Spurious Arcing

This is not the same as accidental arcing, which is when two contacts are too close. This category of arcing is as a result of contacts not close enough or as a result of loose connections. The capture shows a high frequency burst superimposed on the main sinewave when a light switch was operated.

Close inspection revealed that the point where there should have been contacts, that these had disappeared altogether and the switching was done purely by 'flat' copper faces. This was in itself a symptom, the cause being the switch controlled a string of compact fluorescent lamps (CFLs) and the inrush currents had burned the contacts away.

Generally speaking low incidences of this type of fault tends not to cause damage as such, and will manifest as little more than communication disturbances if anywhere near comms lines.

In extreme cases, this type of fault could cause computer crashes and reboots (purely through loss of power if the switch or loose connection happens to be part of the power feed). In persistent cases, premature component failure can result from subjecting them to multiple impulses and in-rush currents.

Power factor correction and/or EMC filter capacitors are especially vulnerable as inductors are present in both instances. With a loose connection on the power feed, the inductors induce huge back-EMF voltages across the capacitors causing them to break down. If the breakdown is permanent (short circuit) then it has been known for such capacitors to voice their distaste at this treatment by blowing up.

It is not just the back-EMF voltages that are of concern here. Just simply discharging the capacitor and then "force feeding" it the incoming supply voltage over and during many cycles will cause excessive currents to flow in the capacitor again stressing the material with possible short circuiting as a result.


Another name often given to loose connections or bad contacts is 'hot-spots'. Many believe that hotspots are simply as a result of high contact resistance which, in turn, leads to high power loss (P=I²R) thus creating heat. This is not erroneous thinking, but there is another aspect too.

What appears to be less known regarding hotspots is they too have a significant level of spurious arcing. A voltage drop across a contact of less than a volt is hardly going to be noticed against 230V so such arcing tends to go unnoticed.

Monitoring across such a contact with a suitable instrument (rather than measuring the output after the contact) reveals a large amount of 'noise' emanating from the contact as a result of the arcing. Such measurement are dealt with later.


On single phase systems, having a loose connection can in extreme cases result in failure of electronic components. However, on 3-phase supplies this becomes even more critical - in fact, it becomes so critical that 3-phase systems should come with a fire risk warning!

If single-phase loads are connected to a 3-phase system, then the same level of impact on the single-phase devices can be expected from any loose connections on the phases of the system (the Neutral is an exception to this, read below).

With 3-phase devices things are a whole lot different! Even a simple device such as a motor is highly influenced by even the smallest of voltage drops on one of the phases. It causes phase imbalance and the motor will immediately attempt to draw all the required energy from the two remaining phases with the distinct possibility of burning out.

This fire risk is not limited to inductive devices such as motors and transformers, but extends to electronic devices too. Items that use 3-phase rectification will have the diodes on the 'solid' phases doing all the hard work (getting hot under the collar as a result) while those on the faulty phase taking a rest (except they may have to deal with back-EMF spikes!).

Loose Neutral:

And this is where the 'fun' can really begin. If loads are beautifully balanced, then life would probably carry on regardless of a Neutral making its way back to the source or not. But life is not like that, and perfectly balanced phases are nothing more than a myth.

In every explanation I have seen as to the effects of a loose Neutral, the writer invariably states that the most loaded phase will tend towards a lower voltage with the least loaded phase being subjected to the highest voltage and therefore the greatest damage on this phase from the over-voltage. This is sometimes true, but not occur as often as you would think!

It is not a well known fact, but a loose or disconnected Neutral can actually result in all phases suffering over-voltage. I have dealt with a number of cases and in most there have been reports of over-voltage damage on all the phases, not just the so-called "heavily loaded" ones.

It's not just my own cases, but have also had many surprised folks call me up asking what the problem is as they measure 400V between phases, but are measuring about 265 or higher Phase-Neutral (they're expecting about 230).

The cause is the phase angle of the current on the separate phases is not co-incidental with the voltage (i.e. perfect balance is not present). It only takes a slight power factor difference between two phases for a large voltage to occur on the loose or disconnected Neutral.

Note, power factor difference! I mentioned nothing about whether one was leading and the other lagging. Simply that one phase is more capacitive/inductive than the others. Revisit "artificial consumption" for a full explanation on the import/export occurrences within a cycle on a load with power factor less than unity.

As there is no Neutral to do this energy "lending and borrowing" with the source, such actions now occur via the other phases. Therefore, all phases are involved with the import/export energy transfer. And if this coincides with the voltage peaks, then all phases suffer over-voltage.

What's even worse is if there is a distinct capacitive power factor on one phase and a distinct inductive power factor on another. The capacitance and inductance forms a series tuned circuit. If this is resonant on an available excitation frequency (fundamental or harmonic), the junction, which in this case is the Neutral, can reach hundreds of volts!

And the reason we should be so concerned; In buildings there is usually no lack of fuel, and oxygen is in full supply. Heat is the only remaining element of the 'fire triangle' not in abundance. Not any more it ain't!

If there is ever a fire risk in a premises fed by 3-phase power, it is a loose Neutral. Devices fed on the phases that are subjected to an increase in voltage will usually attempt to rid the extraneous power being shoved up their mains ports in the form of this previously missing element.

With components such as fluorescent lamp ballasts turning in to heaps of molten plastic and fizzing copper windings, impulse arrestors now absorbing extraneous energy and bursting open like a box of fireworks that had a match thrown in, and electronic components going red in the face with the extra working voltage, heat is no longer in short supply!

One case revolved around a faulty Mains-Generator changeover switch, and I didn't even flinch when I shut the building supply down. Although the client was, at first, highly annoyed, his attitude soon changed to one of acceptance when faced with the alternative of explaining to management why he gallantly pursued the day's takings of a few hundred Pounds but allowed the building to become a heap of cinders!

Not only that. I too could have faced court proceedings with possible professional and public liability claims if I had allowed them to continue to operate in light of the shed-load of burnt-out fluorescent fittings, umpteen failed power supplies (all molten), and, to top it all, even burned building wiring providing evidence of this loose Neutral!

With all this evidence, it is believed that spurious arcing remains the number one cause of power quality issues facing modern industry!

Yet it still amazes me why so few power quality surveys are done (but, then again, one needs to know the survey is being done correctly too!).


| | Ask a Question |

© 30.07.04 / 31.10.04