wannaburn said:
What are you using for the sensor? is that also a TEC similar to a LaserBee I?
I wanted to takle making one for the fun of it some time, TEC seems to be the way to go for DIY LPM's
Depends on which one you mean.
My primary personal meter uses a surplus Coherent thermopile, which turned out to be VERY accurate (i was lucky with the board i got!), only showing 1.96% too much, so i set the gain to 980.4 and now it's as accurate as it gets.
Brtaman's meter uses a Peltier "in reverse", thermal epoxied to a heatsink and coated, just like the LaserBee, yes, but possibly with a different coating (i don't know the coatting used on the LaserBee).. I call it a "Peltipile"..
(
- it was sold to me as a "thermopile", and when i finally saw what it really is, i started mocking the seller a little (since he did trick me), by calling it a peltiplie, and later it just kinda stuck)
On one hand, both a thermopile and a Peltier work on the same basic principle (the Peltier/Siebeck effect), but in the opposite way from each other, each optimised for a specific purpose. A Peltier for PUMPING heat, a thermopile for DETECTING it...
So while you can use a Peltier for heat detection, it will be a little slower in response, and less sensitive. But that also makes the reading more stable, by averaging it out a bit. Luckily a Peltier in reverse works better than a thermopile in reverse would (for pumping heat).. ;D
The secret of a good "Peltipile", is of course in the coating. I use a special coating, which only has +/-0.5% error over the visible spectrum! If it is set in the middle (to be spot on at 532nm), it will show 0.5% too much at 405nm (405nm absorbs a bit better) and 0.5% too little at 660nm. In practical terms, that's an error hardly worth mentioning. (Quite a few DIY meters are up to 7% off, some even more. And even that usually doesn't bother anyone too much.)
So if the meter is perfectly callibrated, the accuracy of a Peltipile meter could be as good as +/- 1.5%! But you would need a perfect calibration reference for that and a perfect calibration laser (with APC and constant temperature TEC thermal regulation for CONSTANT power). The calibration itself takes a lot of time without the special laser. And the accuracy will be as good as the accuracy of the reference, at best - if you do everything just right, or it could be slightly off, if you don't...
But if you have both a good coating, and a good calibration reference and special calibration lasers, the Peltier approach could yield surprisingly good results! Officially however, i won't claim more than +/-3% accuracy. But then again, that is the accuracy of many professional meters.
The only drawback is that PeltiPiles take longer to "climb" to the laser's full power on the display. But at the same time, they fluctuate MUCH less for the same reason. That actually makes them more stable (the slowness of the response averages the result out a little).... If there was a way to create a coating as good as the ones on real thermopiles, they would be just as accurate, only slower...
Anyway, the sensors my meters use...
I made two Coherent Thermopile meters so far, a third is in the making:
- One of them i made for a guy from Taiwan.
- The other is my main meter of course (that was my first meter).
- The handheld will also use a Coherent thermopile, it's my last one, it's also VERY accurate, currently at -1%! I can set the gain to 1010 and it will be "spot on". I am not selling that, i want to keep my last thermopile!
"Peltipile" meters i made:
- Brtaman's meter uses a Peltier on a heatsink with the special coating.
- My secondary meter also uses the same thing - Peltier in reverse on a heatsink....
- Jake's meter will use a Peltipile too...
You want to make your own? Do you know (want to know?) how to make a proper amplifier circuit?
A high precision PeltiPile amplifier is actually quite simple and easy to make, if you're used to "wire-wrapping" circuits together, or if you can etch it.
The circuits required for real thermopiles are much more complicated, if they are to be just as accurate. For example, my main meter now uses a high precision triple op-amp circuit, to eliminate certain errors present in the single op-amp design. Before that it had a single op-amp amplifier, which is now in Brtaman's meter... But since Brtaman's meter uses a "Peltipile", the single op-amp circuit classifies as "high precision" in that application.
This comes from the differences between how a thermopile and a PeltiPile work. But i won't go into details here..
In any case, i have the amplifier circuit schematics, optimised for "Peltipiles", if you want to make your own.... Other than that, you just need a tiny peltier and a good way to coat it.