Fine temperature control (Triac + temperature sensor)

[Gryd3]

I've been doing a little digging online and on the forum and found some basic circuits that utilize a Diac and a Triac to control a load. The Diac/Triac combo is triggered by a very basic RC circuit that usually has a potentiometer/trim-pot in series to adjust the timing that the Diac/Triac fire.

My question involves adding feedback. Instead of being able to adjust the output from 0-100%, I would like to adjust the output temperature. A temperature probe would adjust the timing to fire sooner when the temperature is too low, and fire later when the temperature is too high.

The example circuits I keep finding are attached.

Is what I'm trying to accomplish as easy as connecting a variable resistive temperature sensor to the RC timing portion of this circuit? I don't like the idea of having 110 or 220 on a temperature probe.

I would also like to avoid using a microcontroller and stick to some good ol' fashioned discrete components, but would like to avoid the temperature regulators you find on cheap waffle irons...

Tips, tricks, suggestions, or research topics are welcome.

 

[(*steve*)]

Look up PID controllers

 

[Gryd3]

Look up PID controllers

Done some light reading on these things. I'll dig a little more. I had thought that would be me over-engineering it
I am not concerned with the rate at which the temperature changes, as the probe will be in close proximity to the heat source.
Thanks Steve for the suggestion

 

[KrisBlueNZ]

I don't have much experience with control loops and PID, but I suspect you don't need PID either. At worst, you could do simple on/off thermostat-type control, but I imagine that any proportional system would give better performance than that, without PID.

You could try to control the firing angle of the triac directly from a temperature sensor, if you replaced the resistor in that circuit with an NTC, but I doubt you would find an NTC with a wide enough resistance range over a narrow enough temperature range to give you proper control. In any case your temperature sensor would be potentially live.

I think it's probably best to use a transformer and isolate the control circuitry from the power switch. That requires zero crossing detection and probably a comparator to fire the triac, via a triac driver optocoupler, and probably an op-amp to scale the temperature sensor range to the firing angle. So you've immediately added a transformer and at least three ICs.

See, I tend towards the over-engineering approach myself :-(

If you can handle the temperature sensor being potentially live, or you can ensure that phase and neutral are always connected round the right way, you could avoid the transformer and use a capacitor-fed power supply. That would eliminate the need for the optocoupled triac driver too, and might simplify the zero crossing detection.

 

[Gryd3]

Thanks for the reply Kris.

II doubt you would find an NTC with a wide enough resistance range over a narrow enough temperature range to give you proper control. In any case your temperature sensor would be potentially live.

I'm nervous about making it live, and would probably end up using an isolation transformer if this were the case..
I did not know the resistive sensitivity of the probe would need that kind of characteristic, but could I not compensate with a different value capacitor? Or is this typically a fixed value with this kind of application?

 

[KrisBlueNZ]

It's really just the range of resistance, and a gut feeling. You would want the thermistor to vary your conduction angle over most of the 0~180 degree range over a relatively small temperature range, for reasonably accurate control. So you would need an NTC where most of its resistance range maps to a small part of its operating temperature range. I doubt that such a thing exists. I could be wrong.

The resistance vs. temperature characteristic of standard NTC thermistors is defined by about four values that plug into a formula. From that, you can graph resistance vs. temperature. Not all thermistors work that way though. You'll need to do some research because I don't know much about thermistors either. Start with Wikipedia I guess, and application notes from thermistor manufacturers (see the Digi-Key selection guide next).

Here's a selection guide at Digi-Key that would probably contain a suitable one, if there was one. http://www.digikey.com/product-sear...ColumnSort=16&stock=1&quantity=1&pageSize=250

 

[Arouse1973]

What about using the property of a magnet, that it's magnetism decreases with a rise in temperature and increases with a reduction of temperature, you could then sense this and maybe use this information to adjust your angle? Ultimate isolation!
Just a thought.
Adam

 

[Gryd3]

@adam, that's a unique solution and would not even begin to know where to start implementing it
I think I'm going to have to add some additional components to be able to properly compensate for the thermistor values.

@kris, I hopped on Digi-Key for a quick look, but this would be my first application with a temperature sensor and was surprised by the amount of different kinds available. I'll do more digging and see what I can do.


*On a side note, I've got LTSpice and am having issues figuring out some of the basics to be able to simulate a circuit. Is this something I should power through and learn, or do you suggest a different package?

 

[Arouse1973]

What issues are you having? Both myself and Kris have used it a fair amount. But Harald is the resident boffin I do believe.
Adam

 

[Gryd3]

Very very basic beginner issues. I don't know how to add a power source yet which is preventing me from doing simulations.
I'm sure I can look it up and learn how as I do have the internet at my fingertips, but am curious if there is a better package perhaps that you have used and enjoyed.

 

[Arouse1973]

To add a power source go to components, it's the AND gate symbol on the tool bar at the top. It called voltage. I have used Tina Ti it's free but limited in features.
Adam

 

[Gryd3]

To add a power source go to components, it's the AND gate symbol on the tool bar at the top. It called voltage. I have used Tina Ti it's free but limited in features.
Adam

Haha! I've added components easily enough. I figured the sources would be a different option XD
I take it, it's worth my learning though.

 

[Arouse1973]

Yeah it's a very good package, I changed because I wanted something better for producing documents.
Adam

 

[Arouse1973]

Here is what Tina can do. Although this can be done in LT Spice it's a bit more intuitive in Tina. I selected a variable for the 100K and made it step in value 5 times from 1K to 100K. VF1 is the output plotted against the input waveform.

 

[KrisBlueNZ]

For power sources use "voltage" and "current" under the main components menu, or "battery" under the "misc" subfolder.

Make SURE you put a 0V point (the downward-pointing triangle) on your 0V rail or LTSpice will barf badly.

 

[Gryd3]

Thanks for the tips guys!

 

[Gryd3]

Triac behavior question:
A triac can be triggered with a positive or negative voltage at the gate. What is the reference point for this voltage. If the gate is at 0 and there is a voltage at A1(MT1), then the gate would have an equal opposite voltage present... or is this in relation to A2(MT2)?

I've noticed on my diagram and Arouse's that the Gate enters the Triac from the bottom of the component symbol.

I would like to control this with a some additional components, as mentioned above I will need to condition the results from the temperature probe to be able to control the timing of when the Triac fires. Other than google something done already, I would like to understand the component in greater detail. I'll be going over the wikipedia page a couple more times to see if I can answer my own Question before you guys chime in

 

[Gryd3]

... Never mind, should re-read before posting. The gate voltage is in reference to A1(MT1), but I am still curious if I can leave the gate floating, or if I should tie it to the same potential as the A1 lead.

 

[KrisBlueNZ]

Why would you want to leave the gate floating?

Normally the gate is driven from a low impedance source that's referenced to MT1, which is normally the 0V rail of the circuit. Often there's a low-value resistor, and sometimes also a capacitor, between the gate and 0V as well. The gate normally sits around 0V and a pulse (usually positive) into the gate fires the triac.

In the case of the simple lamp dimmer, the pulse is generated by the breakdown of the diac, which occurs when the capacitor connected to the top of the diac reaches the diac's breakdown voltage.

 

[Gryd3]

Why would you want to leave the gate floating?

Normally the gate is driven from a low impedance source that's referenced to MT1, which is normally the 0V rail of the circuit. Often there's a low-value resistor, and sometimes also a capacitor, between the gate and 0V as well. The gate normally sits around 0V and a pulse (usually positive) into the gate fires the triac.

In the case of the simple lamp dimmer, the pulse is generated by the breakdown of the diac, which occurs when the capacitor connected to the top of the diac reaches the diac's breakdown voltage.

It's not that I want to float it, but is more of a behavior questions. I have never worked with a Triac before, so I am trying to determine it's quarks before I build a circuit with it.
If I want to build this and have temperature control, I will need to build an additional portion of the circuit to monitor and amplify the output from the temperature probe, as I am understanding that it's range is not great enough to significantly alter the firing rate of the Diac to be used in a temperature control.