Building a orange (613nm) Laser from scratch with off-the-shelf parts (and for cheap!)

[farbe2]

You can get the 488nm sharp laser diodes wavelength selected from 475nm-500nm for about 400$ per diode (if you order only a few and not 2k or more. This is definitely much cheaper than the cost of a high level cooling/heating assembly and can be just inserted into the usual diode mounts.

Does not seem like it. You would need two of these diodes, so thats already 800$. If you need to purchase a "few" that means 10?
That would be 8000$ right there. You would still need a driver and a mount + optics.
Even if we assume that you only need to purchase three of these special diodes, you would be at 2400$ excluding driver / optics / mount.

Selecting a few 488nm diodes and using my module would mean <2000€ (I was asking for 1020€ for a module with the expensive 633nm diode) for both modules including driver. I would buy my selected diodes from Phillip, I know he has some and I assume they would not be much much more expensive than the normal price. So assuming 50€ per diode, we would be around 1850€ for both modules.
So cheaper and less work. Also fully sealed, so no worries about dirt.
And these modules would have the advantage of being able to tune the wavelength to some degree.

SURE! these modules are not "AliExpress pointer" cheap, but they are quite cheap for the advantages and wavelengths they offer. Where would you get a 613nm 100mW laser thats tunable between 613 and ±640nm for this price?
If you calculate all the materiales used, you will see that I do not make a fortune with these (if someone would buy them). Actually quite the opposite, the work the modules take, the equipment needed and so on. If I would be running a company, these needed to be at least double the price for work alone!

However I am still interested, where would you get these diodes? I only know topgan lasers that sell custom diodes in low quantities. These are quite expensive and they only offer between 420 and 460nm with 50mW.

 

[Singlemode Laser]

Does not seem like it. You would need two of these diodes, so thats already 800$.

Why I need two? Your setup ( that is actually very cool) has also one diode and not two.

If you need to purchase a "few" that means 10?

You can purchase even a single one. That's why the price is higher. Everything less than a full lot ( i think they package them in 5 boxes with 200 diodes each) requires someone open one and pick it. 1 or 20 won't change the price.

That would be 8000$ right there. You would still need a driver and a mount + optics.

I heard that there should be enough options (DTR is one of many) to get a mount, lens and a driver. It's the laserpointer forum here

Even if we assume that you only need to purchase three of these special diodes, you would be at 2400$ excluding driver / optics / mount.

You can buy 1 diode for little less than 400$ ( my most recent number)

Selecting a few 488nm diodes

From the final batch ( thats the one you only can buy now) you will never get 480nm or 495nm.

and using my module would mean <2000€ (I was asking for 1020€ for a module with the expensive 633nm diode) for both modules including driver. I would buy my selected diodes from Phillip, I know he has some and I assume they would not be much much more expensive than the normal price.

selected from the ones he has his hands on.

So assuming 50€ per diode,

For 50$ I will buy 10 or more 483nm diodes immediately. But you assume and not guarantee.

we would be around 1850€ for both modules.
So cheaper and less work. Also fully sealed, so no worries about dirt.

A to can diode in a DTR mount is less work and also sealed.

And these modules would have the advantage of being able to tune the wavelength to some degree.

The blues tune 1nm for every 20°K. I can tune one 488nm by +/-4nm with a cheap diffraction grating (and I did it and showed it here)

SHURE! these modules are not "AliExpress pointer" cheap, but they are quite cheap for the advantages and wavelengths they offer. Where would you get a 613nm 100mW laser thats tunable between 613 and ±640nm for this price?

You were talking about the 488nm ones. And that's the wavelength I answered to

If you calculate all the materiales used, you will see that I do not make a fortune with these (if someone would buy them). Actually quite the opposite, the work the modules take, the equipment needed and so on. If I would be running a company, these needed to be at least double the price for work alone!

I don't care about how much you spend or earn. It's not in my interest to sell diodes for a profit.

However I am still interested, where would you get these diodes? I only know topgan lasers that sell custom diodes in low quantities. These are quite expensive and they only offer between 420 and 460nm with 50mW.

 

[liveforphysics]

You can get the 488nm sharp laser diodes wavelength selected from 475nm-500nm for about 400$ per diode (if you order only a few and not 2k or more. This is definitely much cheaper than the cost of a high level cooling/heating assembly and can be just inserted into the usual diode mounts.

Singlemode

I am so grateful for you sharing this information Singlemode. 25 wavelengths of blue to green transition for $10k in diodes is a lot, but also priceless for experiments if they offer any output level with a stable-ish single mode or at least repeatable double mode power level.

I would mount all 25 of them on a temp controlled mount, if you have at least +-1nm temp wavelength drift, it would be priceless.

 

[farbe2]

Why I need two? Your setup ( that is actually very cool) has also one diode and not two.

Whoops, i was talking about liveforphysics need to have two diodes/wavelengths close to each other for raman purposes. Thats why i wanted to use two modules for the two diodes for more difference out of the two selected normal 488nm diodes.

I just assumed stuff, because you where saying "a few" and not "one".
This was just a answer on your argument that it would be cheaper to get a special diode than using the cryo technique.
So i calculated what two of the cheap diodes + cryo module and what "a few" diodes of the special ones cost.
Actually i allready made a mistake because you would only need to cool the lower diode and heat the upper one. This would make the heating module way cheaper.

So i still stand my point, if you need +-5nm around 488nm, a cryo assembly for the lower one and a heating assembly for the higher one would be cheaper than a specialty diode.

Granted for different wavelengths e.g. <483 or >493nm, this would not be possible.

So i think we just talked past each other.

The blues tune 1nm for every 20°K. I can tune one 488nm by +/-4nm with a cheap diffraction grating (and I did it and showed it here)

Thats interesting. Do you have a link? Did you take spectrometer readings before/after tuning? I assumed that the 633nm diode is not much lower tunable because it already emits around the bottom of the laser gain bandwidth. Did you capture the LED emission spectrum and compared it to the tuning range?
How did the power compare before/after the grating? Could you drive the diode at the same power as without tuning?

 

[Singlemode Laser]

Whoops, i was talking about liveforphysics need to have two diodes/wavelengths close to each other for raman purposes. Thats why i wanted to use two modules for the two diodes for more difference out of the two selected normal 488nm diodes.

I just assumed stuff, because you where saying "a few" and not "one".
This was just a answer on your argument that it would be cheaper to get a special diode than using the cryo technique.
So i calculated what two of the cheap diodes + cryo module and what "a few" diodes of the special ones cost.
Actually i allready made a mistake because you would only need to cool the lower diode and heat the upper one. This would make the heating module way cheaper.

And the heating will kill the diode pretty fast

So i still stand my point, if you need +-5nm around 488nm, a cryo assembly for the lower one and a heating assembly for the higher one would be cheaper than a specialty diode.

Feel free to stand to every point you like. I am not interested in calculating part costs etc.

Granted for different wavelengths e.g. <483 or >493nm, this would not be possible.

So i think we just talked past each other.


Thats interesting. Do you have a link?

Gan Laser diode wavelengh shift temerature - Google Search

www.google.com

And look at the datasheets for the blue ones here (single mode diodes):

LD (Laser Diodes) | Products | NICHIA CORPORATION

www.nichia.co.jp

and google in general answers most of the questions

Did you take spectrometer readings before/after tuning?

During tuning

I assumed that the 633nm diode is not much lower tunable because it already emits around the bottom of the laser gain bandwidth. Did you capture the LED emission spectrum and compared it to the tuning range?

Should be measured and listed in a datasheet of the diode.

The ASE and the laser gain bandwidth below threshold shifts with temperature as well as above threshold if only temperature tuning is used. This is the only way to change the gain bandwidth

How did the power compare before/after the grating? Could you drive the diode at the same power as without tuning?

Read the results here: https://arxiv.org/abs/1810.11434

Singlemode

 

[liveforphysics]

Nice paper and nice experiment Singlemode. I've got a few diodes in the 460-495nm range to measure nm/degC drift and share results when I get a few hour.

Do you understand how the 375nm multi-mode LD had no shifting with temperature from -20degC to 60degC, aside from getting dim at 60degC and losing one small mode spike at -20degC? I had incorrectly assumed everything had to shift some amount with temperature?

 

[Singlemode Laser]

Nice paper and nice experiment Singlemode. I've got a few diodes in the 460-495nm range to measure nm/degC drift and share results when I get a few hour.

Do you understand how the 375nm multi-mode LD had no shifting with temperature from -20degC to 60degC, aside from getting dim at 60degC and losing one small mode spike at -20degC? I had incorrectly assumed everything had to shift some amount with temperature?

What measurement with the 375nm are you referring to?

Singlemode

 

[liveforphysics]

The dominant energy wavelength (in the multi-mode spectra) was 375nm, and it remained stable across a wide temperature range.

 

[paul1598419]

And the heating will kill the diode pretty fast


Feel free to stand to every point you like. I am not interested in calculating part costs etc.

Gan Laser diode wavelengh shift temerature - Google Search

www.google.com

And look at the datasheets for the blue ones here (single mode diodes):

LD (Laser Diodes) | Products | NICHIA CORPORATION

www.nichia.co.jp

and google in general answers most of the questions

During tuning


Should be measured and listed in a datasheet of the diode.

The ASE and the laser gain bandwidth below threshold shifts with temperature as well as above threshold if only temperature tuning is used. This is the only way to change the gain bandwidth

Read the results here: https://arxiv.org/abs/1810.11434

Singlemode

Great to see you posting here again.

 

[Singlemode Laser]

The dominant energy wavelength (in the multi-mode spectra) was 375nm, and it remained stable across a wide temperature range.

Please provide more information, diode datasheet, your measurement or. s.e.
Otherwise I can only say:
Nothing to add here

 

[farbe2]

And the heating will kill the diode pretty fast

Thats not true, if you keep the junction below the maximum rated operation temperature, there is now problem.
All of the laser video projectors I know of drive there (blue) diodes in small heatsinks that are usually not at <40°C, thats almost always > 60°C. The NUBM31t for example runs quite happy at 70°C according to the datasheet.
The gh04850b2g is happy with up to 60°C case temperature. Yes the MTBF will be lower than running at 25°C but not "pretty fast" by my standards.

I can tune one 488nm by +/-4nm with a cheap diffraction grating (and I did it and showed it here)

I was asking for a link to your thread / post that you show this.

Should be measured and listed in a datasheet of the diode.

No, thats not true. I have seen countless diode data sheets, not one of them showed the LED emission spectrum (before lasing) of the diode. Most diode manufacturers don't even show a spectrum at all for example: Oclaro, Sharp, Mitsubishi and Ushio does not show a spectrum at all. Nichia and Osram include a spectrum that does not show the LED emission from the diode.

... Temperature .... This is the only way to change the gain bandwidth

Thats clear, I was talking about the (grating) tuning ability inside the gain bandwidth. You cant tune a diode to the edge of amplification, the diode will just snap back to its free running wavelength. So my assumption was that you can deduce the tuning rage if you know the gain bandwidth and that the gain bandwidth can be deduced out of the LED emission spectrum before lasing.

I did google that, however I didn't find anything about this. As you have already made experiments, I was hoping that you have some data there.
However, The paper you have posted would indicate that this assumption could be true. The experimenters could not get the diode to lase at lower / higher wavelengths at the same -45dB emission power from free running. This would mean one could in theory at least deduce the grating tuning capability out of the free running power distribution.

The dominant energy wavelength (in the multi-mode spectra) was 375nm, and it remained stable across a wide temperature range.

Thats very interesting. I tried to tune a 405nm diode with temperature and got more tuning range than a 445nm diode for the same temperature range. So the 405nm one showed significantly more nm/k than the 445nm diode.

The HL37013MG that you are most likely are talking about, is from Ushio, so they don't show anything useful in there data sheets...
They will provide lifetime data if you ask them, but thats about it.



Back to Topic:

My 613nm modul is almost finished. Adhesive is set! I currently run a vacuum test to check for tightness. Everything seems good so far.
Keep you updated!

 

[liveforphysics]

The cup and insulation wadding was added because condensation and frost formed faster and harder than I expected for how dry of a day it was.

I gotta see how to shrink the spectrometer plots, because I took streaming data logs of sweeping the temperature vs wavelengths between a couple of devices that aren't time synchronized, because my goal was only see which diodes I had were most impacted by temp.

 

[farbe2]

Nice work!
Good idea to use a ice cup on a heatpipe/vapor chamber to keep the Hot side of the Tec in check!
For a quick test much less expensive than a big heatsink with big fans. Also no Noise! I hate the fan noise while taking measurements, it makes me less concentrated.

Just to mention it:
Be careful with drive the diodes at full current while cooling them, the efficiency will increase a lot thus overdriving them significantly (possibly to failure). I underestimated this effect and needed to modify my driving circuit to lower the current with lower temperature.

For a 405nm diode (don’t know which one I tested) I got around 0,07nm/k from 50C to -50C.
My prototyp Module is perfect for tests like this, I can change diodes quickly and can get 100K swing without problem.
I just need to make a new lid to have a elastomer seal instead of a metal/metal seal.

 

[Singlemode Laser]

Thats not true, if you keep the junction below the maximum rated operation temperature, there is now problem.
All of the laser video projectors I know of drive there (blue) diodes in small heatsinks that are usually not at <40°C, thats almost always > 60°C. The NUBM31t for example runs quite happy at 70°C according to the datasheet.
The gh04850b2g is happy with up to 60°C case temperature. Yes the MTBF will be lower than running at 25°C but not "pretty fast" by my standards.

Your argument won't get better if you just say the opposite of mine. Think what you want about temperature and MTBF. Every diode laser textbook has an exponential relationship between temperature and MTBF. But by your standards it's not relevant, right?

I was asking for a link to your thread / post that you show this.

The data in the paper I posted is from my measurements....

No, thats not true. I have seen countless diode data sheets, not one of them showed the LED emission spectrum (before lasing) of the diode. Most diode manufacturers don't even show a spectrum at all for example: Oclaro, Sharp, Mitsubishi and Ushio does not show a spectrum at all. Nichia and Osram include a spectrum that does not show the LED emission from the diode.

Look in textbooks or get your hands an these data. Scientific Lasers from Toptica have these data for the included diode. This is only for the customers but with searching the internet, and sci bub you can get all the publications you want. This stuff is not brand-new and there is plenty of information.

Thats clear, I was talking about the (grating) tuning ability inside the gain bandwidth. You cant tune a diode to the edge of amplification, the diode will just snap back to its free running wavelength. So my assumption was that you can deduce the tuning rage if you know the gain bandwidth and that the gain bandwidth can be deduced out of the LED emission spectrum before lasing.

I don't get the point you are trying to make here.

I did google that, however I didn't find anything about this. As you have already made experiments, I was hoping that you have some data there.

If I google: ecdl diode laser temperature tuning cooling
I get a paper about a red ECDL that was cooled down to -60C or so and the grating was always able to tune 4-5nm lower than the free running diode.

However, The paper you have posted would indicate that this assumption could be true. The experimenters could not get the diode to lase at lower / higher wavelengths at the same -45dB emission power from free running. This would mean one could in theory at least deduce the grating tuning capability out of the free running power distribution.

Also here I can't follow your point , except that you just want to disagree.

Thats very interesting. I tried to tune a 405nm diode with temperature and got more tuning range than a 445nm diode for the same temperature range. So the 405nm one showed significantly more nm/k than the 445nm diode.

The HL37013MG that you are most likely are talking about, is from Ushio, so they don't show anything useful in there data sheets...
They will provide lifetime data if you ask them, but thats about it.



Back to Topic:

My 613nm modul is almost finished. Adhesive is set! I currently run a vacuum test to check for tightness. Everything seems good so far.
Keep you updated!
View attachment 73585

I am wasting my time by answering to you in this thread. Wish you all the best with your setup.

Singlemode