A Huge Puzzle

[westie101]

Now we can't figure this out, a few of us at work and those affected (an electrician & an electrical engineer).

Late yesterday a colleague was called to a case of high volts with 800V AC being measured at a customer's premises (an ex-job centre now taken over by a safety advice firm that works at Sellafield.) Coupled with a lost neutral

When our standby crew arrived the electrician told them what he had found and added "you think I'm mad don't you?"

So checks were carried out.

At the single phase (Y or L2) meter position was a cut-out, a mechanical meter, a pair of Henley blocks and 3 CUs all switched off and all MCB.s in the off position
2 of the CUs had double pole main switches, the other only a single pole switch (odd I know but true)
Voltage when measured at the blocks was zero L-N and in the normal range L-E

When one of the double pole CUs was switched on, the L-N voltage leaped from 0 to over 500V and then quickly climbed to 800V, with the (mechanical not the test) meter making a very odd sound!!

Further investigation showed that the neutral was open circuit on the incoming supply cable.
Earthing was TN-S

The electrician also added that whilst investigating and checking the internal wiring he found that at one socket that had been reported as faulty, the neutral and earth were both in the earth terminal.
Obviously this "connection" had been carrying all the load current back to our system since the neutral had been lost on the service cable.

Checking on customers past this point showed there was nothing abnormal with their supplies, so the fault was definitely on the service cable

An excavation showed that the service cable had been damaged, ultimately causing the open circuit neutral, at some time in the past (we don't know when) by BT when a duct had been installed.
The cable was disconnected and terminated, a new piece of service cable joint made to the main and the two service cables jointed together to restore supply. (the damaged service cable did not have enough length on it to make a straight joint but just enough to terminate it).

Now we can fully understand the effects of the lost neutral, the effect of the neutral earth connection in the socket.
We could also understand voltage in the 400V range for a lost neutral on the main (very, very rare on a PILSTA cable), or even a faulty joint where the neutral was open-circuit and then had come in contact with another phase

But where did the measured and witnessed by about 5 folk (our engineer & jointing team, the electrician and the electrical engineer employed by the customer) 800V come from as we have absolutely no idea?

Once the supply was restored to normal at about 02:30 this morning BTW all was fine with no issues.

 

[ericmark]

Can't tell you about your situation but I can tell you on the building of T5 a crane failed dropping it's load to the ground.

This of course was serious and the engineer grounded all the tower cranes pending an investigation.

Duly the Italian team arrived and we were called as it was claimed the CAT generator was running at 100 Hz. We watched the gauges and I looked at my boss and we both said diodes. We convinced the Italian he should disconnect one drive at a time and that one of the drives was effecting the other drives and he did locate which he replaced and the fault vanished.

We both tried to convince the Italian this was not good enough and that one drive should not be able to effect another drive to the extent that the crane dropped it's load.

However I was then involved in an industrial accident mauling my hand and my boss was asked to submit his resignation as a result so likely the Italian got away with it.

Returning to your problem I would guess something similar some switch mode supply or a motor drive has likely tricked the meters to giving wrong readings.

My first idea would be look at any solar panels or wind generators although they should close down and fail safe switching off with over voltage means they may not export the power the firms fitting them require to get pay back. So modifying the inverter to stop it failing on over voltage is not unknown.

 

[bernardgreen]

What sort of meter was used to measure the 800 volts ? Low impedance or high impedance and was it one that held the highest peak voltage ?

To get more than the supply voltage would require some form of transformer or other inductive load being switched on and off thus creating a back EMF several times that of the supply voltage.

So my first guess would be a motor with being repeatedly turned on and OFF, possibly getting a "neutral" from a connection to ground possibly via a fault of sneak connection via bonding / earth cables. The back EMF would produce a voltage in excess of the 230 volt supply.

The noise from the mechanical meter could be the result of a load to the motor being chopped on and off.

I would suspect a motor with a defective no volt release that was not fully releasing ( or a self restoring one )

 

[JohnW2]

At the single phase (Y or L2) meter position was a cut-out, a mechanical meter, a pair of Henley blocks and 3 CUs all switched off and all MCB.s in the off position 2 of the CUs had double pole main switches, the other only a single pole switch (odd I know but true) ... Voltage when measured at the blocks was zero L-N and in the normal range L-E ... When one of the double pole CUs was switched on, the L-N voltage leaped from 0 to over 500V and then quickly climbed to 800V, with the (mechanical not the test) meter making a very odd sound!!
Further investigation showed that the neutral was open circuit on the incoming supply cable. Earthing was TN-S
The electrician also added that whilst investigating and checking the internal wiring he found that at one socket that had been reported as faulty, the neutral and earth were both in the earth terminal. Obviously this "connection" had been carrying all the load current back to our system since the neutral had been lost on the service cable. ...
Now we can fully understand the effects of the lost neutral, the effect of the neutral earth connection in the socket. We could also understand voltage in the 400V range for a lost neutral on the main (very, very rare on a PILSTA cable), or even a faulty joint where the neutral was open-circuit and then had come in contact with another phase ... But where did the measured and witnessed by about 5 folk ... 800V come from as we have absolutely no idea? ... Once the supply was restored to normal at about 02:30 this morning BTW all was fine with no issues.

Hmmm. Since it’s clearly impossible, all five of you must be lying

Very odd. On the face of it, it would seem that it is absolutely impossible to get a pd of 800V ‘(or even 500V) between anything and anything else’ from the secondary of one of your 400V transformers, no matter what has happened to the wiring - so the only way one ought to be able to get such a pd is if someone has somehow managed to get a higher voltage source or, at least, a source out of phase with yours into your system (solar panel gone mad?), or an out-of-phase back emf from a highly reactive load.

However, I have to say that my first inclination in an ‘impossible’ situation like this would be to suspect your measurements. No, I’m not suggesting that the five of you were lying or hallucinating but, as we do when DIYers report unexpected voltage measurements, would ask what you (and the electrician) were measuring these voltages with? If it was a high impedance meter, then did you examine the effect of putting some load across the meter? If not, it’s just possible that you were somehow ‘picking up’ induced (or whatever) voltages (given some ‘floating N conductor etc.) which were not a true reflection of what the ‘real’ voltages/pds actually were. A fairly wild suggestion, but about the best I can think of at present!

Do you have any idea of what (if any) load , and what type of load (possibly highly reactive?), was introduced when you switched on one of the CUs?

Kind Regards, John
Edit: seemingly a bit slow again, but maybe I've made a few points/suggestions over and above bernard's!

 

[Adam_151]

To get more than the supply voltage would require some form of transformer or other inductive load being switched on and off thus creating a back EMF several times that of the supply voltage.

Perhaps even the voltage coil of the mechanical meter?

Perhaps the board being switched on (double pole switch on that one) had a fleeting N-E short, perhaps it was that wrongly wired socket and something had burnt out due to the load.

 

[westie101]

Sorry should have said that the voltage effect was seen on 4 separate meters! Certainly the one my colleague used would have been a digital one.

The building is all offices with no major motor loads, and in any case (all MCBs OFF in the three CUs could not have got a supply anyway.

My thought at the moment is something surrounding a UPS, but I don't know if they have any installed

 

[JohnW2]

Sorry should have said that the voltage effect was seen on 4 separate meters! Certainly the one my colleague used would have been a digital one.

'Digital ones' are surely the most likely to be very high impedance. On high voltage ranges, even moving coil meters (if you still use any) are likely to have very high input impedances, unless shunted - a "20kΩ/v" one would have an input resistance of 20 MΩ on a 1000V range.

Kind Regards, John

 

[deadshort]

Are you thinking the perhaps a UPS connected to the CU with the SP main switch ? With the un- switched neutral and the earth/neutral connection at the socket ? The voltage from the UPS was in some way creating a back feed?


Interested :

Kind Regards,

DS

 

[JohnW2]

Are you thinking the perhaps a UPS connected to the CU with the SP main switch ? With the un- switched neutral and the earth/neutral connection at the socket ? The voltage from the UPS was in some way creating a back feed?

There would have to be something very wrong with the way in which the UPS had been installed/wired for that to be possible, wouldn't there?

Kind Regards, John

 

[RF Lighting]

I'm thinking capacitor(s). Maybe all the PF correction caps in the fluorescent lights all added together makes one big cap?

Have seen voltages upto 1000V from a SP&N supply on a planer thicknesser with a knackered starter as a result of two 100μF caps in parallel.

 

[deadshort]

No, not if it was wired by the person who wired the earth and neutral together into the socket? U know, the boffin type, who can't fit a plug, we've all met them, or at least heard them !

Kind regards,

DS

 

[JohnW2]

No, not if it was wired by the person who wired the earth and neutral together into the socket? U know, the boffin type, who can't fit a plug, we've all met them, or at least heard them

Fair enough - but wouldn't that qualify (in the minds of most of us!) as "something very wrong"?

Kind Regards, John

 

[westie101]

earth/neutral connection at the socket

That had been corrected it was after that had been done that the voltage arose.

No, not if it was wired by the person who wired the earth and neutral together into the socket? U know, the boffin type, who can't fit a plug, we've all met them, or at least heard them !

Ex government building so unlikely and no evidence of rewiring after changing hands that I know of.

 

[deadshort]

No, not if it was wired by the person who wired the earth and neutral together into the socket? U know, the boffin type, who can't fit a plug, we've all met them, or at least heard them

Fair enough - but wouldn't that qualify (in the minds of most of us!) as "something very wrong"?

Kind Regards, John

But it is wrong john.

Kind Regards,

DS

 

[JohnW2]

Fair enough - but wouldn't that qualify (in the minds of most of us!) as "something very wrong"?

But it is wrong john.

I'm getting confused. Are you just making some distinction between your "wrong" and my "something very wrong", or are you making some other point? Maybe I get a bit dim on Friday evenings!

Kind Regards, John