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

Has anyone ever ran their diode through a diffraction grating before turning the TEC on, so they can physically see the wavelength shift?
Could be quite a cool visual experiment.
As Fire has done, using the laser diode at its threshold will yield the best results. InGaN is known not to thermal shift as much as AlGaInP/AlGaAs.

To clear up the colour difference I have included notable wavelengths across the a rough observed temp tuning range of a common 635nm diode. It should be noted that these are daytime visuals. In darker lighting conditions they will appear redshifted unless compared with a longer wavelength laser.

As you can see not too much noticeable difference between 607 and 613nm. 607nm is only noticeably more yellow with the bloom. The clearer difference is with the 622nm. It is unlikely even with the night-time redshifting you would mistake 613nm as red. You would have to view a short wavelength laser right next to it, for it to suddenly appear blood orange I would expect.









I'm not sure how RMPC/LDX have achieved such a drastic drop in wavelength with very little temp change.

These are the specs of their diodes at different operating temps.

632-638nm (635) @ 20 degC
627-633nm (630) @ 15 degC -5nm
624-630nm (627) @ 10 degC -3nm
619-625nm (622) @ 5 degC -5nm

The central shift would translate to an average of ~4.3nm per 5 degC which is a lot

My expectation is that with a cheap setup 625nm is more realistic limit. It is possible to get lower, but condensation will start to cause issues. And cooling would become harder to achieve. 613nm may end up being unrealistic unless you can do what planters did.

I don't have a spectrometer but am trying to get one from work. I used another diode running at room temp just for visual comparison. It was easy to see the difference then, but with just the shilled diode by itself it was just looking kind of red...

Yes, that probably is the only way of truly noticing a difference by eye. Having them side by side. For there to be a difference in colour at that part of the spectrum, you would have to shifted it at the very least 5nm. Do you reckon you could manage to get close to 625nm?

Without a spectrometer it is hard to tell.

The RMPC diode seems to be really special.
Where did you get the data from?

The diode must surely be very expensive.
Also its a MM diode with 60u emitter, this would make for a very bad beam.

I think -80 to -95°C should be doable with just a 4 stage tec element. See the data sheet of the tec, it should do 120k delta with 0,5W heat load.
To keep condensation at bay, I will use a vacuum "filled" casing. this would also limit heating losses by convection / conduction. Radiation will be mostly eliminated by using a polished diode mount.

We will see how it goes.

CurtisOliver said:

As you can see not too much noticeable difference between 607 and 613nm. 607nm is only noticeably more yellow with the bloom. The clearer difference is with the 622nm. It is unlikely even with the night-time redshifting you would mistake 613nm as red. You would have to view a short wavelength laser right next to it, for it to suddenly appear blood orange I would expect.

Click to expand...

Personally I think it would be more accurate to say 613nm has more red in it than 607nm having more yellow in it. Semantics. Pictures are difficult to give it justice, on my monitor your 622nm example is closest in color to my 607nm.

CurtisOliver said:

I'm not sure how RMPC/LDX have achieved such a drastic drop in wavelength with very little temp change.

These are the specs of their diodes at different operating temps.

632-638nm (635) @ 20 degC
627-633nm (630) @ 15 degC -5nm
624-630nm (627) @ 10 degC -3nm
619-625nm (622) @ 5 degC -5nm

The central shift would translate to an average of ~4.3nm per 5 degC which is a lot

My expectation is that with a cheap setup 625nm is more realistic limit. It is possible to get lower, but condensation will start to cause issues. And cooling would become harder to achieve. 613nm may end up being unrealistic unless you can do what planters did.

Click to expand...

They did that because they are different diodes. Structure is different. The cooling is likely only to assist achieving the lower WL.

farbe2 said:

The RMPC diode seems to be really special.
Where did you get the data from?

The diode must surely be very expensive.
Also its a MM diode with 60u emitter, this would make for a very bad beam.

Click to expand...

Surely is special. RPMC has the data available on their site. Lowest quote I've gotten was $500, though I'm pretty sure it wasn't through them. Definitely would expect this to have a bad beam.

CurtisOliver said:

I'm not sure how RMPC/LDX have achieved such a drastic drop in wavelength with very little temp change.

These are the specs of their diodes at different operating temps.

632-638nm (635) @ 20 degC
627-633nm (630) @ 15 degC -5nm
624-630nm (627) @ 10 degC -3nm
619-625nm (622) @ 5 degC -5nm

The central shift would translate to an average of ~4.3nm per 5 degC which is a lot

Click to expand...

These informations are interesting, i could not find anything about that by RMPC.

However i think these diodes are not really usable, just because of the 60u emitter size.
For example the 700mW diode has a 40-50u emitter size (someone did tell me that, however i don't have an idea where that is coming from either). This would mean a even worse beam divergence at even lower power. So nogo for me.

I will order TECs and machine a housing.
We will see how it goes.

CurtisOliver said:

Has anyone ever ran their diode through a diffraction grating before turning the TEC on, so they can physically see the wavelength shift?
Could be quite a cool visual experiment.
As Fire has done, using the laser diode at its threshold will yield the best results. InGaN is known not to thermal shift as much as AlGaInP/AlGaAs.

To clear up the colour difference I have included notable wavelengths across the a rough observed temp tuning range of a common 635nm diode. It should be noted that these are daytime visuals. In darker lighting conditions they will appear redshifted unless compared with a longer wavelength laser.

As you can see not too much noticeable difference between 607 and 613nm. 607nm is only noticeably more yellow with the bloom. The clearer difference is with the 622nm. It is unlikely even with the night-time redshifting you would mistake 613nm as red. You would have to view a short wavelength laser right next to it, for it to suddenly appear blood orange I would expect.

Click to expand...

Not that it will help much, but on the gas side, I've set up one of my 594s with my 612 and one of my 633 HeNe lasers, side by side. The 612 look like a tangerine orange.

I've also seen a tube with a secondary line of 604 and it didn't strike me as significantly different from the 612. (bloompyles orange HeNe before he broke it)

farbe2 said:

These informations are interesting, i could not find anything about that by RMPC.

However i think these diodes are not really usable, just because of the 60u emitter size.
For example the 700mW diode has a 40-50u emitter size (someone did tell me that, however i don't have an idea where that is coming from either). This would mean a even worse beam divergence at even lower power. So nogo for me.

I will order TECs and machine a housing.
We will see how it goes.

Click to expand...

It is the item list. But as ZRaffleticket correctly pointed out. They are separate diodes, with just differing operating temperatures.
Based on my own expectations and that seen in reality by FireMyLaser, you will need a very good setup to achieve your goal of 613nm. I won't say its impossible, because it isn't. But it is unlikely unless you can safely drive the temperature very low indeed.

I picked up some of the 633 nm diodes for a project just like this. I haven't gotten around to developing it yet.
I do have access to liquid nitrogen, though. At LN2 temperature, the wavelength shift is noticeable but considering the expense of these diodes, LN2 cooling is not something I want to experiment with very much.

As others have noted, good thermal insulation and a hermetic seal to protect against condensation will be crucial. The guy in the TechIngredients video cleverly uses a resistor to heat the optical window. I had an idea to couple the hot side of the TEC heatsink to the optics to prevent those from developing ice.

A vacuum would be great but it's technically harder to accomplish than including some solid desiccant in the diode chamber and just sealing it with regular ol' air or, better yet, nitrogen-enriched atmosphere (having access to LN2 has its advantages yet again).

I don't think that 120k delta will be possible with a air or nitrogen filled box.
The TEC would be overwhelmed with just the conduction and convection energy losses.
I also think that the desiccant will not help very much, it can't get the air dry enough to make it not condense at the optics.

I will definitely try to make a setup that can be evacuated. I will use a "normal" red diode to confirm the temperature achievable with this.
If this gets me anywhere close, I will purchase the 633nm diode.
For now its waiting for a TEC manufacturer that could send me a single unit for development.
I did write to multiple but no one answered. Its night time in china, so no surprise here.

I have also done some experimenting with this but did not finish it as my design was not looking promising. One issues I ran into was creating the hermetic seal for the diode. The main issue is with the TEC inefficiency and determining the number and power of the TEC units in a stack to get the best efficiency, my design used an equivalent of 3 stacked elements. There are multi-stage TEC elements sold that would be better, but they are a bit more expensive.
There are some very good research papers on this topic for the medical field.

A very good resource to read is: https://aip.scitation.org/doi/full/10.1063/1.4967231
In order to get lower wavleength they also used used and ECDL, External Cavity to bring pull the wavelength down a bit more.

0.15-0.25nm/degC seems much more realistic.
Temp tuning a 635nm to 613nm would require an 88 deg temp change at best. So -63degC minimum. Not an easy feat.
Worst case operating temperature is -122degC based on only 0.15nm/degC.
Lets assume all goes well and you do manage to achieve a goal of -80degC. That would be a 105degC change. I'll use 0.15nm/degC as our base. So you may expect to see a 15.75nm reduction in wavelength. So ~619nm.
The colour difference would at least be noticeable even if off your initial goal.

This paper is nice!

It also states the convection losses with around 0,6W @ 10mbar.
I expect to get around 1mbar or even less. So even less losses.

And I did finde a low power cheap diode that states 0,25nm/k wavelength shift. (it shows a graph)
However this diode is 638nm
This would be good too. 30nm less at 120°C temperature drop would give me 608nm at 5mW. Yes thats not much, however it would still be nice to have.

I did get a quote for a tec. It would be around 160USD for one, 100USD for 10pc.
I will buy the diode first, test if the 0,25nm/k are correct. that should be doable with Ice spray.

If everything works out, I will buy tecs.

Ok heres a link to someone else thats done it before not sure if u have access to the same equipment but maybe u can find some of the info helpful. They brought emission down to 611nm from 638 just by cooling with liq. Nitrogen, when inside the pressure chamber vessel they got it all the way down to 575nm

18LJ said:

Ok heres a link to someone else thats done it before not sure if u have access to the same equipment but maybe u can find some of the info helpful. They brought emission down to 611nm from 638 just by cooling with liq. Nitrogen, when inside the pressure chamber vessel they got it all the way down to 575nm

Methods and Results

Click to expand...

Hey, thanks for that link