Page 3 of 22 FirstFirst 123456713 ... LastLast
Results 21 to 30 of 213

Thread: cylindrical lens pair vs anamorphic prisms?

  1. #21
    Join Date
    May 2016
    Posts
    104

    Default

    You can modify the beam to whatever size you want by adjusting the spatial filter telescope ratio. The price you pay is divergence.
    I rewatched the spatial filtering video and I get this now. Thanks. Although I haven't seen a commercial projector use this technique, so my guess is the stacked beams aren't so big to not fit and maybe they don't care about the beam profiel so much. Sure, Arctos uses singlemode reds and telescopics, but I dont see anything done on the multimode blues and greens besides prisms. Same with the inside pics of Eightonlight projectors. But I will keep this in mind in case I have a situation with too big of a beam or need a cleaner beam than I get.

    Quote Originally Posted by planters View Post
    Mitsu P73s cooled and driven hard
    I'm still not sure if I understand correctly, you mean to overdrive 500mW mitsus to 2W each?


    Quote Originally Posted by planters View Post
    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.
    I do get that concept.
    And in that video you have one diode and make a vertical line (in the near field) more cubical, so you expand in the horizontal, sure. But there's one diode in the video.
    My confusion comes from this: with knife edging the beams seem to be stacked horizontal from what I see.

    Click image for larger version. 

Name:	150320103415696392.jpg 
Views:	27 
Size:	53.5 KB 
ID:	50034
    The mounts are on the same height and mirrors are offsetted in a horizontal axis. When looking from the top, the beams seem side-by-side, not on top of each other.
    If I put the beams on top of each other, wouldn't that mean I have to have a different height for each mount/beam and also mirror?
    Please explain if I understand these images wrong or if vertical stacking is a better idea.
    Before I was sure both of you thought of doing it like in the pics, Im sorry for the confusion and please bear with me.
    Last edited by neskusen; 05-24-2016 at 23:17.

  2. #22
    Join Date
    Aug 2010
    Location
    London or Spain depending on the weather
    Posts
    1,390

    Default


    They are stacking like this precisely because they are using prisms not cyl lenses !
    You will also notice that the gaps are expanded as well (not good)

    Cyl lenses:
    It is much easier to shim a diode mount and a knife edge by a fixed (well known) amount and then precisely adjust the cyl lens in a horizontal plane than to leave the diode mounts in a horizontal plane which would require you to adjust the cyl lens (on it's side) in the vertical plane. (possible but tricky)

    The efficiency of cyl lenses compared to prisms although not quite as noticeable as is suggested earlier is big enough for many of us with experience, to plump for that choice, but for someone just starting out the roughly 15% extra loss of the prisms is compensated by the ease of setup (especially if making a shitload of projectors). Worrying about small losses when you are just beginning to get a feel for the optics, is for many not worth the hassle (price of power is not that high compared to good optics) and you will probably have bigger losses due to other factors which will outweigh the prism/cyl lens difference.

    If however you really want to go the whole hog (Planter style), then good on you but he has a very large amount of grey matter, a well equipped workshop, a ton of experience and a wealth of knowledge! So don't get upset if you can't quite manage to achieve the same results.

    But if you want to get something up and running fairly quickly then prisms will help and you can make all sorts of mistakes without them becoming critical.

    Should you want a good complete (cyl lens) model to aspire to, and are ambitious enough to do the work involved then check out "Jors" (Atenlaser) thread "http://www.photonlexicon.com/forums/showthread.php/24405-8W-RGB-and-New-Modules-New-Company-AtenLaser"
    That will clear up most of your doubts if you look very very carefully at all his pics.

    I also suggest you read the Holy Grail thread which would clarify more of your doubts !


    Cheers

  3. #23
    Join Date
    May 2016
    Posts
    104

    Default

    I was avoiding replying to this thread because I noticed I was sounding rude/disrespectful which was not my intention. Sometimes text messages just aren't enough to communicate what you mean, tone and whatnot can't be put in text. On top of that your message sometimes ends up different than what was in your head. So, sorry.

    All I tried to say was I wanted to make sure I was understanding what I was being thought correctly, because it didn't feel like it. There's nothing worse than figuring out you've misunderstood what you've been thought after you have bought the wrong thing you don't need.

    And about the pics: I did check them, but it's hard for me to make up much from those pics. Too many components, plus the angle and resolution don't help, I can't tell the height of different diode mounts relative to each other.

    So I'll get back to my question and hope I can deliver it better this time.

    My question is, can't I rotate the diodes in their mounts and then jut flip the cylindrical lens 90 degrees along the beam profile? This would spare me from trying to get spacers precision milled to the beam height(s) so I can stack them vertically. Getting parts milled is not the only concern for me, an extra spacer makes worse heat transfer between the housing and baseplate, under which I want to attach TECs.
    Myabe I have the wrong idea, but it appears flipping the cylindricals will allow to just offset the mounts on the horizontal axis and get the same beam in the end, without making spacers for the diode mounts.

    It is much easier to shim a diode mount and a knife edge by a fixed (well known) amount and then precisely adjust the cyl lens in a horizontal plane than to leave the diode mounts in a horizontal plane which would require you to adjust the cyl lens (on it's side) in the vertical plane. (possible but tricky)
    Looks like you answer it here but I don't quite understand.
    adjust the cyl lens (on it's side) in the vertical plane? Why vertical plane? Maybe you mean the vertical plane of the cylindrical lens rather than the optical setup
    The beams are flipped, so the cylindricals are flipped too. What changes?
    Last edited by neskusen; 06-23-2016 at 01:17.

  4. #24
    Join Date
    Aug 2010
    Location
    London or Spain depending on the weather
    Posts
    1,390

    Default


    I think you are still a bit unclear as to the relationship between the beam shape and the cylindrical lenses, and please don't worry about me taking offense, as I have a very thick skin and forget stuff very easily.

    I will try to clarify again. (without smileys).

    The first collimating lens corrects the fast axis as soon as possible (very close to diode) to avoid the near field beam size growing to unmanageable proportions. This has the unfortunate effect of "over-correcting" the slow axis making it very narrow but also very divergent.

    The thin/narrow part of the beam (over-corrected SLOW axis) should pass thru the first cylinder lens so that it gets expanded (making it fat instead of thin/narrow) then through the second cylinder lens so that it gets re-collimated (to focus as small as possible in the far field).

    This narrow part must be perfectly centred with respect to the curvature of the lens to avoid the beam changing its path drastically and requires precise positioning.

    If you flip both the diodes and the cyl lenses as you suggest, both the initially thin/narrow part of the beam "-", and also curvature of the cyl lenses will be in the vertical plane (the wide part will be horizontal parallel to base plate).
    This means you will have to adjust the vertical position of the cyl lenses with a great deal of precision to avoid the beam being "bent" upwards or downwards. The wide part will be unaffected since the cyl lenses will have no curvature in the horizontal plane (only in the vertical plane if on their "side").

    You are quite right that this can be done,...... but it is tricky since you also have to precisely "rotate" the cyl lenses simultaneously with a precise vertical positional adjustment to avoid aberrations and change in beam path.

    Most of us (including Planters) avoid this by mount the diodes so that the wide part of the beam is vertical such that the curve in the cyl lenses is horizontal allowing them to be mounted inside a tube (to allow rotation inside a shaft mount) leaving the fine cyl lens adjustment in the horizontal plane (by sliding and if necessary twisting this mount on the base plate).
    Incidentally you can get precise thickness copper shims on fleabay for your diode mounts to avoid having to mill them yourself.

    I think you feel that the knife edging, requires more precision that the cyl lens positioning, but this is NOT the case at all. Since you will be stacking narrow on narrow to end up double as tall but equally narrow, the "clipping" or "space" where one beam joins or overlaps the other, does not represent a large proportion of the total power or size of the beams, so precision is relative to the larger dimension of the beam.

    Contrary to many peoples instinct it is more efficient to knife edge larger beams than smaller beams as the overlap or join gap losses are much smaller when compared to the overall beam sizes.

    If you use prisms to re-collimate the over-corrected slow axis (narrow part of beam), you can in theory, superimpose the images of the beams in the far field by using slightly different horizontal trajectories allowing narrow tall beams to be stacked side by side instead of one on top of the other and this seems to be what you are trying to do.

    Cylinder lenses will NOT allow this UNLESS you use different lens sets for each beam as they require the beam to be as narrow as possible, AND centered exactly in the curvature of the lens. Whereas prisms do NOT require this centering so that the same prism can be used for all the beams despite their slightly different trajectories. (They will also have slightly different angles with respect to the prisms and so will not superimpose exactly in the far field). Also with prisms the near to far field size relationship (magnification) can be decided at the last moment by adjustments, but with Cyl lenses this is ideally predetermined by the choice of lens pair at the time of purchase.

    However the cylinder lenses are more efficient, and also most people struggle much more to compensate and focus (in the far field) the re-corrected slow axis than the correctly collimated fast axis.

    Hope this helps.

    Cheers

    PS. Yes...Planters DOES overdrive the mitsus, and is able to do so due to his cooling methods. What kills the diode when over-driven is local heat on the diode itself, so the current limits vary with the temperature and are "delicate" to say the least.
    Last edited by catalanjo; 06-23-2016 at 11:37. Reason: punctuation,spelling etc.& adjustments

  5. #25
    Join Date
    Aug 2010
    Location
    London or Spain depending on the weather
    Posts
    1,390

    Default


    @ "So what is the answer to this question? ...
    I assume question ..."you mean to overdrive 500mW mitsus to 2W each?"
    No ..... I think Planters would limit this to "some people who know what they are doing could achieve this" !
    I personally would not recommend that a beginner try this, unless they have a very deep wallet.

    Read the PS. in my last post.

    Cheers

  6. #26
    Join Date
    Aug 2010
    Location
    London or Spain depending on the weather
    Posts
    1,390

    Default


    Using G2 collimators (no correction/filtering) = just under an amp each @ just under 2.5volts (these figures vary between diodes).

    For lots of pics and different currents- https://sites.google.com/site/dtrlpf...m-500mw-diodes

    Cheers
    Last edited by catalanjo; 10-21-2016 at 11:46. Reason: dislexia incremental due to age!

  7. #27
    Join Date
    Feb 2011
    Location
    New Hampshire
    Posts
    3,513

    Default

    I wouldn't want to drive them so hard but I would like to drive them up to at least 800mW each with TEC cooling....but how much current would I need to do that?

    Lets just pretend for a minute I do know what I am doing lol

    Knowing what you are doing implies wisdom and remember the old adage. "Good decisions come from wisdom and wisdom comes from bad decisions". This is an expensive hobby and if you want high performance you should not be afraid to break a few things. Diode life is a pretty high power of the driving current. It isn't even close to linear or even the square. So, if you are interested in driving these diodes harder than spec then you might want to run a few tests. Two amps will just about break them no matter how well they are cooled and this is probably due to damage to the optical cavity. One amp is safe for many hours of use. The question will be where in between these limits you choose to run.

    If you cool them, the efficiency rises very quickly. A diode that produces 500mW at room temperature will make 800 -900 mW at -35C at the same driving current and at a slightly higher driving voltage. The wavelength also blue shifts to a more visible red color and so they will look at least twice as bright.

  8. #28
    Join Date
    Aug 2010
    Location
    London or Spain depending on the weather
    Posts
    1,390

    Default


    Bench supplies allow you to limit current directly, so the voltage is sorted "automatically" regardless of temperature/voltage of diode,.... same as with your driver.

    Only problem you might have, is if like Maltes, your driver starts getting a bit warm due to large voltage drop across it.

    Cheers
    Last edited by catalanjo; 10-22-2016 at 10:47.

  9. #29
    Join Date
    Feb 2011
    Location
    New Hampshire
    Posts
    3,513

    Default

    I have a question for you....the driver that Im using..I know it controls current but if I dropped the temperature to get 800mw's out of them..will the driver supply the right voltage when I do this? Or would I need a bench power supply to control the voltage coming from the driver?
    The driver controls the current and the laser diode itself actually controls the voltage. The driver will limit the current and the voltage drop across the diode determines how much voltage that the driver will have to supply. Drivers typically have an internal voltage drop of 2-3 volts and the spec sheet should identify this drop. As long as the power supply provides at least the sum of the driver drop and the laser drop then you will have no problem. The reason that you might want to limit the supply to not much more than this total is that the additional voltage beyond what is needed will produce additional heat for the driver to dissipate. This is a fine point however, because the increase in voltage drop with cooling is on the order of 0.5V even with significant cooling. So, as long as your power supply exceeds the needs at room temperature by a couple of volts, then you should be fine.

  10. #30
    Join Date
    Aug 2016
    Location
    Minsk, Belarus
    Posts
    67

    Default

    Here's a question I think very relevant to the topic title: isn't one fundamental difference between cylindrical lenses and prism pairs for doing knife edging the fact that with prisms you can stack the "tall" but horizontally thin beams horizontally and have them expand equally while with cylindricals since the beam has to be in the exact center of the lens you have to expand these tall beams stacked vertically? With the latter I don't see how you could knife edge any of the multimode 500mW+ diodes without hitting the scanner mirror size limit with more than two diodes. The blues and greens are about 3mm tall. Stack two of these on top and you already need a big 6mm aperture scanner. I don't see how you could go higher.
    I understand with prisms the losses are much more but isn't this the reason that the commercial several diode modules go with prisms, not just the time saved on calibration with prisms?

Posting Permissions

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