Difference between polarized and non-polarized beam splitters

Thanks, those links lead me to this page on prisms: http://www.nitto-optical.co.jp/english/products/basic_prism/index.html

Edit: I'm going to edit in something I was looking at today, that's the perceived brightness of 660nm compared to 638nm, using RHD's online calculator, found in my signature, I'm seeing 1.4 watts of 638nm is just as bright to the eye as 5 watts of 660nm. Due to this, I'm looking for a higher power single beam output 638nm diode, but only finding a 700mw rated Oclaro diode right now, but the good thing is they are cheap. I could combine them together using a cube to give close to 1.4 watts, if pushed hard. This would give me the same amount of brilliance as the more expensive 5 watt 660nm but at the cost of simplicity.

Here's a RHD comparison of the brightness:



If you look at the way Mateusz Szymański of Opt Lasers or ebay/tomorrowssystem combines two relatively inexpensive 700mw 638nm red diodes, this is how complex it gets, below. Although when you consider the cost of the extra optics & diode holder, the price isn't so cheap.



I see he uses a rotator for the output of one and places the diode holders at different distances, still trying to understand why he did that when you could just turn the diode to adjust polarity for one (there must be more to it than that) and for the distance question, I am clueless.

Another edit: Mateusz got back to me with a answer to my question about the wave rotator, if you don't use it, you have a cross shape, this rotates the shape of the laser diode output. Here was his answer:

"If you don't use it the output beam will be in cross shape. If you use it joined beams can be in the same position for example horizontal. So you do not have cross effect."

I guess the wave rotator not only corrects the orientation of the beam, but rotates the polarity at the same time? The only reason I am questioning this is I thought I once saw a wave rotator which only changed the polarity leaving the beam the same shape.

This is a an idea which might be something I can do, maybe I should find dicro's which will allow me to combine 638nm and 660nm together so I can get higher power red without needing a cube which has more loss? I'm wondering if there are dicro's which are "tuned" to such specific frequencies to allow one to pass through and the other to reflect, maybe these two wavelengths are just too close together to do that. I wouldn't use a 5 watt 660nm and a 700mw 638nm diode this way, just looking at this as a method of using a less expensive 660nm diode together with a cheap 638nm diode to allow me to throw out the losses a cube would produce, only draw back is a lower power 660nm diode just doesn't add as much brilliance to the eye which puts me right back to using a cube, either that or using the high power 5 watt diode to make up for it, arrggg! I'm lost in the wilds walking in a circle

It sounds like the rotator shifts polarity within the rectangle shaped output while leaving the shape and orientation of the rectangle alone, that would make it very useful, that could explain how he gets away with just the 1 cylindrical pair.

OK, that clicked, it must be just changing the polarity leaving the bar output in the same orientation so the two beams overlap, but I'm still left wondering why one diode holder is offset so far away from the cube.

Because of that bounce mirror in the corner.
Measure the distance from the 1st diode to the mirror to the cube, that right turn adds up to the straight length of the number 2 diode.
Otherwise one bar would be more diverged than the other.

oh, duh! Thanks for helping me see that, completely missed it.

Don't feel like the lone ranger.
Also I didn't know that was how those rotators worked, that's fantastic.

PBS losses versus dichro losses aren't that different. Correcting a beam of two 638 beams will ultimately be easier than correcting the beam from two different diodes.

Don't worry about losses so much. The difference between a 5W beam and a 5.2W beam is only distinguishable with a power meter. a 4% increase in power is only a 2% increase in brightness, and the eye just can't detect changes that small. There's no need to spend a bunch of time and money for that extra 4% efficiency if you can't even tell the difference.

RedCowboy said:

Also I didn't know that was how those rotators worked..

Click to expand...

You'll have better luck (in your searches, in your conversations, etc) if you call them by their proper name: half waveplate.

There's another good point, use the same diodes if you are going to correct the beams.

I was comparing the available power and cost between 660 or 650 with 638nm trying to see where the best advantage is for getting the most bang for the buck brilliance to the eye either using a PBS cube or dicro's, or even a combination of them going as far as considering mixing 650 with 638nm by cubing two 700mw Oclaro diodes and adding 650-660nm to the stream through the use of dicro plates, if they can be obtained to reflect 650-660 while passing 638nm through them, which I'm still researching to see if this is possible. If so, I could correct the 700mw cubed Oclaros first, before shooting that output through dicro plates where the 650-660 could be added in for more red power. The 650 or 660nm outputs would need to be corrected before hitting the dicro plate too, but now we are adding a whole lot of complexity for so little gain.

I don't know if this is really worth the trouble to get more red and whether it is really needed, I need to research the power ratios required to make white, it's using the output of the NUBM44 giving so much blue output I thought I needed more red to mix with it, might not be the correct assumption.

I probably ought to stick with combining two 638nm diodes and call it good and perhaps back the power down from the NUBM44 if it is causing so much trouble to balance the powers, or use a PLTB450B or M140 instead which has closer beam properties to the green diode too. Doing optics right when adding the outputs of laser diodes is not for the meekly qualified.

Yes, they exist. PhotonLexicon had a group-buy for 635/660nm dichros a few years back. Back when high powers weren't really available, it was somewhat common for people to build a goofy number of colors into their projectors... for example 660, 638, 532, 473, and 405, or "RRGBV".

They eye has an automatic white-balance. You can get away with a lot. The "proper" ratio depends on the wavelengths used. Less 445 is required compared with 473 for example. 1:1:1 is doable for 445/520/638, but if you were to use 660, you'd need something closer to 1:1:4. Someone on photonlexicon wrote a program called "chroma" many years back to calculate the proper ratio. I'm sure you could dig up a copy somewhere.

Thanks, I saw mention of that there a few weeks ago, but never found a link to it when I looked, might have given up too early.

Yikes...you have me confused ... You want to do a RGB hand held....or a non-hand held. ??

To do a HH RGB.... with good/corrected optics is gonna be another BIG HH !!

In either case....for a RGB...go with a dual Oclarro 63193 set up....exactly as the Matuz set up shows !! PBS combine'em....and then thru a set of 6X cylindricals.

If I were doing a higher power RGB set up...I would use a dual Oclarro 63193 + a single NUBM07E + a single NUBM01T ( or NDG 7475 ) I will do a drawing soon.

If you are sticking with a Blue + Green...I would go with the exact set-up as I'm using on the Cyan Cannon build.

I have not yet seen the NUBM044E correct in an acceptable fashion yet. yes...highest power available....but...a crap beam...which for me is BS !!

SO...the NUBM044E remains " uncorrectable "....as far as I have seen so far...unless it has been done ???? Anybody correct the NUBM044E with perhaps a 2mm EFL collimation lens....then a set of 6X optics ????

If not....I would not use the NUBM044E. And really....when comparing the relative brightness...a NUBM07E @ 470nm @ 4.7W output will appear brighter than a NUBM044E @ 450nm @ 7W !!!

When using the relative brightness calculator....as you used above....the NUBM07E @ 4.7W will be 1.09x brighter than the
NUBM044E @ 7.0W.

Just those extra 20nm make all the difference....and the NUBM07E is very, very correctable.....down to 0.65 mRad !!! very tight !!
Why fight it....seems lika No Brainer to me to chose the NUBM07E.

Beam out

Cyparagon said:

You'll have better luck (in your searches, in your conversations, etc) if you call them by their proper name: half waveplate.

Click to expand...

Yes thanks, I had heard waveplate before but I wasn't sure if it was a 1/4 waveplate or a 1/2 waveplate that would be needed.

I have searched and all I found is a 1/4 waveplates turn the beam 1/4 turn.

So does a 1/2 waveplate flip the beam over. Is the polarization of a beam a north/south deal?

Is that how he is keeping the rectangle, by turning it 180 degrees?

I thought it was either horizontal or vertical but if I understand this correctly now a horizontal beam has a north and south end?

With the 1/2 waveplate, the shape of the beam is kept the same, but the polarization is rotated 90 degrees. Flipping 180 degrees wouldn't do anything anyway; it'd have the effect of returning the polarization back to its original. The polarized beam cubes pass with one polarization, and reflect light with a 90 degree polarization.

Alaska: you can always try and find some on eBay (such as this). Usually they're for HeNe lasers, so the wavelength is not spot-on. I'm not sure how much this affects their efficiency.

Electronic Badger, I've been wondering where you went, long time since I've seen you post, or I'm just not happening upon them. I went to that ebay link, but found some less expensive ones here: Rotators - Opt Lasers

Cyp. I was unaware of the 1/4 or 1/2 rotators, thanks for letting me know what to look for with them.

CDBEAM777 - I want a Cyan Cannon, the unit we have been talking about, and a RGB pointer too, one I can control the mix of colors, one color, or all of them on at the same time. I thought maybe I'd combine the 450nm NUBM44 I have with my NDG7475T and see what it looks like, I have all the parts I need to do that, but after seeing your recommendation to use th3 07E, I want to use that now, but don't have a enough beer money right now to buy another one, my last one died.

Oh I didn't realize OptLasers now carries those waveplates. You'd probably be better off getting those from OptLasers (PM him for the 10% discount) than some HeNe waveplate from an eBay seller. You can get his set of cylindrical optics too, which are pretty decent.

I've bought the cylindrical pair and used them and they're not bad (not in any finished project). His brass mount works, but I'm not a huge fan of his screw-based leveling system. You really want only two or three adjustment screws for stability. Combined, you'll be able to get 1.4-1.8W or so with a pair of Oclaros and the polarized beam-cube + cylindrical optics.

For myself, I don't know if I'd bother with the waveplate if I could get decent a squarish dot. I'd just line overlap the two corrected beams instead, one rotated at 90 degrees. There would be a higher powered center surrounded by the surrounding beam, but it'd probably be sufficient.