I'm trying to repair a broken fog machine (some parts of the control circuit are broken) by Arduino-fying it. My plan is to replace the controller circuit with an Arduino and connect the Arduino to a DMX-shield for control.
The fog machine a couple essential parts:
Two pumps (to pump the fluid into the chamber)
A resistor (for warming up the fog chamber)
A thermistor (for keeping the temperature of the chamber optimal)
A thermostat, that will force the resistors off when the chamber gets too hot (to prevent burning of the fluid)
The part I'm having problems with is the thermistor. Today I tried to read some values off of it with an Arduino Uno, but the values would barely change even with a difference from room temperature to operating temperature (upwards of 200 °C). I also tried it in an oven, but it's resistance only changed only about 4-5 steps on Arduino's 10-bit analog scale (1011 in the oven). The resistor is (the original control circuit of the fog machine is 12v), while the Arduino operates at 5v.
I don't have much previous experience with thermistors in general. Am I perhaps doing something wrong or is the thermistor bad/broken? Are there better options for sensing the temperature of the chamber (better thermistors or other forms of temperature sensors)? I have attached an image of the thermistor assembly.
Are you sure that thing is a thermistor? Looks more like a thermostatic switch. Have you checked it's resistance with an ohm meter? What is the resistance when cold? when hot?
I have thermistors that change quite a few percentage resistance when just held between my fingers.
No, I'm not certain it's a thermistor. Just something I thought it was.
The "thermistor's" resistance is almost zero in all temperatures (only some very slight changes. As I already mentioned, something around 10 units in Arduino's 10bit analog reading at most).
EDIT: Another thing that comes to my mind is that the auxiliary resistor (connected to ground) might have to be a less resistant one.
That's a thermocouple most likely. A K type. If it is the voltage increases as temp increases. Problem being that the voltage is in the mV range so you will need an amplifier.
cbrunnem:
That's a thermocouple most likely. A K type. If it is the voltage increases as temp increases. Problem being that the voltage is in the mV range so you will need an amplifier.
I have used that one before and it works fine. Yeah a thermistor should have a large resistance between the leads. The ones I have used are 10k ohms. I can measure my thermocouple when I get home but you should be able to Google a k type resistance
The color coding at the end of the wire would indicate that it is a T-type thermocouple. (Attached image) What kind of an amplifier should I use with this?
Sorry, I didn't realize to include this kind of a picture in the first place.
With thermocouples, it's not a simple matter of amplifying the signal. The thermocouple is created by a junction between two different metals. ANY two metals have a thermocouple effect. If you add solder to a copper wire, you have a thermocouple. If you then attach a steel wire to that, then you have another thermocouple. Normally this is not a problem because the thermocouples all over your PCB are generating microvolts when most of your circuit is working with whole volts. It is also not a problem because the voltage created at the copper-solder-steel junction at one end of a component such as a resistor is perfectly cancelled out by the steel-solder-copper junction on the other side of the resistor, but only when both ends of the resistor are at the same temperature.
But at the point where you join the thermocouple into the normal copper wires of the PCB, you are creating a thermocouple junction between the special thermocouple wire and your PCB copper tracks. This is sometimes called the "cold junction" because it's at room temperature while the other end of the thermocouple is in whatever hot furnace you are measuring. This does not cancel out like the resistor example, due to the different thermocouple wires.
The best solution I have found is the MAX31855 series of cold junction compensation chips. There are different ones for each different type of thermocouple. This chip should be placed on the PCB as close as possible to the point where the thermocouple joins the PCB tracks, so that it can measure the temperature of that cold junction.
The proper thermocouple plugs and sockets are also specific to each type because they are made from the same metal as the thermocouple. While you can solder the thermocouple directly to a PCB or plug it into a breadboard, the correct plugs are much nicer. Newark is the only place I have found that will sell individual plugs and sockets. There is a version of the socket that will fit a PCB or breadboard.
After all that, perhaps you might be better off with a thermistor or a LM335 sensor.
MorganS:
With thermocouples, it's not a simple matter of amplifying the signal. The thermocouple is created by a junction between two different metals. ANY two metals have a thermocouple effect. If you add solder to a copper wire, you have a thermocouple...
Thanks for that reply! Opened up a few things for me.
The main point here, is that I have to differentiate too-cold and too-hot somehow. Accuracy is not an important figure and neither is absoulute temperature readings. That being said, does any amplifier do the job reasonably well with a similar thermocouple type as rated or do I still have to purchase a specific kind of cold junction compensation chip just to get some kind of relative, not-too-accurate readings out of it?
The thermocouple only measures temperature difference, not temperature. It consists of a probe
with the hot junction and the terminals that you join to the electronics being the cold junction. The
wires from the probe are difference metals all the way, that's essential.
