Triac failure mode

[TomS_]

Hi all. What would be a typical triac failure mode: short, open, or "it depends" ?

A friend came to me with a Dyson fan that is constantly in rotate mode, even when it is "turned off". I had a probe around and it seems that a triac (MOC3083) controlling a synchronous motor is basically shorted, even though I can see the control input being toggled.

I've ordered a replacement triac to see if that will fix the issue, but thought I might ask this question in the meantime while I await it's delivery.

Thanks!

[james_s]

I've only ever seen them fail shorted. I suppose if one was overloaded sufficiently it would vaporize the bonding wires but initially silicon almost always fails shorted.

[Zero999]

Yes, TRIACs normally fail short circuit, which is why they musn't be used for anything safety critical. There should always be another layer of protection, i.e. safety relay + contactor for motors, or thermal fuse for heaters, if an injury, fire or damage to property, could result if it failled on.

The MOC3083 is an opto-coupler with a TRIAC output and won't be used to switch the motor directly. It will be used to switch a separate TRIAC in series with the motor. It's possible the TRIAC, optocoupler, or both have failled short circuit.

[TomS_]

I buzzed it all out, and the MOC3083 is definitely the only thing switching the motor. Its just a small one to swing it side to side rather than turning the blades, although this is one of those fancy bladeless fans anyway.

And in the end it seems that it likely wasnt the cause of the problem. After replacing the TRIAC it was still rotating even when the fan was switched off. So I buzzed out some more and found a MOV and a couple of resistors and a small capacitor in parallel to the TRIAC, and it would seem that the capacitor is the thing that failed short. Just a small 0805 or thereabouts ceramic cap. I removed it and now the rotation function is working as expected.

Im not sure what the value of the capacitor is supposed to be, and I probably dont have anything in my stock that is appropriate to replace it with anyway, and everything seems happy so far with it out, so maybe I'll just leave it out unless someone knows a good reason to put it back in? Mains voltage stuff isnt really my forte, but it was a fun experience none the less.

[james_s]

As has been mentioned, the MOC3083 is not a triac, it is an optically isolated triac driver. Technically it is possible to control something directly with one but I don't think I've ever seen that done, they can not handle very much current at all. Remove it completely, if the thing still runs then there is definitely something else shorted.

[Zero999]

I buzzed it all out, and the MOC3083 is definitely the only thing switching the motor. Its just a small one to swing it side to side rather than turning the blades, although this is one of those fancy bladeless fans anyway.

Perhaps if it's a very small motor, it'll be okay. The reason why I said that is because all the applications I've seen show it switching another TRIAC. Look at the bottom of page 7 of the data sheet, which shows some example ciruits.
https://www.onsemi.com/pdf/datasheet/moc3083m-d.pdf

And in the end it seems that it likely wasnt the cause of the problem. After replacing the TRIAC it was still rotating even when the fan was switched off. So I buzzed out some more and found a MOV and a couple of resistors and a small capacitor in parallel to the TRIAC, and it would seem that the capacitor is the thing that failed short. Just a small 0805 or thereabouts ceramic cap. I removed it and now the rotation function is working as expected.

Im not sure what the value of the capacitor is supposed to be, and I probably dont have anything in my stock that is appropriate to replace it with anyway, and everything seems happy so far with it out, so maybe I'll just leave it out unless someone knows a good reason to put it back in? Mains voltage stuff isnt really my forte, but it was a fun experience none the less.

The capacitor and resistor form a snubber network, which absorbs the high voltage spikes, when the motor is turned off. It will work without it, but the MOC3083 might not last as long. The capacitor value isn't critical. Try 10nF, 400V.

[TomS_]

Technically it is possible to control something directly with one but I don't think I've ever seen that done

Well as they say, there is a first time for everything. :-)

There are no other components involved with switching for this motor. It's quite small, not much bigger than a golf ball, so maybe they figured they could get away without a full blown triac. :shrug:

[james_s]

Well, if it failed maybe they actually couldn't. Either way if there are so few components involved it shouldn't be very difficult to track down what's wrong.

[TomS_]

The capacitor and resistor form a snubber network, which absorbs the high voltage spikes, when the motor is turned off. It will work without it, but the MOC3083 might not last as long. The capacitor value isn't critical. Try 10nF, 400V.

Ah ha, yes. I did a little looking around and started to notice some references to this kind of thing, but the math involved in calculating the values went way over the top of my head.

So something like a flyback diode for relays, only for AC?

I'm much more comfortable in digital electronics. :-)

[TomS_]

Well, if it failed maybe they actually couldn't. Either way if there are so few components involved it shouldn't be very difficult to track down what's wrong.

I believe I did manage to find the culprit, see my post where I attached an image. It seems the capacitor in the snubber failed short, bridging out the triac. When I remove it from the circuit, the motor no longer turns unless turned on by the user.

I haven't actually tested the original "triac" since so it could be perfectly fine, but this feels pretty conclusive.

[Alti]

Just a small 0805 or thereabouts ceramic cap.

Are you sure there are no more caps in series? The varistor kicks in only at some 600V+ and the capacitor might already be charged to opposite polarity when transient happens. These 0805 caps are rarely rated for >200V and there are some special 0805 for 500V. If that is the case, it is not going to be easy to replace. Without a snubber the triac can have problems with turn-off, not turn-on.

[miken]

I have a vague memory of finding an always-on triac in my parents' no-name pillar-type fan, many years back. I want to say it was always in move-back-and-forth mode, so perhaps the regulators didn't think that was a safety risk? Anyway, cheap stuff, I think I replaced the triac and rotate on/off control worked again.

[TomS_]

Are you sure there are no more caps in series? The varistor kicks in only at some 600V+ and the capacitor might already be charged to opposite polarity when transient happens. These 0805 caps are rarely rated for >200V and there are some special 0805 for 500V. If that is the case, it is not going to be easy to replace. Without a snubber the triac can have problems with turn-off, not turn-on.

It was a bit hard to probe due to conformal coating, but I didnt find anything else that seemed to be in parallel or series.

I measured the caps dimensions and it is actually 1206.

[Alti]

(..)However, for igniting a photo-Triac through its embedded LED, its internal structure cannot be vertical (from bottom to the top silicon sides) but requires to be lateral. This is the reason why its voltage drop at ON-state is relatively high and it cannot be used as a power Triac. (..)

Link to ST AN5649.

Dodgy design or an unusual opto-triac rated for direct (small) motor control drive.

[perieanuo]

mostly shorted a1-a2 or g-a2
gettin an 'insulated' gate is rare stuff, i got lots of triacs shorted in textiles industry