I am having 3 triacs and 3 relays that are used to control heater elements in triangle configuration. Atleast the triacs need some kind of surge protection since they can only handle 800V and surges/efts can be atleast 4kV.

I would also want to protect the enviroment since switching 13 Amps in 3 phases is a big change that can cause problems to other devices. There is no current over the relays when they are switched. The triacs are first turned off and only then can the relay be switched on/off.

The switching of triacs currently causes problems with the device that is controlling it. It resets the controller some times. I need to add protection there aswell, but that is another topic. It is still important to mitigate spikes caused by these triacs and relays.

What kind of protection would you suggest? Options that I have been thinking about:

  • Varistors to neutral
  • Varistors in triangle
  • Snubber circuit over Triacs
  • Snubber circuit over Relays and Triacs
  • TVS diodes over Triacs

Description of the system: The Upper relays are always on if the system is heating. There are 3 heating modes for the system.

  1. Low heating where switching relay RSW is at L3. And only 2nd triac is open. In this mode current flows through all 3 Heating elements between L2 to L3.
  2. High heating where RSW is at L2 and all the Triacs are open. Here the heating elements are in triangle between L1, L2 and L3.
  3. Medium heating where RSW is at L2 and either L1 or L3 triac is not open. So there is full power between L2 and L1/L3

enter image description here


1 Answer 1


Triacs will turn on to (maybe) protect themselves with an overvoltage surge and voltage will be dropped across the load. The MOC3063 is only 600V rated so it will probably turn on first and trigger the triac. You could parallel the MOC3063 with a bipolar TVS to trigger the triac at a lower voltage than the MOC3063 rating.

I don't think varistors will do much and are another point of potential failure. They are also have a very soft 'knee' and you might have to get higher voltage optos and thyristors for them to have much effect. If they are conducting regularly then they will wear out faster. If you have a specification to meet on the transient it will be easier to come up with a design. You might need to slow the dv/dt to keep the thyristor from damaging itself (or use a much heftier thyristor). The fundamental problem with protection is that the semiconductor devices are actually pretty fragile compared to possible transients (and compared with things like wires and mechanical contacts) unless you grossly overrate them, which costs money (and likely requires a lot more trigger current).

You are apparently already zero-voltage switching, so noise is relatively low. As you probably know they don't actually switch right at the zero voltage point, however that noise is relatively consistent.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.