TRIAC surprise

[S. Petrukhin]

Hi, friends!

I am trying to make a simple TRIAC control of a 3-phase load with reverse.
Shortened scheme:


There was a problem with the TRIAC opening spontaneously at the time of power on.
I know, but not very well, the vagaries of TRIAC in theory. But I can't figure out what the problem is.

At some time, a short circuit occurs when switching on. If it was possible to turn on without a short circuit, everything works steadily.

For verification, I have deattach cross triacs.
For direct TRIACs, the control pins are raised from the board and a locking resistor is connected to them only. There should be no interference with the control pin.


At the moment of switching on, there is some kind of transition process that gives a brief opening of the TRIAC.
The waveform one phase to neutral:


What could it be and how to deal with it? Please help me.

4 Quadrant Triac are used: https://datasheet.lcsc.com/lcsc/2105241606_KY-BT137S-800E_C2831693.pdf

Full schematic:
 

[KerimF]

Decreasing the resistance between G and T1 increases the triac limit of dv/dt though it increases the driving current which in turn increases the phase at which the triac is on (after zero crossing).

On your schematic, I read 620. If it means 620R (620 Ohm), it is relatively high (it has no significant effect).
The internal G-T1 is somehow like two reversed diodes connected in parallel (this can be checked using a multimeter).
So, the excess driving current for the 620R is 1/0.620 (assuming Vd=1) ~=1.6 mA only

A value of 33R (instead of 620R) may be needed here. The excess driving current for the 33R is 1/0.033 ~= 30mA. If the delayed phase is acceptable in your application, MOC can deliver this added triggering current.
For instance, the resistance, 620R, between a phase and MOC could also be reduced a bit if necessary (lower limit if 240Vrms is 450R with a margin around 25%).

One more point, here R8 is connected to U4 and U5. I personally prefer driving a MOC couple, as U4 & U5 for example on your schematic, by two separate resistors instead of one.

Hope this helps.

Kerim

[Doctorandus_P]

Triac's have parasitic parts too.

A part of those is capacitive coupling between the gate, and the other terminals.

As a result, fast voltage transients over the other terminals can couple enough current into the gate to trigger the triac.

Common remedies are:
1. Use a less sensitive triac.
2. Add a snubber.
3. Use a "snubberless" triack (High dV/dt), low capacitive coupling).
4. Short the gate in an attempt to re-direct this current (I'm not sure if this works).

[S. Petrukhin]

A value of 33R (instead of 620R) may be needed here.

Thanks for the help!
I'm radically reduced the locking resistor to 22 Ohms. The situation has improved, but, nevertheless, a short circuit sometimes occurs when switching on. Very rarely, but it appears.
Maybe I chose the wrong TRIAC for such purposes? I'm thinking of replacing it with the STM T1635-800G. Or use MOSFETs. 

One more point, here R8 is connected to U4 and U5. I personally prefer driving a MOC couple, as U4 & U5 for example on your schematic, by two separate resistors instead of one.
Hope this helps.

U4 and U5 will never turn on together, I think one resistor is enough for them. 

[strawberry]

back in the day ,I put 600..800V MKP10 capacitor across 240VAC . After week or so it went shorted.
input line filter or snubber could help maybe reduce fast/HF signals across triac or/and load
for 3 phase 240V systems usually use 1.2kV devices https://en.wikipedia.org/wiki/Rectifier ~560Vpk

[jonpaul]

inductive load transients and arcs in load

Use a properly rated snubber.

I would start with 5..10 nF 1000V VDE X in series with flameproof 100 Ohm across the triacs

j

[Kean]

Yep, add a snubber across each triac.  For guidance, see ST Micro AN437 or Onsemi AN1048/D.

[S. Petrukhin]

Yep, add a snubber across each triac.  For guidance, see ST Micro AN437 or Onsemi AN1048/D.

An interesting circuit design in an1048-d.pdf (see arrows) 

[Kean]

An interesting circuit design in an1048-d.pdf (see arrows) 

Interesting!