When is a spark an arc? OR - When is an arc a spark?

An interesting discussion!


In the course of refurbishing my lathe (the one with the defective VSD and DC injection brake) I noticed that it would occasionally stop, apparently because of a voltage drop. I heard a sizzling sound from the enclosure, but because it was difficult to locate it whilst the enclosure was closed, it took a while to find the fault.


Eventually I found the loose connexion in a terminal block, which had been a little singed.


Would an AFD have detected this potentially dangerous fault (in terms of fire)?

Well, I have to say that I am still not sold on the concept. To me, it is a solution looking for a real world problem which is not really significant enough to justify a requirement being written into the wiring regulations to fit one.

It would seem that the ideal test lab generated arc is the only one which will actually trigger this device, and that all other real world faults would be missed.

Given that in order to provide protection, one has to be fitted to each individual circuit, then joe public ain't going to be up for chucking a couple of hundred quid extra per circuit for one either. I'm sure these devices have their place, it's just that it hasn't been found yet!

Time to stop flogging the horse and put it out to grass on this occasion methinks.

Come back when they're a fiver apiece and that they'll actually detect an intermittent fault on a hoover flex.

Until then, forget it.

Yes, I agree with the thrust of your comments Mr Johnson. However, the fact that there is a recommendation in a British Standard for their deployment leaves the designer in an unenviable situation. I can give you lots of examples where arcing has been determined to be the ignition source for serious fires so it is not unreasonable that solutions are sought. Whether the AFDD is one of those solutions is debatable, it would obviously be good if they were!

Well, I'd say the standards probably need to change, and the hubris around how great the  devices we currently have may be in a real setting needs to subside, as demonstrably they are not, until we can show something that is actually useful, reliable, and cost effective.  And as part of that rock solid pass/ fail waveforms and test methods and test gear need to be available.  Otherwise it risks becoming as laughable as the little bell pulls the Victorians put on graves so the dead could ring them and attract attention if they woke up. Or perhaps if significant money is involved, then eventually we risk a mis-selling scandal more like PPI (although hopefully not as bad as the fantastic but cautionary tale of the non-working bomb detectors.) My point is that nice presentation, endorsement by apparent authorities and a slick advertising campaign are no substitute for a useful product.


It might just be that detecting an arc from the current waveform of an unknown and rather arbitrary load, really is a very difficult problem, and there are better ways to invest against failures. Personally  I have been involved in the design of systems fitted  with UV detectors and EMP receiver antennas for detecting early breakdowns and potentially expensive arc / flash-over  failures, and the reason this was done, and we rapidly ruled out the option of  some DSP (Digital Signal Processing) on the current waveform was because this was not adequate, given the great variety of fail modes, and the absence of a catalogue of good and bad waveforms.

I would like to see research efforts being redirected into measuring something tangible.

Let us gather empirical evidence that arcing events in a domestic installation do actually cause houses to burn down.

Then, let us break that evidence down into root causes.

For example, what is the number of arcing events caused by supplier's equipment not being regularly inspected and maintained?

Another example might be what the measurable effects are of the use of steel single screw square cage terminals in consumer unit main switches.

Find the root cause and treat that instead of the symptoms. Then you won't have to fit AFDDs in the first place.

Apparently, all those heath robinson youtubers have been getting it wrong because they have unhelpfully been simulating real world arcing events which these devices won't actually pick up.

 

Come back when they're a fiver apiece and that they'll actually detect an intermittent fault on a hoover flex.

Just a thought ... the device has to be practicable.


But at the same time, it won't be if it nuisance-trips.


The problem is, sparks (and perhaps arcs) are formed all the time in an electrical installation, for example when swtichgear and functional switching operates (light switches, relays in appliances and controls, etc.), and when carbon-brush motors are running. It would be no good at all if the device were continuously nuisance-tripping.


Perhaps the real-world arcing events which have been simulated, in fact share the signature of some of "normal" events such as the ones described above, and this is the reason they are not detected by an AFDD.



So, the question to answer is not whether a so-called "heath-robinson" sparking/arcing test operates the device, but whether there are events which are detected by devices to BS EN 62606, which will save lives.

So this video posted this evening is interesting.

Would (or should) an AFDD have picked up this issue earlier?


Start at about 3:29 for the background or 6:00 to get straight to it.
https://youtu.be/xveGg0C_Q88?t=3m29s

A few years ago I went to a house and installed a completely new cooker circuit because the people who lived there wanted an electric cooker instead of a gas cooker.


So I installed the new circuit, then a couple of days later had the lady of the house on the phone ranting that I had messed up their consumer unit, because her husband was trying to use the cheap arcing sparking chop saw he had bought from a DIY store in the garage.


I said it was absolutely nothing whatsoever to do with me or the work I had done, but I would call back around to have a look at it.


The garage had a double socket outlet next to the consumer unit, the only sockets on the circuit, so I swapped the B16 MCB for a C20, problem sorted.


There was still a problem from my point of view, the lady had absolutely no intention of paying me for resolving what she thought was a problem I had caused, when it was absolutely nothing to do with me. Materials and labour should have been around forty or fifty quid with the call out.


So the idea of persuading customers to pay hundreds of pounds for a new consumer unit with devices designed to trip if there is an arc that my actually trip when they plug in cheap electrical appliances is not over appealing.


Andy Betteridge