I'm having problem with triac handling. The plan is to use it to drive traffic light signals. The trigger is fine, but I want to use the hold current as a burnt bulb detector. Meaning, if the current goes under the holding threshold (ie: the bulb's filament is burned), the triac should go off.

What I don't understand, is why the triac doesn't go off after the pulse, even if the triac's output is completely open (disconnected). From my understanding, if there is not current flowing between A1 and A2, the triac should go off. The yellow line is the gate pulse, the blue line is the output of the U5 opto.

What am I doing wrong?

The optotriac is MOC3083

The triac is T1623-600G

The optocoupler is LTV-354T

Schematic scope1 scope2

  • \$\begingroup\$ where is the bulb connected? Red_Out to Neutral? \$\endgroup\$
    – user16324
    Mar 6, 2020 at 19:41
  • \$\begingroup\$ triac snubbing, as in 3083 data sheet? \$\endgroup\$
    – Neil_UK
    Mar 6, 2020 at 19:47
  • \$\begingroup\$ I wonder if you are reading the MOC3083 Opto Isolator instead of the main triac. \$\endgroup\$
    – vini_i
    Mar 6, 2020 at 19:58
  • \$\begingroup\$ Triac is a snubberless, but tried also with standard triac. Bulb is connected between Neutral (L1_NEUTRAL) and L1_RED_OUT. \$\endgroup\$
    – cezane
    Mar 6, 2020 at 23:27
  • \$\begingroup\$ maybe putting the sensor optocoupler in series with the MOC3083 would work better (it'll need a backawards diode in parallel). \$\endgroup\$ Mar 6, 2020 at 23:27

1 Answer 1


enter image description here

Table 1. Extract from the MOC3081 series datasheet.

The opto-triac has a very low holding current. It's likely that you're getting enough leakage through U1 and R13.

I note that you seem to be triggering only every second half-cycle.

From the comments:

Now, what exactly is a Holding Current on a opto-triac?

Basic triac theory tells us that once triggered an SCR (silicon controlled rectifier) - of which the triac is a bi-directional type - will remain on until the current through it falls below the holding or hold-on value. We use this feature on AC dimmer circuits, etc., to our advantage by giving a short pulse to the triac gate when we want to turn it on and let it turn itself off at the next zero-cross.

enter image description here

Figure 1. The upper trace shows the trigger delayed close to the end of the cycle. The resultant effective voltage is low. The lower trace shows the trigger close to the start of the cycle. This will result in close to full voltage. The relationship between phase angle delay and resultant RMS voltage is graphed on the right. Source: Opto-triacs, solid-state relays (SSR), zero-cross and how they work,

For larger triacs the holding current might be 50 mA or so. (It's a long time since I've checked.) For your teeny opto-triac it's only 0.25 mA. Since U5's opto-LED will draw more than that it holds the triac of U1 on once triggered until the next zero-cross. The linked article (by me) may be worth reading.

But I'm lost now, how can I achieve the burnt bulb detection?

Two ways:

  • Sense the current while the lamp is supposed to be on. This will be a bit bulky and will require some sort of current sensor - maybe a Hall effect type.
  • Monitor the voltage across your triac, D5, while the triac is off. If the lamp is OK you'll have full mains voltage. If the lamp is open-circuit you will have none (provided you remove your other monitoring circuit). To do this you can add an opto-isolated zero-cross monitoring circuit (of which there are very many on the web and on this site - and not all of them are good) across the D5. You'll have to figure out the logic of testing and, for example, if you need to monitor the lamps while on then you'll have to switch them off periodically for a half-cycle or two to check them.
  • \$\begingroup\$ Now, what exactly is a Holding Current on a opto-triac? \$\endgroup\$
    – cezane
    Mar 7, 2020 at 8:43
  • \$\begingroup\$ See the update. \$\endgroup\$
    – Transistor
    Mar 7, 2020 at 9:12
  • \$\begingroup\$ Thanks a lot @Transistor for your explanation. You are right, the opto-triac is holding the triac's gate open. I've added a new optocoupler (LTV-354T) to the gate to confirm it. But I'm lost now, how can I achieve the burnt bulb detection? How can I stop the current flowing from opto-triac? I've removed the R13, behavior is the same. \$\endgroup\$
    – cezane
    Mar 7, 2020 at 10:40
  • \$\begingroup\$ Thanks again @Transistor for the solutions. I had in mind the hall effect sensor, but it gets a lil more expensive for all channels. I have tested with a opto-isolated monitor circuit (basically moved the pin 2 from U5 to L1_HOT) and I'm getting the mains whilst the triac is blocked. Two more questions, if you indulge me again. Would this solution still work for LED bulbs (with AC-DC electronics), and, Why the opto-isolated has to be zero-cross (because the current LTV-354T) seems to do the job. \$\endgroup\$
    – cezane
    Mar 8, 2020 at 15:01
  • \$\begingroup\$ It should work for LEDs but they vary in their internal circuits. I didn't say that U5 has to be zero-cross. In fact, it can't be. Zero-cross is only relevant to opto-triacs not to opto-transistors which you will require for interface with your controller. \$\endgroup\$
    – Transistor
    Mar 9, 2020 at 18:49

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.