Neutral Bounce ? ? ( discussion only )

[bernardgreen]

During a thunderstorm last week there a power outtage lasting a few seconds after a lighting strike to ground less than a mile from the location.

It was later discovered that the RCD in an out building had tripped. It would not reset until two items of equipment were unplugged. Investigation found the mains input filters on these two items were showing low impedance between Neutral and Earth terminals.

The outbuiding is fed TT with a pair of ground rods.

The cause of the fault is believed to be a severe potential difference between ground ( at the ground rod ) and the supply neutral as a result of the lightning strike and that this transient punched through the dialectric of the filter capacitors.

The extremely rapid rise of the transient voltage from a lightning strike means the internal inductance of the wound paper capacitors becomes very significant.

Is there a case for fitting fast acting transient absorbers between TT Neutrals and ground rods where the distance involved would create a significant voltage gradient across the ground. ?

 

[ericmark]

I went to a lecture on thunder storms and the effects on electric items and there are in the USA many things done to stop electrical storm damage. However in this country the damage is so rare we do very little.

We were told how sharp edges and tall conductors attract lighting so when fitting aerials one should avoid these so never fit to soot lined chimney and gable ends. Guess where we normally fit them!

We were also told don't earth the aerial this attracts the lighting yet for years we all earthed aerials.

What worries me is diodes in the earth. I know they are designed for the job but fitting back to back diodes in the earth supply to boats to stop electrolysis eating the hull must be a problem with electric storms where it could kill the diode without the boat owner knowing a real problem in fresh water.

With electric storms on the increase I would agree we need a rethink on many of our practices.

 

[ban-all-sheds]

I wonder if French practices might be worth looking at? There's a lot of thunderstorm activity over the Massif Central, and much of that is pretty rural, so TT supplies are likely to be commonplace.

 

[bernardgreen]

There has never been an answer to the question about the top of a lightning conductor. Should lightning conductors have points or balls at the top.

A ball appears to be better at taking a strike but points do tend to create a current that reduces the static charge and this may reduce the severity of the strike when it occurs. Some of the first tall buildings in the City of London had many pointed rods spread over the roofs in the belief this would discharge enough static to prevent the building being struck.

 

[JohnW2]

The cause of the fault is believed to be a severe potential difference between ground ( at the ground rod ) and the supply neutral as a result of the lightning strike and that this transient punched through the dialectric of the filter capacitors. ... Is there a case for fitting fast acting transient absorbers between TT Neutrals and ground rods where the distance involved would create a significant voltage gradient across the ground. ?

I'm sure there is 'a case' (provided the absorbers actually 'work' - do they?), but you'd presumably have to extend that the the L conductor as well, since 'next time (if...) it presumably could be a L-E pd which did the damage.

I would suspect that (unless it involves highly crucial equipment) there's also a case for not doing any of that, on the basis that such events are so rare (at least, in the UK) that it's better to just deal with the consequences if/when they ever arise.

Kind Regards, John

 

[chris765]

I got called to a job a couple of weeks ago that was affected by a lightning strike.

Both sides of a semi detached were affected. They have a loop supply. We believe that the aerial was hit.
In both properties the RCD main switch had burnt/melted. Both properties were fed overhead and should be TT (PME unavailable) but neither had an earth rod.
They had a lead water service which was bonded, and after the strike they had muddy water for some time.

So we assume that the fault went to earth through the water pipe, and had it hit the aerial it probably found it's path to earth somehow. What's strange is the melted RCDs, why would they come into the fault path?

The customer said that UKPN had visited about 4 times after working on the overheads in the field behind, so I wonder if any of the overheads had been struck?

 

[JohnW2]

We believe that the aerial was hit.
In both properties the RCD main switch had burnt/melted. Both properties were fed overhead and should be TT (PME unavailable) but neither had an earth rod. ... They had a lead water service which was bonded, and after the strike they had muddy water for some time. ... So we assume that the fault went to earth through the water pipe, and had it hit the aerial it probably found it's path to earth somehow. What's strange is the melted RCDs, why would they come into the fault path? ... The customer said that UKPN had visited about 4 times after working on the overheads in the field behind, so I wonder if any of the overheads had been struck?

As you say, odd. I suppose that if the overhead L had been struck, it could then have flashed over to N within the RCD/Main switch and found its return path to earth via the supply N - but I might have expected the cutout fuse to have operated if that were the case. Was anything else in the installation damaged?

Kind Regards, John

 

[chris765]

There was no installation damage that I could find. There was however TVs/Sky boxes - infact most things plugged into the aerial system in both properties had failed.

Their central heating programmer had also failed and upon opening it a resistor looked to have over heated.

I normally deal with Neutral faults/surges on the network on behalf of DNOs and this wasn't typical of those types of faults.

 

[JohnW2]

There was no installation damage that I could find. There was however TVs/Sky boxes - infact most things plugged into the aerial system in both properties had failed.

Is it possible that, in the absence of a proper installation earth (other than bonded supply pipes), the lowest impedance path to earth present within those TVs/boxes etc was the final circuit's neutral and that the lightning therefore flashed over to that (within those aerial-connected devices) and took the route via the supply neutral back to earth at the tranny (via the RCD/main switch on the way)?

Kind Regards, John

 

[bernardgreen]

I suppose that if the overhead L had been struck, it could then have flashed over to N within the RCD/Main switch

The sensor toriod in the RCD is the first intentional inductive component the supply meets as arrives at the property. For mains AC voltage changing as a sine wave at 50 Hz ( 0 to max in 5 milli seconds ) this inductance creates virtually zero impedance to the current. The current from the lightning strike however increases very rapidly, typically less than a tenth of a millisecond, and for this the toriod presents a significant impedance and the resultant voltage spike will be in the thousands of volts so voltage break down is very likely.

Also the motor effect of the current may be enough to mechanically damage the internals of the RCD as the current carrying parts move relative to each other .

( Motor effect = current in a conductor that is in a magnetic field will tend to move the conductor at right angle to the magnetic flux. The magnetic field can be created by current in an adjacent conductor which will also be affected by the motor effect )

Because the current is very short duration the main fuse may not have time to rupture before the current ceases.

 

[JohnW2]

I suppose that if the overhead L had been struck, it could then have flashed over to N within the RCD/Main switch

The sensor toriod in the RCD is the first intentional inductive component the supply meets as arrives at the property. For mains AC voltage changing as a sine wave at 50 Hz ( 0 to max in 5 milli seconds ) this inductance creates virtually zero impedance to the current. The current from the lightning strike however increases very rapidly, typically less than a tenth of a millisecond, and for this the toriod presents a significant impedance and the resultant voltage spike will be in the thousands of volts so voltage break down is very likely.

All true, but does that not still beg the question (particularly in the absence of any major damage elsewhere in the installation) as to where the "current from the lightning strike" would have been flowing to, after going through the toroid? Just through the capacitance of the wiring?

Kind Regards, John

 

[bernardgreen]

An old and often effective protection from lightning strikes on over head wires entering a property is to bring the cable down below the entry hole and then do a U turn back up the hole. Then fit an earthed conductor close to the U on the incoming side. The impedance of the U bend at the frequency ( rise time ) of a lightning strike is often high enough to create a voltage spike which results in flash over the to earthed conductor. If flash over does occur then the majority of lightning current goes to ground outside the building.

Sometimes the impedance was created not by a U ( half turn ) but by a spiral coil of several turns.

The U bend also creates a rain drip that will prevent rain water that runs down the cable from entering the building.

 

[JohnW2]

An old and often effective protection from lightning strikes on over head wires entering a property is to bring the cable down below the entry hole and then do a U turn back up the hole. Then fit an earthed conductor close to the U on the incoming side. The impedance of the U bend at the frequency ( rise time ) of a lightning strike is often high enough to create a voltage spike which results in flash over the to earthed conductor. If flash over does occur then the majority of lightning current goes to ground outside the building.

Again, does that not only work if there is a very substantial current flowing through the impedance of that U-bend (to somewhere)?

Kind Regards, John

 

[bernardgreen]

where the "current from the lightning strike" would have been flowing to,

The clue may be

They had a lead water service which was bonded, and after the strike they had muddy water for some time.

Examination of the bond cable may show signs of a massive peak current. Not heat damage as the duration is too short for any significant rise in temperature.

It is not impossible that the lead pipe could have been damaged by the current hence muddy water.

 

[JohnW2]

where the "current from the lightning strike" would have been flowing to,

The clue may be

They had a lead water service which was bonded, and after the strike they had muddy water for some time.

Why should such a current flow through the RCD?

Examination of the bond cable may show signs of a massive peak current. Not heat damage as the duration is too short for any significant rise in temperature.

As a matter of interest, what "signs of massive peak current" (or, indeed, what types of damage) would you expect to see if there had been 'no significant rise in temperature'? We are certainly talking about an adabiatic process, such there would not be enough time for heat to redistribute but AFAIAA any/all damage to the conductor (including melting/'fusing', if we got to that point) would be due to local temperature rises, wouldn't it?

Kind Regards, John