Contact sparks

sophiecentaur

As has been mentioned already, this sounds just like a petrol engine ignition system - using spark plugs (which last for millions of strikes). But if you want a low voltage arrangement, the problem is different.

If you could cut off the current each time, before the contacts weld, that may achieve what you want. This wouldn't be too hard to achieve with a pulse from a fat power transistor in the circuit, to control the current. The pulse length would need to be increased gradually to maximise the energy supplied to each strike of the spark.

Alternatively, if you look at the way the old carbon arc lamps (cine projectors) used to operate, you might get some ideas. The arc was struck by bringing the carbon electrodes together and then separating them until the current was what was required. There was then a servo'd system, involving a motor which slowly brought the carbon rods together at the appropriate rate to maintain that current as they eroded.

Because the arc that you need is a non-linear resistance, I think you might consider controlling the separation of the electrodes in some way. This would not be difficult - just not what you had in mind - involving a solenoid, which could achieve the movement. If you don't actually separate the electrodes, once the arc has struck, you will risk the contact welding together permanently (for fat electrodes) or melting so that the spacing is too great (for a thin electrode) for another arc to strike. This is exactly what you do when you manually 'dab' two wires together to stimulate a spark.

You will need to think 'laterally' for the best, suitable solution.

maimonides

I worked with a test rig for circuit breakers as an intern long ago. I don´t remember wether it was pneumaticlly driven or spring loaded; the electrical part was a capacitor discharge triggered by a SCR.
You could scale that down for "household use". Triacs and SCRs have pretty good pulsed current ratings (rule of thumb: 10x continous current) and you can easily get 25A continous or higher rated types. Use some big relay and trigger the relay and the scr at the same time. If you start with a 30V 5000u capacitor you´ll probably see some effect.
(You´ll want to add some charge control for the capacitor, a current limiting resistor and a few other things, but you get the idea).
As has been mentioned, relay contacts are not designed for massive sparking/arcing. This is why the current and voltage rating for non-ohmic and DC loads are considerably lower than those for ohmic AC loads. (The zero crossing in AC will extinguish the spark). So you´d better expect frequent failure, unless you use a higher powered mechanical drive.
I don´t see what you want this for. Is it for a millisceond spark or a continous arc? The ignition system in a car has become very much more reliable with the elimination of mechanical elements, so why not use a proven system?

Curl

I want just a milisecond spark for an experiment I'm doing. I'm done trying the "car ignition" way because I don't want to carry around a 20lb ignition coil. I want something that's small and I just want some sort of spark, doesn't have to jump through air or anything it just needs to work repeatedly so I can run the experiment over and over.

Any ideas on materials? If I build some triangle/cone type electrodes out say, brass, and touch the tips, then discharge a rather large capacitor on that. Would that work?
I'm not sure how carbon electrodes would work in this case, it seems to me like they wouldn't make a spark at all. I'm more interested in what makes the "spark" in the first place, is that just hot metal being eroded out by the large current or what?

Bobbywhy

It is always useful here on Physics Forums when the original poster describes in the most detail possible what she wants to do. It saves members from guessing and speculating. Thank you, Curl, for making clear what you are asking for. May I suggest you use a simple voltage multiplier driven by a simple oscillator? Here are some references:

http://en.wikipedia.org/wiki/Voltage_multiplier
http://www.kronjaeger.com/hv/hv/src/mul/
http://www.allaboutcircuits.com/vol_3/chpt_3/8.html

Curl

Ok that is a good suggestion. The only voltage converters I learned about can only step up the voltage 20x practically but these seem better. The problem is that I don't have an oscillator, and even with one, how would I use this voltage multiplier to achieve what I want? I are you saying that I can step up a voltage enough to jump a spark gap?

Danger

You can do it the way that I did as a child. Get an ignition coil made for a 60's or 70's car and wire it in series with a doorbell buzzer, using a motorcycle or power drill battery for power.

maimonides

The capacitor discharge method will only work if you open the contact while current is flowing. You need a drive for that. (and some kind of synchronization)
If you are only after a spark: why don´t you use a piezo gas lighter? Much simpler than everything we discussed so far.
I also suggest you recheck the weight of an ignition coil used in a modern car.

Bobbywhy

If you read the three references provided you would discover answers to all the above doubts/questions.

sophiecentaur

There are so many ignition systems available for small petrol engines - lawn mowers, mopeds etc. We're not talking "20lb". These things are not much bigger than a box of matches. Scrap yards must be full of them. All you need is a 12V supply and an input to replace the timing signal from the flywheel.
And what's wrong with using a spark plug for your spark gap? They use just the right material for the job and you won't ever manage to melt it.
The only reason that it might not be suitable would be if you needed a lot of energy in your spark. You haven't mentioned how many J you want.

CWatters

There are even spark ignition systems for radio control model aircraft.

PaulS1950

The circuit that you are desribing sounds like an ignition source for a flammable or explosive device. I won't help with that - sorry

Paul

Integral

Curl,
Have you researched old school automotive ignition systems? They do exactly what you want. you just need the battery, coil and points from the distributor.