Page 1 of 4 1234 LastLast
Results 1 to 10 of 37

Thread: A few "almost there" shots of the new board.

  1. #1
    Join Date
    Jun 2005
    Location
    SoCal
    Posts
    490

    Default A few "almost there" shots of the new board.

    So after months and months of toiling, we've almost started...

    Some of you may have known I've been poking at building my own
    16-bit board from scratch... I've been a long-time pangolin user,
    but I don't like where Pangolin is heading with the new "improvements"
    in the new QM2000 board... I figured that by the time their system
    was able to do what I wanted to do, I'd have to wait a few more
    generations.

    But after we finished building the prototype of the new 16-bit board
    that I unveiled here on PL, we've inadvertently started doing laser shows
    again. We'e working with some popular artists here in LA/Hollywood
    and I haven't been getting much free time or sleep.


    A 2W DPSS based whitelight

    The latest being a show at the MGM Grand in Las Vegas in January...
    We'll be bringing 4 - 40W Yags w/ 18k XYZ scanners, 1 - 15W JenLas
    white light w/ 2 sets of cambridges, and that 2W white on a cambridge
    you see above... All controlled w/ our custom board...

    Kiva just finished the PCB design for the new board with some improvements
    we though of while running some shows we did earlier this year.



    So on Tuesday we sent a small run off to the PCB fab (A&C in Los Angeles,
    who btw did an incredible job on the board) who actually got them done
    in less than 24 hours!

    So without much futher ado...





    Yesterday, these boards got dropped off at the assemblers so I'll see
    them again on Tuesday, all populated with chips...

    Anyways, back to my scheduled programming.

  2. #2
    Join Date
    Feb 2005
    Location
    Florida
    Posts
    802

    Default

    Nice Yadda, Cant wait to get mine !!!
    "My signature has been taken, so Insert another here"
    http://repairfaq.ece.drexel.edu/sam/laserfaq.htm
    *^_^* aka PhiloUHF

  3. #3
    Join Date
    Jul 2005
    Location
    UK
    Posts
    249

    Default !!!

    Sign me up....
    Regards,

    Aidan

  4. #4
    Join Date
    Jan 2005
    Location
    Norway
    Posts
    1,301

    Default

    hmm..

    is there any 16 bit board ready for selling ?
    I`m ready to upgrade my laser-scanner....

    But maybe a 16bit will be quite expencive ???





  5. #5
    Join Date
    Jun 2005
    Location
    SoCal
    Posts
    490

    Default An update... the boards are back!

    We got the boards back from the assemblers!



    They came back very nice! Here's one posing in from of my custom scanhead...



    The boards for flashing the unit aren't finished yet, so we had to rig a
    quick job on phenolic to flash the SX micro on it...



    Kiva flashed the board with the debug microcode... and we ran some diags...
    Then discovered a bug!!! Whoops! The surface mount version of 74x138's
    we used had two pins flipped from the through hole version we
    had tested the board with! Darn manufacturers!!! Unfortunately
    the traces ran under the chip, so we did some creative rerouting...



    The next time it passed the diag tests, but we discovered that the
    ungrounded inputs on the SMT 7404's caused significant RF noise
    in the circuit! So we fixed that too...



    Now everything passed the diags with flying colors, so we flashed the
    latest microcode onto the board and...



    Perfect!!! The performance exceeded our highest expecations!
    We zoomed in all the way on the scope and got PERFECT traces!


  6. #6
    Join Date
    Jun 2005
    Location
    SoCal
    Posts
    490

    Default

    Hi liteglow! Not for sale yet... I am still planning on one more
    pre-production run, and one final verification run before possibly
    releasing the real thing... There's still a few more features I
    want to squeeze in there.

    This isn't the buy/sell forum so I shouldn't be posting numbers,
    but since you can't buy them at any price right now, I'll slip up and
    give some info out. We're not planning doing even mid-volume
    manufacturing so the cost of production is pretty foul...

    Tempering the parts and production costs, is the fact that we all have
    very nice day jobs so profit isn't really a motive... If we can get some
    better deals on manufacturing costs, the prices will definitely come down...
    Maybe if enough people want the board, we can do larger runs...

    We built this laser board for ourselves to be able to do things that
    other professionals told us was "impossible", but since we do far
    scarier work during the day at our regular jobs, we chose to disbelieve
    and made a few proofs of concept... they worked so well, we made
    a "professional" board for ourselves and now we starting to figure
    other folks might want to use it to make their ideas possible as well...

    We're planning on releasing in March and unless we can get the
    costs of manufacturing down, we're estimating $1850 for a retail
    package, and $1250 for an OEM version... Established PL members
    will be eligible for a discount so start some rousing discussions
    already

  7. #7
    Join Date
    Jan 2005
    Location
    Norway
    Posts
    1,301

    Default

    Hi Yadda !

    I`m scary impressed over your work !!!!!
    Damn did work at NASA or something like that ?

    I understand while i read your post that the 16 bit board will be expencive.
    As i understand that this is NOT a commond thing to use

    If I get it right, I can still use my 8-bit with better scanners and be happy with that
    And maybe someday in the future I can plan a visit to your house and se the amasing work you have done maybe that will be cheaper for me haha...


    If I maybe someday get my hands in a powerful RGP (dpss) laser and maybe rent it out, i can earn some money and maybe buy pangolin and better stuff

    Since I live in Norway i have N E V E R in my live seen a LIVE lasershow
    In Norway i think actually there is ony ONE (or 2) peoble that OWN a Full Colour laser for "show use"

    Maybe I will be the next person who knows...

  8. #8
    Join Date
    Jun 2005
    Location
    SoCal
    Posts
    490

    Default

    Heh... thanks for the compliments!

    Actually that's one of the things we're thinking about... we want to
    make low-cost boards and that's one of our goals, but this was
    more like scratching a really bad itch I've had for the past 10 years...

    The bulk of the cost is the components, and if one designs a board
    with less "accurate" components (to say only 0.0006V accuracy)
    then prices come dramatically down!

    We just got sidetracked/distracted and designed this like it was
    a medical/scientific component instead of entertainment... mostly
    since that was a lot more fun...

    8-bit is actually a fine accuracy for most shows! I wouldn't feel the
    least bit ashamed of running a full show with a decent 8-bit board...

  9. #9
    Join Date
    Nov 2005
    Location
    Massachusetts Institute of Technology
    Posts
    80

    Default

    'We built this laser board for ourselves to be able to do things that
    other professionals told us was "impossible", but since we do far
    scarier work during the day at our regular jobs, we chose to disbelieve
    and made a few proofs of concept... "

    What can you do with this board that you can't otherwise?

  10. #10
    Join Date
    Jun 2005
    Location
    SoCal
    Posts
    490

    Default

    Hi eric! Just got back from a New Years event!

    I'm planning on doing a full writeup whenever I get some time,
    the easiest way to explain is that I'm a long-time satisfied pangolin
    user who is intimately familiar with their product line and the
    workarounds one has to implement to make their product do what
    you really want it to... But every so often, you will run into
    a brick wall (usually client instigated) which is officially titled
    "The Pangolin Way" which is what their support folks will utter if you
    try to do something that pangolin doesn't handle or at the very
    least isn't very graceful about...

    The biggest reason I'm going through a lot of work to make my own
    gear is because I suspect the laser industry is heading in the wrong direction.

    I don't want a better video projector, I already have very nice video projectors.

    I don't want a better intelligent light, I also have very nice intelligent lights...

    I want my laser shows to show off to the strengths of the actual laser itself.
    Unfortunately, as software and hardware "improve" it's getting surprisingly
    hard to do that...

    In any event, to prevent this becoming an decade long writeup, especially
    since I want to go to sleep soon, I'll give one example and then shut up.
    This example was chosen because it shows how a different hardware design
    can actually liberate software... so...

    I guess the thing thats always bothered me the most about modern
    laser projection systems is how they handle color and color modulation...

    This is a two part problem. One deals with color space and
    the other deals with "point based color lookahead" which is what
    professional packages use to compensate for the disparities
    in the color modulation speed vs scanner speed.

    The color gamut on laser shows is actually being held back by the computers
    that are being used to run the shows... Computers present color data
    using Red Green & Blue, so it was quite natural for systems engineers to
    take the existing standard and apply it to the laser systems. Of course
    everyone was aware that most non-dpss lasers generate more than those
    3 lines of light... The solution which was implemented provides
    RGB to multiple color line (eg. 8-channel) color conversion...
    The easiest way using simple linear math on a CIE chart.



    For example, Pangolin supports full 8-bit modulation on all 8 color channels,
    however, if you examine their file formats, you'll note that the color
    is stored as an 24-bit value. If you do the math (8x8bit = 64-bit), you'll
    note that it doesn't add up... However, on a computer you'd be hard pressed
    to see the difference between 16 and 24-bit, much less 24-bit and 64-bit color.

    The problem arises from engineers applying the same "rules of thumb" of computer
    color depth to lasers... because lasers use fully saturated "alternate" colors lines.
    It's unfortunately really easy to see the difference between the RGB approximation
    of 488nm vs the true argon 488nm laser line... In fact, it is 100% impossible
    to display a fully saturated alternate (non RGB) color using RGB. This is
    why viewing the CIE chart on a computer monitor shows some interesting color
    banding. printed out in real life, the colors should be a smooth gradient.
    This problem carries through to the RGB to 8-channel color conversion...
    Acknowledging this problem laserists have implemented "boosting" the colors
    such that 100% green + 100% blue = 100% cyan... This is shifting the problem
    because the CIE chart displayed on this system would look pretty bad! A
    computer showing an image using the same color system would show pictures
    that look surprisingly like old faded color photographs taken with polaroids...
    Using some correction algorithm would result in an actual "usable" color depth
    closer to 16-bit approximation on an RGB monitor!!!

    Gah! All these compromises! TOO MUCH THINKING!

    Well, thats what we thought, so we just said... lets store all 8 color channels
    as 8-bit values in both software and hardware. Hardware engineering wise,
    it was slightly painful due to the bandwidth requirements, but we did it and
    it works...

    So now, with the new board, you can decide how you want to address the colors,
    we have several different algorithms ranging from manual control of each color
    at any given point in time to the exact same RGB-to-N-channel conversion that
    everybody else uses... My personal favorite mode is color boost mode... since
    I really like beam shows... What this does is fully activates say 488nm "cyan"
    then also activates the red and blue lines to approximate cyan... this makes
    really bright colors that pop out.

    OOH... I was going to go into the second part, but I'll just skim it...

    The second part is color modulation... basically professional systems use
    per point lookahead to handle the disparities between the color modulation
    vs scanner speed.

    The problem is two fold inside here as well, but simply put, the actual
    laser modulation doesn't usually fall exactly on a point and on many
    color modulation technologies, each color has a slightly different modulation
    speed... and the modulation speed (slew rate) is dependant on time, however
    per point based modulation lookahead means that you will have to tune for
    each and every scan-speed you look at!!!

    Basically, the standard industry solution involves a hell of a lot of
    "anchor" points which "pauses" the laser at a point long enough for the color
    to be reliably considered changed.

    In my opinion, that really blows... why buy 30+kpps scanners when you're going
    to be spending most of your time waiting for color to catch up?

    Simply put, we fixed this in hardware, throw a test pattern up, then tune each
    color so that the rise and fall of the colors is precisely matched up...

    Then if your system was tuned correctly, you could even use ZERO anchor points
    for color changes!!!

    The second issue (which is retuning color lookahead based scanner speed),
    isn't considered a problem by most laserists since they choose a scanrate, then
    convert all their frames to their scanrate...

    Well another feature we added to this board causes it to raise its ugly head...
    we implemented variable scanspeeds within a given frame. This lets us just
    simply stop thinking and throw images generated by other laserists into a show
    and just tune the scanrate for each individual "laser object" inside the software.
    And we just fixed the second issue using hardware as well.

    hmm I overshot my time limit. Sigh... G'nite and Happy new years!

Posting Permissions

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