2
\$\begingroup\$

After a successful first design of my AC triac dimmer, I decided to improve it. I did this by adding a snubber circuit, so I could control inductive loads as well. I also added a EMI filter to reduce the noise it is outputting. The MOC3052 is controlled by a MCU with a dedicated zero cross circuit.

enter image description here

However, after assembling the circuit and testing it for a minute or so at 20% dimming, R4 started arcing. After that, the resistor seems to be broken because I couldn't measure any resistance over it.

Arcs

After measuring it with the scope, I figured out the voltage drop over the resistor is around 5.5v. However, this are small peaks and not a continues load. Could this be the reason the resistor is killed and started arcing? Right before failure, I can see the resistor degrading because the voltage over the resistor is rising in a quick tempo.

I also tried searching online for answers, however, none of the snubber circuit examples are mentioning a high wattage resistor. After simulating it in falstad (with a diac trigger circuit just for making the simulation more easy), I could see the peak is around 0.01ms long.

I'm unsure what else could lead to those issues. I'm also unsure how to solve this, or how to make a better design.

\$\endgroup\$
4
  • \$\begingroup\$ Is C3 on the other side of the PCB? \$\endgroup\$ Commented May 16, 2021 at 17:14
  • \$\begingroup\$ Yes, C3, C4 and L1 are on the other side of the board. \$\endgroup\$
    – jancoow
    Commented May 16, 2021 at 17:17
  • \$\begingroup\$ Have you looked at what happens if you use your actual 100 μH inductance in that simulation? It's some strange behavior, though I don't think it's the cause of this problem. \$\endgroup\$
    – Hearth
    Commented May 16, 2021 at 19:11
  • \$\begingroup\$ As you state, there are high peak currents because the switched waveform has high frequency components. In addition to the power dissipation in the resistor, it might also have caused the series capacitor to fail. Surface-mount components do not have much thermal mass to absorb short-term overheating. \$\endgroup\$ Commented May 16, 2021 at 22:01

1 Answer 1

2
\$\begingroup\$

When your resistor starts spitting out flames, it means the voltage across the resistor exceeded the maximum rated voltage and/or over heating. I'm guessing you're using 0603 resistors which means the peak voltage across the snubber resistor is exceeding the 75V rating. The snubber resistor will see the peak voltage of your input voltage. I would use a ceramic carbon resistor (Ohmite OX & OY series) for your snubber resistor since they can handle high joule loads. There are MELF film resistors made by Vishay (CMB series) that can handle pulse loads.

You have another problem. C4 & C1, in conjunction with L1, is causing ringing on the output. This causes ringing peaks about 2x the peak voltage on the output. You need to de-Q the inductor by placing a resistor across the inductor (around 10 ohms).

This circuit is easy to simulate in Spice. You can use a switch component for your triac to simplify things.

\$\endgroup\$
6
  • \$\begingroup\$ What I don't understand is how the resistor can see those peak voltages. I put the scope probes on both sides and the largest difference I measured was 5.5v. Also, I'm using 1206 high voltage resistors which can handle 500v and 0.25w: lcsc.com/product-detail/… \$\endgroup\$
    – jancoow
    Commented May 16, 2021 at 22:36
  • \$\begingroup\$ Yes, unfortunately I also saw those ringing issues. I saw voltages up to 432 Volts on the output! \$\endgroup\$
    – jancoow
    Commented May 16, 2021 at 22:39
  • 1
    \$\begingroup\$ The resistor you're using isn't rated for pulse loads. Snubbers take a beating. The glitch across the resistor is caused by the snubber capacitor charging up when the triac is off. The capacitor is discharged thru the snubber resistor when the triac turns on which puts the full line voltage across the resistor for a short moment. I suggest doing a Spice simulation so you can safely understand what's going on. \$\endgroup\$
    – qrk
    Commented May 17, 2021 at 17:11
  • \$\begingroup\$ Thanks, but still, I never measured those full line voltages across the resistor with the scope! I get that the resistor can't handle those peak loads and that is the reason it breaks. Lesson learned! \$\endgroup\$
    – jancoow
    Commented May 17, 2021 at 19:22
  • \$\begingroup\$ Are you using a differential measuring technique? Takes two probes to do this. \$\endgroup\$
    – qrk
    Commented May 17, 2021 at 19:28

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.