<...posted this question in General and Communication & Electronics forum but was not quite convinced with the answers...>

So does that make this one a duplicate thread, then?

So, what happens when one of them conducts and the other one doesn't?

If one MOC device is faster or slower than the other one, such that one device fully turn off first before the other, then you will see one device taking the full voltage stress momentarily. The question is whether the speed variation is small compared to the overall switching time. If this switching speed variation is available from data sheet, then you can easily determine whether the above unequal voltage stress situation will happen. If this data is not available, then the next best thing to do statistical measurement. My opinion is that this is too much effort, better to look for another circuit techniques .

YOUR LOGIC is Perfect

Go ahead and use it.

IN A FEW DAYS YOU WOULD HAVE PROVEN YOUR OWN JUDGEMENT -CORRECT.

And MEANWHILE you WOULD HAVE DONE NO HARM TO YOUR application/assets .

The logic is ALMOST ok. He forgot the time aspect. While this may work for a few shots, one of the switch will be stressed and will eventually fail.

You need to add balancing capacitors to your resistors to share the dv/dt as well. This can unfortunately produce leakage current that could trigger your main thyristor or make it re-trigger when it should turn off. This will reduce its noise immunity.

You need to add back to back zener diodes to protect each switch.

You need to take into account that the reliability of this circuit will be greatly reduced compared to using a single switch properly rated. Don't use it for potentially dangerous applications.

Once you do all this, it might work relatively well, most of the time.

Once you do all this, it might work relatively well, most of the time.

It will WORK .

ALL the time.

Hi Marcot, Bravo 88,

I think that you are forgetting one very important fact; MOC3083 has zero-crossing circuits built into it with a specific inhibit voltage. For a Fairchild MOC3083, this inhibit voltage is 20 (Max) which means that the opto-triac in the device will not be turned before the voltage applied across it falls below 20 V. So for two devices in series, the applied must fall below 40 in order that one device of the two, with more sensitivity (Or higher inhibit voltage) to turn.

This tells us that with balancing resistors in place, in order for "both" devices to start conducting, the voltage across them must fall below 20, and for the first one to turn ON the voltage must fall below 40.

This also tells us, irrespective to the differences in sensitivity, timing (Or propagation delay) or inhibit voltage, one one MOC3083 will be seeing only a maximum of 490 assuming a +5% fluctuation of the 660V supply.

Regards,

Job.

Thank you for the details on the operation for this chip. It certainly look good on paper.

I design industrial electronics and tend to operate on the conservative side. You would be surprised by how easy it is to trigger a thyristor device with a high dv/dt. The output triac could be turning on or even simply leak more even when the inhibit is supposed to prevent it to do so.

Once the MOC3083 leaks in to the main gate, it makes the main thyristor more sensitive to the dv/dt that it experiences as well. Operating at a high temperature makes it even more sensitive.

This is why, I said that it will probably work most of the time but it is likely to fail in extreme conditions. Play safe.

Good luck.

Hi Marcot,

I buy the your suggestion of turning on the triac with a high static dv/dt (s). Being trained as a mechanical engineer, I am not familiar with commuting of triac. I will appreciate it very much if you can answer the q questions are,

1) If I use MOCZ500 with a dv/dt(s) of 2,000V will it help the situation?

2) Will two devices in series connected in series double the static dv/dt to 4,000 volts?

3) Since both triacs within the devices are in series, if the spike is high enough, won't both of them turn ON simultaneously?

4) If the spike is large just enough to turn On one and not the other, obviously the full spike will be applied the one which did not conduct. Do you think that the duration of the spike will be long enough to damage the device?

5) In the above case how can there be a conduction since one of them is OFF?

6) Some one has suggested to look for other alternatives; I know that there are but do not know what they are. There are zero turn-on solid-state relays in the market with control voltage 5 to 32 Vdc which must be using some opto-isolators in them. Do you know what is being used in them to drive the triac?

Regards,

Job.

Job,

Here are my answers following your questions.

1) If I use MOCZ500 with a dv/dt(s) of 2,000V will it help the situation?

It should help. The usual units for dv/dt are V/uS

2) Will two devices in series connected in series double the static dv/dt to 4,000 volts?

It improves the total dv/dt but does not usually double it as you will not have identical parasitic components.

3) Since both triacs within the devices are in series, if the spike is high enough, won't both of them turn ON simultaneously?

There is no way to tell for sure. This type of thyristor activation tends to be very stressful and even destructive for the device. Try to avoid it.

4) If the spike is large just enough to turn On one and not the other, obviously the full spike will be applied the one which did not conduct. Do you think that the duration of the spike will be long enough to damage the device?

It could be. It depends on so many factors that you are talking about probabilities of the worst case happening. We normally try to design the equipment to handle the worst case or have a sacrificial component like a fuse to absorb the resulting fault when the components are over-whelmed. The objective is to keep all the components alive while the fuse blows.

5) In the above case how can there be a conduction since one of them is OFF?

It will probably be forced on from avalanching processes. See #3.

6) Some one has suggested to look for other alternatives; I know that there are but do not know what they are. There are zero turn-on solid-state relays in the market with control voltage 5 to 32 Vdc which must be using some opto-isolators in them. Do you know what is being used in them to drive the triac?

Once you work at high voltage and high current, you should use pulse transformers to drive the gates of back to back SCR. This is more reliable and SCR are more rugged than TRIAC. The cost is higher but the result is usually much better.

You have to understand that in general, opto-couplers should not be used in applications over 240VAC. Their packaging and electrical clearances are not large enough for higher voltages. I have tried them many years ago but the results were not good enough for a reliable product and we switched to pulse transformers. In general, cheap and high voltage don't mix well.

Solid state relays are designed and tested by people who are experienced in the field (I hope so anyway...). These relays are also potted in silicone to improve the insulation between components. The inside is not exposed to humidity and dirt which helps a lot. Nevertheless, they are usually restricted to 480VAC or less line voltage. Some are available for 600VAC but I would not use them in my products. I had bad experience at higher voltage.

You may find exceptions to my prudent ways, you may make a proto that works well with optos but don't bet anybody's life on a product.

Have fun experimenting and play safe.