MOV to protect triac

[metRo_]

Hi,

I'm control a 230VAC blinds motor with the following triac https://pt.mouser.com/ProductDetail/panasonic/aqh3213/?qs=oR3PEIV06LdjjKqjb7eWLw%3D%3D&countrycode=DE&currencycode=EUR since I don't know which motor I have installed (I suppose it is a 150W motor) I over dimensioned the triac. I have used the triac for some time and didn't notice any problem. The triac didn't even increase the temperature during the 30s the motor is working.

However on a recent pcb revision and following the triac datasheet I installed a MOV to prevent the degradation of the triac. I choose the following one https://pt.mouser.com/ProductDetail/panasonic/erz-e05a361cs/?qs=jYgs00e1acgtsvUGcxtYiQ==&countrycode=DE&currencycode=EUR

As soon as I start the motor the MOV started to burn.

Why did it happen?

I have attached the circuit. M_DOWN and M_UP are the outputs that connect to Motor. Motor turns one direction when Live is connected to one terminal and another direction when Live is connected to another terminal like this:

[nuclearcat]

I think it is better to post circuit, how everything is connected, so people wont waste time guessing.

[metRo_]

I think it is better to post circuit, how everything is connected, so people wont waste time guessing.

I update the post with the circuit

[nuclearcat]

My theory:
1)On disconnect of inductive load there is short spike that MOV cannot adsorb.
2)When motor is turned off it generates some back emf, which is higher than your Vpeak (328V) and your MOV start to work as motor brake.
(Especially when phase of generation and in mains are shifted at 180 degree.
E05 varistor is only 0.25W rated, it means it can handle spikes with specific energy OR steady max 1.52mA current.
And it start conducting more than allowed current at approx 400V so it cannot hold it long time.
Maybe bigger MOV required)
But start from adding RC snubber, it is recommended in datasheet of your SSR.
(i'm not expert in this question, its better to listen if someone will say more qualified opinion)

[Siwastaja]

Datasheet specifies minimum (worst case) unit conducts 1mA at just 324V.

Portugal is 230V, don't know about the specs but likely +6% well possible, i.e. max 243V. Peak voltage is sqrt(2)*243V = 340V, clearly over 324V.

1mA at 340V is 0.34W. This conduction should happen at low duty cycle though, so it likely doesn't explain all the dissipation you are seeing.

In any case, you have dimensioned the part wrong. To protect at 230V nominal power line from spikes, you must pick a part guaranteed to stay off with the maximum normal line voltage, which will be at least 240Vrms. The next part in series, rated at 250VAC, sounds about right, or even the next one at 275VAC.

Roasty MOV is a typical design failure. Yes, they are not precision devices, so you can only take the biggest edge of large spikes. More important is to choose the part so that it does not clamp under normal conditions because they have limited lifetime.

Now looking at the 275VAC rated part, max clamping voltage is 710V @ 25A, so it would be great if the TRIAC would be rated at 710V peak at least.

This may or may not be the reason. I would start by measuring the line RMS voltage.

Inductive kick may be another. Assuming the switches can handle it (i.e., are designed to switch motor loads; I didn't check), you can move the MOVs to the power input side.

[bdunham7]

More important is to choose the part so that it does not clamp under normal conditions because they have limited lifetime.

Unless I'm not seeing straight, his design is using the MOVs to snub the triac, not protect it from line surges.  That seems inappropriate somehow.  Since they are burning up, but only when the motor is turned on, I'd assume something is happening not directly related to line voltage peaks, but rather related to the interaction of the triac and the motor.  I'd be curious as to which MOV was burning when a particular triac was activated.  Time for some (careful) scope work.

[trobbins]

I'm with Siwastaja. 

Line voltages can move around (especially if domestic PV is on-line or ubiquitous in the local region), so without logged data that aspect can't be dismissed as the root cause of the MOV failure - the small 5mm part has a very low continuous power dissipation capability, so even very soft but continuous conduction at the peaks of the voltage waveform are likely to cause MOV degradation.

Even when a MOV seems appropriate for the application, without waveform assessment to confirm that the MOV is not normally entering soft conduction (eg. from triac induced inductive spikes), it would be appropriate to periodically test each MOV for degradation to ensure it is still capable of protection (ie. check for MOV leakage resistance with a 500Vdc insulation resistance meter).

[bdunham7]

I'm with Siwastaja. 

Line voltages can move around

I'm not disagreeing that the 230VAC-rated MOV is probably inadequate at Euro-line voltage.  I'm pointing out that according to the OP, the MOV is burning up precisely when it is not subject to line voltage.  Or, if the MOV snubbing the non-active triac is the one burning, perhaps there is some interaction with the motor subjecting it to much more than line voltage. So moving up to a 250VAC or 275VAC version may not help much, or worse make the problem latent, waiting until later to burn down someone's house.

OP: Can you measure the voltage with a scope or DMM (carefully--there may be more than you think) across the non-active triac as you run the motor in one direction or another?

[Siwastaja]

But OP doesn't have a TRIAC! They just used wrong term in the opening post. The link refers to an SSR - which likely contains a TRIAC, but isn't a TRIAC.

Instead of using a MOV for snubbing a motor load, simply use an SSR rated for the motor load so they have taken care of any snubbing required. The selected part might already be such, I didn't check.

Then use MOV to protect the SSR from high line peaks, this is often done because SSRs have worse overvoltage ratings than mechanical relays. Here it's important to make sure it never clamps under slightly high line conditions as I explained above. Also  a large size is beneficial. Remember to use a fuse before the MOV, it can be a fast type. With this combo, survival against line peaks (lightning transients, for example) is possible.

[nuclearcat]

But OP doesn't have a TRIAC! They just used wrong term in the opening post. The link refers to an SSR - which likely contains a TRIAC, but isn't a TRIAC.

Instead of using a MOV for snubbing a motor load, simply use an SSR rated for the motor load so they have taken care of any snubbing required. The selected part might already be such, I didn't check.

Then use MOV to protect the SSR from high line peaks, this is often done because SSRs have worse overvoltage ratings than mechanical relays. Here it's important to make sure it never clamps under slightly high line conditions as I explained above. Also  a large size is beneficial. Remember to use a fuse before the MOV, it can be a fast type. With this combo, survival against line peaks (lightning transients, for example) is possible.

This SSR require separate snubber circuit according the datasheet.

[Siwastaja]

It also seems to be a part designed for the Japanese 200Vrms mains voltage (according to the front page), with just 600V absolute maximum peak rating. At 230Vrms, it's a bit iffy.

It may not be the best part for the job...