Using Triacs on Transformers

[johofz]

Hello dear EEs out there,

Comming from a Firmware-BG I'm starting to get more into electrical engineering.
My current project is turning a table top oven into a reflow-oven. The first iteration was pretty good for my needs. It did what I needed it to do in therms of controlling the temperature of the oven. I even was able to use the oven at work for temperature related experiments.

The only thing that bothers me is the following:
I use triacs and phase-angle controll to controll the power of the ac-loads. These are the heating-elements, a simple light and a fan, that is driven by a transformer. The loads all have thier own triacs. Some of you pros out there might already guess where the problem lies. Once I started the fan, I can't turn it off anymore by software. I suspect that the Transformer, being a huge inducter, keeps on pushing current through the triac preventing it from turning off.

Is there a way to prevent this? Some sort of free-wheeling diode maybe? I wasn't able to find good information on that sofar, so I thought I ask here on the forum. I might want to contoll AC-Motors at some point in the future and I woud guess they behave simillary to the Transformer.

Help on this would be much appreciated!

Greetings!
Johannes

[Zero999]

Why use phase control for heating elements? PWM, also known as burst control, should be used instead.

Why not power the heating elements from mains, rather than off a transformer? That would get rid of the need to switch an inductive load. If you want to use a lower voltage, for safety reasons, then how about a 12V electronic transformer, designed for halogen lamps, or using DC?

A freewheeling diode won't work, because it's AC. Use a snubber network, consisting of a resistor in series with a capacitor, to dampen the high voltage spike.

True AC motors are generally unsuitable for phase control, to vary the speed, other than the shaded pole type, used in small fans. Universal motors, which have brushes and are work off either AC, or DC, can be speed controlled with phase control. A TRIAC can be used for simple on/off control of an AC motor though, as long as it can handle the current surge, when it's turned on and snubber is added to damp the turn off voltage spike.

[johofz]

Why not power the heating elements from mains, rather than off a transformer? That would get rid of the need to switch an inductive load. If you want to use a lower voltage, for safety reasons, then how about a 12V electronic transformer, designed for halogen lamps, or using DC?

It's only about the fan. Light and heater are connected to mains directly.

Why use phase control for heating elements? PWM, also known as burst control, should be used instead.

I guess it wasn't clear enouth from my original post. Sorry for that. The heating elemets are connected to mains directly so PWM is not possible. Or do I miss something? I thought phase angle control is the equivalent technique for AC-loads.

A freewheeling diode won't work, because it's AC. Use a snubber network, consisting of a resistor in series with a capacitor, to dampen the high voltage spike.

The triacs I use are "snubberless" triacs from ST: https://www.mouser.de/datasheet/2/389/cd00002267-1795764.pdf
I thought that means that I don't need a snubber network on my pcb since there is one in the triac itself. Maybe that is the mistake.

[Zero999]

Why not power the heating elements from mains, rather than off a transformer? That would get rid of the need to switch an inductive load. If you want to use a lower voltage, for safety reasons, then how about a 12V electronic transformer, designed for halogen lamps, or using DC?

It's only about the fan. Light and heater are connected to mains directly.

Why use phase control for heating elements? PWM, also known as burst control, should be used instead.

I guess it wasn't clear enouth from my original post. Sorry for that. The heating elemets are connected to mains directly so PWM is not possible. Or do I miss something? I thought phase angle control is the equivalent technique for AC-loads.

PWM is easy for AC and is different to phase control. It involves turning the load on and off, at a much lower frequency than the mains, often under 1Hz, so is only useful for heating.

A freewheeling diode won't work, because it's AC. Use a snubber network, consisting of a resistor in series with a capacitor, to dampen the high voltage spike.

The triacs I use are "snubberless" triacs from ST: https://www.mouser.de/datasheet/2/389/cd00002267-1795764.pdf
I thought that means that I don't need a snubber network on my pcb since there is one in the triac itself. Maybe that is the mistake.

Yes, it should be fine, especially for a small motor.

Is the fan DC and there's a rectifier on the transformer? If so, then why not use a transistor to switch the fan on the DC side?

[drussell]

Why use phase control for heating elements? PWM, also known as burst control, should be used instead.

I guess it wasn't clear enouth from my original post. Sorry for that. The heating elemets are connected to mains directly so PWM is not possible. Or do I miss something? I thought phase angle control is the equivalent technique for AC-loads.

PWM is easy for AC and is different to phase control. It involves turning the load on and off, at a much lower frequency than the mains, often under 1Hz, so is only useful for heating.

Yes.  Because it is a heating load and the element will have significant thermal mass, you can switch it on and off at a very low speed (compared to what speed you would use for say, directly dimming LED lights,) even on the order of seconds like Zero999 points out and simply use the slow rate of temperature change of the heating element due to the thermal inertia to your advantage to smooth things out and still get very tight "regulation" if you wish by varying that average power. 

You can still easily get close temperature regulation at 1 Hz as compared to, say, a typical mechanical thermal switch / thermostat where the heater on/off cycles are on the order of many seconds or minutes.  You don't need kHz though, and you don't even need to be "regulating" it on every half-cycle to get "smooth" heat regulation, not even close.

[drussell]

The only thing that bothers me is the following:
I use triacs and phase-angle controll to controll the power of the ac-loads. These are the heating-elements, a simple light and a fan, that is driven by a transformer. The loads all have thier own triacs. Some of you pros out there might already guess where the problem lies. Once I started the fan, I can't turn it off anymore by software. I suspect that the Transformer, being a huge inducter, keeps on pushing current through the triac preventing it from turning off.

It's only about the fan. Light and heater are connected to mains directly.

So, are you just trying to turn the fan on and off, or are you trying to regulate the speed?

Is the fan DC and there's a rectifier on the transformer? If so, then why not use a transistor to switch the fan on the DC side?

Yeah...

How is this fan set up and what precisely are you trying to accomplish here, what is the behavior you are trying to achieve? 

I guess I wasn't understanding what the original question really was.   

[perieanuo]

old power supply designs used triacs (primary side) to reduce transformer output to minimise heat dissipation on final transistors, I saw some french supplies doing this, but also US-ones.
it's veeeery tricky to deal with the triacs on primary side of transformer, but not impossible
but that's usually not a begginer's area of competence
if you really want, there is some possibilities, but imho adapting your specific transformer didn't worth the time, try to solution on secondary side, much more simple
see https://www.eevblog.com/forum/beginners/is-a-triac-on-the-primary-side-of-a-transformer-a-good-idea/
[edit]see also https://electronics.stackexchange.com/questions/329071/triac-controlled-voltage-preregulator-mains

[richard.cs]

PWM is easy for AC and is different to phase control. It involves turning the load on and off, at a much lower frequency than the mains, often under 1Hz, so is only useful for heating.

I did this for a kiln once, controller switched full mains cycles, a variable number of cycles on out of ~100 so about 0.5 Hz. The modulation visible in in all lights in the house (including CFL) was impressive, especially given there was no shared cabling downstream of the meter. I used a fairly dated Telefunken chip, I can't remember the part number but it and its clones were common in this application.

Edit: it may have been Telefunken, later Temic, later Atmel, U208B or U2008B. I think it was a U208B, the one with the extra zero is a similar chip but with load current sensing. Of course in the modern world you would use a microcontroller.

[Siwastaja]

Use off-the-shelf SSR (Solid State Relay) modules which internally use a triac, but which have zero cross detection and basic protection features built-in. These are opto-isolated, typically with constant current LED drive circuitry accepting any DC control signal from 3 to around 30V; you drive it easily from any arduinorasperrypi.

The fact they detect zero crossing and switch full mains cycles mean you get rid of EMI, buzzing sound, etc. The downside is, this limits the maximum PWM frequency, but with heaters that have thermal mass, it's no problem. 1Hz would be typical. (With mechanical relay, you would use like 0.1Hz or even slower to limit switching cycles that wear the contacts; with SSR, this isn't a problem.)

Such modules also limit your exposure to having to deal with mains, to only connecting two wires like any switch, and offering opto-isolated (safe) interface. This is great for beginners and people who actually shouldn't be messing with mains at all.

[David Hess]

The triacs I use are "snubberless" triacs from ST: https://www.mouser.de/datasheet/2/389/cd00002267-1795764.pdf
I thought that means that I don't need a snubber network on my pcb since there is one in the triac itself. Maybe that is the mistake.

Snubberless TRIACs have higher di/dt and dv/dt ratings which helps prevent false triggering but you may still need to add an external snubber for correct operation with a difficult load, like a transformer.

[johofz]

Thank you all for the detailed answers. It helped a lot already!

I decided to use phase-angle-control for the heater because I wanted to learn about the technique. It may be over the top for a heater application in a reflow-oven, but for smaller thermal masses like soldering Irions, It seems apropriate to me. That is something I wanted to do in the future at some point.

I have full galvanical Isolation through opto-couplers and even slots in the PCB. I made sure to seperate mains and logic as far as possible. Power for the DC-Logic side comes from a AC-DC module with internal isolation.

At first I thought beeing able to control the fans speed would be nice, hence the triac instead of a relay. But it turns out to be either on or off so if I can't fix that I will just use a relay for the next iteration.

I'm not very experienced with transformers. The one I have here looks like it does not have a rectifier, but I could be wrong. According to it's labeling it is a YJ48-230. To me it looks like a cookie-cutter AC-Motor with a Transformer. I'm currently doing a revision of my pcb. When I'm done I'm posting scematics and my solution (probably a relay).

[drussell]

I'm not very experienced with transformers. The one I have here looks like it does not have a rectifier, but I could be wrong. According to it's labeling it is a YJ48-230. To me it looks like a cookie-cutter AC-Motor with a Transformer.

Are you sure it's not just a typical shaded-pole motor rather than some sort of transformer?

[Zero999]

A Google image search for YJ48-230 shows lots of shaded pole motors, rather than transformers.
https://www.google.com/search?q=YJ48-230&client=firefox-b-d&source=lnms&tbm=isch&sa=X&ved=2ahUKEwj4uqjrvsfvAhUQhlwKHUudA_4Q_AUoA3oECA0QBQ&biw=1920&bih=916

It's possible to phase control the speed of a shaded pole motor driving a fan, but it isn't every efficient.

[drussell]

A Google image search for YJ48-230 shows lots of shaded pole motors, rather than transformers.
...
It's possible to phase control the speed of a shaded pole motor driving a fan, but it isn't every efficient.

Indeed...

Well, shaded pole motors are always inherently super inefficient anyway but are useful for many things due to inherent characteristics whether in a clock motor or things like a fan where even if something stalls completely the design of the motor makes it barely draw any extra current in locked rotor condition...  It's just one of those odd-ball really old-style designs that is still useful in some applications...

... anyway not surprising that this is probably a shaded pole motor not some sort of transformer -> DC -> Fan arrangement which all makes more sense now as to what the OP was trying to do...

[johofz]

Are you sure it's not just a typical shaded-pole motor rather than some sort of transformer?

 

In my defense It does look alot like a transformer to the untrained eye.

Is the problem with inductive loads and triacs, that the current is out of phase with the voltage? Or is the di/dt of the snubberless triac not sufficient?
I might have found the reason for my issue. I use a optoisolator to trigger the triacs. The optoisolator does not have a snubber-network...

[drussell]

Is the problem with inductive loads and triacs, that the current is out of phase with the voltage?

Essentially yes, if the TRIAC never "sees" zero volts it will continue to conduct just like it would with DC, it won't shut off until it actually "sees" a zero crossing.

If it were a transformer, the voltage/current angle would be determined mostly by the secondary load as transformers do a pretty good job of reflecting the secondary load back to the primary, so if it is purely resistive it will "look" resistive on the primary side of the transformer, (which is why we were curious as to what the circuitry of the fan actually was to know what type of load it would "look" like electrically.)

[Terry Bites]

PWM isn't burst control.

[Terry Bites]

Use solid state relays that are inherently zero cross switching. Not all are so check first.
Also see rambling under this Topic: Is a triac on the primary side of a transformer a good idea?  (Read 4910 times)
Triacs arrrgh!

[Zero999]

PWM isn't burst control.

Yes it is. Here's a reference to burst control. Whilst it isn't referred to as PWM, its description matches. The difference is the on/off times can only be in units of half a cycle (a full cycle is better though), so if the frequency is kept constant, then it gives discrete power levels, rather than a continuous range.


http://www.crydom.com/en/tech/newsletters/solid%20statements%20-%20ssrs%20switching%20types.pdf

Use solid state relays that are inherently zero cross switching. Not all are so check first.
Also see rambling under this Topic: Is a triac on the primary side of a transformer a good idea?  (Read 4910 times)
Triacs arrrgh!

It turns out the TRIAC is not controlling a transformer, but a shaded pole motor driving a fan, of which he wants to control the speed, so a random firing TRIAC/solid state relay is required i.e one which doesn't have zero-crossing.