Calibration Laser weekend project

[justinjja]

I chose a linear driver because I believe it will be more stable for my application.

Remember, the main reason a linear driver drifts is because the set resistors.
They have a rating something like 200ppm/degree c
That means that when they go from room temp to 75c
a 1% error is created

the second reason a linear driver drifts is because the IC warms up,

Neither of those apply to me because my set resistors, and my IC are temperature controlled.

@ Laserbee
if your still reading this thread,
how stable is your high power FAP laser without optical feedback?

 

[lasersbee]

@ Laserbee
if your still reading this thread,
how stable is your high power FAP laser without optical feedback?

After about 30-60 minutes to stabilize the whole system while
tweaking the Current to the TEC in the FAP I can hold 15 Watts
to within 2-3 mW over ~60 seconds.
That's more than long enough and precise enough to do our tests...

At one point (when I find the time) I will use the installed Optical
sensor to fine adjust the FAP output once the output is set to a
specific power.


Jerry

You can contact us at any time on our Website: J.BAUER Electronics

 

[Bionic-Badger]

Your mixing up thermal transfer with satiability,
My diode could be sitting at 60c, and if it stays right at 60c, the diodes power output is going to be stable...

Actually, I think it's the opposite. No offense intended, but your calibration methodology is fundamentally flawed. You're confusing thermal stability with optical stability, expecting that the former will imply the latter, which is not always the case. Temperature stability is not even the goal for the calibration source. It is for the thermopile sensor, but not for the radiation source. The radiation source requires consistent optical output.

The only way you're going to achieve optical stability is with optical feedback that controls the output via current regulation or something else. Optical feedback is therefore not adding another variable. It's the only variable that really matters. In fact, you don't even know if your laser is stable at all without feedback for the output you're tracking. Thermal stability doesn't guarantee that, and the output can change over time even from something such as diode degradation while the same thermals are in place.

The only thing your thermals should be doing is keeping the diode sufficiently cool so that there isn't thermal runaway. It doesn't matter if your diode is thermally stable, at least not beyond the normal operating range, as long as the output is being regulated by the optical regulation.

I would also completely ditch that Aixiz module mount.

If my emphasis wasn't sufficient before: it's a piece-of-shit.

Things' 10% surface contact is, in fact, a generous assessment. You're essentially making contact with your Aixiz module on a tangent of the round surface which could mean virtually no contact at all. You risk having to overcompensate with the cooling apparatus, causing frost/water condensation, and also preventing the system from reacting to temperature changes--which could kill your diode no matter how much power you push into the Peltier. If you don't want something from China, buy some other mount, but ditch that piece-of-shit module -- that Aixiz shouldn't even purport to be a heatsink -- because you're going to wreck your diode.

 

[Meatball]

Actually, I think it's the opposite. No offense intended, but your calibration methodology is fundamentally flawed. You're confusing thermal stability with optical stability, expecting that the former will imply the latter, which is not always the case. Temperature stability is not even the goal for the calibration source. It is for the thermopile sensor, but not for the radiation source. The radiation source requires consistent optical output.

The only way you're going to achieve optical stability is with optical feedback that controls the output via current regulation or something else. Optical feedback is therefore not adding another variable. It's the only variable that really matters. In fact, you don't even know if your laser is stable at all without feedback for the output you're tracking. Thermal stability doesn't guarantee that, and the output can change over time even from something such as diode degradation while the same thermals are in place.

The only thing your thermals should be doing is keeping the diode sufficiently cool so that there isn't thermal runaway. It doesn't matter if your diode is thermally stable, at least not beyond the normal operating range, as long as the output is being regulated by the optical regulation.

I would also completely ditch that Aixiz module mount.

If my emphasis wasn't sufficient before: it's a piece-of-shit.

Things' 10% surface contact is, in fact, a generous assessment. You're essentially making contact with your Aixiz module on a tangent of the round surface which could mean virtually no contact at all. You risk having to overcompensate with the cooling apparatus, causing frost/water condensation, and also preventing the system from reacting to temperature changes--which could kill your diode no matter how much power you push into the Peltier. If you don't want something from China, buy some other mount, but ditch that piece-of-shit module -- that Aixiz shouldn't even purport to be a heatsink -- because you're going to wreck your diode.

Thank you. I was waiting for someone to mention it.

He is completely right. If you want to regulate temperature fine, go ahead and monitor it and compensate accordingly, but if you want to regulate optical output, you have to directly monitor that light the whole time that thing is in operation. There's not much of a way to get around it. Optical feedback is much more necessary than even temperature regulation.

I would be assert to say that it would be better to start with a good optical feedback loop which can regulate output without any temperature regulation. A good error amp would be able to do so without issue. Temperature should just be a side note for the sake of diode lifetime or wavelength control.

 

[Things]

Adding optical feed back is not hard at all, as Meatball has said, and it'd mean your laser actually remains 100% stable, 100% of the time, and from the instant you turn it on. The 445nm lasers are about as unstable as you can get in diode laser form, keeping it at a stable temperature will help, but no where near to the point of calling it a calibration laser.

It may seem harsh, but really we're just trying to save you time, money and effort, especially if you're actually intending to use it to calibrate LPMs. There is a reason calibration lasers fall into a league of their own, you need to put in special considerations to actually keep it stable. As Badger said, cooling the diode with the TEC will just prevent it from failing, in terms of stability it won't do much at all without optical feedback.

 

[Seoul_lasers]

Adding optical feed back is not hard at all, as Meatball has said, and it'd mean your laser actually remains 100% stable, 100% of the time, and from the instant you turn it on. The 445nm lasers are about as unstable as you can get in diode laser form, keeping it at a stable temperature will help, but no where near to the point of calling it a calibration laser.

It may seem harsh, but really we're just trying to save you time, money and effort, especially if you're actually intending to use it to calibrate LPMs. There is a reason calibration lasers fall into a league of their own, you need to put in special considerations to actually keep it stable. As Badger said, cooling the diode with the TEC will just prevent it from failing, in terms of stability it won't do much at all without optical feedback.

Totally correct. 445nm lasers are all over the place with their output.
Highly variable depending on temperature and still no where near stable, not enough to do what you are intending..
You'll need optical feedback to adjust the output continuously.
Notice the spec on this 445nm laser. <3% over 8 hrs.

Beijing Laserwave OptoElectronics Technology Co. Ltd.

 

[ARG]

I tried this myself with a 445nm laser that only had temperature feedback, it didn't stabilize very well. Optical feedback is a must if you're using these to calibrate LPM's.

IR is where it's at, it stabilizes much better than 445nm in my experience. You don't even need anything fancy, this is what MarioMaster used to calibrate the Alphas from an old b&w laser.

I do the same thing, it's solid as a rock in terms of stability.

 

[justinjja]

I'm sure your right, an 808 probably would be easier,
but who on lpf wants an lpm calibrated at 808?
I would rather my lpm be as accurate as possible at 445nm.

Laserbee said he has achieved over 99.9% stability over a shot period of time,
without optical feedback (15w/15w+-3mw)

long term degradation / stability isn't going to be an issue.
When I'm calibrating, I wont be relying on this laser alone,
Ill have my main LPM sitting right next to it...

and please stop harping on the aixiz heatsink!
I was able to get my diode temp 3c below ambient,
there was no condensation,
the aixiz heatsink works, period.

 

[ARG]

Are your LPM's not linear across wavelengths? Calibrating at 808nm shouldn't be an issue.I think more people on LPF want an LPM calibrated with a stable laser rather than a 445nm laser.

In my experience M140's are only stable enough for calibration at 500mW or under without optical feedback. How do you plan to curve adjust for the higher powers?

Also, why not just use these if you knew the aixiz holder had a big gap? Better thermal transfer and thermal control would help with thermal stability.
Hot 2pcs SK12 Size 12mm CNC Linear Rail Shaft Guide Support | eBay

 

[Meatball]

IIRC, an LPM CAN be calibrated with multiple wavelengths in a single calibration. The difference you see between a stable 445 and a stable 808 will not be anything more significant than the "noise" that will appear on your sensor rail anyhow. Its that 808s are designed for stability in that they normally function as pump diodes, and are also routinely well heatsinked like Jerry's 445 probably was - not to mention the readily absorbed 808 lambda.

It makes sense to do calibration with the highest knowable test point (mW) attainable, but the difference in resolution should still be negligible given you can perform the calibration properly at a lower (possibly more stable) energy.

Right, so if degradation is not an issue, then why worry about a temperature regulation at all? As long as the output is stable, that's all you need right? The diode could start at 20C and increase to flaming hot over time, but as long that optical power is stable, who cares?

Just monitor a percentage of the deflected beam power off a photodiode, and (very carefully) adjust LD current.

I'm sure your right, an 808 probably would be easier,
but who on lpf wants an lpm calibrated at 808?
I would rather my lpm be as accurate as possible at 445nm.

Laserbee said he has achieved over 99.9% stability over a shot period of time,
without optical feedback (15w/15w+-3mw)

long term degradation / stability isn't going to be an issue.
When I'm calibrating, I wont be relying on this laser alone,
Ill have my main LPM sitting right next to it...

and please stop harping on the aixiz heatsink!
I was able to get my diode temp 3c below ambient,
there was no condensation,
the aixiz heatsink works, period.

 

[Bionic-Badger]

I'm sure your right, an 808 probably would be easier,
but who on lpf wants an lpm calibrated at 808?
I would rather my lpm be as accurate as possible at 445nm.

I'd rather have a TEC coating that is broadband. It shouldn't matter if you're using 808nm instead of 445nm. Otherwise you might as well just calibrate some photodiodes to whatever wavelength you want.

Laserbee said he has achieved over 99.9% stability over a shot period of time,
without optical feedback (15w/15w+-3mw)

Lasersbee is using a CO2 gas laser that must have its gas kept at certain temperatures to achieve output stability. It's the nature of that kind of laser, just as how certain lasers have very high inherent wavelength stability. This does not correspond to your setup: a diode laser with output stability that is critically dependent on the input current not the temperature.

and please stop harping on the aixiz heatsink!
I was able to get my diode temp 3c below ambient,
there was no condensation,
the aixiz heatsink works, period.

No, I'll keep harping on that piece of garbage because it is garbage. I don't think people should waste their money on it, let alone on more laser diodes because they died from insufficient heat-sinking. If you don't wish to heed those warnings and risk your equipment that's your choice.

 

[pschlosser]

Are these hard to build? And can they be built without dependence on a calibrated LPM?

I would welcome building a laser diode-driven reference laser for LPM calibration purposes.

I did a search on the net for output stabilized reference laser schematic and didn't find much in the way of DIY projects.

 

[justinjja]

I'd rather have a TEC coating that is broadband. It shouldn't matter if you're using 808nm instead of 445nm. Otherwise you might as well just calibrate some photodiodes to whatever wavelength you want.



Lasersbee is using a CO2 gas laser that must have its gas kept at certain temperatures to achieve output stability. It's the nature of that kind of laser, just as how certain lasers have very high inherent wavelength stability. This does not correspond to your setup: a diode laser with output stability that is critically dependent on the input current not the temperature.



No, I'll keep harping on that piece of garbage because it is garbage. I don't think people should waste their money on it, let alone on more laser diodes because they died from insufficient heat-sinking. If you don't wish to heed those warnings and risk your equipment that's your choice.

I disagree with basicly every thing you just said wow...

Even broadband coatings are not perfect,
that's why when you buy a professional meter,
they calibrate it at a specific wavelength

The laser Laserbee talked about is a diode laser, google it...

there is a 5c difference between my diode and my aixiz heatsink
there is 3W passing into the aixiz heatsink
1.6c/W is not a bad thermal junction.

@pschlosser, that is a chicken or egg question lol
You need an LPM to make this laser...

 

[robmobz]

@pschlosser, that is a chicken or egg question lol
You need an LPM to make this laser...

Not necessarily, all that you need is access to an LPM on one occasion and then you can use this to calibrate your next one and then use that for all future comparison AFAIK.

NOTE: I have never done this but it seems to make sense that way to me assuming that there is no significant drift between getting it metered and the first calibration using it and that that LPM would keep the same readings after that for recalibration.

 

[pschlosser]

@pschlosser, that is a chicken or egg question lol
You need an LPM to make this laser...

I asked, because if it were possible to do without a calibrated LPM up front, it would make a great project for many of us. Being able to calibrate our own LPMs, without the need to ship them off to somebody else, or acquisition of a calibrated LPM, would be very attractive, indeed.

 

[Bionic-Badger]

I disagree with basicly every thing you just said wow...

Even broadband coatings are not perfect,
that's why when you buy a professional meter,
they calibrate it at a specific wavelength

No they're not perfect, but many are designed to operate over the visible and well into the NIR. If they don't the surface probably isn't that good for red either.

The laser Laserbee talked about is a diode laser, google it...

You are correct, that is a diode laser. I retract that statement.

there is a 5c difference between my diode and my aixiz heatsink
there is 3W passing into the aixiz heatsink
1.6c/W is not a bad thermal junction.

Perhaps you got lucky, or maybe the manufacturer has improved the tolerances? I still don't trust any "heat sink" that has a visible gap around the majority of the exterior. That's no more than a stable mount to me.