# Looking for information on triacs

I’d like to learn about function and use of triacs. Namely to replace spots where a electromechanical relay would be put.
I’m afraid a lot of the stuff I’ve seen/ tried to read in regards to triacs isn’t registering in my brain. I’m guessing it’s not so simple as just being a mosfet for AC (put a signal at the gate, triac opens, the end).

Namely, I have some of these

and I’d like to control them with an mcu (Arduino, ATTiny, ESP8266, etc).

What kind of signal does the ‘gate’ take? How do the grounds connect if the mcu is on a DC circuit and AC is going through the triac?

Not necessarily looking for a schematic to copy, I want to learn and understand what goes where and why. I need kindergarten level explanations. Point me to resources or if this is actually simple enough to explain here, please do.

INTP:
Point me to resources

You might want to study how an SCR works first. (An SCR works in one-direction... DC.) A TRIAC can be triggered with a positive or negative voltage on the gate.

SCRs & TRIACS are non-linear, switching, latching devices. Once "triggered", you can remove the gate signal and current continues to flow until something else stops it. With a TRIAC in an AC circuit, current flows until the next AC zero-crossing.

When used on AC power line voltages with the Arduino, the Arduino has to be optically isolated from the line voltage. There are [u]special opto-isolators[/u] designed for TRIACS, and there are two basic types... There are normal random-phase opto-isolators for dimming/speed control. If you are making a dimmer, you also need to detect the zero-crossing and that signal has to be opto-isolated or transformer-isolated too.

And, there are special zero-crossing opto-isolators that will only turn-on as the AC passes-through zero. Zero-crossing isolators reduce switching noise on the AC line.

Namely to replace spots where a electromechanical relay would be put

Most AC solid state relays are made with optically isolated TRIACS, although some are made with back-to-back, complimentary, MOSFETs. So, if you try to use an AC solid state relay with DC (and if it's made with a TRIAC) it will latch-on and then won't turn-off.

You would find it easier to use premade and tested 'solid state relays'

Allan.

DVDdoug:
With a TRIAC in an AC circuit, current flows until the next AC zero-crossing

Or until the current drops below the TRAICs holding current.

allanhurst:
You would find it easier to use premade and tested ‘solid state relays’

Allan.

And safer for yourself and the arduino if you are inexperienced.
Look for optoisolated devices.

DVDdoug:
So, if you try to use an AC solid state relay with DC (and if it's made with a TRIAC) it will latch-on and then won't turn-off.

I'm reading that a TRIAC will continue conducting even if the gate signal is removed. What kind of signal is needed to stop the TRIAC conduction? Is this a specifically timed event at 'zero-crossing' as the only chance to turn off the TRIAC?

In this circuit (though I'd leave out the snubber bits for just a bulb), I'm guessing it uses both the optoTRIAC as well as a second TRIAC since the optoTRIAC would have a limit of say, 1A over the 4&6 labelled pins. (Is the optocoupler considered an optoTRIAC in itself or no?)

But what if I'm controlling one of those LED bulbs that replace incans? E.g., a 60W equivalent uses 9W. That's well below a 1A current (9W / 120V = 0.075A).

Could I simply just use the optoTRIAC and leave out the TRIAC on the right side of the pic altogether?

1/ Thed only way to stop a triac conducting is to stop the main current flowiing through it. At a zero-crossing, for example.

2/ An opto-triac is a low power device and won't handle big currents - look at the datasheet.

Allan

1/ Thed only way to stop a triac conducting is to stop the main current flowiing through it. At a zero-crossing, for example.

Nearly.

There are two criteria,

1. The current must be below a drop out current
2. The voltage must be below a drop out voltage
These two must happen at the same time to make the triac stop conducting. Sometimes with inductive loads the phase angle between current and voltage is such that these two conditions do not happen at the same time. In this case the triac will not turn off and once triggered will stay permanently on. The solution is to use snubber circuits to get the phase angle close enough to prevent this happening.

Could I simply just use the optoTRIAC and leave out the TRIAC on the right side of the pic altogether?

If the opto triac can handle the current then yes.

Am I correct in concluding that an optoTRIAC or SSR cannot replicate an electromechanical bistable relay in that it draws 0 current, only needing a short signal to change states?

In other words, TRIACS (optos, SSRs) need a constant current to hold a state just like electromechanical non-latching relays?

In other words, TRIACS (optos, SSRs) need a constant current to hold a state just like electromechanical non-latching relays?

No.
While you can apply a constant signal to a gate and it will stay on, you only need a short pulse to turn it on. Then it will only turn off under the conditions I said in reply #9. This happens naturally with AC every cycle, but you can also use SCRs in DC circuits as well.

In the past ( the 60s ) I used them in latching circuits that turned something on ( normally a relay ) with a pulse and kept it on until some other signal commutated the current flowing and turned it off. These sorts of things were used in industrial control.

Grumpy_Mike:
While you can apply a constant signal to a gate and it will stay on, you only need a short pulse to turn it on. Then it will only turn off under the conditions I said in reply #9. This happens naturally with AC every cycle, but you can also use SCRs in DC circuits as well.

Thank you for the help thus far. Are you saying the TRIACs turn off every zero crossing on their own once the gate signal has dropped below voltage and current threshold?
I'm not understanding how it's latching, because it sounds like, yes, it 'latches' for that 1/120th of a second before actually shutting off, which isn't very useful.

Or are you suggesting to get the latching action, I'd have to purposefully throw in an inductive load to offset voltage and current on the AC so they're not synced, and have some way to sync them every time I want to turn it off?

INTP:
I'm not understanding how it's latching...

You see two transistors on top of each other.
When the lower transistor gets base current, it's collector current starts to flow,
which provides the top transistor with base current, so it's collector current starts to flow,
now the lower transistor gets base curent from the top transistor.
They both keep each other in an ON state, untill power is removed.
Leo..

Are you saying the TRIACs turn off every zero crossing on their own once the gate signal has dropped below voltage and current threshold?

Yes exactly that.

yes, it ‘latches’ for that 1/120th of a second before actually shutting off, which isn’t very useful.

That is only your opinion, it is I fact very useful. Not only with DC circuits like I said above but also with AC circuits which is their most common use. For example delay the turn on by a bit from the zero crossing point and you have a phase control. Anyway that is the way these four layer devices work.

Or are you suggesting to get the latching action, I’d have to purposefully throw in an inductive load to offset voltage and current on the AC so they’re not synced, and have some way to sync them every time I want to turn it off?

Not a bit of it. That is normally seen as a fault condition.

Wonder if applying a 10kHz 10% duty cycle pulse stream to the gate would cause flicker?

edgemoron:
Wonder if applying a 10kHz 10% duty cycle pulse stream to the gate would cause flicker?

On what circuit?

If you mean cause it to conduct and then not then no. Triacs don't give a flying fig what happens to the gate once it is conducting. And just one cycle of 10% 10KHz is enough to turn most triacs on. But see the data sheet for the real minimum pulse width.

Grumpy_Mike:
Triacs don't give a flying fig what happens to the gate once it is conducting.

This bit is what's tripping me up. I assume you mean "doesn't give a flying fig what happens to the gate [as long as it stays above voltage and current thresholds]"?

If I have a MOC3021 and a BT136 TRIAC, then anytime the MOC gets DC through the sending side of the optocoupler, the TRIAC turns on, but when that signal stops, TRIAC turns off. It's a momentary switch action.

Is the latching mentioned above only specific kinds of TRIACS (two-transisitor model?) or is there something I can do with these components to get latching?

INTP:
If I have a MOC3021 and a BT136 TRIAC, then anytime the MOC gets DC through the sending side of the optocoupler, the TRIAC turns on, but when that signal stops, TRIAC turns off. It’s a momentary switch action.

The triac will turn off when the load current through the triac is too small, or the supply is interrupted (zero crossing).
Power regulation with a triac is done by turning it ON at a certain time of the sine wave.
Leo…

If I have a MOC3021 and a BT136 TRIAC, then anytime the MOC gets DC through the sending side of the optocoupler, the TRIAC turns on, but when that signal stops, TRIAC turns off. It's a momentary switch action.

No that is not right. The bit "but when that signal stops, TRIAC turns off." is not what happens. Once the gate gets a triggering signal and the Triac is on that is it, it stays on until the current & voltage
FLOWING THROUGH THE TRIAC ( not the gate ), drops below the holding points.

Is the latching mentioned above only specific kinds of TRIACS

No it applies to ALL triacs.

I assume you mean "doesn't give a flying fig what happens to the gate [as long as it stays above voltage and current thresholds]"?

Once triggered you can remove any signal from the gate or keep it on, the result is the same, the Triac is on.