Temperature Controller-Variable

[Boki]

Hello to you all. Well, this is my first post and I already need some help/suggestions.

I'm trying to design a temperature controller, which will have to control
a 20-35W , 110VAC ceramic heater.
Ceramic heater is attached to aluminum element, and has to keep its temperature at 44.50°C
when measured inside the element. Special RTD temp.sensor is used for accurate readings and it's
connected to Rabbit 3000 microcontroller.
If the temp varies for 0.1°C, micro will detect that and will send signal to driver which controls ceramic heater.
I'm trying to come up with interface which will listen for the commands from the Rabbit microcontroller and
adjust amount of current flowing through the heater.
Ultimately, to increase or decrease current to achieve desired temperature, but never to shut it off completely.

Any kind of suggestions or help is more than welcome.

I have already searched forum for similar topics, but could not find anything that I could use.

 

[stevez]

What does the microcontroller do now - turn heater on/off, send PWM output to heater, etc? What do you want it to do that it doesn't do now? What do you expect that this will accomplish?

I've helped people where I used to work with controller problems that might be like this. The controller turns the element on or off and the resulting temperature swings are tolerable for most applications but other times less swing is preferred. Other times the controller is proportional only and there has to be an error to get any output.

 

[Boki]

What does the microcontroller do now - turn heater on/off, send PWM output to heater, etc? What do you want it to do that it doesn't do now? What do you expect that this will accomplish?

I've helped people where I used to work with controller problems that might be like this. The controller turns the element on or off and the resulting temperature swings are tolerable for most applications but other times less swing is preferred. Other times the controller is proportional only and there has to be an error to get any output.

Once the micro is initially powered, it has to turn on heater, and try to maintain required temperature.
I need a controller which is not going to turn the element off, but just decrease or increase the current, I will have less swing in temperature.
What I don't know is how to design circuit which will control AC heater and also talk to micro.
From what I know so far, I'll have to use a TRIAC or SCR to control the ac mains power to the heater.
And also I should isolate the microcontroller from the triac using an optoisolator.
I need to see similar schematic or something that I could modified if necessary.
cheers stevez

 

[stevez]

Boki- my sense is that you have a complete system right now that will control temperature - you have the sensor, a controller with working software and some means of turning the heating element on/off.

I took a fast look at a Mouser catalog page on the controller. I am a beginner when it comes to that kind of stuff but would think that your controller could directly (or nearly so) a triac circuit in a way that is sufficient for your needs. The software or programming might be the answer - might be.

If you turn a heating element on or off with a switch there will be some temperature swing - depending on how much heat, thermal mass, time, etc. If you shorten the on/off cycles (increase the frequency) you can get to a point where the resulting swing in temp is negligible. Another approach would be to reduce or increase the "on" time - I think this is Pulse Width Modulation - to achieve the same thing. I may be mistaken but this is within the capability of many microcontrollers. If it can be done with your Rabbit then all you need is the means to handle the high power end.

Varying the power to the heating element on a continuous basis can be done as well - it would seem that this is also within the capability of a the Rabbit. You would use one of the analog outputs to generate a 1 to 5 volt signal and drive the appropriate power electronics - presumably transistors rated appropriately.

The software might be the biggest problem. You'll need to generate an output based on the difference between the setpoint and the actual temperature. Before dealing with the software you have to decide on what kind of control you want - proportional, proportional/integral, etc. It would not surprise me if folks who've worked with basic stamps or other microcontrollers have already worked this out.

I hope I am of some help. Send me a message directly if I can clear anything up.

 

[Boki]

Stevez, here is the cct. schematic.
It's not working properly.