Variaring power through a Triac to control speed etc

[TimB]

Many moons ago I developed a speed controller for a motor. To control the speed I used a timer to decide when to turn on the triac following zero cross.

Today I was watching a video and they did not do that. They turned the triac on at the zero cross and let it do the full half wave. To control the power they omitted various half waves.

My question is, is that normal? I want fine control over the power going to a heater. I was thinking of just using the timed turn on. Now I'm thinking should I use the full half wave or a combination of both.

Any thoughts?

Cheers

Tim

[John Lawton]

Hi Tim,

why not? IMHO nobody in their right mind would use phase control for large amounts of power because of the EMI issue. Pulse density control is the system I described in my post here: https://protoncompiler.com/index.php?msg=6070. I used a varying bit pattern using 16 bits to control the amount of power delivered to the heater. It worked fine as the thermal mass absorbs the individual 50Hz half-cycle pulses.

[TimB]

Thanks

I went in the end for a MP740784 Link



It's a version of the Atten ST-862D that is a copy of the Quick 861DW. Eevblog did a tear down of it and it has a stand alone power board, So I aim to insert my own control board and use the parts for another project[
My aim is to get a controlled air temp of ~0.1c using a PT100 in the control PID loop. The issue is that its fast reacting so my have to slug the air temp to be able to get finer control.

Tim

[shantanu@india]

Tim,
That form of cycle control is called integral cycle control which is quite common in electrical heating. You just generate a PWM to drive a zero cross opto and the triac and the heater gets the required number of cycles to maintain the required temperature.
Don't ever try this for motor speed control because electrical machines need smooth flux for their operation... interrupted waveforms can be damaging

[Giuseppe]

In fact it is mainly used only for heating elements and by activating the triac right on 0 there is no EMI

[m.kaviani]

Many moons ago I developed a speed controller for a motor. To control the speed I used a timer to decide when to turn on the triac following zero cross.

Today I was watching a video and they did not do that. They turned the triac on at the zero cross and let it do the full half wave. To control the power they omitted various half waves.

My question is, is that normal? I want fine control over the power going to a heater. I was thinking of just using the timed turn on. Now I'm thinking should I use the full half wave or a combination of both.

Any thoughts?

Cheers

Tim

the
look! I was working in the industry about 10 years ago. all ovens had worked with the SSR that had a MOC series chip inside to control zc.
sometimes the current was a maximum of 40A per rail. if your load is greater than this is so better to control with a phase control device like two back-to-back SCR.
the circuit in Vaccum cleaner is based on simple DIAk, the important is price.

[TimB]

It seems the ST-862D is built on the one board. So it's going to be a pain to control the tirac. I really liked the seperate board in the Quick 861DW.

My issue is that I want really fine control. It may be with some way to provide a stabilising the heat eg putting something in the flow to slug the air temp from changing rapidly it will make stable enough that I can achieve my 0.1oc stability.

I will go for chopper triac control as I get more granularity with it. Its a shame the board is all in one. But at least I can try out hardware ideas on stabilising the stability before I design my own control board.

Tim

[okmn]

If I'm not wrong, you want to drive a resistive load with a phase control method and you say it should be sensitive.

Why don't you control the analog PI with the op-amp?by watching the zero crossings.

I think there can be no EMI problem with resistive loads.

In 2016, I made a half wave dc magnet driver for an iron and steel factory (following the zero transition) for 5 DC magnets consuming 100 Amps in DC24v.

i used a 2x thyristor module (it was siemens SKK series..)....

It was controlled by current with 4-20ma coming from PLC and still the system is running 24/7 ... if it helps I can share something

[See_Mos]

I am late viewing this post as I have been out of action for a couple of days.

Burst fire with zero crossing is the correct way to control heating elements. You can get the ON time down to one cycle if needed. Thermal inertia of the element and other factors will smooth out any irregularities.

Last week i repaired some three phase heating controllers using 95 amp SCR modules connected back to back.  They were from a motor baking oven, big enough to drive a van into.

I have also recently repaired some high current controllers used for big LED arrays on 120 volt supplies.  These use forward control where the LED is turned on at the zero crossing point then turned off part way through the cycle.  Good for LED control and very low EMI but needs series connected MOSFETS and opto couplers

[SeanG_65]

I could be wrong, but you need a zero crossing detector, then set up a timer for a 0-25ms delay. Turn on a thyristor at the start, then turn OFF after the timer ends. You can do this on both positive and negative half cycle. This will allow you to control the power going into whatever device you are using. Dont forget to opto isolate the thyristor drive.

[TimB]

I am late viewing this post as I have been out of action for a couple of days.

Burst fire with zero crossing is the correct way to control heating elements. You can get the ON time down to one cycle if needed. Thermal inertia of the element and other factors will smooth out any irregularities.

Last week i repaired some three phase heating controllers using 95 amp SCR modules connected back to back.  They were from a motor baking oven, big enough to drive a van into.

I have also recently repaired some high current controllers used for big LED arrays on 120 volt supplies.  These use forward control where the LED is turned on at the zero crossing point then turned off part way through the cycle.  Good for LED control and very low EMI but needs series connected MOSFETS and opto couplers

Thanks

The issue is that there is not a lot of thermal mass in the system and I need really fine control. The design is a high loss airflow so it constantly has power applied to it to keep the heat going in. One phase is like a PWM with a freq of 50hz and a duty of 0-50. I need a feedback loop that is working with <0.1oc and a time of <1 second.

I decided to rectify the mains to DC then use zero cross and a mosfet to turn on at zero and off the desired time. It's not linear but gives me finer control than 1 in 50 hz.

[Giuseppe MPO]

I decided to rectify the mains to DC then use zero cross and a mosfet to turn on at zero and off the desired time. It's not linear but gives me finer control than 1 in 50 hz.

If you have decided to rectify the mains voltage you do not need zero crossing, you can use some capacitors to decrease the voltage ripple and directly manage the mosfet with a PWM.

[See_Mos]

I decided to rectify the mains to DC then use zero cross and a mosfet to turn on at zero and off the desired time. It's not linear but gives me finer control than 1 in 50 hz.

Once turned on the triac will conduct until the end of the half cycle when the current is close to zero'  True you will get 1 in 100 instead of 1 in 50.

Giuseppe idea works well but you need a big capacitor.

Forward control doesn't need the big capacitor.  If cost is not critical then I can dig out the information that I have but I will need to email it as most of it belongs to the OEM

[Giuseppe MPO]

Once turned on the triac will conduct until the end of the half cycle when the current is close to zero'  True you will get 1 in 100 instead of 1 in 50.

In fact TimB doesn't use TRIAC but want to use a MOSFET

Giuseppe idea works well but you need a big capacitor.

The capacity depends on the current that is used and the final ripple that is desired which, in this case, is not very relevant

[See_Mos]

Correct, I missed the bit about using a MOSFET in post 10