I believe that it can be done but triggering ON both triacs may become trickier than you might think. For if the gate current for triac A decays away before triac B starts to conduct then you'll not have any conduction for that cycle.

Hello Redfred,

I think that I did not said it right. Only one triac is used. But two opto-couplers connected in series is connected to the gate of the triac. Triac's Vdrm is 1200V so for a 660 V application one triac will be sufficient.

Thanks for the response.

Job.

Sometimes I try to be too subtle in my answers here. I forget that often an engineer of another discipline will be looking for guidance in electrical engineering here. Forgive me.

If what you are doing right now on this design is considered a "one of" prototype, not to be repeated, using available parts to see if your idea works. Then I say "go for it" and put two of these TRIAC optocouplers in series to provide sufficient breakdown voltage protection. Make sure you have the correct current protection (circuit breaker, fuse, thermal breaker) in your load current path in case things go wrong. For a final, production design you might consider a solid state relay package, MOSFET drivers, bigger TRIAC device to control your load. Without knowing your anticipated load, I cannot recommend which configuration will likely be your preferred topology.

Snel here recommended another plausible solution of just dividing down with an attenuator the voltage this optocoupler will see in triggering the next part in your chain that actually controls your ultimate load. Without knowing the whole circuit though I cannot recommend any topology or components. The second triac (TPDV1240) you propose using in your reply to Snel is a beefy 40A capable device. You're apparently trying to control something with a lot of power. Component cooling and switching times may make this trickier than you think.

Why don't you try it, A Vdrm of 800 V might work OK.

Try with a dummy load.

Hi Garth,

We have to consider peak of 660 X 1.05, at least, which will be 980 Volts across the triac driver whose Vdrm is only 800 V. Incidentally, I am a mechanical engineer who has only negligible knowledge in electronics and that is why I am looking for help from some experts in commutation of triacs.

Thanks for trying to help.

Regards,

Job

Hi thankan,

just to check that we understand your question correctly:

You say MOC3083 is specified to be used in conjunction with triacs having a Vdrm up to 800 V. This means the driver can take the corresponding peak voltages as a power source (just for the driving circuit). The gate will be driven correctly by the MOC for alternistors with Vdrm>800 V, so you don't need to to connect MOCs (or drive circuits) in series. Just step down the feeding voltage for the MOC (a voltage divider).

If this is what you need, we can go into design details.

brgds

Snel

Hi Snel.

I will be using a TPDV1240 alternistor whose off-state voltage is 1,200V (Peak). The application is for switching 660VAC. When not in conduction, the peak voltage across the MOC3083 will be 933 V (Nominal), which is 133 V above the specified off-state voltage of the MOC3083.

A single MOC3083 will be good in application up to 480 VAC (678 Vpk).

Regards,

Job

Hi all,

just checked the data sheets now. I had expected the MOC to require an external capacitor but it's all integrated (zero-cross driving means you have to generate a current pulse when feeding voltage = 0, so you have to store energy somehow; not so with a simple optocoupler). So connecting 2 MOCs in series does not seem advisable, you would need balancing resistors (pin 6 to 4) which in turn would mean a leakage current thru the triac gate (bad for reliability).

So start with figure 8 on the Motorola data sheet: hot-line switching typical app ckt. Replace R=360 ohms (from HOT and A2 to pin 6) with 1000 ohms. Add another R=470 ohms from pin 6 to A1. Take care not to invert A1 & A2 on the alternistor. You will build up a voltage divider of around 3:1 or 660:240, with an equivalent resistance close to the original 360 ohms as "seen" by the MOC.

MOC can drive 1 A vs. 200 mA required by the alternistor (OK; internal capacitance is low inductive, a steep driving pulse is desirable - BTW, keep the gate wires short and don't twist them). I assume yours is a resistive load, if you need a snubber use a low inductance resistor, 39 ohms in the example. An alternistor (2 chips) is supposed to be more reliable than a triac (1 chip). 3-quadrant-only is a good sign, too.

Start with a lighter load and good luck!

Brgds

Snel

Hi Snel,

It is a good thought; but I think that there is a small problem. MOC3083 is a Zero crossing optocoupler which will turn ON only the line voltage is below 20volts. This means that your gate drive will only be 20 V / 1K = 20 mA max. The Trigger requirement for TPDV1240 is 100mA. So in order for triggering the alternistor, the resistor connecting MT2 to # 6 pin of the triac coupler must be as low as 20 ohms. I can be wrong.

Regards,

Job.

Hi thankan,

you're right, but the max current would be around 50 mA as sourced not thru 1K but 1K // 470 ohms. We could bring this to 100 mA by using 470 & 220 ohms for the divider. (Can't guess the meaning of that *note about 360 ohms on fig. 8) But I don't feel comfortable with such a weak trigger pulse for such a big triac. Now I realize there is no capacitance in the whole trigger ckt, not feasible to implement such large caps monolithically. If I were your alternistor I would demand at least one big gate pulse if not a pulse train.

Can you check the current waveform thru the gate? It would be safer to see it before connecting a load. Look for a short risetime.

An alternative would be to use a small, gate-sensitive triac to fire the big one (kind of a Darlington with triacs). Hope you can get a shorter but amped-up pulse. Too weak a pulse may damage the TPDV1240.

brgds

Snel

Hi Snel,

If I understand correctly, the following is what you are suggesting (I do not know how to attach a drawing for ease of explaining);

1) a 220 ohm resistor (Say R)is connected to the junction of M2 terminal of the alternistor and one line (Say line 2).

2) Other end of R is connected to #6 pin of MOC3083.

3) a 470 ohm resistor is connected to the junction of R and #6 pin of MOC3083.

4) #4 pin of MOC3083 is connected to the gate of alternistor.

My problems with this arrangement are:

a) The Voltage divider formed by R and R1 with an overall resistance of 690 ohm will dissipate 631 watts. The power or the 470 ohm resistor should be 430 watts and that of 220 should be 201.

b) The gate drive will still be 20 V / 220 ohms = 90 mA. As an engineer, I do my designs considering worst case conditions. The specified worst case trigger current of a TPDV1240 is 100mA and I happened to get one which requires 100mA for triggering, the solid-stare will not turn on.

Regards,

Job.

Hi thankan,

you're right again! A resistive voltage divider is not feasible. A step-down transformer is bulky and you may have phase shift. Coming back to the original idea of connecting the outputs of two 3083s in series, you would need balancing resistors dissipating around 30 watts each.

Can't you use e. g. 240 volts from another line at the same phase? You would feed your driving circuit with it and drive the alternistor's gate thru a small trigger xfmr.

brgds

Snel

Hi Snel,

If you refer to non zero crossing triac driver application notes by Motorola (And may br Fairchild also), you will see that two MOC 3011 triac drivers are connected in series to drive a triac to switch a 240 V load. The balancing resistors used in this case are one Meg ohm resistors. 240 VAC applied across 2 meg ohms. Power dissipation in this case is only 0.03 watts.

If I use the same two balancing resistors, for 660 VAC load, the power dissipation will only be 0.22 Watts across the two(0.11 W across one). So one meg ohm, 1/4 W resistors will be More than sufficient to do the job, if Motorola and Fairchild are not wrong (Please see AN-3003 in Fairchild data).

Have a nice weekend

Regards,

Job

I still maintain that it's dangerous to use triacs in series, especially optically-coupled ones. They have quite a long delay from applying current to the LED to firing of the triac. This delay time is not the same from part to part. What's more, the manufacturer doesn't give any min/max specs for this delay, nor even a plot of delay vs LED drive current. There's no guidance.

When a one triac fires first in a peak-voltage situation, the entire voltage is impressed across the second triac. Perhaps it'll simply self-trigger in response, but over time will it suffer and fail from this abuse? It looks dangerous to me. For example, in AN-3003 they suggest using two MOC3011 in series for 230Vac, but it would have been far better to suggest using a single MOC3022 with its 400V rating.

The big boys in high-voltage triac power control design create a generous gate trigger current independently from the instantaneous AC-line voltage, using separate low-voltage circuitry, and then apply that with a transformer or other drive scheme to the gate of the power triac. There are multiple ways to sense the AC line zero crossing and create the right time and duration for an efficacious gate pulse. At low AC line voltages, and for low power applications, the simple circuits using optically-coupled zero-crossing triacs may suffice, but you generally won't find those simple circuits used in high-power triac systems.

If an 800V optically-coupled zero-crossing triac for use in a simple approach doesn't have a high enough voltage rating, then surely it's time to switch to one of the standard high-voltage approaches.

Hi WH,

GA from me.

Job, remember your sound principle about safety margins. If your triac needs 100 mA to trigger, give it say 1 A, short and steep. Repeating the pulse is even better, you have to avoid "overhead firing" i. e. switching by anode voltage and/or dv/dt without a [decent] gate pulse.

nikce weekend to all

Snel

Dear Winfield, Snel,

The following is my logic:

1) The possibility of having two MOC3083 is next to impossible. The specified inhibit voltage range is very wide from 5 V to 20V indicating that the possibility on one device firing way ahead of the other is high.

2) I will be using a 22 ohm resistor in place of the 180 ohm, which will provide a trigger current from 150 to 900mA in a pass voltage range of 3.3 to 20 V which will ensure the triggering.

3) Now, Let the inhibit voltage of one of the MOC3083 "A" equals 15.26 V and the other one "B" equals 15.30, The chances are "B" will turn ON before "A", when the instantaneous voltage drops to slightly below 30.60.

4) Even though "B" turns on, the triac will not be triggered since "A" is still OFF.

5) Just before "B" turns ON, voltage across "A" will be slightly above 15.30.

6) Soon after "B" is turned ON, the voltage across "A" will go high to a value slightly below 30.60.

7) "A" will remain OFF until the voltage across if falls slightly below 15.26 at which point "A" will turn ON.

8) Since "B" is already ON, when "A" comes ON both MOC3083 will conduct to give required the trigger current.

I think that with a 660 VAC supply, the voltage across any one of the MOC3083 will not be over 500V even in the worst condition; Please review and advise.

Thanks and regards,

Job.

Hi job,

your logic looks OK, a higher trigger current sounds good too.

Now proceed to breadboarding... get a scope, looking at the waveforms can't hurt you.

brgds

Snel

Thank you Snel,

I appreciate it very much that you took time to go through my logic.

What I am planning is to apply a DC voltage across the series connected MOC3083.

I Expect both will be OFF when the applied voltage is higher than 60V (Of-course; I will be connecting these in series with a 2 K resistor). Gradually decrease the voltage while monitoring the voltage across both of them with two volt meters. Does it sound OK?

Hi Winfield,

Do you have any comment?

Regards,

Job.

Like others here, I'm very uncomfortable with your idea to use triacs in series to exceed their official peak voltage rating, zero crossing or not. (MOSFETs have a nice property of "safely" avalanching when you exceed their peak rating, so that one can carefully play the series game, but I think that triacs, like BJTs, are not safe to use that way.)

You'd be better off using a divider scheme like Snel suggested (it can employ a "storage" capacitor to boost the available gate-trigger current, like Fairchild shows in their MOC3083M datasheet). Or you can use a small power transformer to create a proper within-spec AC voltage regime for the zero-crossing triac, and then you'll have a floating drive circuit and signal that you can tie to your high-voltage triac or TPDV1240 alternistor. You could also consider transformer coupling the zero-crossing triac's high-current output capability to fire your high-power triac. Really high-voltage triacs / alternistors, etc., are often triggered with gate pulse transformers.

I note that Littlefuse / TECCOR recommends, in AN1007, using ON Semi's UAA2016 zero-crossing trigger IC. ON Semi pitches this part as a temperature controller, but it's really a beautiful zero-crossing gate-turnon pulse generator, with programmable pulse width (with a resistor they call Rsync), and up to 90mA gate drive. (You'd trick it by setting it up so the "temperature" is always too low.) It's rather low-cost too, under a buck. If I'm not mistaken, you'd use that part with an ordinary triac rather than an alternistor.