Digitally control resistive heater

Go To Last Post
5 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

So I've recently been asked to see if I could come up with a simple solution to making a crude microprocessor controlled voltage supply.

The problem:
I have a resistive heater that I need to control with a micro-controller. The heater has a resistance of 5-20 Ohms depending on how warm it is, and I need to supply it with less then an amp (typically ~0.5A, but I'd like to leave myself some wiggle room). I figure I need to supply it with 3-20V DC, and would like to work off a 24V DC supply. I also need to sense the voltage and current being supplied to the heater, on the order of once a second.

My solution:
I figure that the voltage and current sensing are easily enough accomplished with an ADC and a voltage divider (for voltage) and a low-side shunt resistor (for current).

My concern is finding a simple design for a power supply that I can control with a micro-controller. The idea I've had is to use a high current adjustable linear regulator (like a LM338) where I use an op-amp in place of the rheostat. I figure if I configure the op-amp to be a 6x non-inverting amplifier I can then use the micro-controller to run a 0-5V DAC into the op-amp. I can then feed the output of the op-amp through a resistor into the LM338's "adj" pin. I only need to go down to 3V (0.5A at 6 Ohms) so I can tolerate a non-rail-to-rail op-amp and the voltage drop across the resistor won't be a problem.

So, before I go and order parts, can anyone see potential problems with this? A quick spice simulation suggests that it would work but I figured I'd throw it up for the experts.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Analog approach...

Nothing wrong with that, but a lot of work, and BIG heat sinks for the regulator or pass transistor.

An alternative approach is to use PWM, Pulse Width Modulation, of the power to the resistor. The V+ goes to the resistor which goes to a N FET which goes to ground.

You pulse the NFet On and Off. When it is On, it is in saturation, fully on, and its Drain-Source resistance is very low, perhaps milliohms, or just a few ohms, depending on the model selected. Very little power is "wasted" in heat in the transistor.

Zero power is provided by 0 pulse, i.e. not pulsing any power to the resistor.

Feed it a 50% duty cycle pulse train, and it will be at 1/2 of full power.

Just turn the transistor on, no pulse, gives 100% power.

The width of the pulses gives you the full scale of power settings.

Just an option for you to think about.

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you want the heater to be a certain temp, you could put a temp sensor on the heater and read that with the micro. Then do the PWM approach. You could even put a PID control in the software. Works real well.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The heater could be run direct from 2v V AC using opto-coupled TRIAC (MOC3020) using inherent zero-crossing switching. No rectification, filtering, regulation etc. As jaydhall, control with PID control loop. Greater efficiency and lower cost & complexity.

Charles Darwin, Lord Kelvin & Murphy are always lurking about!
Lee -.-
Riddle me this...How did the serpent move around before the fall?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You are far better off with an independent temp sensor. Yes, you can tell the temperature by measuring the element resistance, but that value is "contaminated" by lead resistance. Its not easy to separate the two, especially since lead resistance also varies with temperature, but differently than the element.

I also suggest some sort of PWM, whether with TRIAC in an AC system or conventional PWM in a DC system. Your efficiency will generally be higher if you can do it directly in AC, but at that low voltage and relatively modest currents, DC may be your mode of choice.

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net