Triac Pulses for AC switching

[dknguyen]

I don't have much experience with triacs. Suppose I am using an MCU controlled triac to switch AC current. I am NOT using phase-control, but rather am using the other kind...I don't know what it's called. It's where you let a continuous AC wave pass for a periodic intervals amount of time and then block it for every other periodic interval. I think it's called zero-cross control/switching? I forget.

Why is a momentary pulse ALWAYS used after every AC zero crossing to trigger the triac rather than a continous pulse throughout the half cycle? Sure, this is all that is required since the triac's conduction state will be self-sustaining as long as the AC signal has the same polarity as when the triac was switched on, but can't a continous pulse that remains active across AC half cycles also be used?but can't a PWM signal (that remains active during entire AC half cycles) be used to trigger the triac? The conduction time of the triac for a single PMW cycle would the active time of the PWM pulse plus the time until the next AC zero crossing (due to the 60Hz of AC powerlines and the self-sustaining triac conduction).

Is there some overbearing heat dissipation problem I am completely unaware of if a continous pulse is applied throughout the half cycle rather than just at the zero crossing?

 

[Hero999]

What are you controlling?

What you want to do is called burst control, and is no good for lamps or motors, it's only useful for heaters.

Zero crossing is used because it elliminates noise by not switching when current is flowing. The reason the pulse needs to be sustained with burst control is because it spans multiple cycles otherwise the triac would turn of when the cycle passes through 0V.

 

[dknguyen]

I'm controlling a resisitive heating element in an oven. I think you misunderstand my question. I don't mean sustaining the repetition of pulses, I mean actually sustaining single, long pulse so that it holds the triac gate high through 100% of the conduction time, rather than sending a trigger pulse after every zero crossing.

I know all is needed is a quick trigger pulse at the beginning of each half cycle, but what if I extended this pulse to be on throughout the ENTIRE half cycle(s). Is there some problem with this simpler, less efficient approach that I am not aware of? (Everyone seems to use repeating burst pulses after the zero cross. No one, and I mean no one seems to just bridge all the pulses together to make a long pulse that is always on throughout the half-cycles).

 

[Nigel Goodwin]

I'm controlling a resisitive heating element in an oven. I think you misunderstand my question. I don't mean sustaining the repetition of pulses, I mean actually sustaining single, long pulse so that it holds the triac gate high through 100% of the conduction time, rather than sending a trigger pulse after every zero crossing.

I know all is needed is a quick trigger pulse at the beginning of each half cycle, but what if I extended this pulse to be on throughout the ENTIRE half cycle(s). Is there some problem with this simpler, less efficient approach that I am not aware of? (Everyone seems to use repeating burst pulses after the zero cross. No one, and I mean no one seems to just bridge all the pulses together to make a long pulse that is always on throughout the half-cycles).

It's common to just hold the gate ON as long as you want power to be ON, it's not usual to trigger every cycle for burst fire, there's no point. It's not always even done using zero switching, as it's only turned on at fairly long intervals there's much less scope for interference (as opposed to triggering every cycle).

 

[philba]

I agree, I've never seen what you described. I've used zero crossing opto/drivers so there is no interference at all. I've used the moc3033

 

[Hero999]

I'm controlling a resisitive heating element in an oven. I think you misunderstand my question.

A perfect application for burst control

I don't mean sustaining the repetition of pulses, I mean actually sustaining single, long pulse so that it holds the triac gate high through 100% of the conduction time, rather than sending a trigger pulse after every zero crossing.

I understand, you're talking about holding the gate on even though it's already triggered.

I know all is needed is a quick trigger pulse at the beginning of each half cycle, but what if I extended this pulse to be on throughout the ENTIRE half cycle(s). Is there some problem with this simpler, less efficient approach that I am not aware of? (Everyone seems to use repeating burst pulses after the zero cross. No one, and I mean no one seems to just bridge all the pulses together to make a long pulse that is always on throughout the half-cycles).

Because it's easier to keep the gate high, it makes the design far more simple. Once the triac is triggered it makes no difference whether the gate's held high or low so it's normally more convenient just to keep it high, rather than messing around designing a circuit to take it low once it's triggered.

 

[dknguyen]

Okay. It just seemed in the all the app notes that they were using a very short trigger pulse at the beginning of each zero crossing that would turn off for the rest of the half-cycle. I was just wondering if there was a practical reason for this or if it was to demonstrate the self-sustained conduction of the triac, as opposed to keeping the pulse high all the time.

It's common to just hold the gate ON as long as you want power to be ON, it's not usual to trigger every cycle for burst fire, there's no point.

Yeah, okay. It's just from app notes you would get the impression that trigger pulses every half cycle are the norm, rather than just holding the pulse for as long as you want to conduct.

 

[zevon8]

Sometimes engineers will do things just because they can, or they are bored. Just a quick question tho, you're not looking at the gate current pulse are you? It repeats every cycle, and is commonly shown on the graphs.

Often repetitive pulse control is used because it can be easy to generate with a tap off of the mains voltage. This is common in industrial electronics. Resistor, pot, cap and diac. Synchronized pulses and variable firing. For coarse control, as in your application, or as is done in many microwave ovens, etc, holding on then off is the usual way.

 

[dknguyen]

Sometimes engineers will do things just because they can, or they are bored. Just a quick question tho, you're not looking at the gate current pulse are you? It repeats every cycle, and is commonly shown on the graphs.

Oh yeah, look at that. I AM looking at the gate current. I think they just do this because the triac is triggered by current rather than voltage, and aren't these current trigger pulses pretty much representative of the applied gate voltage required to driive that trigger current into the gate? Or are they actually applying a voltage throughout the half-cycle and the conducting state of the triac actually prevents anymore gate current from flowing into the gate?

 

[Nigel Goodwin]

Yeah, okay. It's just from app notes you would get the impression that trigger pulses every half cycle are the norm, rather than just holding the pulse for as long as you want to conduct.

The only reason to do that is for phase control, it's possible in some cases that they use the same hardware for both?, in which case you would trigger every half cycle - but for just burst fire there's no need or advantage to doing so.

 

[stevez]

Is it reasonable for me to think that I could use what I'll call PWM, that is not synchronized to the AC, to control the flow of AC via a triac? It might not be ideal - but might serve adequately in some cases. I haven't given it much thought but I'd be thinking about AC at 50 or 60 Hz and pulse durations that would span mutliple cycles. It seems that I had the same impression as others - that things had to be switched each half-cycle.

 

[Nigel Goodwin]

Is it reasonable for me to think that I could use what I'll call PWM, that is not synchronized to the AC, to control the flow of AC via a triac? It might not be ideal - but might serve adequately in some cases. I haven't given it much thought but I'd be thinking about AC at 50 or 60 Hz and pulse durations that would span mutliple cycles. It seems that I had the same impression as others - that things had to be switched each half-cycle.

The only reason to switch ON at zero crossing is to prevent interference, with burst fire you only switch on occasionally, so it's not so big a problem. If using burst fire you also only need to switch ON once, at the beginning of the burst, and leave it turned on as long as you want it - once you turn the gate drive OFF, it will turn off automatically at the next zero-crossing point.

There's no need, or reason, to send a pulse every half cycle.

 

[stevez]

I have no practical experience with triacs. This thread gives me the opportunity to ask a question.

In what I'll call a traditional or common, simple triac circuit it appears to me that the triac isn't on or can't be on 100% of the cycle because the diac has to fire first, to turn the triac on. Is my understanding correct?

 

[dknguyen]

Is it reasonable for me to think that I could use what I'll call PWM, that is not synchronized to the AC, to control the flow of AC via a triac? It might not be ideal - but might serve adequately in some cases. I haven't given it much thought but I'd be thinking about AC at 50 or 60 Hz and pulse durations that would span mutliple cycles. It seems that I had the same impression as others - that things had to be switched each half-cycle.

That's what I wanted to do, user existing PWM functions on my uC-using just a continous pulse for as long as I want conduction (rounded to the next half-cycle).

 

[Nigel Goodwin]

I have no practical experience with triacs. This thread gives me the opportunity to ask a question.

In what I'll call a traditional or common, simple triac circuit it appears to me that the triac isn't on or can't be on 100% of the cycle because the diac has to fire first, to turn the triac on. Is my understanding correct?

The diac is only used in phase shift control, no need for it with burst fire.

 

[stevez]

I'll adjust the question a bit - when a diac is used, is the triac "on" for the full half cycle when the control is adjusted for maximum power or voltage - or is there a short delay before the diac is able to trigger the triac to an "on" state?

 

[dknguyen]

I'll adjust the question a bit - when a diac is used, is the triac "on" for the full half cycle when the control is adjusted for maximum power or voltage - or is there a short delay before the diac is able to trigger the triac to an "on" state?

I am not sure why you seem to be interchanging diac and triac in the same sentence, diacs and triacs are two different devices...maybe I am misunderstanding something...you're using a diac to trigger a triac? I'll just assume you are talking about something like that.

There's always a delay when anything happens, even when switching a MOSFET on and off...just depends on how short it is. So from what I can gleam from your words, for all intents and purposes the triac is on for the full half cycle...unless you are switching a sinusoid that has a REALLY frequency approaching the switching propogation delay of the diac-triac, which I think is in the us or MHz range. I would think it wouldn't be a problem in most cases. I don't know of any power application where a MHz or even kHz AC is used.

 

[stevez]

One more adjustment, for the sake of understanding.

I have never designed or built circuits using triacs. I have seen what I'd call relatively simple designs where a diac is used with a variable resistor and other components - along with a triac. Why the circuits are designed this way is unnknown to me.

It appears to me that as the AC waveform goes from zero to some point (determined by the variable resistor and possibly the diac characteristics) the diac will fire which in turn, turns the triac to an "on" state - until the AC voltage drops to zero and the process repeats itself. Again, that's what it looks like to me.

If I am correct in my understanding then it would seem that the triac would not ever be "on" for at least the period it takes for the AC to rise high enough for the diac to fire. I wasn't really thinking about propogation delays though that certainly is a part of it.

I can see where a typical schematic would be helpful but I don't have one at hand. The light dimmers I've had at home adjust nicely from nearly zero light output to some point near full brightness - then there is a jump to full on where I presume that the switch simply bypasses the triac (that I presume is there but dont' know that for sure).

 

[zevon8]

Yes you are pretty much correct stevez, I will add the following:

Triacs often do not play nice in that conduction for each half cycle is not symetrical. Turn on in either direction may lead or lag each other slightly. This can cause all sorts of grief, least of which is electrical noise. Depending on the load, it can be fatal for the triac, exceeding dv/dt ratings. Diac's have a threshold voltage at which point they conduct heavily, ususally between about 30 or 40 volts. This is why they are used in phase control, it helps guaratee a switching point for the Triac, providing a gate pulse at the same level in both polarities.

Burst fire is simple since phase control is not an issue, you're basically just making a solid state relay or switch. Power on, power off, with seconds or minutes between each. Once you get into complex loads, or desire accurate control of the power you need phase control.

Once you get into really serious power, you walk away from Triacs anyways, and go with SCR's in antiparallel and use descrete drive control.

 

[dknguyen]

I wasn't really thinking about propogation delays though that certainly is a part of it.

Sorry, I should have said the switching time of the triac/diac is in the MHz or us range. THe propogation delay is more general and includes these switching times as well as other delays.