Poor Man's Laser Power Meter?

Hi guys, quick question...
I want a laser power meter but I'm on a shoestring hobby budget.
Can I use a peltier plate as a power meter?
I've seen Thermopile power meters on the market, and thermopile = thermoelectric effect exploiting sensors, right?

and / or with the right tinting, could two photodiodes of different technologies be used? (eg. silicon / germanium, each has a spectral response curve, correlating between them would give you power?)

Aside from making my own, does anyone have a good set of metrics for estimating laser power?

You can use a TEC plate (a small one, with a lot of junctions, is better, like the 10x10mm or 15x15mm ..... the usual 40x40mm commercial ones are not good for low powers, they will still measure the power correctly, but you can't expect to use them for anything less than 50 or more mW with precision) ..... then you need a big "thermal reference" mass (a heatsink is ok, more is massive, better it is), and a simple op-amp circuit, with a pair of op-amps in cascade, for amplify the very small signal from the TEC and convert it in something that can be measured from a DMM or a panel instrument.

But the more important thing is the coating of the TEC reading face ..... normal ceramic plates of the TECs does not accept too much a simple hand of paint, so you need to find a decent way for get the paint permanently sticked on it ..... i solved the problem with car-repair products (first layer of epoxy paint, then a layer of matte black "anchorage" paint (don't know the name in English), then the final black matte paint for high temperature.

You need to find, for the last layer, the more possible black AND "matte" paint, the one that reflect back the less possible light in absolute ..... there are a lot of "black matte" paints, but not all them are ok, and some still reflect a lot also after drying.

An alternative can be to ask someone that have a paint oven cabin (again, car repair centers have them), to try to paint the TEC plate on one side with the more black matte possible of their "dust" or "oven" paint, but i can't say if they have any decent black matte, or if the paint, after, stay on the ceramic without "peel away" with the time ..... i only know for done experiments, that simply paint the ceramic plate don't last too long, the paints "peel" and detack, after some time ..... it vary form paint to paint, but first or after, it happens ..... the only ones that are in place from years that i've tried, are the ones i've done using the "anchor" and epoxy layers.

Anyway, this is valid for the paints that i can find here in Italy, ofcourse ..... maybe in some other countries are available better products, that don't require all this work ..... you must try with the products that you can find in your area.

This thread:

DIY thermal lpm under 50$

may be interesting for you:yh:

The problem with power meters based on TECs is the slow response time and to get a good calibration.

Photodiodes can only be used for quite low laser powers (without attenuator). And calibration is even more demanding, since depending on the wavelength. Interesting idea to use two different diodes, but in my opinion it's too complicated and probably not very accurate.

A quite cheap power meter is the Radiant Alpha (but they are Out of Stock at the moment:yabbem: Radiant Alpha

@ DP: oh my gosh, I feel so unoriginal all of a sudden! Thanks very for the links, and the mini-analysis : ) I'll try starting there and if I can't make do with components in my junk box I might very well buy one of those radiant alphas.. they certainly hit an attractive price point : o

@HIMNL9: very cool, you made one! Thanks for the practical construction advice, can I ask how well your setup is working for you so far? And what kind of power levels / concentration are Too Much for the paint layer, even with the epoxy?

I wonder if I could use a thin plate of anodized black aluminum thermally connected to the TECplate and skip paint entirely?

Being sincere, i made some of them, not just one, in the past ..... i already have one for sale here, and i'm checking all the others, cause i've decided to sell them too (i have too much things around, and also i need to get some money for pay the legalized criminals here aka "tax office" )

Using anodized aluminium, if the anodization is the matte black one (very deep black, i mean, not a light anodization) is possible, but you need to get a very thin aluminium layer and thermally glue it on the reading face with the more flat and thin possible thermal glue ..... cause anything that increase the thermal mass of the reading face, also slow down the reading process, and decrease the minimum resolution capability ..... it's a pity that is not possible to anodize directly the ceramic, cause the real black anodization is better than any paint, in light absorption.

About the paint resistence, it ofcourse vary from paint to paint, and also, very thin layers resists more than thick layers (can appear as a nonsense, but this is what i've tested ..... maybe is simply cause a thin layer transfer the heat more quickly to the ceramic plate, where a thick layer, i'm not sure) ..... i'm using a flat matte black paint that is normally used for exhaust tubes in motorcycles, as last layer, that it's rated high temperature resistant, but only cause it's the only paint for high temperature that i can find here ..... i think different countries have different products, so i cannot give you any specific brand.

Anyway, for focused beams there's no resistance (a beam MUST NEVER BE MEASURED FOCUSED, anyway ..... focused beams can damage also commercial thermopiles) ..... defocused beams, indicatively a good painting work can resist to a pair of watts defocused over a 12/15 mm diameter spot, without report damages.

@HIMNL: Wow, reading through that post DP linked me to, I realize I was asking old questions, thanks for letting me know about the anodized aluminum... I take it you tried that method? I wonder how thin a plate they can anodize...
Nifty lamp setup / idea! did you end up getting it working as well as you hoped?
I think I'll go with your paint method for a first try, it sounds like you're getting good results...

-Having said that, is all I have to do:
hit any surface [with some mass and as much light absorbancy as possible] for a known period of time
-determine the temperature difference of the mass, post-heating vs. pre-?
(-calibrate)
in which case could I just substitute for temperature sensors??? That somehow sounds wrong, but I cn't figure out why right now


p.s.
taxes suck! at least you're not living in Sweden, right? : ) And for what you have to pay you get to live somewhere so beautiful!

More easy, if the "cold" (reference side" of the TEC plate) is stable enough (i mean, you use a decently big heatsink and left it stabilize at ambient temperature, and not touching it with your hands when you measure the laser), the output read from the TEC is already proportional to the power you shine on it, without have to do too much calculations ..... only, you need to keep the laser on the plate enough time for let all the heat spread to all the surface and the reading stabilize, and then that what you read is the value, and the calibration is just to regulate the circuit for made your instrument show that value.

In poor words, you need a laser with a power that you already know, and stable (a good red laser is ok, for this, if you know exactly the power) ..... shine it on the read side, then left it on enough for let the reading stabilize (usually, for 10x10mm or similar TECs and lasers around 100mW, 10 to 15 seconds are enough ..... when you see that the reading stop to increase, then the output of the TEC is proportional to 100mW, and you can regulate your instrument for read 100mW.

This is ofcourse very simplified, is just for show you that there is no need of calculation, cause all the work is done from the TEC itself.

About the anodized aluminium, yes, i've tried with a thin fin of a heatsink, and it work, but unfortunately it was around 1mm of thickness, and it added a lot of thermal mass to the reading side, slowing down all too much ..... i don't have any thinner anodized aluminium, and really don't know where to get it ..... probably the better solution can be using something like 0.1 or 0.2 mm thickness, but don't have any idea if this is possible to find around, sorry .....

Anyway, there are some users here that i think have already experimented with anodizations, so they must know better than me how much thin can be a surface for have the possibility to anodize it without ruin it .

That makes a lot of sense, many thanks! Calibrating the output is no problem, though I definitely appreciate the simplification : ) It's the sensor cost & read time that I'm worried about... a peltier seems like a good idea, but they're too slow, everybody says... I'm wondering if I can use a temperature sensor with a good, thin black metal plate... a pair, specifically, for ambient / hot-side difference... because the mass being heated for every test is the same, can I simplify this down to temperature difference per time unit = power?

Thanks again for the insights!

Incidentally, with a typical TEC setup, would a shroud over the laser attached over the tec that's painted as reflective a white on the inside as possible help with accuracy much? Or does a good coating (Cool raindrop method, btw!) absorb enough that it doesn't matter?

You're not trying to cool the cold side of the TEC, only keep it in thermal equilibrium with the area around the TEC. This is so that the light incident on the hot side of the TEC causes a temperature, and therefore a voltage, differential (which is measured). You can actually get negative readings if you cool the cold surface below the ambient temperature. Best to just have a large heatsink or something that will keep the cold side at a constant temperature.

About the speed of the TEC plates, yes, they are more slow than thermopiles, but this is due mainly to the thermal mass of the TEC, compared with the one of the thermopile ..... and, anyway, consider this: a 10x10mm 96-junctios TEC require at least 10 or 15 seconds, for stabilize, but also a good thermopile (like as example the Ophir and Coherent ones that comes with some of the meters sold here), still needs from 4 to 8 seconds, depending from the laser power, for give a stable reading ..... not speaking about some Scientech, that needs also the double, and all these ones are professional units

poor man's meter would be the Radient Alpha

Coherent Light said:

poor man's meter would be the Radient Alpha

Click to expand...

No ..... for buy it, you need money, and if you have money, you are not a poor man .....

(just kidding, ofcourse )

LOL no HIMNL's right : ) Instead of spending $100+ on a power meter I'd rather put it towards an ever more powerful blue, or at very least towards making my own meter so I totally understand its operation. And I'd build a datalogger onto the system, USB is virtually free if you're digitizing the output anyway...

Though as I said, the price point is attractive (so's the one himnl's selling) and pre-calibrated is very tempting. Thanks for the recommendation, CL!

So I think I can use a significantly thinner plate / direct paint on sensor if I make a shutter apparatus, anyone seen this tried? I should be able to divide the power down physically, then multiply my measurement back up.

If I can do this using only temp sensors, if I put one sensor on a heatsink far from the illuminated part of the device for reference and the other on a thin thin black plate that I'd hit with the laser... if the temperature difference alone will do it, can I then bring the front plate back to the same temperature as the reference with a momentarily attached peltier? (I know keeping it physically connected to that plate would ruin my speed increases)

I'm crazy busy with work but as soon as I have a free evening and > 6 hrs to sleep I'll test it out, I'm guessing that if I hit my test apparatus with 125mA, then 250mA with the same laser, I should get close-to-2x values (I expect there'll be some non-linearity)...... I'll see how close this gets me!

Thanks again for the input, guys!

Incidentally, can TECs accurately measure power by rate-of-change (derivative)? If so might this be a faster way of getting data? Or would variance in beam size hitting the unit change the heat dissipation characteristics too much..? I know this is a very technical question but I thought I'd throw it out there : )

Well, as for using a temperature sensor, it's not so easy, cause the difference of temperature caused from a beam, also a 100mW beam, is not so high, and the thermal mass of the reading plate is also an important parameter ..... just as example, i had tried this way, in the first times, with thermocouples junctions (got ten of thin wire thermocouples like these ones and mangled them for get the jnctions with few millimeters of wire), glueing them on the back of a 10mm aluminium disc took from a black anodized fin, with thermal epoxy ..... the assembly was working, but slower than a small TEC, and not for low powers, and it had 10 junctions in serie ..... with only one, it must be almost insensitive (and, BTW, you don't need a reference temp sensor ..... once you set the zero on the ambient temp, the variation is already proportional to the power, without need a reference)

About the derivative, it's probably not useable, cause the derive is not linear ..... it's a curve that depend from the thermal mass, original temperature, power of the beam, air temperature, adsorbtion coefficent, intrinsic thermal exchange rate of the TEC junctions (and number of them), thermal exchange from face to air (that also change if the TEC is sealed or open-air) and so on ..... i think it ends in a too much complex calculation, just for avoid to wait 15 seconds ..... where instead, once the temperature of the reading face is stabilized from the beam power adsorbed, the reading is precise enough for direct reading.