

I remember reading a report about an office fire in Los Angeles in the 1980s which was probably caused by a combination of harmonic distortion (from the fluorescent lights), a lost neutral as it was undersized (probably at the floor feeder) due to overload and the resultant floating voltages being sufficient to cause a fire (I'm not sure of the mechanism but it was probably arcing caused by the spacing between socket busbars being inadequate for the now elevated voltage).

It was anticipated that the neutral currents would be quite low (building was designed and built in the 70s) so to keep costs down thinner (i.e. undersized) neutrals were used. To combat this problem, codes were revised and the neutral conductors enlarged.

Fortunately I've never encountered a lost neutral but as I have a somewhat (overpriced) posh Alienware desktop I am a teensy bit paranoid although a lost neutral and resultant floating voltage isn't too probable.

I'm not using a surge protector as an overvoltage protection device, incidentally as a) it's not designed for that and b) it'll only "protect" above the required voltage (which may never happen, but there'll still be damage done). It was one of those random thoughts that occurred to me rather than something that's been/is being done.

The smoking monitors sounds pretty painful. Was it a fault on the 3ph grid (DNO/Nat Grid) or a fault in the building's transformer (I'm assuming that it was in an office building with 3ph) that caused it? Were they able to replace the caps or were the monitors terminal (pun intended )?

I remember reading of a case on a forum of the CNE/PME in a street cable being damaged. Unfortunately (for the radio ham), a radio ham lived on the street. There was no real interruption to the electrical supply despite the damaged cable(s). The entire street's current return path back to the transformer ended up being the copper earthing rod for the radio setup, which didn't do it much good. It apparently got a bit hot to say the least! I can't find the article (or remember which forum it was on) but it was quite interesting reading. There's quite a few details missing as I can't remember them.

 

[Lucien Nunes]

The smoking monitors sounds pretty painful. Was it a fault on the 3ph grid

These were HD video monitors on a film set. I used to run the electronics dept of a film equipment supply company and got to see a remarkable amount of kit that had been subjected to somewhere between 230 and 400V. The reason was that a lot of power distro on set is by heavy single-core flexibles with single-pole plugs, up to 400A circuits are common, and if somebody pulls the neutral by mistake or puts the plugs in the wrong order, odd voltages are inevitable! These days there are Powerlocks with mechanical interlocking that ensure the neutral goes in first and comes out last.

Usually we were able to repair smoked gear - because of the replacement cost it was well worth rebuilding power supplies etc. Ironically the worst damage from faulty power was ELV - a digibeta field editor that required about 8A at 12V was given 20V AC by a damaged power pack. That cost about £10,000 of parts to fix IIRC.

 

[Sparkgap]

Never bothered looking inside surge protectors but would they be sensing LN or LE? If the latter then they'd just see 240V and not pick up excessive voltage caused by N failure.
Had a primary school which lost neutral and had to evacuate when a couple of fluorescent fittings burst into flames. Problem was electric kitchen pulling the N point right down on one phase and causing excessive volts on another.

 

[Lucien Nunes]

Normally L-N, L-E and N-E. In the case mentioned, the supply was SP+N TN-C-S so if the L-N volts wandered off the L-E would have been very close behind it hopefully putting the L-E protection into conduction. But as I mentioned, whether that would happen would depend on the clamping voltage of the particular surge protection.