Page 3 of 3 FirstFirst 123
Results 21 to 29 of 29

Thread: X/Y Audio Laser Control - Like an oscilloscope

  1. #21
    Join Date
    Feb 2014
    Posts
    14

    Default

    Yes please!

    000000

  2. #22
    Join Date
    Mar 2006
    Posts
    2,478

    Default

    Quote Originally Posted by adoxo View Post
    I can't find any information on protecting the galvos from overload. Can anyone link me to a relevant thread or share some advice?

    I can apply a low-pass filter in my audio software but still won't feel totally safe…
    Bandwidth limiting the signal will keep the galvos in small step response because it's like using an interpolator, it avoids instant jumps. By the way, try a slew rate generator, it may be a better match as well as being more efficient than a filter.

    The inner square and circle of the ILDA pattern test for best combination of scan speed and angle (one must decrease as the other increases). You can simulate these in PCM output as follows: Draw the circle with sines in quadrature, each with frequency 1/12 of your sample rate, and use the sample rate itself to set speed, with a starting value in KHz equal to the nominal point rate (kpps) of your scanners. (Layla 24 will allow sample rate setting from 2000 to 100000 in 1Hz increments if you use the WDM driver). The draw speed of the square is not critical (just make it slower than the sine waves), but the size is; make its two quadrature squarewaves have amplitude exactly 3dB LOWER than the amplitude of the circle's sinewaves. In this case, having set the speed by sample rate, start with low amplitude scan and increase angle cautiously until the circle sits neatly in the square, touching its sides, as it should in the ILDA pattern. This square sets your scan angle, While it is narrower than the outer square of the ILDA pattern, this is because you're not using large step response so you can't make fast uninterpolated jumps. This is not as limiting as it seems, you're trading speed for accuracy and linearity already, this is just part of that.

    Hold out for Pangolin's Scannermax 506 scanners when they get their full public release, because they have an unusually large scan angle before leaving small step response, so are an ideal match for this situation, as well as likely being more accurate than most other scanners, and likely cheaper too.

    EDIT:
    A buffer circuit won't help. I saw this mentioned in posts that weren't there when I started this one...
    This is because you won't know for sure what the limits are, it will depend on your bandwidth limiter, scan speed, and scan angle. Worse yet, if you try clamping the voltages, you'll cause hard nonlinear distortion, and likely force the scanners into large step responses by doing that. The intention that aimed at protecting them might harm them if you do that. It's better to derive a test wave from the small-step test in the ILDA pattern like I described above.

    More edit:
    It may be possible to use a dual op-amp with its slew rate deliberately slowed to match the fastest slope of a 1KHz sinewave at amplitude set for your maximum scan angle. This is crude, but if you put that in your analog signal path it will probably help, but designing the circuit and choosing the op-amp for correct action might not be as easy as it looks. It's better to understand the limiting combinations of scan speed and angle, and the small step response test pattern, so you can figure out whatever safe combination you want.
    Last edited by The_Doctor; 02-22-2014 at 07:51.

  3. #23
    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

    Quote Originally Posted by adoxo View Post
    I can't find any information on protecting the galvos from overload. Can anyone link me to a relevant thread or share some advice?

    I can apply a low-pass filter in my audio software but still won't feel totally safe…
    Don't worry about it. Almost no one does. DC accurate 3.5 Khz low pass filters are hard to build. The oversize
    limiting can be done with a master size potentiometer or a clipping zener.

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

  4. #24
    Join Date
    Mar 2006
    Posts
    2,478

    Default

    Quote Originally Posted by mixedgas View Post
    Don't worry about it. Almost no one does. DC accurate 3.5 Khz low pass filters are hard to build. The oversize
    limiting can be done with a master size potentiometer or a clipping zener.
    My post was a lot longer, but there was a reason for that.

    Master size pot still leaves a person wondering how far they can push it. Clipping may leave a person pushing too much if they trust that alone to protect them, resulting in a full scale high slew rate signal that is damaging for galvos the same way it would be damaging for tweeters in loudspeakers.

    It's tempting to want a simple answer, but I remember well that when I wanted one in the past you were often the first person to tell me there wasn't one.

    As most audio for galvo control is easily bandlimited in the system that generates it, it means the signal is like interpolated jumps, so always small step repsonse, right? Bill's talked about this, you've talked about it, I can't beleive it stops being true now, but if you drive the signal into a zener and clip it hard, that does change to large step reponse, and it won't be either useful or desirable when using audio generators as a source. So if it stays in small step response, the best way is to test for that the same way the ILDA pattern does, because in that part of the test the kpps is directly limited to frequency, which is appropriate for audio generators.

  5. #25
    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

    Let it go.... There is little reason to be concerned at this stage.

    It is very hard to kill a galvo, and they will visually and audibly protest long before you get to that point. In most cases of excessive frequency they will just roll off, unless the tuning is very wrong. With grossly excessive damping in the tuning, you might get some severe oscillation at with large, full field, square waves. It is very, very, unlikely the O.P.s galvo amps shipped with that level of excessive damping.

    If the O.P. configures his master size potentiometer and maximum source levels correctly, he will be hard pressed to drive a galvo amp into clipping or oscillation.
    He will quickly find out that driving the galvo full field diminishes the artistic product anyways, and he'll see his images "square up" and develop barrel distortion if he approaches the limits. If he sees the image "bloom" over a very narrow range with certain frequencies, he may be setting on a torsional resonance. Those frequencies are very high for most galvos.

    Watching the screen, and listening to the scanners, it becomes readily apparent when things are under strain.

    All he needs to do at first is to keep the image size down. If his galvo mount is well heatsinked, if his amplifier is well heatsinked, and his scan angle is limited to 30-40% of full mechanical field, he's going to be fine. The easiest way to deal with this is the image size potentiometers, often a simpe 10K or 100 dual pot is all that is needed per axis.
    On better amps he could turn the image size pots down as well.

    I could come up with a high frequency sine or square waveform that will quickly heat the rotor and stator, but the waveform has little artistic value. The resulting fine filigreed shape is not pleasing to the western eye, anyways.

    The purpose of the Zener clamp is to protect the input circuits of the amp from high end audio sources. 5V, 1 watt Zeners placed back to back with a 100 ohm series limiting resistor is fine for this purpose.

    Steve
    Last edited by mixedgas; 02-23-2014 at 20:52.
    Qui habet Christos, habet Vitam!
    I should have rented the space under my name for advertising.
    When I still could have...

  6. #26
    Join Date
    Mar 2006
    Posts
    2,478

    Default

    Quote Originally Posted by mixedgas View Post
    Watching the screen, and listening to the scanners, it becomes readily apparent when things are under strain.
    Fair enough, That's how I learned it in practise too. So long as people do pay attention to the change in sound from nice noise to nasty, and watch the first signs of visual disturbance I think they'll be ok. If clipping were placed at the right level then it might even enhance this early warning rather than cause real trouble. Just got to heed the warning.

    EDIT:
    One small point I think worth adding, because post 19 mentioned coding... It may be worth making a slew rate limiter, all it has to do is something like: if change in position this sample is greater than N, change by N (where N is set at the step size for fastest small step response for the galvos). Then output this value while remembering it to compare with the intended value next cycle. Much easier to do than a filter, and uses very little CPU. If this is done, the software can then throw any transition it likes with little fear of overload.
    Last edited by The_Doctor; 02-23-2014 at 21:11.

  7. #27
    Join Date
    Feb 2014
    Posts
    14

    Default

    Thanks for your input everyone. Lots of detailed info here, mostly over my head but I'll certainly take it all and chat it over with my colleagues.

    I owe you all one!

  8. #28
    Join Date
    Oct 2013
    Location
    New York City
    Posts
    67

    Default

    Let me see if I have this straight and I don't mean to be caviler in my simplifying...

    1. If one is going to try and connect an audio source, let's say a VCO to a projectors galvos, the frequency range is most important. 2. Whether one is going through a sound card DAC or ILDA connector having some Zener diodes for a bit of padding is a good idea but not necessary. 3. In most likely case scenarios there will be visual signs that the signal is too much long before they kill the galvo. -I am assuming that this is true for slower ones as well.

    Is there a "safe" range? for LFO's and VCO's?

  9. #29
    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

    Quote Originally Posted by Revmutt View Post
    Let me see if I have this straight and I don't mean to be caviler in my simplifying...

    1. If one is going to try and connect an audio source, let's say a VCO to a projectors galvos, the frequency range is most important. 2. Whether one is going through a sound card DAC or ILDA connector having some Zener diodes for a bit of padding is a good idea but not necessary. 3. In most likely case scenarios there will be visual signs that the signal is too much long before they kill the galvo. -I am assuming that this is true for slower ones as well.

    Is there a "safe" range? for LFO's and VCO's?
    Do not exceed +/- 5V on a single ended input and conservatively avoid frequencies above 2.5 KHz small signal on the galvo inputs. The 30K galvos will have a few resonances and will roll off like a low pass starting at about 1.5 to 2 Khz depending on configuration. They do not have a transfer function that is exactly low pass, its very complicated.

    A back to back 5V Zener pair on the input is a good idea if your using VCAs etc and do not have 100% control over your synth.

    "Forcing" the galvos to respond to higher frequencies by increasing the input signal is to be avoided.

    Running at known resonances is very bad.

    You will find out there is severe trade off between bandwidth and angle. At small angles you will have greater bandwidth. At full scan field what you can do will be maxing out at very low frequencies.

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

Posting Permissions

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