I want to grab data from cheap infrared thermometers and was wondering if there are existing projects that tap into a devices LCD output in order to read it's data into an arduino.
For the price these things sell for on ebay it's hard to come up with an easier alternative that will not end up at least 15 times more expensive.
This is quite tricky to do because the multiplexing waveform of an LCD is quite complex. It normally involves multiple voltage levels which requires a crude A/D on the front end.
There is some text on how to do this here that I wrote back in 1987:-
http://www.doc.mmu.ac.uk/STAFF/A.Wiseman/Acorn/BodyBuild/BB87.html
I was actually thinking more along the lines of tapping into the parallel or serial LCD input bus to the controller, but now I realize it might not actually have a separate controller.
Maybe I'll just figure out how to interpret the sensor itself.
Yes these things are often one big chip, to keep the price down. There is no separate units to tap signals from.
Maybe I'll just figure out how to interpret the sensor itself.
Something to keep in mind: Most of the time, these temperature sensors have a couple of "limitations" (one may not be a limitation, depending on your need):
They have a limited range - generally "body temperature" range (70 degrees F - 120 degrees F) or "cooking temperature" range (0 degrees F - 200 degrees F); some will fall outside that range (sometimes you'll find body temperature limited by the controller and not the sensor, other times on cooking probes you'll find it can go way higher or lower, and some are designed for candy/deep frying monitoring, which have much higher temperatures) - so keep this in mind.
They typically will have non-linear response curves.
Number 2 is the "killer" here; with a non-linear response curve, trying to figure out the exact temperature based on its resistance would be nearly impossible. You can get crude readings if you are willing to take the long time to figure out a rough response curve and interpolate, this can be done by:
Freeze the unknown probe and a known good and accurate probe in water. Measure the resistance with a DMM, note it - this is "freezing" (0 degrees C). Put the ice in a pot on the stove, and heat the ice, monitoring both probes, and note the change in temperature with the change in resistance. When the water is boiling, note the final temperature and resistance (100 degrees C) - this is "boiling".
Unfortunately, you can't go beyond these two extremes easily in a homebrew manner, but if you think your probe can go higher, you could chill some oil down in the fridge, note the reading, then heat that up on the stove (know what your smoke point is for the oil, and have a fire extinguisher rated for grease fires handy) and continue to take readings; with the right oil and known probe (candy thermometer works best, use avocado oil for highest smoke point), you can get up around 500 degrees F in temperature. Beyond that, you are starting to get into Type K thermocouple territory anyhow.
Even so - with all that work and time (watching water boil), you still will only get a crude approximation of the curve (which may be OK for your application); thermistors that you buy will generally have a datasheet available for the specs, and most importantly, the response curve (sometimes along with the algorithm of the curve to fairly accurately determine the temperature from the resistance). Unfortunately, trying to find out what the part number is for a thermistor from a disassembled temperature probe from the store is likely an impossibility.
Thanks for your extensive reply, but I think you are referring to a thermocouple or something along those lines (PTR/NTR kind of resistor).
I have infrared non-contact thermometers in mind. They sell for a dollar on ebay (not including the $20 shipping), so this is an interesting source of sensors. Especially since my application environment has the potential to be destructive.
I want to mount an array of 3 or 4 sensors in the wheel well of a race car. Tyre temperatures across different parts of a tires surface are a great source of information for the correct setup of alignment and tire pressure.
I'll just have to wait for the first batch to come in so I can perform some exploratory surgery on one. ;D