My colleague David and I need to soak our scintillation detectors at known constant temperatures if we are ever to graduate. Our thesis advisers need their lagers at 6 to 9 degrees Celsius, ales at 7 to 11 degrees Celsius and the occasional stout at 13 degrees Celsius.
This is what I came up with. Enjoy!

Step 1: The Task

The lab already had a picnic cooler, supposedly for climate-testing of the radiation detectors (?!). The cooler came equipped with a Peltier pump that can run off AC mains or 12V DC and can either heat or cool (depending on the switch setting) the box interior. Unfortunately, there is no thermostatic control built in, so the internal temperature either approaches approximately the ambient temperature less 20 degree Celsius or heats up to +50 Celsius and stays there (probably limited electronically).

This simply screams for some micro-controller driven decision making.

Aside from a picnic cooler, the final bill od materials consisted of:

  • an arduino board
    (we used a pro mini 328/5V/16M, simply because they are the cheapest Arduino clone on eBay @ $2.83),
  • 2 relays (5V/1ch relay modules from eBay at $0.99 each),
  • 2 momentary push-buttons (stolen from colleague Klemen)
    [Klemen comments: "Borrowed not stolen. One day I will have them back."],
  • an electronics thermometer (we used a DS18B20, eBay @ $1.04),
  • 5 resistors (see schematics, later on),
  • an LCD 2x16 display module (HD44780 compatible module, eBay @ $2.26),
  • a 7805 voltage regulator (eBay @ $0.90) and its two decoupling capacitors,
  • an electrical enclosure box from the hardware store ($6.00 for the cheapest one!),
  • many, many, many... wires (depths of supervisor's drawer).

As far as the thermistor, that's reasonable to go for a calibrated module versus calculating against a precalibrated source and software converting values. As far as the FETs go, I believe you would use an H-Bridge configuration similar to how you drive a motor backwards and forwards.
Aha, <br>I think you're right. We could go for something like a L298 based dual H-bridge and use both channels in parallel. Thanks for the great insight!
I like the project, but I had a few suggestions. Instead of using relays, you could use logic level n channel FETs. This would eliminate the need for an additional 7805 and caps. Instead of using a thermometer circuit, you could try building a simple one from a thermistor and a voltage divider. I don't think the former suggestion is cost efficient, but the latter most likely is.
Hi, thanks! <br> <br>Regarding remark 1, we cannot figure out how to wire the FETs so that we could <br> * run the peltier in both directions <br> * never possibly short the power supply with faulty software. <br>Like you, I'm also not sure about pricing of low impedance FETs with 5 amp rating and robust to inductive kick of both fans? The relays cost us $1.98 and give off very satisfying audible clicks. :-). <br> <br>Regarding remark $2: <br> * the 18B20 thermometer cost $1.04 S&amp;H incl in quantities of 1. We buy bags of 10, so they run only $.50 or so. They are very accurate and come calibrated. With an NTC resistive divider, we'd have to calibrate the system ourselves. We were simply too lazy to go for that. ;)

About This Instructable




More by mvencelj:Baobab Germination - the Angle Grinder Method Hack a picnic cooler (into a thermostatic chamber) 
Add instructable to: