Why do capacitors create a spark when shorted?

[themadquark]

Why exactly is it that a spark is created when capacitors are shorted? Even with a few volts capacitors will spark when shorted, with my limited somewhat basic knowledge this has puzzled me. Any help is greatly appreciated.

 

[Simon Bridge]

They do not spark when they are shorted - but just before they are shorted, when there is still a small gap.
The capacitor sparks for the same reason that spark-plugs do - or any situation where a spark occurs.
http://en.wikipedia.org/wiki/Electric_spark
http://en.wikipedia.org/wiki/Electrical_breakdown

 

[meBigGuy]

I disagree.
"While lower voltages will not, in general, jump a gap that is present before the voltage is applied, interrupting an existing current flow often produces a low-voltage spark or arc. As the contacts are separated, a few small points of contact become the last to separate. The current becomes constricted to these small hot spots, causing them to become incandescent, so that they emit electrons (through thermionic emission). Even a small 9 V battery can spark noticeably by this mechanism in a darkened room. The ionized air and metal vapour (from the contacts) form plasma, which temporarily bridges the widening gap."

Capacitors supply very high currents when shorted.

 

[Simon Bridge]

Please provide the citation to go with the quotation.
I think the passage is talking about the sort of spark you can get lifting a knife switch, or separating the claws on jumper-cables or somesuch - to break an existing electrical circuit. You also get it from brushes or steel-wool contacts.

But here we are talking about shorting a capacitor. i.e. closing a circuit via a very low resistance.

A spark on separation requires some resistance between the terminals - so the discharging takes a reasonable time - so there is still a current to be interrupted. If the resistance is so significant, then the connection is not properly called a "short"... or you'd have to be awful quick.

But let's say this is the case -
Then OK - it is possible to get a spark just before electrical contact and again just when the skrewdriver (or whatever) is being lifted. Like if you whip the contacts with the metal.

There is also the situation where you get a spark shorting a dielectric, lift the skrewdriver, and bring it close again and get a second spark.

For every simple-sounding model there is a complication or exception.

 

[meBigGuy]

No decent citation --- just some guy quoting some guy.
I think the mechanism is actually vaporizing of the contact area. You don't have to separate them, they do so them selves.

Hard to imagine that the arc from 1V would have the dimensions you see when shorting a very large capacitor charged to 1 V. Theoretically you could get a 1V over the air arc, but the dimension is lower than the surface roughness (0.3uM?).

 

[Simon Bridge]

That's right - if you lightly rub two contacts together you get a series of low voltage sparks sometimes - it is imperfections in the surface and bits of dust.

Probably not worth the quotation marks if you are only quoting "some guy quoting some guy".
The passage you have quoted comes from wikipedia.
http://en.wikipedia.org/wiki/High_voltage#Sparks_in_air

Do you have a reference for a 1V arc from a capacitor?
Note: It is not so much the voltage that is important as the gradient of the voltage.
If there is material in the gap which has a lower breakdown potential than air, then a spark is more likely.

 

[dlgoff]

Not the best reference but for switch contacts,

Where the voltage is sufficiently high, an arc can also form as the switch is closed and the contacts approach. If the voltage potential is sufficient to exceed the breakdown voltage of the air separating the contacts, an arc forms which is sustained until the switch closes completely and the switch surfaces make contact.

http://en.wikipedia.org/wiki/Contact_bounce#Arcs_and_quenching

 

[psparky]

To add to this, let's look at the three phase super high voltage powerlines 138KV and above. (or even lower voltages like 30KV...or any voltage really when refferring to open air wires.)

If you look at the towers going down the highways, you will see that the three phases are separated by roughly 10 feet or so. (give or take, just guessing) The proper spacing for powerlines is listed in the codebook.

If a 138KV wire were let's say just 1 foot away from another phase, that arc will jump blowing up the nearest fuse or safety switch, however you want to call it.

These arcs also make huge explosions and destroy anything within a 10 foot radius or so.

Which also leads us into arc flashes in typically worn out switchgear. Again, if there is a high voltage and a gap, and some guy is in there...he is typically either burned badly or dead.

I believe there is a formula that you can get exact arc jump measurements.

Please keep in mind all measurements and voltages above are just approximate guesses to get the point across.

Like they said above, the spark plug in a gasoline motor is the most common arc. You could take a short drive and have over 100,000 sparks just like that. In fact, not to long ago, they used capacitors to create this spark. (then replace with HEI systems and cars now use even more sophistocated, reliable systems to make their spark.

 

[Simon Bridge]

Yes there are lots of ways to get sparks.
http://tesladownunder.com/
... I think the question is well and truly answered and now we are just having fun :)

 

[psparky]

Yes there are lots of ways to get sparks.
http://tesladownunder.com/
... I think the question is well and truly answered and now we are just having fun :)

Absolutely...explaining that was fun.

 

[meBigGuy]

Do you have a reference for a 1V arc from a capacitor?

I'll have to wimp out on that one also. I was thinking I've made them, but now I'm not 100% sure I was dealing with only 1V. Short the bank and get a "flat" spark radiating from the contact point. Years ago. Definitely a "pretty low" voltage.

 

[Simon Bridge]

See if you can duplicate it - a number of mechanisms for a low V spark have been proposed but without the reference there is no way to know which is apropos to what you observed.

I also have not found a reference for the term "flat spark", nor for a spark being described as "radiating".
However, it is not so much the PD between the contacts that counts as the electric field (equal to the gradient of the potential).

i.e. the breakdown potential for dry air, between spherical contacts, is about 30kV/cm ... a 1V PD over 1cm gap is 1V/cm, over a 1mm gap that's 10V/cm or 10kV/cm at micron separations.
If there are sharp-ish protrusions on the contact, like a speck of dust or a tiny scratch in the metal, then the field gets more concentrated about the point. Stuff like dust or fine metal can glow brightly at those sorts of fields too.

But my point here is that even low voltages can result in high electric fields if the distance is small.

 

[meBigGuy]

I'm thinking the high current melts the metal. The heat at a tiny high current connection would be substantial and would cause a disconnecting arc. You even have inductance working for you. You're not buying it?

 

[Simon Bridge]

That's fine - it's one of the available mechanisms discussed before.
Peter Terrin (from just across the ditch from me: link post #9) demonstrates sparks from steel wool.
It's a mix of melting and regular incandescence.

Like I said - there are lots of ways to get sparks.