Combine triac with a relay to drive an AC load

[VinzC]

Hi.

TL;DR: I only want to understand the principle. There may be a better way to design this, I'm only interested in *understanding* if/how a triac can work together with a relay the way I described it.

I'm designing a power stage for two heaters on my modded oven. Because of available space I can't install a solid state relay. I've also come to prefer zero crossing instead of using relays only. So my choice went to a MOC3043, doubled with a triac, as the datasheet recommends (see https://www.onsemi.com/pdf/datasheet/moc3043m-d.pdf).

The switched currents (AC) will be between 5 and 10A (measured). Since I fear (and am almost certain) the triac I've chosen (MAC9NG, see https://4donline.ihs.com/images/VipMasterIC/IC/LFSI/LFSI-S-A0012699119/LFSI-S-A0012723662-1.pdf) will blow up, even if mounted on a heatsink (which can only be small due to lack of space), I've thought of adding a relay (G5CA is the one I have, see https://components.omron.com/us-en/asset/53866) in parallel with the power triac (see the red outline in the picture below) to turn ON after being certain the triac conducts¹.

Regardless of whether it's a good idea (I suppose it's not), I'd like to know (because I want to understand) what happens when the relay contacts turn OFF: will the power triac remain ON or will there (potentially) be a spike when the contacts are OFF, say in the middle of a peak? I can find valid argumentation in both cases (i.e. "there will be a spike" and "there won't be any") which translates my misunderstanding of the MOC + triac pair.

So the question is: assuming the MOC is turned ON, what happens when the relay is switched OFF? Will the current be maintained by the triac or will it be cut off, defeating the purpose of switching ON/OFF at zero crossing only?


Thanks a lot in advance.

¹ In that shape, the relay is always switching off first and then the MOC is turned off after a given delay, which is greater than the rated time for the contacts to be released.

[DavidAlfa]

It will remain off, as there wasn't any current flowing through it in first place (Shorted by the relay).
In any case, the worst case scenario would be it conducting for 1/2 wave, turning itself off at zero cross.

[VinzC]

That was one of my two assumptions indeed. Makes sense. So the idea is pretty terrible and totally defeats the purpose of switching on zero crossing.

On a side note, is it worth switching the heaters on zero cross or does a relay do the job? After ovens and microwaves do use relays, too, right? (At least mine does, I hear 'em.) What is the compelling reason for wanting to switch on zero cross¹?

¹ Apart from wear out, which I suppose is the biggest deal.

[DavidAlfa]

For just 5-10A and resistive load (Oven heater), a good relay will have a long life!
They stress out a lot more with inductive loads.

Alternatively you could make a SSR using a pair of MOSFETs, giving almost zero on-state power loss.
https://www.allaboutcircuits.com/news/littelfuses-gate-driver-promises-fast-turn-on-speed-in-solid-state-relays/

[langwadt]

For just 5-10A and resistive load (Oven heater), a good relay will have a long life!
They stress out a lot more with inductive loads.

Alternatively you could make a SSR using a pair of MOSFETs, giving almost zero on-state power loss.
https://www.allaboutcircuits.com/news/littelfuses-gate-driver-promises-fast-turn-on-speed-in-solid-state-relays/

need pretty low RDSon, a few tens of mOhm and ~600V FETs at 10A to get "almost zero on-state power loss."

[VinzC]

For just 5-10A and resistive load (Oven heater), a good relay will have a long life!
They stress out a lot more with inductive loads.

That is really interesting and relieving indeed! I've measured those quartz tubes and each are about 100 ohms (cold), which yields 2.3 - 2.4 A and I'll account for 5 for safety. The relays I have are made by OMRON and I bought them from Digikey so I'm pretty sure I can trust them.

Thanks a lot for the tips.

[DavidAlfa]

need pretty low RDSon, a few tens of mOhm and ~600V FETs at 10A to get "almost zero on-state power loss."

Right, I later realized 10mOhm RDS is rarely a thing in 500V+ fets.

[Bikkel]

you do not shown how you drive / power the relay. Also keep in mind that when the relay opens there is also sparks going on. I have seen a lot of relays drive with freewheel diodes which is not a great idea. I would fire the triac some cycles before switching on the relay and fire the triac again when the relay opens. This way your triac dissipation is minimal and the wear on your relay is also minimal as the contacts never see a high voltage.

[DavidAlfa]

Remember that relay contacts get clean by arcing (Maybe not all types).
So no arcing at all might also reduce the lifespan as the contacts will oxidize.

If mcu-controlled, it might be a good idea to make some sort of wear timer, so every n cycles the relay is switched "hard" to clean the contacts, skipping the triac.
You could also measure the relay switching time, and calibrate the trigger.
So if it needs 5ms to switch, measure the time between two zero crossings (I.E. 20ms) and activate the relay 5ms earlier (15 ms after zero cross), so the voltage will be very low when switching.

That way you can skip the triac entirely.
Not sure how the switching time changes with temperature and aging.

[langwadt]

datasheet says that relay is good for 300000 cycles of 10A @ 240V resistive

[VinzC]

Thanks for your suggestions.

I've made up my mind after measuring again the currents involved and I kept my initial design with triac & MOC. The currents involved are 2.5A maximum per heater so the heatsink dimensions are quite reasonable and will fit the oven compartment.

[VinzC]

you do not shown how you drive / power the relay. Also keep in mind that when the relay opens there is also sparks going on. I have seen a lot of relays drive with freewheel diodes which is not a great idea.

Ah? It's how I've been driving relays for years now ... That is interesting; how do you suggest to drive these? With a push-pull stage I suppose?