Building a Reference Laser

[Warske]

This is an attempt at building a diode laser that can be used as a power meter reference.

The concept was already discussed (at length!) in the "Power Meter Calibration and Comparison" thread http://www.laserpointerforums.com/forums/YaBB.pl?num=1217029972/0 but I want to try to take it a step further.

In that thread, the idea was to build two lower power (120 mW and 60 mW) lasers that didn't push the diode. Large heat sinks and stable power supplies were used. The lasers were to be used in a room that was heated or cooled to 77 deg F, and the power would be read exactly 30 seconds after the laser was turned on.

Although it wasn't mentioned (that I saw), presumably if you wanted to do a re-test you would have let the laser cool back down to room temperature first, which could take a while depending on how careful you were being.

In this thread, I want to look at ways to make the output power of a diode laser more stable so that it can be more easily used as a power meter reference. One of the goals is to keep the cost down.

The obvious methods are temperature compensation, temperature control (heating and/or cooling) and optical feedback, AKA automatic power control (APC) regulation.

I want to look at APC first, because I think it will be easier. If that works well enough, I'll just stop there.

This thread is a branch from the "Laser power metering service!" thread http://www.laserpointerforums.com/forums/YaBB.pl?num=1238760659/0 where Adam (Diachi) generously offered to use his new LPM to test lasers for forum members.

From that thread...

[quote author=Warske link=1238760659/0#18 date=1238793307][quote author=lasersbee link=1238760659/0#14 date=1238785082]So what is the problem with IR... the LaserBee I Thermopile Sensor can easily see
the IR wavelength.... :
Jerry

Nothing wrong with IR. I quite like it myself. But I've seen so much fear and loathing of IR in this forum I thought that was reason enough not to ask someone to test an IR laser for me.

Also, some thermal LPMs that are home made or are commercial but old may be a bit suspect in the IR (Dusty? Someone sneezed on it?). If you see a black coating on the detector, you have some assurance it is actually somewhat flat across the visible spectrum. If it wasn't it would appear to be colored, after all. But to know how it is doing in the IR, you actually have to measure its reflectivity at that wavelength. Of course you must have done that for the LaserBee.

Anyway, I was starting to think of putting something together, and if I keep it simple I could provide the details on this forum so someone else could build it if they want. Seems like some of us might be interested in a simple, inexpensive, stable laser for testing. Nothing fancy. If so I think it probably best to keep it in the visible. Maybe an LPC-815. I wonder if the IR diode in there has the detector in it...

You can stick a bit of glass at a 45* angle in the beam that reflects some of the light into a photo diode, that's how it's done in commercial lasers

Sorry I'm unable to help you with metering. If you are in the US there are a number of members there who I'm sure would be willing to help you out

-Adam

Adam, I forgot to thank you for your generosity and inspiring me to take the extra step. You did a good thing.

I tried the glass in the beam. I used the skylight filter from a camera. I was really surprised that the readings I got were very noisy. The readings kept bouncing around. No, the train wasn't going past at the time... On the other hand, the readings I got shooting the photodiode through an ND 8 filter were very stable. I tried it a couple times with different materials and got the same results.

I am leaning toward drilling a hole in the side of the aixiz module and using feedback to the current regulator. That should be pretty straight forward and take care of the first order temperature effects.

I don't want to hijack your thread. Would you like me to start a new one, or would you prefer it remain here. Its almost on topic!
[/quote]


Hmm that's a little strange, but if the other ones worked that's alright.

If you drill a hole in the aixiz make sure you take the diode out ( don't want to drill through the diode now do you ? ) also make sure that the photo diode doesn't get in the way of the beam as this will give you two problems:

A) you lose a whole bunch of power
B) the diode dies from back reflections.

You may want to keep the diode temperature at a constant also, as temperature shifts will effect power and wavelength, and since you are using an optical means to measure the power, wavelength shifts will effect your readings. You could use a TEC with temperature feedback and control to keep the diode at a stable temperature.

It would be good if you started a new thread in the tutorials, help and repairs section, or experiments and modifications. More people would be able to read it and learn from your experiences

-Adam[/quote]
Adam, thanks for the excellent ideas. My hope is that the optical approach will be "good enough" without adding temperature control also. I'll do some tests to find out. And here is the new thread...

Hey Warske....
yeah... another DIY LPM build is always welcome in the community...
The trick is and has always been... how to calibrate the DIY LPM afterit is built...

So true... Maybe a Reference Laser would help.

There is another DIY LPM that was posted that ran current though a resistor to calibrate
the LPM..... here is a link...

http://www.laserpointerforums.com/forums/YaBB.pl?num=1226965719/all-

Thanks for the link. I had looked at that earlier, and it seems like a very worthwhile project. Some folks were put off by the soldering, I think. The idea I have in mind involves scissors, tape, glue, sewing thread... that sort of thing. More like a crafts project, I suppose...

===============================

To get started on this project, I pulled apart an LPC-815 sled, grabbed the red diode, and also pulled the IR diode to see if it had a photosensor in it. The answer: NO. But the IR diode itself is sensitive to light, and maybe I can use that.

So far, I have drilled a 9/16" hole in the side of a new AixiZ case, that being the right size to fit the IR diode. I pressed the red diode in, and powered it up. See the pics below...

Next I'll test the IR diode to see if it will work well enough as a photodetector. Ideally it won't saturate and won't be too temperature sensitive.

 

[diachi]

thats a very nice start you've made, you sure do things quickly !

Does enough light get into the closed can detector diode in order to get a reading ?

Fair enough not wanting to use temperature control, it just adds to the expense, but it would make the laser ULTRA stable  

I'll keep an eye on this thread, nice work so far  

-Adam

 

[lasersbee]

Very nice idea.... I agree with you 100 % until you get to the IR Laser Diode...
Would it not make more sense to use a Photo Diode or similar Photo Detector
that was designed to detect light... I may be wrong and you found a unique way
of using an IR Laser Diode...

BTW... you can also use a standard RED LED as a detector for Red Laser
beams.... :

Jerry

 

[...]

I would suggest firmly against using the ir laser as your power reference, there are many better alternatives (for reason's I don't feel like getting into).  The simplest, as that people have noted, would be to press a red led in there (I bet a 3mm one would fit pretty well), or better yet steel a photo diode out of some ir controlled gadget.  Considering that you are trying to make something stable here, I think it would be worth the time to use a decent detector, I am sure you can find something decent (probably even in the drive somewhere).  Note-do not use a CDS cell, since they are incredibly nonlinear and not stable worth a crap.  

Also, make sure that the beam splitter is held firmly in place in the can, if it shifts it will of course throw off all of your measurements.

A final comment, the apc circuit will keep the power stable, but it won't do anything about the wavelength shifting around.  Now for a thermal sensor this won't be an issue, and if you are only using your photodiode detectors to the 1mw place i think that your plan to just run it for a constant amount of time would work fine (although the norm is to use a munch longer time period, usually 20 minutes, to let everything warm up to operating temp and then take your reading).  If you want to read down better than 1mw, i would bet that you need a more carefully designed standard that incorporates temperature control and has a way to clean up the beam so that scatter doesn't affect the detectors.

Good luck, it sounds like a noble project!  If it were me I would make them for as many wavelengths as I could, and have a kit that we could send around to do a say 10 point calibration (808 780 650 630 405 and 50mw and 100mw) so you could really calibrate your meter.

 

[Warske]

...Does enough light get into the closed can detector diode in order to get a reading ?

Fair enough not wanting to use temperature control, it just adds to the expense, but it would make the laser ULTRA stable  

You're right, the IR diode isn't sensitive enough.
I'm starting to warm up to the idea of temperature control (sorry!).
Thanks for all your great input.

...I agree with you 100 % until you get to the IR Laser Diode...
... you can also use a standard RED LED as a detector for Red Laser beams.... :

I so liked the idea of using the IR diode as a detector.  There are so many nice things on the sled, its tempting to try to use them.  I figured it would work like your LED as detector idea, which I had forgotten about (thanks!).  I had remembered cutting the tops off metal can transistors to make photodetectors.

For the heck of it, I tried a red LED with a 9vdc bias. It worked just fine (see pics), although the current is on the low side:

58 pA dark current
286 nA with the laser running at 100 mW out of the lens.

The signal is about 5,000 times the dark current, which is nice and comfortable.
I also checked to see that signal wasn't saturating.

So far, so good.  Meredith sells an APC Laser Diode Driver Kit for $15.00 (LDD-15 http://www.mi-lasers.com/cgi-bin/shopper.cgi?preadd=action&key=LDD-15) but its 150mA output current is a bit low, and it wants 0.2mA minimum monitor current.  That is where a real photodiode would work better.

So then I looked up the temperature dependence of a silicon photodiode, and it looks like its quantum efficiency varies by 0.09% per deg C at 660 nm (see pic).
http://home.sandiego.edu/~ekim/photodiode/pdtech.html

If I want this thing to work between 60 F and 80 F ambient temperature, that gives me a 1% change in sensitivity.
0.09 * (80 - 60) * 5 / 9 = 1

There is also a change in the dark current, such that it could double, but that doesn't look like an issue.

And as Adam pointed out, there a shift in laser wavelength which will affect the reading because the photodiode response curve is not flat around 660 nm.  I'm not sure this is a large effect, but I haven't looked it up yet.  That may be the next thing to do.

A 1% change is somewhat more than I was hoping for.  I can look at temperature compensating the photodiode, or for that matter, the laser.

Or it may really be simpler to control the temperature.  If I decide to go that route, I am leaning toward using a resisor to heat things up rather than a TEC to cool things down.  Maybe run the laser at 90 or 100 F.  If I only ask it for 100 mW, it should be OK.

I would suggest firmly against using the ir laser as your power reference
... press a red led in there (I bet a 3mm one would fit pretty well)
...Note-do not use a CDS cell
...make sure that the beam splitter is held firmly in place...
the apc circuit ... won't do anything about the wavelength shifting around.  Now for a thermal sensor this won't be an issue...your plan to just run it for a constant amount of time...let everything warm up to operating temp...incorporates temperature control and has a way to clean up the beam so that scatter doesn't affect the detectors.
...make them for as many wavelengths as I could, and have a kit that we could send around to do a say 10 point calibration (808 780 650 630 405 and 50mw and 100mw) so you could really calibrate your meter.

You're right about not using the IR laser.  Also about the 3mm LED.
I had looked up cadmium sulfide photocells.  Radio Shack actually sells them.  From what I can tell, the temperature effects vary with the amount of light, and that is going to be tricky.  So right on that, too.
No beam splitter though, as you can see from the pics.  Sorry I didn't make that clear.
I need to determine how much wavelength shifting I would be looking at ...
The idea of running a constant amount of time was on the "Calibration" thread.  I had hoped that optical feedback would remove the need for warm-up considerations.  Not looking to go ultra stable here, just mostly stable.
Cleaning up the beam is an excellent point.  I have been getting a LOT of side scatter, and I had planned to mount the laser in the back of a project box and run the beam out through a small hole in the other end.  Does that sound about right if I can keep it all nailed down?
I hadn't considered the possibility of the laser power meters being non-linear, requiring different power outputs from a reference laser.  Perhaps a calibrated ND filter would work for that.

If this works out, I would be happy to figure out a way to send it around.  Doubt I have the ambition to do the different colors, though!

Thanks everyone for all the great feedback!

 

[Warske]

Wavelength vs. temperature effect

As several people mentioned, there is a shift in laser wavelength which will affect the reading because the photodiode response curve is not flat around 660 nm.  To make a long story short, there is something like a 0.33% change in the photodiode output with a laser temperature change of 20 deg F.  If you want to check my math:

To find out how large this affect is, I found a datasheet for an Index-Guided AlGainP Laser Diode SLD1133VS http://www.datasheetcatalog.org/datasheet/sony/a6802822.pdf
The datasheet shows that from 25 C to 50 C there is a wavelength shift from 654.6 nm to 659.0 nm  respectively (see pic).  This works out to a 2 nm wavelength shft with a 20 deg F temperature change.  Of course, I don't have a spec sheet for the red LD from the LPC-815 sled, so there is no guarantee that it has the same shift.

To determine how this shift would affect the output of a silicon photodiode, I used the response curve from http://www.laserpointerforums.com/forums/YaBB.pl?num=1217029972/448#478 and found the slope of the curve at 660 nm (see pic).  This shows that a 2 nm shift around 660 nm will change the output by about 0.33%

Latest design

I am thinking of mounting the aixiz case (epoxied in with Arctic Alumina) and an aperture on a 5" length of 3/4" aluminum angle stock (see pics).  Mounted in a Radio Shack project box and powered by a wall wart.

The heater resistor, temperature controller IC, LM317 regulator and current set resistor would be thermal epoxied to the aluminum as well.

Rather than use optical feedback to control the output, the photodiode signal could be made available to check for changes in optical power output.  A reference voltage could be made available for comparison, so that a DVM would nominally read zero volts between the photodiode output and the reference voltage when everything was working properly.

I am thinking of setting the focus at 10 feet and gluing the lens in place.  This could allow a diffraction grating to be used to compare the wavelength between two similar lasers.  Knowing the wavelength difference, it should then be possible to use a large area photodiode, or photodiode and integrating sphere, to compare the power between the two lasers, and potentially match them.

An alternative idea is to set the focus to about 12 inches, so that a spot size can be chosen depending on the distance from the laser.

All of this is seeming like overkill, but once the details are worked out, it shouldn't be too hard to implement, I hope.

Thoughts?

Parts suggestions?

I'm thinking of ordering some photodiodes and a temperature controller chip from Digikey tonight:

OP906 $0.91 photodiode
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=365-1084-ND

PDB-C134 $2.14 photodiode
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=PDB-C134-ND

TMP01FPZ $5.76 temperature controller
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=TMP01FPZ-ND

I haven't worked with these parts, so if you have any other suggestions for cheap red-sensitive photodiodes or a cheap but effective temperature controller, please let me know.