Which thermopile (TEC) is better for power meter?

[Mikos]

I want to improve my idea on laser power meter. My first idea has been thermistor in combination with power resistor (see my previous topic). It works, but thermistor has pretty slow reaction time (it takes more than minute to stabilise) and also is too small (I didn't find any big thermistors).

I would like to use thermopile (TEC, also known as Peltier coolers) instead of thermistor. I have found that professional calorimeter-based laser power meters also use thermopiles as sensors. TEC should have much better reaction time than thermistor and dependence between heat input power and output current should be linear. TEC should have also much better accuracy. I can use power resistor (with potentiometer for power regulation and multimeter for measurements) as heat source to determine multiplication factor and calibrate it.

There is great guide for DIY thermopile-based laser power sensor here on Photonlexicon forums:
Part 2 - Laser Power Meters

I know I would still need another calibrated laser power meter (or laser sources with known output power) to determine correction factor (because of imperfect absorption of laser light), but with proper coating (activated charcoal powder and/or lamp black) error wouldn't be that bad even without it, calibration by power resistor should be enough (at least for my needs).

What I need to know now is which TEC is best for my needs. I have access to these TECs (you can also look at datasheet):

TEC1-01708: Size 15x15x3.3mm (WxDxH), Imax 8,5A, Umax 2,06V, Qmax 9,2W, couples = 17
TEC1-07103: Size 30x30x4.9mm (WxDxH), Imax 3.3A, Umax 8.1V, Qmax 16,4W, R = 1.80 ohm, couples = 71
TES1-12704: Size 30x30x3.2mm (WxDxH), Imax 3,9A, Umax 15.4V, Qmax 33,4W, couples =  127
TEC1-07108: Size 30x30x3.8mm (WxDxH), Imax 8.5A, Umax 8.6V, Qmax 38,5W, R = 0.85 ohm, couples = 71
TEC1-03114: Size 30x30x4.6mm (WxDxH), Imax 14A, Umax 3.75V, Qmax 29,2W, couples = 31

I assume that smaller area is better (because of thermal isolation), am I right? But also number of thermocouples per area is really important (for better accuracy). So I am deciding between TEC1-01708 (15x15 mm, but only 17 thermocouples) and TES1-12704 (30x30 mm and 127 thermocouples). Also I don't know if lower voltage (but higher current) models or higher voltage (but lower current) models are better. I will be measuring current, so I assume that higher current models are better, or is it irrelevant?

 

[Bionic-Badger]

Re: Which thermopile (TEC) is better for power met

From what I've seen of the LaserBee module, their TEC is a tiny, maybe 1cm x 1cm TEC (maybe even half that size). I don't know if this was to keep costs down, or that it helped with the thermo characteristics of the device, but the ones listed above are pretty big, especially the 3x3 cm^2 ones. They also appear to be teflon coated, which may hinder conductivity. For example, here is the large one with 127 couples on it.

On another note, does anyone know how much heating is produced, or rather the temperature changes (in K or C), by a 5mW difference on a well-absorbing black surface? I was wondering what kind of sensitivity these devices need to have.

 

[Mikos]

Re: Which thermopile (TEC) is better for power met

On another note, does anyone know how much heating is produced, or rather the temperature changes (in K or C), by a 5mW difference on a well-absorbing black surface?  I was wondering what kind of sensitivity these devices need to have.

It depends on specific heat capacity and mass of the target. These TECs are extremely sensitive. It even register heat from your body if you put palm of your hand near them.

 

[Bionic-Badger]

Re: Which thermopile (TEC) is better for power met

Well I just bought some cheap-ass small peltiers off ebay, so I'll see what kind of response these things have to lasers. If it sucks I can use it to cool down a laser or something.

 

[Mikos]

Re: Which thermopile (TEC) is better for power met

Well I just bought some cheap-ass small peltiers off ebay, so I'll see what kind of response these things have to lasers.  If it sucks I can use it to cool down a laser or something.

Great! ;-) I am going to buy TEC and electronic parts for thermopile amplifier next week.

Just remember that you must attach other side of peltier to some big heatsink (to get accurate readings). Also you must paint it black with active carbon powder or lamp black (you want as great absorption in visible spectrum as possible, it must be definitely matte, not glossy).

Response will be IMHO in range of microvolts, so standard multimeter isn't enough for measurement, you must build operation amplifier (here is good one). Or you can try to measure current (in range of hundreds to thousands microamps), but I am afraid it will be less accurate.

 

[MarioMaster]

Re: Which thermopile (TEC) is better for power met

[quote author=Bionic-Badger link=1218028835/0#3 date=1218193134]Well I just bought some cheap-ass small peltiers off ebay, so I'll see what kind of response these things have to lasers. If it sucks I can use it to cool down a laser or something.

Great! ;-) I am going to buy TEC and electronic parts for thermopile amplifier next week.

Just remember that you must attach other side of peltier to some big heatsink (to get accurate readings). Also you must paint it black with active carbon powder or lamp black (you want as great absorption in visible spectrum as possible, it must be definitely matte, not glossy).

Response will be IMHO in range of microvolts, so standard multimeter isn't enough for measurement, you must build operation amplifier (here is good one). Or you can try to measure current (in range of hundreds to thousands microamps), but I am afraid it will be less accurate.[/quote]

you may also want to get a small piece of aluminum or copper to mount to the face of the TEC to ensure the heat is spread evenly across all the junctions - if you only heat the central junctions for example your reading may be inaccurate as the non heated junctions will reduce the output

 

[Bionic-Badger]

Re: Which thermopile (TEC) is better for power met

Yup, I have nearly all the parts I need to experiment with. Heatsinks from computers, flat black paint, ADCs, op-amps, and all that good stuff. I've also got some other stuff on the way to see how they perform. If there are some usable results, I may be forced to buy a reference meter to calibrate with. Won't that *cough* be a sad day...

 

[Mikos]

Re: Which thermopile (TEC) is better for power met

you may also want to get a small piece of aluminum or copper to  mount to the face of the TEC to ensure the heat is spread evenly across all the junctions - if you only heat the central junctions for example your reading may be inaccurate as the non heated junctions will reduce the output

This would IMHO get you worse results than without it (especially worse reaction time). Thermopiles should have same output independent from which thermocouple in thermopile is heated (if you aim laser on center or on the edge, output should be same). Absorber should be as thin as possible for measuring power (and there is already thin ceramic plate on TEC). Volume absorbers are used only for measurement of laser pulse energy, not average power of CW laser.

Black paint with high concentration of activated carbon powder or lamp black should be IMHO enough...

Btw. there is great description of laser calorimeters from Scientech:
Measuring Laser Output

 

[Mikos]

Re: Which thermopile (TEC) is better for power met

..., flat black paint, ...

It should be carbon-based paint (with high concentration of carbon particles). Maybe it would be better if you make it yourself from activated charcoal powder. There are great instructions:
http://www.photonlexicon.com/forums/showpost.php?p=20935&postcount=12

 

[Bionic-Badger]

Re: Which thermopile (TEC) is better for power met

Hmm, I'll have to check that out. Actually, since this is mostly for experimentation, and I'm receiving three of the same peltier type, I can try out the various paints, etc. to see how they perform. This paint has a high emissivity at around 0.97 though it was measured in the infrared region. I'll have to see how it performs as a broadband absorber.

For that carbon deposition I should try picking up a chemical vibrator thing at my university if they still have them to deposit it evenly. Never thought about ever needing one before though. Sounds like a fun project regardless.

 

[MarioMaster]

Re: Which thermopile (TEC) is better for power met

[quote author=MarioMaster link=1218028835/0#5 date=1218223252]you may also want to get a small piece of aluminum or copper to mount to the face of the TEC to ensure the heat is spread evenly across all the junctions - if you only heat the central junctions for example your reading may be inaccurate as the non heated junctions will reduce the output

This would IMHO get you worse results than without it (especially worse reaction time). Thermopiles should have same output independent from which thermocouple in thermopile is heated (if you aim laser on center or on the edge, output should be same). Absorber should be as thin as possible for measuring power (and there is already thin ceramic plate on TEC). Volume absorbers are used only for measurement of laser pulse energy, not average power of CW laser.

Black paint with high concentration of activated carbon powder or lamp black should be IMHO enough...

Btw. there is great description of laser calorimeters from Scientech:
Measuring Laser Output[/quote]

also keep in mind that TECs are NOT thermopiles, ceramic is not great at conducting heat - it it just used because it doesn't conduct electricity otherwise the whole junction would short, the piece of metal i was suggesting was to ensure all the peltier junctions in the TEC are evenly heated so the measurement stays accurate

 

[Kenom]

Re: Which thermopile (TEC) is better for power met

[quote author=Mikos link=1218028835/0#7 date=1218275181][quote author=MarioMaster link=1218028835/0#5 date=1218223252]you may also want to get a small piece of aluminum or copper to  mount to the face of the TEC to ensure the heat is spread evenly across all the junctions - if you only heat the central junctions for example your reading may be inaccurate as the non heated junctions will reduce the output

This would IMHO get you worse results than without it (especially worse reaction time). Thermopiles should have same output independent from which thermocouple in thermopile is heated (if you aim laser on center or on the edge, output should be same). Absorber should be as thin as possible for measuring power (and there is already thin ceramic plate on TEC). Volume absorbers are used only for measurement of laser pulse energy, not average power of CW laser.

Black paint with high concentration of activated carbon powder or lamp black should be IMHO enough...

Btw. there is great description of laser calorimeters from Scientech:
Measuring Laser Output[/quote]

also keep in mind that TECs are NOT thermopiles, ceramic is not great at conducting heat - it it just used because it doesn't conduct electricity otherwise the whole junction would short, the piece of metal i was suggesting was to ensure all the peltier junctions in the TEC are evenly heated so the measurement stays accurate[/quote]

This is exactly the reason for using a SMALL peltier.