Triggering a triac

[dr pepper]

I'm not new to thyristors and triacs, however I'd like to pose a question.
My current project a power controller uses a bta136 600b triac, these need a higher gate current than the tic xxx series and have been problematic in the past.
My software waits till the phase angle matches the setpoint then triggers the triac, pretty standard so far, my program applies the trigger current then leaves it there untill just before the next zero cross on the ac cycle where its removed.
The trigger angle appears to jitter a few degrees, I'm thinking maybe sometimes the triac stays on for an extra cycle due to gate capacitance or something.
I've seen designs that apply a very short pulse to the gate, only a few uS, and others that aplly a series of high freq pulses.
Which of these 3 would be the most reliable way of triggering an insensitive gate triac?

 

[MrAl]

Hi,

By using a short pulse you can use a higher peak current and still limit gate power dissipation. The pulse doesnt have to be too short though.

 

[dr pepper]

Ok so the idea is to give the gate a good short kick rather than a longer lesser push.
I researched a little and found 100uS seems to be a reccomended duartion, should I set gate current to the triacs max?

 

[KeepItSimpleStupid]

One of the problems is to KNOW just what kid of load you are driving?

Resistive is relatively easy, bit with these CFL's type things, nearly nothing is resistive anymore. It's also ESPECIALLY important when doing phase angle firing.

You could turn on at one voltage zero crossing and suddenly a current zero crossing shows up and you load is turned off. Oops!! For reliability, your best to trigger, burst wise, for an entire period or at least close to one.

Watch the quadrants that the triacs trigger in and the different thresholds.

For higher reliability, back to back SCR's are recommended. I know there is a case where, they should be fired together, but I forget why. I THINK it's a surge thing. Higher reliability might also result in using a triac trigger transformer rather than opto-isolated.

 

[Mike odom]

definitely watch the quadrants. Most triacs don't like to be triggered in IV. The best way is to trigger the same direction as the load current.

If you're leaving your trigger on until just before the zero crossing, you could be triggering an extra cycle due to capacitance loading on the gate. Most of my early triac drivers used an rc for the gate current to ensure it went off. Lately I've been using zero crossing opto isolated triac drivers with triac outputs to get the 'double triac' effect. Although these won't help in your 'power control' circuit.

 

[KeepItSimpleStupid]

Also don't forget the snubbers. Triacs like to turn on if dv/dt of the external load gets exceeded.

In one of my use of an OEM triac controller (Used SCR's back to back), I made sure we used the current limit option (It was driving a variable transformer) and the 3AG (Say 3 Amp) fuse was sized considerably less than the 25 Amp I2t EXPENSIVE fuse. The heating elements were exposed in a vacuum chamber.

I inherited that one. Later I replaced the entire mess with a programmable temperature controller and a DC programmable power supply.

 

[Mike odom]

oh yeah, snubbers, was going to comment on that one also...

a lot of designs call for a snubber on the main triac, and also on the gate triac... make sure your phase angle isn't doing something funny with your gate snubber...

 

[MrAl]

Ok so the idea is to give the gate a good short kick rather than a longer lesser push.
I researched a little and found 100uS seems to be a reccomended duartion, should I set gate current to the triacs max?

Hi,

I dont think i would go with the max, but definitely over the min. Some testing is in order too if you want to be sure.

 

[KeepItSimpleStupid]

If the spaces sya min x; max y; asa process parameter, then their should also be an absolute maximum. There also may be a typical value.

Robust design would say use the max as this is the maximum value that the triac needs to trigger. If you pick a value between min and max then some parts may not work. You don't want to go there in a production environment. Then all your triacs become "selected". Also pay attention to gate resistor tolerances.

Think of sizing the resistor for an LED and using Vf. 90% of the time you can use 20 mA and 2.1 V or so, but don;t count on it.

 

[MrAl]

Hi,

The min is the minimum current required to always trigger the triac. The max is the maximum current the triac gate can take. We dont like to run at max because it's a constant unnecessary stress for the part.

 

[dr pepper]

The snubber is capacitive, only a small amount of phase shift appears to be going on.
The triac seems to 'like' 100uS trigger pulses the most, quad 4 doesnt seem to an issue, and I spose running the device towards its max isnt gonna be so bad with a short pulse.
I'm trying to avoid opto's but then I could just copy a design allready proven if I did.