Triac Blowing

[DillenM]

Good Afternoon

I am busy building a static VAR compensator for a project and i need to use TRIACS to switch in my capacitor banks. The system is 3 phase 400 VAC and consists of 3 motors with varying loads. There is a system that determines the power factor and switches in the capacitors. Today all my TRIACS blew when I put a 1.1kw motor on. I am using Q6030LH5 TRIACS and MOC3021 optocouplers.
The thing is, the switching circuit was not connected to the microcontroller yet. Therefore the switches should be completely off. The capacitors are connected as shown on the attachment.
The optocoupler and TRIACS are connected as shown in the other attachment. The resistor used between the optocoupler and triac is not 220 ohms but 100 ohms and the other resitor i not 330 ohms but 2 kilohms.

I hope this makes sense and this is urgent.

Kind Regards
DillenM

 

[dr pepper]

I maintain a couple of pf machines, they use contactors.
If your triacs trigger inputs were not connected then maybe noise or static triggered them, also depending on your circuit the triacs switching a capacitive load will be subject to very large current spikes which could easily fry them, maybe you need some protection, how about some charge resistors to keep the caps charged while not in circuit, doing this would also require some safety circuitry to prevent caps from being charged after shut down.

 

[spec]

Good Afternoon

I am busy building a static VAR compensator for a project and i need to use TRIACS to switch in my capacitor banks. The system is 3 phase 400 VAC and consists of 3 motors with varying loads. There is a system that determines the power factor and switches in the capacitors. Today all my TRIACS blew when I put a 1.1kw motor on. I am using Q6030LH5 TRIACS and MOC3021 optocouplers.
The thing is, the switching circuit was not connected to the microcontroller yet. Therefore the switches should be completely off. The capacitors are connected as shown on the attachment.
The optocoupler and TRIACS are connected as shown in the other attachment. The resistor used between the optocoupler and triac is not 220 ohms but 100 ohms and the other resitor i not 330 ohms but 2 kilohms.

I hope this makes sense and this is urgent.

Kind Regards
DillenM

Hi DM,

You don't say what opto coupler or what TRIAC, so it it impossible to diagnose the failure mode definitively but, unless the TRIAC optocoupler is zero switching, that TRIAC circuit is pretty much guaranteed to blow the trigger junctions of the main TRIAC:
(1) The current through the trigger terminal will be (240 * 1.414)/330 = 1.028 Amps which is way to high for most TRIACs
(2) There is no resistor between the trigger and neutral.
(3) A shaping capacitor may be required
(4) the TRIAC optocoupler may be having a rough time too.

spec

 

[DillenM]

Hi Spec

The optocoupler is a MOC3021 and the TRIAC is a Q6030LH5. What do you mean a resistor between the trigger and neutral? Remember my switches are connected between phases (red and white, white and blue, red and blue). Its three-phase.

Hi dr Pep

Someone told me that if i put a few coils in the wire before it goes into the switch that will limit the inrush current. Do you think that might help?

 

[spec]

POST ISSUE 2 of 2017_01_27

Hi DM

As Mr Al's post #7 has answered your question, I will just leave this post with the datasheet links for reference.

spec

DATASHEETS
(1) https://www.ti.com/lit/ds/symlink/moc3020.pdf
(2) https://www.littelfuse.com/~/media/...r_qxx30xhx_qxx35xx_qxx35xhx_datasheet.pdf.pdf

 

[spec]

The optocoupler is a MOC3021 and the TRIAC is a Q6030LH5. What do you mean a resistor between the trigger and neutral? Remember my switches are connected between phases (red and white, white and blue, red and blue). Its three-phase.

Sorry DM: it was late for me and I had a hard day breaking concrete.

I meant a resistor between gate (trigger) and main terminal 1 (MT1)

Hopefully I will give a more thorough reply in post #5 above.

spec

 

[MrAl]

Good Afternoon

I am busy building a static VAR compensator for a project and i need to use TRIACS to switch in my capacitor banks. The system is 3 phase 400 VAC and consists of 3 motors with varying loads. There is a system that determines the power factor and switches in the capacitors. Today all my TRIACS blew when I put a 1.1kw motor on. I am using Q6030LH5 TRIACS and MOC3021 optocouplers.
The thing is, the switching circuit was not connected to the microcontroller yet. Therefore the switches should be completely off. The capacitors are connected as shown on the attachment.
The optocoupler and TRIACS are connected as shown in the other attachment. The resistor used between the optocoupler and triac is not 220 ohms but 100 ohms and the other resitor i not 330 ohms but 2 kilohms.

I hope this makes sense and this is urgent.

Kind Regards
DillenM

Hi,

Let me see if i have this right ... you have a 400vac system and using 600v triacs?
Sorry to say, that doesnt sound right, right off the bat.

Peak voltage at 400vac rms is 566v peak already. At a minimum high line calculation it would be 622v which is above 600v obviously.
So the system already does not sound like it is designed for minimum reliability.

Now add to that the fact that if it switches out, the capacitor could have +622v DC across it while the line swings down to -622v which temporarily places over 1200v across the triac.
If we forget about high line, we still end up with 1120v across the triac, which obviously exceeds the rating of a 600v device.

So it would make sense to use a 1200v triac.
Of course zero cross detection would stop bad surge currents from damaging the device too.

Other things to look out for would include:
1. Check for high voltage spikes when the motor disconnects.
2. Check for large peak gate current that exceeds the device rating.
3. The triac has enough current during turn on to fully turn it on.

 

[DillenM]

MrAl to me that makes the most sense... Ill try to replace them to 1200v triacs and hope it works... ill also add a resistor betweeen gate terminal and T1... any ideas of the size of that resistor.. also do you think a snubber ircuit is required?

 

[spec]

ill also add a resistor betweeen gate terminal and T1... any ideas of the size of that resistor

2.2K would be a good value.

spec

 

[DillenM]

OK thanks everyone. Attached i have my new design. What do you guys think?

 

[Les Jones]

I think you also need to add some zero crossing circuit to each triac so it will only switch on when the votage across it is close to zero. In the worst case the triac could be switching on when the voltage across it is close to 1200 volts. You don't say what the value of the capacitors is but there will be a lot of energy to disipate first discharging it to zero volts ( C x V'2)/2 And the same amount of energy then to charge it in the opposite polarity. The current will only be limited by the impedence of the power supply. I think the alternative would be to use triacs with a very high current rating. (At least seversl hundred amps.)

Les.

 

[MrAl]

OK thanks everyone. Attached i have my new design. What do you guys think?

Hi,

I think a snubber would be a good idea, as well as zero cross switching.

Also, i see you have a 16 amp triac spec'd on your drawing. If you can get a higher current rated triac for almost the same price, i'd go with the higher rated part as long as it does not impact the total cost too much. If this is a one-off project though, i would not hesitate to spend twice as much or even more for a higher current triac knowing that i will be getting a more hardy part to start with. If this project is that important to you, do it justice by getting good parts that can stand up to the current and voltage demand and doing the design as close as possible to recommendations by manufacturers who make the parts. Often you can find recommended designs for these parts. In other words, do everything possible that makes the design better. Once you do everything possible you should get a good working system or else the very concept of doing it that way is somehow flawed.

It also does not hurt to roll your sleeves up and get in there with a good scope. Heck, even a not so good scope
In power designs of any kind (those that deal with line voltages and/or creating power supply like voltages and currents) there is no chance in heck that a manufacturer of a power system will allow that system to go out the door without some rigorous testing. That testing would included looking at waveshapes during the complete operation of the device, which includes various things like looking for over current surges and looking for over voltage spikes. Once it is found that the transient behavior is acceptable, the system would then go on a life test which could be as simple as an overnight non stop run at the full current rating of the system.
I am telling you this because just building a system that looks like it should work is never enough...it has to be tested and known issues are searched out to find out if there could be a problem before it happens. That's mostly because some design flaws do not show up right away but take time to progress, and some fail right away and it's hard to know why without looking at the waveforms in various places in the system. For example, if you see a 600v spike across a 600v device, you know there is a good chance it's going to fail sooner or later. The design schematic may not tell you that though, you need to look at that with a scope. That gives you the confirmation of everything you thought was true is true or is not true in the theory of operation.

What this means is if you still have problems after this, you should get a scope and start looking for problems before they happen. This may involve trying to operate at a lower voltage so that nothing blows right away, if the failure happens soon after turn on. Sometimes that is not possible but if you can make that happen you have a better chance of finding the problem. I dont know if you have a scope though, or can get one somewhere for this.

 

[dr pepper]

If you dont have zero crossing then a choke would be a good idea, you'd need more than a few turns though if its air cored.