Driving Triac using Opto from A microcontroller

[UroBoros]

I need to control a a load (heater) using a triac. I need to use something like pulse skipping modulation to control the heater. The Microcontroller board and Analogue part will be working from a +12-0 -12V transformer based power supply. I think I can use the DC before filter with a transistor arrangement to trap the zero crossing point for microcontroller. MCU drives the trac using a zero crossing type opto for isolation. I have no plans to connect the neutral or ground from outside to my system. My understanding is that when opto is used no such live interaction is needed. Hope my plan is clear.

My confusion is, on searching I read about cold side switching , hot side switching, phase side etc. Also in some cases only one resister is connected to the gate of Triac via opto. Some circuits use another resistor from the other AC terminal also.
Also in this situation is the load connecting side or phase neutral connection is important? Will triac act like a conventional switch for AC or 'polarity' is important?
Thanks

 

[betwixt]

The 'polarity' is not important as is switches once per AC cycle anyway. The conventional way of connecting the Triac is in the phase (live) side of the circuit although it will work equally well in either leg of the circuit. Switching the phase side has a slight safety advantage in that the rest of the circuit is not live unless the Triac is conducting.

The second resistor is not essential but I would recommend using it. When the opto-coupler is not conducting the output side appears like an open switch so the Triac gate is left floating and susceptible to picking up static and interference, the resistor helps by keeping it tied in an 'off' state.

If you don't do the crossing detection at the Triac itself, be careful that you are checking it without phase shift. For example, if you are using the voltage from a low-voltage transformer on your microcontroller board, even if it is fed from the same AC supply, there may be a shift in the signal phase and the zero crossing time might be shifted. If you know the amount of shift you can compensate in software of course.

Alternatively, if you trigger your Triac with an opto containing it's own zero crossing switch, (MOC304x for example) all you need to worry about is the relative on and off periods to control the heater, timing at AC cycle level is done for you.

Brian.

 

[UroBoros]

Thanks. Clear About the resistor and 'polarity'.
I plan to use a zero crossing detector built in opto coupler because compensating phase shift and all seems extra work which can be avoided.
But to keep track of the wave form timings(which may vary slightly) should I use a zero crossing detector circuit and divide on time and off time strictly based on the periods measured from zero crossing circuit pulses? Or Simply set the on off timings in multiples of 20 milliseconds?(My 50 hz AC period). In this case the AC frequency changes will not be compensated and on off ratio may be changed in the AC?
Is it something to be considered or just a zerocrossing opto and a internally gerated 20 millisecond timing count is all that needed?

Thanks

 

[betwixt]

You probably don't want to use 20ms timing at all and you might be able forget about synchronizing to the AC supply altogether.

When turned on, the opto will not trigger the Triac for up to 10mS anyway, depending on when the next zero crossing occurs and the Triac will not turn itself off until the voltage across it is close to zero. You would be better to use longer timing periods so that an 'on period' is several AC cycles and not just one. Look at the type of heater you are using and decide on the timing based on the thermal mass of what you are heating. For example if you are heating a small object you can probably set the switching period to a second or more. If it's a 100L water tank you can probably time it by several minutes.

Brian.

 

[UroBoros]

Thanks a lot. I got the idea.
I was thinking about tracking the waveform and providing 'accurate' make and breaks.
btw. My object to be heated is a metallic block about 6 inches long and say 4 inches wide, it will be heated by two cartridge style heaters of 200w each inserted into holes drilled into this metallic block. This piece of metal block will be used to heat and seal a plastic sheet made pouch. The temperature to be maintained is around 150 D.Celsius. I think short bursts will give more control. But I have no experience in this. guidelines will be helpful.
Thanks a lot.

 

[nandhu015]

If you want to maintain the heat with low tolerance then use PID.

How do you measure the heat. Better have more than one sensor.

 

[betwixt]

Nandhu15 is correct. With a block of metal that size it will take several seconds of heating power to make a noticeable difference in temperature. You can use PID to give the most accurate control but if absolute accuracy and stability isn't too important and you can afford say +/- 2C, you could simply use a thermostat principle. Read the temperature and if it is out of allowed range, either turn on or turn off the triac control.

Brian.

 

[UroBoros]

Thanks.
+/-2C is not sufficient. My target is +/-1C.
My plan is to implement a PID based heating.
I am using PIC18 controller ( They have an application note for PID heater with code).
Sensing Temp is via a PT100. Still not finalized on the PT100 conditioning circuit.
Thanks again
Roy

 

[betwixt]

+/- 1C should be easy to achieve using PID methods. At 150C you will need to use a thermistor to sense the temperature so your feedback will be analog, it is too high to use a digital sensor, they usually can't go beyond 125C

Remember that using zero crossing control your switch on latency can be 10mS and when on, the switch off time is also 10mS, in other words you can't switch on or off until a zero crossing occurs. As the opto-triac internally controls the AC cycle timing, I would suggest you follow a strategy such as dividing the time slots into 200mS intervals, so your on periods are multiples of 10 full AC cycles. The more cycles you use the less important it is to synchronize your control with the AC waveform but the less precise the control becomes. Considering the mass of the block you are heating, you could probably use even longer than 200mS but other factors, such as radiated heat loss and conduction into the plastic pouch need to be evaluated to give a more precise figure. Remember that PID is a feedback based system so it will adjust to demand by itself anyway.

If this is a commercial product and cost is important, you can probably do this without the PIC and use a comparator and timer instead.

Brian.

 

[FvM]

According to the thermal mass of the application, a several seconds pulse period seems appropriate. Please notice, that a longer period will result in a better reproduction of the controllers manupulated value, because it's quantized to 10 ms. Exact control of fast acting heated zone needs a synchronisation to mains zero crossings to be aware of the actual output power, but it's not necessary with a several seconds period.

 

[UroBoros]

Thanks a lot. The optimum 'period' of the pulse skip modulation may be in seconds and I may have to watch the response of the heating block, especially when it makes contact with the plastic sheet so that temperature 'dips'. But considering the nature of the object, to start with, say we can choose 5 seconds. if I choose more period; after a contact with plastic sheet, more time will be needed by hater to react.
The rest I have to zero in by testing in real situation.

Thanks a lot