Hi All

Can anyone shed any light on avoiding inductive spikes when switching inductive loads with a triac. I understand that these voltage spikes can jam triac's open or in worse case scenario damage circuitry.

Any help given would be much appreciated.

cheers

 

RC snubbers are usually used to help suppress this.

http://www.fairchildsemi.com/an/AN/AN-3008.pdf

 

Those spike (if they are the ones I am thinking of) are due to stray inductance. Try to keep all connection to/from a power device as low inductive as possible (either by busbar or extreamly short harness lengths)


This is assuming these are the spike seen across the device at turn off.


Try not to use snubbers if you can get away with it. All they do is shift the losses (since the voltage spike will increase the turn-off losses). Try to reduce it by other means first.

Try improving the layout of the power components - you are usingf decent busbars for the DC-link arn't you?

Next have a play around with the gate-drive. If you can slow it down a bit (at turn-off) you will ok increase the switching losses of the thyristor but you will help it. - do not slow it down loads it still needs to be as fast as possible, just take the edge off it

with an improved power layout and a slightly slower turn-off then start looking at a RCD snubber circuit.

The reason I say try to reduce the spike by other means first is, every reduction in the spike you can do before having to resort to s snubber decreases the size of the snubber

 

Thanks for the replies. Just a couple of questions

In the attached circuit diagram, is C1 regarded as a snubber?

By busbar do you mean connector's and/or wiring or the rail on the circuit board?

Thanks

Attached Images

circuit_3.jpg (23.5 KB, 448 views)

 

You need some noise filtering in that ther circuit.

simply put a diode across the inductive load matching it polarity.

Looks like you got the op amps set up for open loop gain, so any stray voltages will be amplified alot.

__________________
Your website has been blocked

 

Can't do that... The load is apparently an AC load - A parallel diode would cause a short circuit on one side of the AC cycle.

 

His application is with a triac on a AC supply

 

Why is he using a TRIAC in an AC system then - it will not turn off?

 

It will turn off at every half cycle, at the zero crossing point where current will fall to zero (assuming no gate signal). Triacs are specifically designed for AC systems!.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

true BUT what he doesnt say is what the load current is. The thing aboyut TRIAC's and THYRISTORS is there needs to be a significant period of no current flowing for them to stop ocnducting - I just dont think the zero-cossing in itself is long enough for minority cariier decay thats all

 

Yes, because he is using the first 2 opamps as comparators where you need the gain, but for any comparator with slow rising signal inputs, you need some hysteresis to prevent multiple switching's at the trip point, as I have mentioned in one of my previous posts on this subject.

 

I though I posted something about voltage suppression using zeners, hmmm where did it go...Any who I guess I digress, I can't remember all that I wrote to many beers, mabey I didn't hit the submit button??

I do remember that I said I forgot that Traics and such use AC.

Ah well if you think zeners can supress these induction spikes then go here:

http://www.onsemi.com/pub/Collateral/1N6267A-D.PDF

these are unidirectional so they can be wired to deliver in-phase current rated on both sides of the wave.

__________________
Your website has been blocked

 

The device will quickly self commutate once through zero due to the reverse current, with absence of a gate signal

 

That's how they work!, the load current doesn't really affect it.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Thanks to all those that answered