Can a TMP36 as a LPM Meter Sensor?

So I can get perfectly good temperature reading from a TMP36 temperature sensor (its in a small TO-92 package). And I noticed that most LMPs go by the heat generated right? So that makes me wonder if I can use the TMP36 to sense heat and get the mW off that.

The only problem is, I dont know what the temperature to mW ratio is.

So my questions are

Can a TMP36 be used for a Home made LPM?

What does one degree C equal in mW?

------------------------------------------------------------

I know theres no direct conversion, but there has to be some correlation because thermal LPMs use peltier arrays or thermopiles.

I will be testing it using a rough 1 watt laser so my hope is it doesn't burn the TMP36 out lol. And I dont expect any better than +-5%

TMP36 Data sheet

No, this will not (really) work. One problem is that the TO92 package is normally air cooled, so its temperature per amount of power will vary with air flow and such. The relationship will also not be linear either.

Power handling will also be very low, those packages are made of plastic that will start to melt/smolder if you point a 50 mW laser at it with a small beam diameter.

You cannot use a thermometer to measure power. While they are somewhat related, power and temperature are not directly convertible. It would be like trying to measure volume with a pressure gauge.

The closest device that can convert power/energy to temperature is a calorimeter setup, but calorimeters are not suited for very low powers, especially the DIY variety.

I know most agree it doesn't work perfect but I was using a thermocouple setup in front of black grip tape(had stone and black carbon) only to compare lasers.

I was using altering the method posted on LPF of using an IR temp gun reading ambient temp, then the temp of laser focused to the T/C junction.. subtract ambient and multiply the temperature rise by 3.3mw.

Obviously not perfect, but it was good for comparing my 2w builds to each other to ensure consistency between them and no bad diodes.

You can buy TEC or a single TC online to use for this, it's cheap but it's really reliable or accurate. I will be able to compare this method to the Hyperion Argum.. and the Ophir head I'm buying next very soon.

TEC is the way to do it if you're going DIY ... still need something to calibrate it with though - otherwise the best you can do is compare relative powers, no absolute powers.

Personally I'd just buy something off the shelf ... if I were to include the cost of my time spent DIYing something it'd work out the same price or more expensive than some options anyway. Unless I was just doing it for the sake of building an LPM...

https://www.laserbeelpm.com/laserbee-3.7w-usb.html <-- these actually use a TEC.

94Z28 said:

...subtract ambient and multiply the temperature rise by 3.3mw...

Click to expand...

Bullshit. The actual physics (for blackbody radiation) is not linear. It's about as far from linear as you can get - a FOURTH power.



Your 3.3mW/degree "constant" will change drastically with emitting area, emissivity, conduction to the surface, and probably 7 other variables I don't feel like listing.

Cyparagon said:

You cannot use a thermometer to measure power. While they are somewhat related, power and temperature are not directly convertible.

Click to expand...

Exactly, you need to measure the flow of heat from some absorbing surface into a heatsink.

Similar how you would measure electrical current across a shunt - you measure temperatures at both sides of this thermal connection (of known resistance) and the calculate the flow rate from the difference.

This is the operating principe of all thermal laser measuring systems. The specialized heads like the ophir ones contain a plate inside with a ring of of thermocouples between it's laser-heated middle and the outside heatsink. TEC's do a similar thing: their thermal resistance is determined, and their output voltage is proportional to the temperature difference between sides. As one side is on a heatsink and the other is heated, the produced voltage indicates heat flow and thus laser power.


Heating a black body -could- be used, if it were placed in a hard vacuum free from any heat transfer by other means (conduction, convention). You could think of something like a small black ball dangling of a thin wire in an evacuated glass bell jar or such.

In such a system the temperature of the black body (relative to ambient) would give the exact laser power after calculation, but these systems would be very slow, very impractical to build, require complex calculations and probably be pretty fragile too.

I have proposed a "Poor man's powermeter" outline in: http://laserpointerforums.com/f67/l...logging-capture-device-98338.html#post1449466
The idea is to use a very thin oxidized aluminium foil to collect the heat of the beam (thermal collector).
Why: :tinfoil: foil is cheap, in my country can be obtained from eaten chocolate, there are DIY manuals for oxidizing aluminium, AL has excellent heat conductivity so it's expected to withstand narrow beam without melting. Foil is thin so it should heat up and cool down quickly. I propose using contactless sensor from Melexis to measure the temperature to retain fast reaction. It is being sold for about $5.4. The tricky part is the mechanical construction. It should account for a fixed thermal resistance between the thermal collector and the cooling radiator (can be just the box of the device).
The thermo-optic sensor's package must also be tied to the cooling radiator with the less possible thermal resistance, because the sensor measures the difference between it's package and the monitored surface with 0.5% accuracy. That's it. Then it could be calibrated in several fixed points and the results can be interpolated between the calibration points.
Anybody keen on to experiment ? :thanks:

I've done some experiments with thin strips of brass/copper material and absorptive coating.

It does work to some degree, but you need to eliminate practical things like air flow/convection to make them work reliably.

Also, you need to calibrate such set-ups which is problematic unless you already have a reliable meter or known source.

I've tried to tackle the calibration issue by actually attaching a small SMD resistor on the back of the area that you shine the laser onto. This way i could calibrate the unit for a given thermal input (say 100 mW, regulated 1 volts into a 10 ohm resistor).

In my experience this approach works pretty well, but resulted in VERY long measurement times. It took about 30 seconds to over a minute to get a stable reading. This is fine if you have a laser with stable output power (direct diode) that long duty cycle. It makes this type of meter unsuitable for fluctiating DPSS lasers or ones that have a host so small you cannot run them for a few minutes straight.

@Benm +rep - the calibration with resistor is what I've missed Thanks !
As for the slow update - did you use the sensor I proposed ? Which thickness was your strip ? I have estimated that the commonly available Al foil has around 18μm thickness. I target low power lasers 80..1800mW.
The sensor is contactless, basically we convert the heat to a secondary stream of IR radiation, which is then captured via the sensor's IR lens from the back side of the strip.
The sensor's FOV must cover the area which gets irradiated with the LASER and should not "see" outside of the foil. It's ~ 90 deg. for these sensors.
Yes, the air flow must be reduced, so some wind shielding is a must, but if you look at the Ophir's sensors, they also have this (the beam enters via a hole).

The material i used was thicker than that. I didn't measure it exactly, but i reckon somewhere in the order of 50 to 100 uM including the black coating on it.

As for the heating resistor and thermal sensor: i connected those directly onto the back of the foil, adding quite a bit of thermal mass to the system (the components are tiny smd but compared to the foil itself they contribute a lot).

Using a contactless IR system would help a lot, but then you cannot really do the resistor calibration that easily anymore. You could still do it (put the heating resistor onto the foil and take a 'slow' measurement since you know it's power is constant, but yo could not leave it on there to re-calibrate whenever you want to.

If you go this route, i'd suggest using a disc instead of a strip, with the IR temperature meter behind the disc in an enclosed space to block airflow. On the measuring side of the disc you could build a setup similar to the ophir sensors, ecasing the whole thing inside a heatsink with only a small hole to shine the beam into.

Also note you will need 2 identical temperature measurement devices, one for the center of the disk/strip and the other for the heatsink itself. The heatsink -will- heat up above ambient temperature. At 80 mW this will not be that noticable, but running up to 2 watts it may get warm enough to feel that by hand depending on how big it is.

As for measurement speed: With a foil that thin and contactless temperature measurement it may actually be decently fast. I would not expect accurate results in a second, but if you get to >95% of the end result in 5 seconds that is already very good (and you can do some tricks in software to predict the endpoint, i believe ophir does that too, and all the TEC based meters certainly do).

On the heatsinking: The sensor I mentioned is differential - e.t. tie it's body to the heatsink and you're all set. (But we assume constant thermal resistance heatsink-ambient).
Now, when the plot is already drawn, uh oh some volunteers with free time are needed
Nowadays I'm working on a web site so little to no time for the hobby :-(
Infact, I'll have the I2C protocol on my MCU working for my main DIY/hobby project, so more or less I'll have done most of the work required to wire the sensor.

If it measures the temperature difference between what it's looking at and its mounting point that would be fine. We can assume that the heatsink has the same temperature all over, or at least that thermal resistances are fairly constant regardless of laser power. Since these are all relative measurements thermal gradients within the heatsink are not problematic as they are predicatble.

Do you have a datasheet for this contactless sensor, and perhaps some place i could obtain it? Given the low cost you mentioned this might be interesting to fool around with a bit - and if it's I2C interfaced it should not be a problem reading it with an arduino or something like that.

Yes, I have what's needed. I work with ARM MCUs, M0 w USB would do the job. Or the famous "blue pill". If I remember, with I2C it can refresh with more than 100Hz.
Here is the information:
Datasheet @ attachment # 1
And here is where you can buy it: mlx90614esf-baa @ eBay

Those look pretty nice!

Speed and accuracy are certainly there, though the package isn't overly practical. I guess the best way to mount it is to drill a 8mm hole through the heatsink and press-fit it in in such a way that it 'looks at the hotspot' on the foil as best as you can.

Price seems good in any case, i guess i'll order one or two to play around with.

Yes, I cared for the prices - it should be affordable. The foil should be somehow connected to the same radiator. Ideally, the foil'd hang in vacuum but since we search reproducible solution, I imagine it inside a pipe (the wind shield). I wonder whether the pipe should be thermo insulator (probably PVC) or thermo-conductive and directly connected to the radiator. Actually part of it.
I'm not a fan of press-fitting of hi tech semiconductors but in this case, I can't offer easier and better solution.
But it's still possible not to press it that hard. I think that if it fits almost tightly, some amount of thermo conductive compound could fill the gaps. Or the radiator might have a side cut for additional elasticity - to allow the hole to expand when forced to.

For LASER diodes, there are Chinese hosts which do not require press-fitting but rely on a nut to hold the diode in it's seat.