Power controll on 230V with zero switching and PWM?

[asrock70]

Have me

500W load on 230V 50Hz.
for switching we use a triac controlled by MOC3081 (integrted Zeco control logic)
We want to control the power from the MCU in the range 0-100% with a step of 0.5%

The MOC3081 only allows half-wave control
There are only 100 half-waves in one second.
For regulation with a step of 0.5%, we need to be able to control with a precision of one half wave every two seconds

And here I am clear, if we wanted to drive with an accuracy of 0.25%, we would have to control with an accuracy of 1 half-wave in 4 seconds.

Question is.
Is it possible to control a triac in this way using PWM without the MCU (STM32) knowing when the mains voltage passes zero?


If so, what is the relationship between the power and the set parameters of the PWM output?

If we want a power of 25%, we need 50 of the half halves to be switched on. What should the PWM setting look like?

 

[KlausST]

Hi,

With a zerocross MOC you don't need line zero cross information.
Line frequency is quite stable and accurate, thus use a PWM with 10ms step size (Xtal accurate) and you are done.

PWM duty cycle is equal to power ratio.
5% duty cycle = 5% power ... and so on...

Klaus

 

[FvM]

Running a half- or fullwave-switcher without mains synchronisation may result in unreliable switching when the switching edge is coinciding with mains zero crossing. I would suggest a zero crossing detector.

 

[asrock70]

Thank you for the information.

1. how is it, for example, with the MOC3083 switching only whole periods or half periods?
2. Zero crossing detector for MCU, maybe, but
- one of two high-resistance resistors is simple, cheap with low consumption, but still not galvanically separated
- anything else such as an AC optocoupler is relatively large and consumes power
I'm not saying I won't do it that way, I'm just trying to find out I can't do without it

The synchronization problem should be solved as follows.
Use a PID control loop

1 . I measure the temperature
2. I compare the measured and desired temperature
3. According to the size of the error (temperature difference) set positive pulse width in 1000ms slot. Full power = 1000ms ON , zero power 1000ms Off
4 After 1000ms goto to 1.
in 1s there will be only 1 pulse and regardless of the position with respect to the mains voltage, the uncertainty will be only in one wave or half-wave

Am I right?

 

[carpenter]

1. how is it, for example, with the MOC3083 switching only whole periods or half periods?

Interesting question.
I have found datasheets from several manufacturers , veryone writes about ZCD
20V MT1-MT2 Voltage above which device will not trigger.
Unfortunately, none of the manufacturers specifies what is MT1 and what is MT2
In the figure, the terminals are marked only as MAIN TERM.
The conclusion of the information from the datasheet is ambiguous

 

[KlausST]

Hi,

MT = Main Terminal.
Since it's an AC device, the order of MT1 or MT2 does not matter.

Klaus

 

[carpenter]

This is YES, but if I use any designation in the technical documentation, it should be unambiguous and precisely described. What is MT1 and MT2 is not described anywhere in the DS.
I found AN-3004 and here the ZCD is described better and apparently I have two ZCDs, one for each "thyristor". Then there is nothing to solve, it switches for every half-wave

PID and control loop

In principle it will work, the question is speed regulation?
Of course, we do not know the inertia of the whole regulated system and so only general reasoning.

What will ultimately control the PID is a variable that can range from 0 to 1000 with a step of 10 (0ms heating off 1000ms heating max power, 10ms half-period network). The control status of the system can take one of 101 values.
If we want to give the control loop the possibility to control at the smallest possible step, then with a control range of 1x per second and it will take about 100 s. Of course, PID can regulate faster (after larger quantities), but if the heating is able to raise the temperature by, for example, 10C in 1 s, such regulation will be poor.
It is therefore up to you to increase the speed and measure the temperature 10 times per second. However, we return to the beginning, the question of what frequency of PWM to use?

1Hz with the fact that I can change the pulse width with a frequency of 10Hz or use a higher one?

 

[KlausST]

Hi,

1Hz with the fact that I can change the pulse width with a frequency of 10Hz or use a higher one?

Makes no sense.
I'd synchronize the PID with the PWM. You may run the PID multiple times per PWM (oversampling style) but update the PWM only once.

Klaus

 

[BradtheRad]

Demo of SCR operation. Gate timing is determined by a resistor-capacitor combination. By adjusting values you can trigger a thyristor to turn on near the beginning of the waveform, or wait until the end.

A triac operates in a similar manner, with an AC supply.

 

[FvM]

Unfortunately, none of the manufacturers specifies what is MT1 and what is MT2
In the figure, the terminals are marked only as MAIN TERM.

Can't agree. Which datasheets did you check? Vishay e.g. does specify

--- Updated Jan 24, 2021 ---

If we want to give the control loop the possibility to control at the smallest possible step, then with a control range of 1x per second and it will take about 100 s. Of course, PID can regulate faster (after larger quantities), but if the heating is able to raise the temperature by, for example, 10C in 1 s, such regulation will be poor.
It is therefore up to you to increase the speed and measure the temperature 10 times per second. However, we return to the beginning, the question of what frequency of PWM to use?

I think it's the wrong approach. You'll choose an appropriate sample period for the time discrete controller according to the plant time constant and switch full mains periods in a sigma delta modulator manner.

Switching half waves results in DC load current and isn't complying with power quality regulations. Unfortunately, additional constraints for possible modulator patterns of higher power heaters are set by flicker requirements (IEC 61000-3-3).

 

[carpenter]

Yes Vishay specifies and the change does not refer to MT1 and MT2. Specify ZCD only as Zero cross inhibit voltage see IL410 and comapare with
MOC3081 from LiteOn or MOC3081 from ISOCOM or MOC3081 from Motorola

 

[schmitt trigger]

The distinction between MT1 and MT2 is relevant in discrete Triacs (with a physical gate), as the gate firing pulses require to be referenced to MT1.

But in an optically triggered Triac, as far as I know can be used interchangeably.