Combining a third beam

Hi all,

I am beginning to collect all the materials for my next build which will be a triple diode the only problem is combining a third beam. I have been doing a lot of research on the combining of light but I have not been able to find anywhere that can explain principles of combining a third laser beam of the same wavelength(without using knife edging). The options I have looked at are using a two way mirror to let pbs combined beam pass through while the third bounces off in the direction of the pbs beam(it is knife edging but cleaner imo). The problem with this is that the reflectivity of the mirror from one side largely depends on the amount of light emitted. The other thing that I have been debating is using a 2nd pbs and using a quarter wave plate to turn the third beam into circularly polarized light, although I am also 50% sure that the beam from the 1st pbs will diffract and enter the 3rd diode and kill it.

What do you guys think? How would you do this?

Also, I am going to use 3 NDB7875s.

:thanks:

Without knife-edging? Why?

You can't do it with PBS cubes due to polarization. You have the pass diode and the reflect diode, there is no usable third state. Though turning it may do something. I'm no optics expert, I'd just knife edge....

Edit 3: For anyone reading this thread in the future, take a look at my reference thread here: http://laserpointerforums.com/f51/reference-guide-how-combine-lasers-77449.html#post1113793

upaa27 said:

I have not been able to find anywhere that can explain principles of combining a third laser beam of the same wavelength(without using knife edging).

Click to expand...

That's because it's not possible AFAIK.

upaa27 said:

The options I have looked at are using a two way mirror to let pbs combined beam pass through while the third bounces off in the direction of the pbs beam(it is knife edging but cleaner imo).

Click to expand...

Got a datasheet for this mirror?

upaa27 said:

The other thing that I have been debating is using a 2nd pbs and using a quarter wave plate to turn the third beam into circularly polarized light.

Click to expand...

It doesn't work like that. The light entering the PBS has to be at opposite polarization to combine. The polarization of the light exiting the first PBS won't be uniform, a wave plate won't be able to change that. You'll just end up splitting the light from the first PBS back into two beams if it enters another PBS.



You would be best off knife edging all three diodes for the most uniform beam.

However, you can buy dichros with a very narrow reflection/transmission gap. If you can find a high WL diode and a low WL diode you may be able to get a dichro and combine the light that way. This probably wouldn't work though.
-The WL difference between the diodes would have to be pretty big.
-The cost of such a dichro will be at least $100.
-The wavelength gap between the diodes will narrow as the diodes heat up.
-You'll have to bin a ton of diodes to find three suitable for this.
Edit: This may actually work if you use one of those 460nm diodes instead of all NDB7875.
Edit 2: I stand corrected about the cost, turns out it's under $100. See RHD's post.

Why not use one NDB7A75 with corrective optics and one NDB7875?
The NDB7A75 once corrected should be similar in output as the NDB7875. Then PBS cube em.

I don't see why you couldn't combine at least 4 blue beams without resorting to knife edging. It would just take some planning, and a bit of hunting for the right parts.

You should be able to send 2x NDB7A75 beams into a PBS, and then 2x NDB7675 beams into another PBS. Take the outputs of both PBS cubes (spectro'd first to ensure that the former is around 450nm, and the latter around 470nm) and combine them with a dichro like this one: Optical Filter Dichroic460DRLP (or similar, as long as there is a sharp curve around 460).

ARG said:

That's because it's not possible AFAIK.



Got a datasheet for this mirror?



It doesn't work like that. The light entering the PBS has to be at opposite polarization to combine. The polarization of the light exiting the first PBS won't be uniform, a wave plate won't be able to change that. You'll just end up splitting the light from the first PBS back into two beams if it enters another PBS.



You would be best off knife edging all three diodes for the most uniform beam.

However, you can buy dichros with a very narrow reflection/transmission gap. If you can find a high WL diode and a low WL diode you may be able to get a dichro and combine the light that way. This probably wouldn't work though.
-The WL difference between the diodes would have to be pretty big though.
-The cost of such a dichro will be at least $100
-The wavelength gap between the diodes will narrow as the diodes heat up.
-You'll have to bin a ton of diodes to find three suitable for this.
Edit: This may actually work if you use one of those 460nm diodes instead of all NDB7875.

Click to expand...

Thanks for the confirmation that 2 pbss will not work. I will look into a dichro like that thanks for the info.

VisibleGreen said:

Why not use one NDB7A75 with corrective optics and one NDB7875?
The NDB7A75 once corrected should be similar in output as the NDB7875. Then PBS cube em.

Click to expand...

That would definitely work the only problem is the price of just 1 NDBA7A75. I will look into this for sure though. Thanks!

rhd said:

I don't see why you couldn't combine at least 4 blue beams without resorting to knife edging. It would just take some planning, and a bit of hunting for the right parts.

You should be able to send 2x NDB7A75 beams into a PBS, and then 2x NDB7675 beams into another PBS. Take the outputs of both PBS cubes (spectro'd first to ensure that the former is around 450nm, and the latter around 470nm) and combine them with a dichro like this one: Optical Filter Dichroic460DRLP (or similar, as long as there is a sharp curve around 460).

Click to expand...

I will look into this. Only problem is that the host I am making is a triple diode handheld and not a labby or tabletop one so I don't think that I could fit 4 in although I could just pbs 2 and dichro the remaining beam. Thanks for the link that is one great looking dichro.

At the moment knife edging has become a bit more of an option due to your replies however with the recent discovery of dichros that can combine beams within such a small wl range I might go with that. I am thinking either pbsing the first two beams and knife edging the third just for divergences sake or picking up one NDB7a75 and dichroing it with NDB7675 2 pbsed. Thanks for the help from you guys!

rhd said:

I don't see why you couldn't combine at least 4 blue beams without resorting to knife edging. It would just take some planning, and a bit of hunting for the right parts.

You should be able to send 2x NDB7A75 beams into a PBS, and then 2x NDB7675 beams into another PBS. Take the outputs of both PBS cubes (spectro'd first to ensure that the former is around 450nm, and the latter around 470nm) and combine them with a dichro like this one: Optical Filter Dichroic460DRLP (or similar, as long as there is a sharp curve around 460).

Click to expand...

How would I get the NDB7A75 to operate at 470? By increasing the heat? On the datasheet it says the max wavelength it will output is 455.

upaa27 said:

I could just pbs 2 and dichro the remaining beam. Thanks for the link that is one great looking dichro.

Click to expand...

From experience, it works out well. Just make sure to do the dichro or PBS before the bounce mirror.

One big question is would the Dichro pass enough of the light from the third diode to be worth it? I am just thinking that the extra power lost from an extra optic may eat up most of the benefit of a 3 diode handheld.

IsaacT said:

One big question is would the Dichro pass enough of the light from the third diode to be worth it? I am just thinking that the extra power lost from an extra optic may eat up most of the benefit of a 3 diode handheld.

Click to expand...

Math time!

Reflection band 415 to 455 with average of 95%, Transmission band 465 to 650 nm with an average of > 90%

Click to expand...

Lets say the two 445nm diodes combined put out 5W post PBS, then 250mW will be lost from the dichro reflecting the light. The 3rd diode would then have to deliver over 275mW post bounce mirror to make the added power worth it.

MMmmmmm ? What you really want are three (3) NDB7A75's. Just knife edge em !!

See TRIDENTIS BUILD-Three (3) 445 Combiner-Knife edge


http://laserpointerforums.com/f65/tridentis-6-7w-blue-beauty-83781.html

Yes....the beam to beam to beam alignment, at several miles out, will not be perfect. But then, the beam to beam to beam alignment on any method of optics combination will have some alignment issues, at great distances ! That is just the way it is.

I predict that with the divergence of the NBD7A75, you will need a form of corrective optics. Maybe Anamorphic prism's or a set of Cylinderical lenses.

I think you may need the correction factor of the Cylinderical lenses vs the Anamorphics. Not sure as I have not yet experimented with a NBD7A75.

The Tridentis has Anamorphics after the knife edge arrangement. But the LD's are the 9mm variant, not the NBD7A75. I have a concern about all three (3) beams passing thru the Cylindericals, for really, only one beam will be dead center of the Cylindericals and the other two ,ever so slightly off the center axis.

I have yet to experiment with this so you may be limited to an Anamorphic Prism pair. The down side to this is about 3~4% more power loss with the Anamorphics. There is that pesky optical trade off again !

Good luck !

IsaacT said:

One big question is would the Dichro pass enough of the light from the third diode to be worth it? I am just thinking that the extra power lost from an extra optic may eat up most of the benefit of a 3 diode handheld.

Click to expand...

Yes it is very worth it if all goes well this build should be able to pull of 11w or more of power counting in power loss due to reflectivity(ndb7875s outputting ~3.4w and ndb7a 5w so PBS is going to be right on the first diode so pretty much no loss there and the PBS I am using is a thorlabs one so it is very high quality so I expect 5>% of loss there so 0.95*3.4=3.23. Dicho will have 5% loss so 0.95*5=4.75. 4.75+3.23+3.4=11.38). Although go a bery short time and the heatsink I sent to sinner to build is a pretty big one.

So estimated 11.38w of blue in a handheld. Although in my experience the ndb78s usually output a bit more than 3.4 when driven at 2.4a

IsaacT said:

One big question is would the Dichro pass enough of the light from the third diode to be worth it?

Click to expand...

Not with the setup that everyone else has been talking about. For example this (which I'm having trouble following):

upaa27 said:

Yes it is very worth it if all goes well this build should be able to pull of 11w or more of power counting in power loss due to reflectivity(ndb7875s outputting ~3.4w and ndb7a 5w so PBS is going to be right on the first diode so pretty much no loss there and the PBS I am using is a thorlabs one so it is very high quality so I expect 5>% of loss there so 0.95*3.4=3.23. Dicho will have 5% loss so 0.95*5=4.75. 4.75+3.23+3.4=11.38). Although go a bery short time and the heatsink I sent to sinner to build is a pretty big one.

So estimated 11.38w of blue in a handheld. Although in my experience the ndb78s usually output a bit more than 3.4 when driven at 2.4a

Click to expand...

The NDB7875 and the NDB7A75 are 450-ish nm diodes in practice. I don't know how you would accomplish combining 3x diodes, with just a dichro and pbs, if they all tend to scope at about the same wavelength.

The combination I proposed was:

...send 2x NDB7A75 beams into a PBS, and then 2x NDB7675 beams into another PBS.

Click to expand...

That ^ would work.

rhd said:

The NDB7875 and the NDB7A75 are 450-ish nm diodes in practice. I don't know how you would accomplish combining 3x diodes, with just a dichro and pbs, if they all tend to scope at about the same wavelength.

...Take the outputs of both PBS cubes (spectro'd first to ensure that the former is around 450nm, and the latter around 470nm) and combine them with a dichro like this one: Optical Filter Dichroic460DRLP (or similar, as long as there is a sharp curve around 460).

Click to expand...

I was under the impression that somehow you would change the wavelength to one of the beams to 470.

I think that I might just knife edge with a pbs and mirrors if this if the pbs with dicro setup wont work.

Is there anyway to change the wavelength to 470?

upaa27 said:

I was under the impression that somehow you would change the wavelength to one of the beams to 470.

I think that I might just knife edge with a pbs and mirrors if this if the pbs with dicro setup wont work.

Is there anyway to change the wavelength to 470?

Click to expand...

You would need to use a 470nm diode to begin with. You can't just change a wavelength.

upaa27 said:

Yes it is very worth it if all goes well this build should be able to pull of 11w or more of power counting in power loss due to reflectivity(ndb7875s outputting ~3.4w and ndb7a 5w so PBS is going to be right on the first diode so pretty much no loss there and the PBS I am using is a thorlabs one so it is very high quality so I expect 5>% of loss there so 0.95*3.4=3.23. Dicho will have 5% loss so 0.95*5=4.75. 4.75+3.23+3.4=11.38). Although go a bery short time and the heatsink I sent to sinner to build is a pretty big one.

Click to expand...

There's going to be a lot more loss than that. Count your bounce mirrors too, and a wave plate if you're using one.

I highly recommend stanwax's cubes. The efficiency I got with one of them was near 100%.