Advice on Triac circuit for heater phase control

Thread Starter

ra5040

Joined Sep 26, 2018
42
Hi,

I have simulated this circuit using LTSpice and it appears to work fine:


However, the datasheet (https://www.mouser.com/ds/2/239/MOC302-1175440.pdf) shows the following circuit:



I've attempted a simulation to try to understand it (attached). The simulation sort of works (works better with the load on the -ve rail). But the resistor and capacitor values I've chosen are taken out of very thin air! The fact is that I don't understand the circuit, but assume that the additional components (compared to my circuit) are there for a reason.

Some of the load is carried by the MOC3020, but this seems a bit redundant as its maximum power dissipation rating is only 300mW, which isn't much help to the power triac.

I'm particularly interested in the capacitor as I suspect that it is there to stop false firing on the triac???

Help appreciated!

Robert
 

Attachments

MaxHeadRoom

Joined Jul 18, 2013
28,617
Generally heater control does not require phase angle control (MOC3020), the burst firing/zero crossing method is used instead, MOC316x,MOC308x etc.
See Fairchild AN-3006. Also Picmicro AN958
Max.
 

ebp

Joined Feb 8, 2018
2,332
The optocoupler is not intended to carry any of the load current, just trigger the triac. The load is normally in the main terminal 2 circuit.

The triac is triggered by current into the gate via the optocoupler. Once triggering has occurred, the voltage between the main terminals of the triac drops to something in the 1 to 1.5 volt range, depending on the specific triac. This voltage is low enough that current through the output of the optocoupler and into the gate will drop to a very low level because there simply is no longer enough voltage available between MT2 and MT1 via the coupler output and the gate-MT1 path to keep current flowing through the PN junctions involved. The effect is that you get a current "spike" of short duration into the gate.

The peak of the current does depend on where in the half cycle triggering begins. The resistor value may be chosen to be sure there is adequate gate current near the beginning and end of the half cycle while still adequately limiting current if triggering were to occur right at the peak. Usually the compromise is quite easy with triacs that don't require very large trigger current. For 240 VAC RMS the peak is around 340 V. The peak allowable current through the trigger device output is 1 A (spec is for 20 ms, which is far longer than the current would flow, but let's take 1 A as the maximum even for shorter time). A resistor of something around 470 ohms (a standard value) would probably be about right. The triac you are using doesn't require much gate current for triggering, so a higher value would still be acceptable. Having chosen the value you would then calculate the minimum instantaneous voltage at which triggering would occur using the triac trigger current specifications (e.g. at 5 degrees, the voltage would be sine(5°) x 340 = 29 V, so 470 ohms would allow about 62 mA gate current, well above the minimum spec for the triac you are using).

Triacs, including the output of the optocoupler, are susceptible to being turned on by very rapid change in voltage across the main terminals. With resistive loads, this is very rarely a problem because there is no mechanism to produce a fast rise in voltage. With an inductive load, a very fast rise in voltage can occur right after the triac turns off because "natural" turn-off occurs when the current drops below the triac's minimum "hold" current. With a purely inductive load, this happens when the line voltage is at its peak and the sudden appearance of the peak line voltage across the triac could cause it to turn on again. This is often dealt with with a "snubber" that slows the rate of rise across the triac. That is what the capacitor and extra resistor are for. Some triacs, such as the one you are using, have high tolerance for fast rising voltage, so they can often be used without a snubber. The optocoupler output is susceptible, so a snubber may be required anyway. With a heater as the load, you very likely would not require a snubber unless something else is putting some really nasty voltage spikes on your power line.

ap notes of interest
https://www.fairchildsemi.com/application-notes/AN/AN-3003.pdf
www.vishay.com/doc?84780
 

ebp

Joined Feb 8, 2018
2,332
when comparing phase angle control to burst mode control for heaters:

- phase angle control will produce some amount of radio frequency interference which may require "filtering" with an inductor (inductor doesn't so much filter as slow rate of change of current at turn-on so the noise isn't generated in the first place)
- burst mode control doesn't produce RFI if done using zero cross firing opto triggers
- burst mode control can cause annoying flickering of some types of lights if the controlled load is high - slight line voltage drop each time heater turns on causes light to dim a little, then light brightens when load turned off); usually not a problem, but really annoying it it occurs
- burst mode control can be totally asynchronous with the AC line when zero cross opto triggers are used, so there is usually no need for zero crossing detectors or the like
 

Thread Starter

ra5040

Joined Sep 26, 2018
42
Generally heater control does not require phase angle control (MOC3020), the burst firing/zero crossing method is used instead, MOC316x,MOC308x etc.
See Fairchild AN-3006. Also Picmicro AN958
Max.
I've had a look at the AN-3006 and at first sight it doesn't seem much different to the MOC3020, except that it isn't apparently designed for phase control. My intention with the MOC3020 was not in fact to do phase control, but to turn on/off whole half-waves at the zero-cross ... which is what I mean in the diagram by 'knocking out half-wave ... (firstly because phase control isn't necessary, and secondly because switching at zero, or thereabouts should avoid big current spikes ... and presumably also the need for a snubber).
 

Thread Starter

ra5040

Joined Sep 26, 2018
42
The optocoupler is not intended to carry any of the load current, just trigger the triac. The load is normally in the main terminal 2 circuit.

The triac is triggered by current into the gate via the optocoupler. Once triggering has occurred, the voltage between the main terminals of the triac drops to something in the 1 to 1.5 volt range, depending on the specific triac. This voltage is low enough that current through the output of the optocoupler and into the gate will drop to a very low level because there simply is no longer enough voltage available between MT2 and MT1 via the coupler output and the gate-MT1 path to keep current flowing through the PN junctions involved. The effect is that you get a current "spike" of short duration into the gate.

The peak of the current does depend on where in the half cycle triggering begins. The resistor value may be chosen to be sure there is adequate gate current near the beginning and end of the half cycle while still adequately limiting current if triggering were to occur right at the peak. Usually the compromise is quite easy with triacs that don't require very large trigger current. For 240 VAC RMS the peak is around 340 V. The peak allowable current through the trigger device output is 1 A (spec is for 20 ms, which is far longer than the current would flow, but let's take 1 A as the maximum even for shorter time). A resistor of something around 470 ohms (a standard value) would probably be about right. The triac you are using doesn't require much gate current for triggering, so a higher value would still be acceptable. Having chosen the value you would then calculate the minimum instantaneous voltage at which triggering would occur using the triac trigger current specifications (e.g. at 5 degrees, the voltage would be sine(5°) x 340 = 29 V, so 470 ohms would allow about 62 mA gate current, well above the minimum spec for the triac you are using).

Triacs, including the output of the optocoupler, are susceptible to being turned on by very rapid change in voltage across the main terminals. With resistive loads, this is very rarely a problem because there is no mechanism to produce a fast rise in voltage. With an inductive load, a very fast rise in voltage can occur right after the triac turns off because "natural" turn-off occurs when the current drops below the triac's minimum "hold" current. With a purely inductive load, this happens when the line voltage is at its peak and the sudden appearance of the peak line voltage across the triac could cause it to turn on again. This is often dealt with with a "snubber" that slows the rate of rise across the triac. That is what the capacitor and extra resistor are for. Some triacs, such as the one you are using, have high tolerance for fast rising voltage, so they can often be used without a snubber. The optocoupler output is susceptible, so a snubber may be required anyway. With a heater as the load, you very likely would not require a snubber unless something else is putting some really nasty voltage spikes on your power line.

ap notes of interest
https://www.fairchildsemi.com/application-notes/AN/AN-3003.pdf
www.vishay.com/doc?84780
Many thanks for taking the time to write such a comprehensive and useful note (tutorial)!. I need time to digest it, but from what I understand, the snubber isn't really needed for a purely resistive load, but it would be a good thing to have anyway (it can't do any harm and may prevent some false firing).
 

Thread Starter

ra5040

Joined Sep 26, 2018
42
when comparing phase angle control to burst mode control for heaters:

- phase angle control will produce some amount of radio frequency interference which may require "filtering" with an inductor (inductor doesn't so much filter as slow rate of change of current at turn-on so the noise isn't generated in the first place)
- burst mode control doesn't produce RFI if done using zero cross firing opto triggers
- burst mode control can cause annoying flickering of some types of lights if the controlled load is high - slight line voltage drop each time heater turns on causes light to dim a little, then light brightens when load turned off); usually not a problem, but really annoying it it occurs
- burst mode control can be totally asynchronous with the AC line when zero cross opto triggers are used, so there is usually no need for zero crossing detectors or the like
Very interesting ... I didn't realize that there was such a thing as a zero-cross opto-trigger! (If you could recommend one it would be much appreciated!).

My intention was not to do burst-mode but to selectively turn on half-waves to give a reasonably smooth feed (for example 010101010... for 50%). As the control will come from a microcontroller this is quite easy to do, as long as the firing is done through the microcontroller (which means that zero-cross detection will be required). However perhaps it's possible to get the zero-cross signal from the output stage of the zero-cross opto-coupler?(I realize that the signal would be too late, but it could be used to synchronize the controller timer). That would be really great because it would save on circuitry and complexity.
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
I've had a look at the AN-3006 and at first sight it doesn't seem much different to the MOC3020, except that it isn't apparently designed for phase control. ).
The AN-3006 states the MOC3020 is a Random phase opto, not a zero-crossing type as the MOC316x/MOC308x series. It shows an application and opto versions for each.
The Picmicro app sheet also shows the Triac use in a heat controller zero cross mode.
 

Thread Starter

ra5040

Joined Sep 26, 2018
42
The AN-3006 states the MOC3020 is a Random phase opto, not a zero-crossing type as the MOC316x/MOC308x series. It shows an application and opto versions for each.
The Picmicro app sheet also shows the Triac use in a heat controller zero cross mode.
Thank you ... I'll check this out.
 

Thread Starter

ra5040

Joined Sep 26, 2018
42
DodgyDave ... I've tried your circuit, below. Looks good, I think.



However, I think R3 is too low (gives 1.4A) and R8 gives 2.4A or thereabouts. Is there any reason why these resistors are so low-value?
 

Thread Starter

ra5040

Joined Sep 26, 2018
42

Attachments

Top