Page 2 of 22 FirstFirst 12345612 ... LastLast
Results 11 to 20 of 213

Thread: cylindrical lens pair vs anamorphic prisms?

  1. #11
    Join Date
    Feb 2011
    Location
    New Hampshire
    Posts
    3,513

    Default

    Thanks but again, what does doing well mean? (around how much losses with 3x and above?)
    It has been a couple years since I did the actual test, but the difference is very significant. I believe the loss for the lens pair was around 5 % and the loss for the prisms was around 25%!

    But does that mean that the middle beam in a knife edge won't be corrected as much/the same way as the leftpost and rightmost beams?
    Actually, it's worse. As the beams move off axis the beams develop aberrations. The largest is called coma where part of the beam is focused at one angle and the other part of the beam is focused at another angle. The result is a tear drop shaped beam. You MUST place the beams on axis. So, PBS the beams first and if you want more than two diodes then stack the beams vertically to keep the stack on axis.

    Most people rely on the physical edges of the lenses to be aligned with their various optical axis which works most of the time although many here choose to correct the "rotate" axis" and take care great that the lens is "facing" the beam as best they can.
    I think this adjustment is worth it and always do it.

  2. #12
    Join Date
    May 2016
    Posts
    104

    Default

    Quote Originally Posted by catalanjo View Post

    They should be stacked vertically (thin on top of thin) so all beams pass as nearly as possible thru the centre of the lens.
    ...
    If you stack the thin beams side by side as you seem to be suggesting, they will not be collectively thin anymore so when you make them fat, they will become VERY fat and so will the narrow spaces in between them ... they will not all fit on your scanners unless you use massive mirrors, which would make them a tad slow, and in short, it would be a ghastly mistake!
    I did not know this. Actually I've been told otherwise, that 3x3 stacking is a possibility. Maybe the guys were wrong. I'll try to find the discussion (not in this forum).
    But there's also this werbsite, which is dead now but is archived
    https://web.archive.org/web/20150126...ode-array.html

    Browsing my reference images, all the knife edging are done "side by side". That said, what is vertical axis is determined by how the diode is angled in the housing, right?
    Click image for larger version. 

Name:	150320103259590965.jpg 
Views:	99 
Size:	52.9 KB 
ID:	50001
    Click image for larger version. 

Name:	150320103259705952.jpg 
Views:	84 
Size:	45.2 KB 
ID:	50002
    Click image for larger version. 

Name:	150320103415696392.jpg 
Views:	82 
Size:	53.5 KB 
ID:	50003

    This does raise the question for me, how on earth do some build a 10W red module when the strongest diodes are like 0.7W each. Is the red's diameter in the near feald much smaller than for the greens and blues?

    Quote Originally Posted by catalanjo View Post
    PS. Cyl lenses against prisms controversy ....losses not different enough to be important
    Quote Originally Posted by planters View Post
    the difference is very significant. I believe the loss for the lens pair was around 5 % and the loss for the prisms was around 25%!
    Dammit, lads, I want to believe both of you but it's hard

  3. #13
    Join Date
    Feb 2011
    Location
    New Hampshire
    Posts
    3,513

    Default

    First of all, if you have watched my videos then you have seen how I approach these issues. You can "believe" me. The center axis we are referencing is on the axis where the lens has its power. You stack additional diodes parallel to and on the axis where the cylinders do not magnify. But, you can test our recommendation quite easily. Set up a single diode with a lens pair and see what happens when you slightly rotate one lens relative to the other and see what happens when you pass the beam through the pair progressively further and further off axis. This is fun to do and a good investment before you spend a lot of time and money on a "10 W" red.

  4. #14
    Join Date
    May 2016
    Posts
    104

    Default

    Quote Originally Posted by planters View Post
    First of all, if you have watched my videos then you have seen how I approach these issues. You can "believe" me.
    I sure have I remember how you explain why a circular cylindrical lens pair and a corresponding mount with circular hole is easier to use because the rectangular ones are not perfect and a little off axis and cause some abberation.
    But I don't remember anything about losses with prisms and also stacking the beams vertically, as catalanjo says, I've always seen them stacked horizontally with the cylindricals of course rotated the right angle so only one axis wil be affected as you explained. I'll rewatch the videos and come back.

    PS. But the page I linked to and some Arctos projectors I've seen use telescopic lens and several both vertically and horizontally stacked beams for the reds. Maybe they were built before the Oclaro and Mitsu multimodes came out and those diodes dont have a short and long axis.

  5. #15
    Join Date
    Feb 2011
    Location
    New Hampshire
    Posts
    3,513

    Default

    I only meant that the information I provide is based on something I have tested and that it should seem like I would plausibly test myself.

    You should run the test yourself because it IS fun and provides an opportunity for you to judge the magnitude of the trade offs.

    But, let me try to explain why the off axis problem occurs so you don't have to treat this as "it just is". You are correct that it will be less pronounced if the input beam is more parallel. Nevertheless,the rays that pass through the lens are bent more strongly as they enter the lens further from the axis. That is no secret and is obvious. However, when a diverging beam (all beams diverge to some degree) enters off axis,, because it has a finite width, the side of the beam that enters closer to the axis will experience less refractive power than the opposite side of the beam that enters further out. Because the divergence of the input beam is presumably symmetrical about its axis then you have introduced an asymmetrical aberration to the output beam. This would not occur if the input beam was perfectly parallel, which it is not.

    You can produce that 10W red, but practically this will require you to use 2 W multimode reds and therefore you will want to avoid off axis aberration. The power loss with the prisms is close to what I stated above and this means 2.5W of loss for your 10W red. So, I would stay with the lenses as well.

    Another consideration that has been covered before is whether to use many low power single mode reds (like Arctos) or to use a low number of high power red (Like I for one do) and add the beam correction elements these require. There have been threads about this, but the major difference will be the costs of the diodes, mounts, prisms, collimators FOR EACH DIODE. This can add up if you need 50 diodes.

  6. #16
    Join Date
    May 2016
    Posts
    104

    Default

    Quote Originally Posted by planters View Post
    I only meant that the information I provide is based on something I have tested and that it should seem like I would plausibly test myself.
    Sure. I could have worded what I said myself better.

    But, let me try to explain why the off axis problem occurs so you don't have to treat this as "it just is". You are correct that it will be less pronounced if the input beam is more parallel. Nevertheless,the rays that pass through the lens are bent more strongly as they enter the lens further from the axis. That is no secret and is obvious. However, when a diverging beam (all beams diverge to some degree) enters off axis,, because it has a finite width, the side of the beam that enters closer to the axis will experience less refractive power than the opposite side of the beam that enters further out. Because the divergence of the input beam is presumably symmetrical about its axis then you have introduced an asymmetrical aberration to the output beam. This would not occur if the input beam was perfectly parallel, which it is not.
    It makes perfect sense. The confusion came from catalanjo saying vertical instead of horizontal. Or maybe I'm wrong in saying horizontal instead of vertical here, in which case vertical relative to what? Because we clearly mean the same thing.

    You can produce that 10W red, but practically this will require you to use 2 W multimode reds and therefore you will want to avoid off axis aberration.
    2W reds? Which company?
    The problem I see with multimode reds is the limited number of mitsus or oclaros one can stack before the beam is too fat to fit any mirror. My scanners cant fit more than 6mm.

    The power loss with the prisms is close to what I stated above and this means 2.5W of loss for your 10W red. So, I would stay with the lenses as well.
    Well, I'm glad I made this thread before ordering few prisms. Cause now I'm going to go with cylindricals. Now I can buy the correction optics from the same seller who sells everything else and who doesnt have prisms so this is convenient.
    However, I don't want to sound skeptical, but just for reference for future builds, do you think perhaps the prisms you had were just bad quality (bad AR coating)? Like imagine the faulty 6w blue diode you got. The reason I'm asking is Lightspace uses prisms and Arctos uses prisms. One possibility is the time saved on building a projector with prisms justifies the power loss, but have you considered the other possibility that the prisms you checked few years back just weren't good enough or their quality has improved?

  7. #17
    Join Date
    Feb 2011
    Location
    New Hampshire
    Posts
    3,513

    Default

    2W reds? Which company?
    Mitsu P73s cooled and driven hard, PSB into pairs and stack three pairs vertically. The six diodes will produce more than 10 W and you will loose some with the down stream optics. These diodes are not expensive and minimizing the number means you limit the all up costs and when a better diode becomes available then you have fewer to replace. These usually are released before you burn out the last generation.

    My scanners cant fit more than 6mm.
    You can modify the beam to whatever size you want by adjusting the spatial filter telescope ratio. The price you pay is divergence.

    Lightspace uses prisms and Arctos uses prisms.
    The prisms do not have an off axis penalty. You can add as many diodes as you want, but the beam dimensions/divergence will be the same. I can't prove this, but I believe the cylinder pair will actually produce a little better performance because you can compensate for any residual defocus in the expanded axis. The prisms produce no focusing effect.

    I measured the loss, but everyone that has posted their experience has noted the effect. I have replaced a couple of prism pairs as I have upgraded a couple of projectors and the performance has improved notably, but it is possible that my prisms were defective.

  8. #18
    Join Date
    May 2016
    Posts
    104

    Default

    Quote Originally Posted by planters View Post
    Mitsu P73s cooled and driven hard, PSB into pairs and stack three pairs vertically. The six diodes will produce more than 10 W and you will loose some with the down stream optics.
    Wow, you mean to overdrive a 0.5W diode to output 2W?

    These diodes are not expensive and minimizing the number means you limit the all up costs and when a better diode becomes available then you have fewer to replace.
    I agree.
    On that note what do you tink about the Oclaro 63193? Seems a tiny bit better profile and rated for 200mW more but maybe won't handle overdriving to 2W https://sites.google.com/site/dtrlpf...w-638nm-diodes

    You can modify the beam to whatever size you want by adjusting the spatial filter telescope ratio. The price you pay is divergence.
    I will respond about this after I rewatch some of your videos. I don't remember noticing telescopic lens on multimode modules anywhere else . I have some questions but maybe I'll find the info in the videos.

    I measured the loss, but everyone that has posted their experience has noted the effect. I have replaced a couple of prism pairs as I have upgraded a couple of projectors and the performance has improved notably\
    OK

  9. #19
    Join Date
    Aug 2010
    Location
    London or Spain depending on the weather
    Posts
    1,390

    Default


    Quote Originally Posted by planters View Post
    ........and stack three pairs vertically...............
    Tread softly with the "vertical stacking" advice E...... you might cause "confusion" like I did...
    Cheers
    PS. @"short and long axis" .....ALL laser diodes have a fast and a slow axis! ... difference is just that the fast axis in multimode diodes are much faster!
    Last edited by catalanjo; 05-24-2016 at 17:11.

  10. #20
    Join Date
    Feb 2011
    Location
    New Hampshire
    Posts
    3,513

    Default

    If they watch my spatial filtering video they should understand this. Even if they don't get the fast and slow concept the axis I expand is in the horizontal and stacking in the vertical will work. But, I understand the risks.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •