Hi, I have a project where I need to control the temperature in a box, by a computer. I have found a control board with 8 bit analog input and output. I will measure the temperature using a thermocouple or alike, and control the temperature with heat resistors. The problem I have is that the analog output from the board, 0 to 5V, output resistance 1K5, can only supply like 50 mA.
I need more power than that, say 20 W. So I need a circuit that can be powered and that takes the 0-5V input and have a proportional output, allowing more power. Is this something simple?
Hi Magnatro,
There are two transistors in your circuit, each one powered with 5V?
Where is the output? Sorry, but my knowledge on these things are pretty limited.. Can you perhaps suggest a transistor that can deliver like 20Watts?
Hi Magnatro,
There are two transistors in your circuit, each one powered with 5V?
Where is the output? Sorry, but my knowledge on these things are pretty limited.. Can you perhaps suggest a transistor that can deliver like 20Watts?
OK, seeing the 20 watts has me curious. The 20 watts is your heater load? That is what I am guessing anyway. So maybe you have a 5 volt signal and when on you want the heater on? If that is the case 20 watts at what voltage in other words what is the load? Maybe if you explain in detail what you are trying to do we can give you some good workable answers.
OK, seeing the 20 watts has me curious. The 20 watts is your heater load? That is what I am guessing anyway. So maybe you have a 5 volt signal and when on you want the heater on?
Partly right, 5Volt is max voltage from my analog output. The analog output will vary from 0-5V, heating with more or less power.
If that is the case 20 watts at what voltage in other words what is the load? Maybe if you explain in detail what you are trying to do we can give you some good workable answers.
Ron
Totte; If you have a higher voltage avalibe for the heaters you can get the same power with less current. In magnatro's circuit the 1st transistor is like a predriver and should be powered from your logic supply. The 2nd transistor is the power driver and should be powered off a high currant supply and can be of a higher voltage to lower your currant. Andy
Hi Andy,
That makes sense. I will be looking around for transistors. I am also trying out LTSpice, which someone suggested for me to simulate the circuit.
Thanks!
Totte
Remember that as you change the heater current, the power transistor dissipation will vary, with a maximum of 5W (when the transistor resistance is equal to the load resistance) so it must be mounted on an appropriate heat sink.
A couple of other things to keep in mind: If you are using a power resistor/heatsink combo that is rated at 20W, don't run it at 20W - run it at 18W, 19W max instead - you never want to run a component at its maximum levels (especially in something like this, where there could be a risk of fire). Another thing to keep in mind is that it will take a fair amount of time for that resistor to heat up and cool down. Controlling it via PWM likely won't be too effective. Instead, mount it to a heatsink with a fan, turn it on full (with the fan running), monitoring the temperature with a tuned PID loop, and when you reach your cutoff point, turn off the resistor, but leave the fan running until you get to your set point temperature, then turn off the fan. As long as your insulation is adequate, the cycle time from on to off should be large enough so that the fan/resistor heater combo is only working occasionally, and not constantly. Over time, the temperature will average out.
Hi Crosh,
I will use a PID loop for the controlling part. I will not be able to use a fan however. I have seen heating resistors that comes as sheets, and that is what I will put on the walls in the enclosure.
Any ideas on what transistors to use?
Totte
Here is a hack. If you put both Q1 and R1 inside the box, the power in the box is proportional to the A/D voltage; 5V in produces 20W.
R1 can be broken up into many resistors, such that the effective resistance is 24Ω, and the power rating is >25W. The 2n3055 (Q1) would have to be on a heatsink, AND inside the box.
The purple trace in the plot is the total power dissipation in the box vs A/D voltage (x-axis of the plot).
That is why I suggested putting it inside the box to be heated.
Actually, here is a plot of power dissipated by the transistor and the resistor(s) on the same scale vs the input control voltage. Note that the transistor dissipation (Green trace) peaks at about 6W, which is not a huge heatsink. The total dissipation inside the box (Blue trace) is linear, while the individual dissipations are not.
If you built the experiment into a aluminum project box that was placed inside a styrofoam beer cooler, the transistor can just be bolted to the aluminum box. The heater resistor(s) could be the kind with the screw-down flange, like these.
I owned a HP Frequency Counter that had a heated crystal oscillator box that used this trick.
Thanks for your circuit, Mike
I have some questions, in your diagram, the UI represents the output of my analog device, right?
I don't understand what R4 and R5 is doing?
And V2 is powering the Q1 and Q2. Is that 24 Volts? I don't have 24 volt... how to deal with that?
Apart from that, thanks for the suggested transistor models!