2
\$\begingroup\$

The circuit is a MCU triggered, capacitor based fan speed controller.

The circuit works fine when a single triac is connected but when both triacs are wired and triggered alternately, any of the triac fails randomly by shorting. Initial load is a 60W incandescent bulb.

The triac gate takes close to 50mA through optocoupler. This seems much higher as the triac has senstive gate which can be triggered by a low power MCU.

Kindly help. Fan speed Controller

MOC3041 Datasheet https://www.fairchildsemi.com/datasheets/MO/MOC3041M.pdf

\$\endgroup\$
4
  • 1
    \$\begingroup\$ This looks like a circuit to burn out resistors. What is it intended to do? \$\endgroup\$
    – Spehro Pefhany
    Aug 11, 2015 at 13:52
  • \$\begingroup\$ Apart from the circuit deficiencies why use 2 triacs at all. Why not combine the logic signals and drive a single triac? \$\endgroup\$
    – Kevin White
    Aug 11, 2015 at 16:28
  • \$\begingroup\$ @ Spehro Pefhany: Its the digital replacement for the rotary type capacitor ceiling fan controller. Each Triac is connected in series to a capacitor of different value; say 1uF,2.2uF,3.3uF etc. \$\endgroup\$
    – Jack
    Aug 12, 2015 at 6:40
  • \$\begingroup\$ @ Kevin White: Each Triac has a series capacitor of different value to step down/Step up fan speed. \$\endgroup\$
    – Jack
    Aug 12, 2015 at 6:46

2 Answers 2

Reset to default
1
\$\begingroup\$

For one, I'm surprised that this circuit functions at all, even with one triac in place. The power triacs (BT136-600D) connections should be flipped, with the AC hot wire connected to MT2 and the AC neutral wire connected to MT1 as can be seen in the datasheet.

As can be seen here

Triacs are thought of as completely symmetrical devices but in reality their slight asymmetry makes them not function correctly when the terminals are flipped. This could possibly be the solution to your problem. Additionally, depending on what other components you have hooked up to your microcontroller, it is possible that you are attempting to draw to much current from it. The datasheet for the MOC3041M says that the max current draw of the phototriac is 15 mA, this can easily be supplied by a single pin of most microcontrollers but there is still a maximum total for the MCU which could be exceeded by adding an additional MOC3041M.

\$\endgroup\$
4
  • \$\begingroup\$ The terminals of the Triac are connected wrong in the schematic but not in the actual circuit and it works fine with 1 Triac in wiring. The problem happens when MT1,MT2 terminals of the Triac which is not triggered also get the AC mains when the other Triac is triggered. You can see that MT1 of both Triacs are connected togather. Is there another way to do it? Or a different circuit other than using SSR? \$\endgroup\$
    – Jack
    Aug 12, 2015 at 6:43
  • \$\begingroup\$ I guess I'm not fully understanding your problem, are you saying that both triacs turn on when one is triggered? If so you may want to use a separate snubber circuit for each triac rather than one for both triacs. Also, sharing the resistor R2 between the phototriacs seems unwise. You should attempt to separate the triac circuits as much as possible to make the triggering mechanisms as independent as possible (i.e. separate snubbers, resistors etc...) \$\endgroup\$
    – nreath
    Aug 12, 2015 at 12:33
  • \$\begingroup\$ to nreath: goal is to replace the rotary type capacitor regulated ceiling fan speed controller. In the circuit MCU controls instead of rotary. The trouble is when both Triacs are wired and triggered separately, any one of them shorts. The Triac driving part of the optos are given input through R2 as any one will be conducting as per trigger from MCU. When power terminal pins of Triacs are swapped, it conducts slightly even without trigger to opto. Please refer this: electronics.stackexchange.com/questions/185292/digital-replacement-for-rotary-capasitor-based-fan-speed-controller \$\endgroup\$
    – Jack
    Aug 12, 2015 at 13:15
  • \$\begingroup\$ Spontaneous triac triggering is often caused by a sharp rise in dv/dt or di/dt, both of which are strongly associated with inductive loads (such as the motor which you are attempting to control). The snubber you have with R1 and C2 should help with this but improper selection of values can easily nullify this effect. Snubberless triacs are made which do not operate in the 4th quadrant and have high tolerance for high dv/dt and di/dt characteristics. The following is a good and cheap example. st.com/web/en/resource/technical/document/datasheet/… \$\endgroup\$
    – nreath
    Aug 12, 2015 at 14:52
0
\$\begingroup\$

I suspect there are times when both triacs are turned on, either because they both have gate current or you are turning on the second before the previous one has had time to turn off.

The 440K resistor across the caps will help but that is a time constant of more than a second so won't help much if you alternate them faster than that.

If so and there is any charge on C1 or C3 you will have a loop with both triacs and both capacitors all in series with no current limiting. Extremely high currents could develop and ultimately damage one of the trials.

You need to leave some dead-time between the two triacs being active.

What is the purpose of alternating them?

\$\endgroup\$
1
  • \$\begingroup\$ to Kevin White: As you suggested correctly, I did alternating at 1 second and lesser durations. Alternating is just to test the working of triacs. Every time before failing, the triacs functions around 5 seconds. About leaving dead time: I am trying for a smooth step and a 0.5uF cap will connect in series in between the steps to make the speed change more smooth. So lesser timing of triac firing is more suited. Is there a better way of doing this? Many thanks for your thoughtful suggestion. \$\endgroup\$
    – Jack
    Sep 14, 2015 at 11:43

Your Answer

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

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