Let's Make Robots!

Help Me Hack It:The Non-contact Thermometer (or, WWG/ND*?)

I'd posted a request for reports of any experiences with one of these:

I managed to pick one up on eBay for a hair under U$10.  After the week waiting for it to arrive and then the week building the HAL-met, I'm back to this.  Which is a distraction from K-9, but who says you have to do everything in a serial fashion, right?  The thing wasn't terribly accurate-it said Sarah and I were only about 85° F and the dogs were in the 70s.  Still, it was accurate enough to tell a human from a wall, so the project (which if you missed it is to use temperature as a means of detecting human beings) So anyway, I opened it up:

I opened that bitch up.

YAY!  Just two wires going to the PCB, one is black and one is pink.  Gotta be an analog system, right?  So the question is: is it a (semi)conductor or  a resistor?  One way to find out.  I soldered Dupont plugs onto the leads and hooked it straight into A0 with the AnalogReadSerial sketch directly out of the Examples folder.  The result was something like 215, no matter where I pointed it.  Not a conductor.  So next step, if it's a resistor, what's the optimal voltage divider?  Easy way to find out:

I didn't have any 84.5 KΩ resistors laying around, but as you can see there was an 82 KΩ-close enough for a voltage divider.  I plugged that in, and this time I got all 600s, no matter where I pointed it.  I reversed the positions (putting the fixed resistor on voltage and the sensor between ground and A0) and got 20s.  No variation.  This looks less and less like an analog sensor.  I figured maybe I'd fried it, so I soldered headers onto the original PCB and plugged it in.  It still works.

So the next question is, what the hell kind of sensor is it?  In the original post, ChuckCrunch pointed me to this:


...which looks an aweful lot like this:

(trust me, it does.  There's a reflective surface that's throwing off my camera's ability to focus on the sensor mounted behind it.)  Sparkfun's website claims that it works via i2c, but I don't think that's a go with just two wires when one is pretty obviously (likely?) ground.  Here it is from behind:

Again, the reduction from 3 to 2 dimensions is making things difficult here-there are actually four pins to play with, but only two are used apparently.  Googling the TP45 and 070510 as well as a few other terms turned up nothing unless you're interested in the average Temperatures in Texas in early May.  

So anyway, I'm open to suggestions to what to do for the next step.  Did I miss an obvious analog possibility?  Is there a digital protocol that might work over only two wires, with one of them being ground?  Should I try the pin-out and i2c protocol from Chuck's find?  (The point here is to hack it into something robotically useful.)


(*"What would Gareth or Nils Do?")


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

i2c and 10-bit PWM output provides a resolution of 0.14°C,

if it's the same as i pointed out , if you had a scope you could see if it's a PWM signal

if you have a set of computer speakers you may be able to hear the signal if you connect the output of the sensor to the input of the speakers "might" depending on frequency

pulseIn()  might be better

Those are both great ideas. I actually just invested in a combo oscilloscope/logic analyzer (I didn't think my bid had any chance but then I won.) Thanks!

tell us what you find out  :D.

Can you check the wires going FROM the pcb when "using" it instead and that way see what it is sending/recieving to/from the sensor? maybe do a logic analyze of whatever being sent. Good luck

One Wire Protocols. Obviously not i2c, if one wire is indeed ground. Of course, the reflector may be a ground, so, all you need to do is figure out what the 2 wires actually carry with respect to data. :)