Results 1 to 10 of 10

Thread: Using RGB projector for R&D with homebrew ILDA controller - failure modes?

  1. #1
    Join Date
    Jun 2017
    Posts
    4

    Default Using RGB projector for R&D with homebrew ILDA controller - failure modes?

    Hello everyone,

    I am working in a university lab, where we do computer vision, machine vision, a bit of consulting, lots of research and some development.

    Since we needed a flexible tool for experimenting with laser illumination, we bought a RGB laser scanner projector with ILDA input and so far, it turned out to be pretty awesome toy. (Yes, we use goggles and we put big warning signs on the door). The laser is Ibiza 1000, http://www.lajbi.com/ibiza1000-1w-rgb-laser-ibiza-light 1 W total, 20 kpps, the channels are TTL. Laser proudly declares to be made in China.

    We drive the ILDA input using the modified 5.1 channel USB audio card (output capacitors removed so it behaves like a plain DC DAC) and a custom in-house-build hardware that transforms the DAC outputs to differential bipolar (X,Y) and unipolar (R,G,B) signals according to ILDA interface standard. We generate signals to drive the galvos and laser outputs using Octave (a Matlab-like) scientific application.

    Here comes the question:

    Laser has a warning "To protect the scanner system set the speed to 20 kpps in the scanning software, such as Pangolin, before connecting the ILDA input". But, alas, we do not use proper laser software, so no setting the scanning speed.

    Given that is nearly impossible to calculate allowed signal bandwith (Khz) from the stated 20kpps capability I have the following questions:

    1) What are the failure modes if we overdrive the scanners with the frequency that is too high? I have a degree in electrical engineering, and can understand that applying 20Khz signal to a motor that can handle 100Hz will result in heating of the coils. Is that the only problem we are facing?
    2) What is the telltale sign that the scanners are overdriven? Is the fact that the pattern looks ok (as designed) a robust indicator that we are not overdriving the scanners? Or can damage occur even if galvos follow the pattern without trouble?
    3) Is there an (open, published, public...) algorithm anywhere that verifies the point-stream for compliance with the set kpps? After all, professional lasershow software must have the means to verify the data before generating signals, so we could do it too.

    Thanks,

    Janez Pers
    Faculty of Electrical Engineering
    University of Ljubljana
    https://vision.fe.uni-lj.si/people/JanezP.html

  2. #2
    Join Date
    Mar 2012
    Location
    Akron, Ohio USA
    Posts
    2,197

    Default

    Check out LaserBoy.

    You are already most of the way there. You have a modified sound card DAC.

    That (among other things) is what LaserBoy was made for.

    It takes (or makes) ILDA frames and makes wave files that you can play through that DAC.

    It has all of the optimization stuff built in and the default values work very well at 48KHz DAC into 20KPPS scanners.

    You can control the velocity of the scanners and add dwell in the corners of your art.

    It also has the benefit of being free and open source.

    Plus I am here to answer all of your questions about it.

    James.
    Last edited by james; 06-30-2017 at 10:41.
    Creator of LaserBoy!
    LaserBoy is free and runs in Windows, MacOS and Linux (including Raspberry Pi!).
    Download LaserBoy!
    YouTube Tutorials
    Ask me about my LaserBoy Correction Amp Kit for sale!
    All software has a learning curve usually proportional to its capabilities and unique features. Pointing with a mouse is in no way easier than tapping a key.

  3. #3
    Join Date
    Jun 2017
    Posts
    4

    Default

    Quote Originally Posted by james View Post
    Check out LaserBoy.

    You are already most of the way there. You have a modified sound card DAC.

    That (among other things) is what LaserBoy was made for.

    It takes (or makes) ILDA frames and makes wave files that you can play through that DAC.

    It has all of the optimization stuff built in and the default values work very well at 48KHz DAC into 20KPPS scanners.
    Oh wow, thank you. And GPL! I will certainly look into it.

    Janez.

  4. #4
    mixedgas's Avatar
    mixedgas is offline Creaky Old Award Winning Bastard Technologist
    Infinitus Excellentia Ion Laser Dominatus
    Join Date
    May 2007
    Location
    A lab with some dripping water on the floor.
    Posts
    9,890

    Default

    20K/30Kpps = N /2500 = around 1.667 KHz small signal bandwidth while scanning the ILDA 30K test pattern at eight degrees. . Some of us who experiment with these things have determined the small signal bandwidth with 30K galvos at 8 degrees optical angle to be about 2.5 KHz for small angle jumps, based on the circle in the square in the ILDA test pattern. Some engineers who design galvos have quietly confirmed that for the 1990s technology that is in most Chinese copies, 2.5 KHz is a good number at small angles. Eight degrees is the ILDA specified test angle for the main test pattern.
    ~
    Scanner bandwidth falls off rapidly as scan angle increases. The wider the angle, the slower the scanners, and the model is not simple. Its also dependent a bit on the previous waveform thru the scanner.
    ~
    On a modern system, the cloned galvos use coils are about three ohms, and are just a few turns of magnet wire within say 100 microns of the permanent magnet rotor.
    The main danger is driving the coils too hard, the wire expands, the rotor scrapes the coils, and usually the rotor is grounded thru the bearings, unless you have a good scanner with ceramic bearings. The classical commercial amps that your clones are copied from had an log-antilog multiplier chip as a "coil temperature calculator", that drove a fet to reduce the gain of the scanner amplifier if the user pushed the input signal hard. That was the first thing the makers of low cost galvos removed from the circuit board on the copies. That low cost chip is was out of production by the mid 2000s anyways. T
    ~
    The good news is test point two on the amplifiers is nearly always the high end of a 0.1 Ohm resistor that goes to ground and carries the coil current as part of the feedback loop. You can monitor the coil current there with an oscilloscope. You can also look at the tail end of the differential amp that processes the optical feedback signals.
    ~
    The second issue is resonance frequencies. trying to "push" the rotor hard results in vibrations trying to tear the shaft apart or out of the bearings.
    The galvos will visibly protest in the projected image and you'll hear clicking and oscillation at the base of the mirror as you get close to destruction.
    You can watch for that in the feedback signal, but you would have to qualify that by testing a scan pair to destruction for a given model.
    ~
    With the low cost, low performance clones of the Cambridge 6800 made with inferior magnetics (hence the 0.707 times the performance of the Prototype at 20K)
    It would be cheaper to just keep a spare pair of 100$ galvos and amps laying around then to spend graduate student and professor time working on a model.
    ~
    The ILDA scanner tuning standard, was designed by a team of artists and engineers to insure the PII or PID loops on the driver amplifiers from any two given users were tuned the same, so that projected graphics looked the same when interchanged by users. It is a more a data transfer standard based on a test pattern's visible appearance then a measure of the galvo's performance. The whole driver is the X-Y waveform's repetition rate. It was never envisioned as a means to make a very specific measurement of galvo performance. There are 12K, 20K, 30K, 60K, and 90K galvos in the world.
    ~
    The team that designed the ILDA Test Pattern used to measure "PPS" never intended the pattern to make detailed measurements, and no engineering relationships have ever been published for it. It was based off a hybrid of test patterns used by a few laser show companies at the founding of ILDA. its designed so that an artist who is not an engineer can tune his or her galvos to a common image performance standard, not quantify them as an engineer would. An engineer would look at small and large angle jump time, and settling time.
    ~
    On classical laser show software, we scan the pattern at a known angle, and vary the waveform update rate, which is calibrated in DAC updates per second on older software, while modern software calculates an optimized vector. The "circle in the square" potion of the pattern is designed to be a rough measure of 3 dB rolloff, everything else is for measuring overshoot or undershoot. So you change the scan rate while projecting the standard pattern at the specified angle until the sides of the circle just touch the inside walls of the square, and when they just touch, you look at the PPS readout in software and you know the "speed" your tuned for. It is very useful for laser shows when the galvos are driven in their ballistic region, it does not directly translate to a specified engineering performance measure.
    ~
    Think of it as a TV test pattern for setting up a broadcast camera. It really just ensures your complying with a loosely defined set of conditions. It lets a non-engineer with no test gear qualify the performance of their hardware, in respect to image interchange in a laser show vector graphics environment.
    ~
    The ILDA test pattern is a series of X-Y pairs. Each time you update one pair of data, you have sent the signal to the scanner to update one point. If you do 1000 updates in one second, you have scanned 1 KPPS.
    ~

    I can send you a private message with links to papers on galvo performance in a imaging environment, however there is not a well known set of conditions to avoid destruction. There are plenty of papers on optimizing the drive waveform, many by Florian Duma. Some early papers by Pierre Brosens go into the mechanical limits on speed.
    ~
    So:

    1. Do expect the galvo bandwidth to decrease with increasing scan angle.
    2. Avoid shaft resonances.
    3. The issue is waveform and mirror inertia dependent.
    4. Galvo manufactures specify a small angle seek time and a large angle seek time on their datasheets, and some will provide you with a frequency vs scan angle roll-off curve, but only if asked. Try not to force the galvo to exceed the seek times.
    5. There is no easy answer to your problem.
    6. Stator housing temperature matters, try to keep your galvos cool.
    7. Galvo coils are a thin wire with an upper instantaneous current limit.
    8. Listen and watch for distortion in the image as a clue to galvo stress.
    9. Monitor your test points for waveform compliance with an oscilloscope.
    10. Don't push the coil so hard as to partially demagnetize the rotor.
    11. For a given set of Chinese copies with a given mirror size, you'd have to test to destruction to learn, the model is not published.
    12. IF you need really high performance, buy better scanner amplifiers with notch filters in the feedback loops to remove the resonances.
    13. Some galvo datasheets have a maximum specified RMS current, but not the clones/copies.
    ~
    If your an EE, you'll learn much by looking at the waveforms. If your interested, send me a PM, and I can send you some info on what to look for at the test points. I can't host the PDF here, it is copyrighted information by a major manufacturer of Galvos.

    Steve
    Last edited by mixedgas; 06-30-2017 at 10:47.
    Qui habet Christos, habet Vitam!
    I should have rented the space under my name for advertising.
    When I still could have...

  5. #5
    Join Date
    Mar 2012
    Location
    Akron, Ohio USA
    Posts
    2,197

    Default

    If you are already making wave files, LaserBoy can open them.

    It can show them to you on your computer monitor in real time and it can open them as a frame set for closer analysis.
    Creator of LaserBoy!
    LaserBoy is free and runs in Windows, MacOS and Linux (including Raspberry Pi!).
    Download LaserBoy!
    YouTube Tutorials
    Ask me about my LaserBoy Correction Amp Kit for sale!
    All software has a learning curve usually proportional to its capabilities and unique features. Pointing with a mouse is in no way easier than tapping a key.

  6. #6
    Join Date
    Jun 2017
    Posts
    4

    Default

    Quote Originally Posted by mixedgas View Post
    1. Do expect the galvo bandwidth to decrease with increasing scan angle.
    2. Avoid shaft resonances.
    3. The issue is waveform and mirror inertia dependent.
    4. Galvo manufactures specify a small angle seek time and a large angle seek time on their datasheets, and some will provide you with a frequency vas scan angle roll-off curve, but only if asked.
    5. There is no easy answer to your problem.
    6. Stator housing temperature matters, try to keep your galvos cool.
    7. Galvo coils are a thin wire with an upper instantaneous current limit.
    8. Listen and watch for distortion in the image as a clue to galvo stress.
    9. Monitor your test points for waveform compliance with an oscilloscope.
    10. Don't push the coil so hard as to partially demagnetize the rotor.
    11. For a given set of Chinese copies with a given mirror size, you'd have to test to destruction to learn, the model is not published.
    12. IF you need really high performance, buy better scanner amplifiers with notch filters in the feedback loops.

    If your an EE, you'll learn much by looking at the waveforms.

    Steve
    This is excellent, and gives plenty of info, thanks. It is true, it is not worth my time to study scanner models in detail, so I was after a simple solution. I was instinctively checking for a few things you mention. And yes, we are focused to small paterns (so far). Definitely I am on the alert for any kind of galvo noise and any amount of distortion. I also created my test patterns to check whether galvos are keeping up with the patterns I am projecting. And you are right, we have all the equipment that is needed to analyze the signals, but I hoped there will not be a need to go into such detail. Basically I don't want the unit to crash and burn 2 days before some important test or demo, that's all

    Temperature probes on galvos is an excellent idea. If it is not too much trouble, I would appreciate the (links to) publications on galvo performance.

    Regards,

    Janez.

  7. #7
    mixedgas's Avatar
    mixedgas is offline Creaky Old Award Winning Bastard Technologist
    Infinitus Excellentia Ion Laser Dominatus
    Join Date
    May 2007
    Location
    A lab with some dripping water on the floor.
    Posts
    9,890

    Default

    PM will be sent this evening when I get back from a lab session.

    Steve
    Qui habet Christos, habet Vitam!
    I should have rented the space under my name for advertising.
    When I still could have...

  8. #8
    Join Date
    Mar 2012
    Location
    Akron, Ohio USA
    Posts
    2,197

    Default

    Do you mind if I ask what sound device did you use to make your DAC?
    Creator of LaserBoy!
    LaserBoy is free and runs in Windows, MacOS and Linux (including Raspberry Pi!).
    Download LaserBoy!
    YouTube Tutorials
    Ask me about my LaserBoy Correction Amp Kit for sale!
    All software has a learning curve usually proportional to its capabilities and unique features. Pointing with a mouse is in no way easier than tapping a key.

  9. #9
    Join Date
    Jun 2017
    Posts
    4

    Default

    Quote Originally Posted by james View Post
    Do you mind if I ask what sound device did you use to make your DAC?
    This one:

    https://www.amazon.de/dp/B005M19I0K/

    Found somewhere on the net the instructions how to remove capacitors for the similarly looking box. This one has nice large through-hole electrolytic capacitors that are trivial to solder off and substitute with resistors. And it is dirt cheap

    Janez.

  10. #10
    Join Date
    Mar 2012
    Location
    Akron, Ohio USA
    Posts
    2,197

    Default

    This one:

    https://www.sabrent.com/product/USB-...igital-output/

    is nice too. It also used a C-Media chip that is actually the updated version of the one you have.

    It is fully 8 channel so you have room for stereo audio tracks in 7 & 8.

    I've made countless DACs for people all over the world with this device.

    If you watch eBay you can get it for less than $20.

    I had a PC board made and I sell a very nice kit for the Correction Amp that works with just about any computer sound device.

    http://laserboy.org/forum/index.php?topic=561.0

    James.
    Last edited by james; 07-01-2017 at 09:10.
    Creator of LaserBoy!
    LaserBoy is free and runs in Windows, MacOS and Linux (including Raspberry Pi!).
    Download LaserBoy!
    YouTube Tutorials
    Ask me about my LaserBoy Correction Amp Kit for sale!
    All software has a learning curve usually proportional to its capabilities and unique features. Pointing with a mouse is in no way easier than tapping a key.

Posting Permissions

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