Shop OBEX P1 Docs P2 Docs Learn Events
LDI, PCBs, And The Propeller (New Photos Uploaded 02/19/2019 - 8:05 PM) - Page 3 — Parallax Forums

LDI, PCBs, And The Propeller (New Photos Uploaded 02/19/2019 - 8:05 PM)

135678

Comments

  • jmg
    I'm talking about correctly exposing with either one-pass, or two. ie tolerating one unit, or two units of (mJ/cm2), without suffering over-exposure effects.

    I guess with more SW work, you could take explicitly single-exposure paths, and either double run those, or double the power.

    I would imagine there is some wiggle room, but if the resist is anything like the last stuff I used, when exposing with incandescent bulbs, you have to be pretty darn close.
  • jmg, two overlays or more. Some pads have three or more traces terminating there, so 3X exposure is likely. Of course, the pads are usually bigger than the traces, so trace burn-in could be fenced in by the pad size. Still, though, I would prefer a raster approach. At least with raster plotting, backlash is less of an issue.

    -Phil
  • jmgjmg Posts: 15,140
    jmg, two overlays or more. Some pads have three or more traces terminating there, so 3X exposure is likely. Of course, the pads are usually bigger than the traces, so trace burn-in could be fenced in by the pad size.
    For positive plotting yes, but less an issue for negative plotting.
    Another appeal of negative plotting, is drill dots are easier to add, than in Positive plotting.
    ..
    Still, though, I would prefer a raster approach. At least with raster plotting, backlash is less of an issue.
    Certainly true if you flyback plot, so always scan in one direction. - but that adds another speed cost...


  • Okay.... This I know.... Those gears sure are increasing the complexity and cost of the design. I think I will call Stock Drive and Misumi tomorrow, and see if they can help get me back on track with a belt design.

    Or perhaps, I should just find the closest match to 600 dpi and export my images at steps per inch.

    The overall diameter of the largest gear is 2.401 inch. That is huge for a small machine. Not only are the gears costly, but to provide rigid support and house the gears.... well.... let's just say that a belt drive would be much more economical and have a much smaller footprint.
  • Get that word backlash out of here.... I want to scan back and forth.... Don't be jinxing me :)
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2017-04-26 02:50
    idbruce wrote:
    Get that word backlash out of here.... I want to scan back and forth.... Don't be jinxing me
    Sorry, pal. It's real and can't be ignored, no matter how tight your mechanical system is.

    -Phil
  • jmgjmg Posts: 15,140
    idbruce wrote: »
    Or perhaps, I should just find the closest match to 600 dpi and export my images at steps per inch.
    If you microstep, then hitting some magic DPI alignment is less critical ?
    Does not sound like you are pushing upper speed limits here ?

  • If you microstep, then hitting some magic DPI alignment is less critical ?
    Does not sound like you are pushing upper speed limits here ?

    I only want to microstep, when I want the highest quality that the machine can produce, because of the processing time involved.

    It really is not that critical if I set steps per inch myself, because like I said earlier, EAGLE allows you to set the dpi of an exported image. However, I would like to stay with the norm, if at all possible.

    I notice there are barcode machines named Zebra, in which they offer replacement belts and pulleys and claim 600 dpi.
  • potatoheadpotatohead Posts: 10,253
    edited 2017-04-26 03:33
    For what it's worth, a high precision machine I'm involved with is using belts and steppers. Precision is .0005" repeatability .002"

    If one is careful to engage in the same direction, at the cost of some spurious movements, repeatability is near .001" it could do 600dpi, carefully programmed.

    The belts are the kind that have teeth and mesh with pulleys that have cutouts. Active mechanism is supported by two precision linear guides, light lube.

    The stepper motors have a gear on them, capable of engaging the belt and there are several steps per .001" of travel.

    It's capable of raster scanning, say 5x5 area in minutes.

    Maybe that gives you some idea on where a belt system could take you Bruce.

  • potatohead

    Sounds nice.

    Yea, I realize a belt system is the way to go, but finding one that will give 600 dpi is no treat, but I think I am on my way to a solution.
  • kwinnkwinn Posts: 8,697
    Ink jet printers are producing images and printing while traveling in both directions, and they use belts so it is possible.

    Early dot matrix printers used a "picket fence" strip and optical sensors to trigger the print head solenoids. Perhaps something similar could be done for bi-directional scanning. An optical sensor that is blocked on each end of the scan and open over the area of the board. By traveling a fixed number of steps beyond the open area before reversing direction any play in the belt would be compensated for.
  • Okay... Using GT2 timing belts and pulleys, in conjunction with a set of small gears, I believe I can achieve a theoretical 600 dots per every 1.004 inches.

    I believe I can live with that for my own personal needs.

    Now let's say that we wanted to make 3.2 inch * 4.0 inch board with just a single copper plane, that copper plane should come out to be 3.2128 inch * 4.016 inch.

    Do you guys think that would be acceptable for most prototyping needs?
  • potatoheadpotatohead Posts: 10,253
    edited 2017-04-26 16:43
    Get a better motor. Doing that will improve precision by a digit.

    Or a better gear.

    Both are a part of the machine I described. You are on a good path here, IMHO.

    Get several steps per .001"

    Doing that will put a lot of your problem into software assuming the mechanism is stable otherwise.
  • idbruceidbruce Posts: 6,197
    edited 2017-04-26 17:32
    Now if I go with a similar setup....

    GT2 timing belts and pulleys, in conjunction with a set of small gears...

    Only for this calculation, instead of exporting an image out of EAGLE at 600 dpi, or multiples thereof, I export them at 598 dpi.

    This would give me a theoretical 598 dots per every 1.001 inches.

    So this time, a 3.2 inch * 4.0 inch board should come out at 3.2032 inch * 4.004 inch
  • idbruceidbruce Posts: 6,197
    edited 2017-04-26 21:38
    And then one other scenario and setup.....

    If I only go with GT2 belts and pulleys, and eliminate the small set of gears....

    I could export images out of EAGLE at 299 dpi, or multiples thereof......

    Which would theoretically give me....

    EAGLE export image at 299 dpi
    FULL step mode - 299 dots per every 1.001 inches.

    EAGLE export image at 598 dpi
    HALF step mode - 598 dots per every 1.001 inches.

    EAGLE export image at 1196 dpi
    QUARTER step mode - 1196 dots per every 1.001 inches.

    EAGLE export image at 2392 dpi
    EIGHTH step mode - 2392 dots per every 1.001 inches.

    Under each of these scenarios, a 3.2 inch * 4.0 inch board should come out at 3.2032 inch * 4.004 inch

    EDIT: Although not 600 dpi, or multiples thereof, this would be the simplest, cheapest, smallest footprint, and maybe, just maybe, the most accurate.

    EDIT: I believe this is going to be my route. It is not necessarily the best route, but the simplest. So instead of being a 600 dpi machine, run in FULL step mode, it will be a 598 dpi machine run in HALF step mode. And instead of having a theoretical pixel size of 0.0016666666666666668 in. * 0.0016666666666666668 in., it will have a theoretical pixel size of 0.0016722408026755853 in. * 0.0016722408026755853 in.
  • With the design of this new machine, being approximately 99% complete, I went ahead and ordered a bunch of parts for it this morning.

    Most of the parts should arrive within a week, but I bought a couple of things off ebay, and those parts have to be shipped from China, so of course those parts will take much longer to arrive. By the time the parts arrive from China, I would imagine that I will be ready to bolt them on and take the new machine for a test drive. However I will still need to order and obtain the PCB material and the dry film photo resist, before any serious testing can commence.

    From the drawings, it appears that it will be a "STOUT", "CUTE", and "SEXY" little machine. I am trying something that I have never done before, and that is to use self-adhesive PTFE tape for a portion of the PCB bed to slide upon. I just hope that the self-adhesive tape pans out and works well. All in all, I am very happy with the design, but ultimate testing, may reveal some hidden flaws.

    Anyhow... not including the PCB material or the dry film resist, I will have approximately $350 tied up into this experimental machine. I hope this little baby will work as good as it looks on paper.

    Wish me luck!
  • idbruceidbruce Posts: 6,197
    edited 2017-05-01 11:03
    Hello Everyone

    Out of curiosity, I decided to dive further into the actual capabilities of a Blu-Ray laser diode and optics. I am really not sure how this information can or will be applied to my project, but It has definitely opened my eyes a little. As it pertains to this discussion, there are two key pieces of information, which are:

    Blu-Ray Track Pitch = 32 microns
    Blu-Ray Minimum Pit Length = 15 microns

    Even though the machine that I am building is theoretically capable of achieving 9568 dpi in 1/32 microstepping mode, the Blu-Ray optics won't come anywhere near that. If I break down the two previous Blu-Ray specs. into dpi, here is the outcome:

    Blu-Ray Track Pitch
    32 microns = 0.0012598425197 inch
    1 inch / 0.0012598425197 inch = 793.74999999057422 dpi

    Blu-Ray Minimum Pit Length
    15 micron = 0.0005905511811 inch
    1 inch / 0.0005905511811 inch = 1693.3333333401065 dpi

    Out of these two specifications, I believe the "Minimum Pit Length" is the most important, because the pit width would definitely have to be in between two tracks of the pitch.

    So about the best anyone could hope for with an unaltered Blu-Ray laser diode and optics is 1693 dpi, as applied to a laser direct imaging machine. In order to achieve this type of resolution with an unaltered Blu-Ray laser diode and optics, I would imagine a person would have to reverse engineer the Blu-Ray sled. It is worth noting, during my research, I encountered two groups of individuals, that were heavily into reverse engineering the Blu-Ray sled, but only specific models were evaluated.

    As for my project, instead of attempting any reverse engineering of the Blu-Ray sled, I decided to attempt going with the bare minimum, which would be just the laser diode and the collimating lens. I created a housing which is similar to most laser modules, in the fact that it has a housing for the laser diode, with an adjustable optics barrel. However, unlike standard laser modules, instead of having a threaded adjustable barrel, my optics barrel is adjusted with a set screw.

    Considering that the machine that I am building will theoretically provide multiples of 299 dpi, I would assume that the best I could possibly hope for with the Blu-Ray optics would be 1196 dpi, however I will initially be striving for 598 dpi. As previously mentioned somewhere in this forum, I believe the hardest part of this project will be the focusing of the optics. In order to effectively achieve 598 dpi, I will need a laser spot size somewhere in the vicinity of 0.0017 in. (rounded) * 0.0017 in. (rounded) or 42.52 um (rounded) * 42.52 um (rounded).

    I believe I am going to need access to a microscope, in order to be even partially successful with this project.
  • Don MDon M Posts: 1,647
    edited 2017-05-01 20:43
    How are you going to focus the beam? Blu Ray and other type players have a focus assembly much like a voice coil in a speaker.
  • jmgjmg Posts: 15,140
    idbruce wrote: »
    ...

    So about the best anyone could hope for with an unaltered Blu-Ray laser diode and optics is 1693 dpi, as applied to a laser direct imaging machine. In order to achieve this type of resolution with an unaltered Blu-Ray laser diode and optics, I would imagine a person would have to reverse engineer the Blu-Ray sled. It is worth noting, during my research, I encountered two groups of individuals, that were heavily into reverse engineering the Blu-Ray sled, but only specific models were evaluated.

    As for my project, instead of attempting any reverse engineering of the Blu-Ray sled, I decided to attempt going with the bare minimum, which would be just the laser diode and the collimating lens. .

    Keep in mind that a key benefit of the 'reverse engineering the Blu-Ray sled' is that the auto-focus capability is retained.
    You only need one of the supported models, to test this.

  • @jmg
    Keep in mind that a key benefit of the 'reverse engineering the Blu-Ray sled' is that the auto-focus capability is retained. You only need one of the supported models, to test this.

    I agree 100%, however I am not anywhere near that point yet.

    Out of the two groups that were reverse engineering, one group was testing a sled in which there were three diodes bundled into one can, which I am not enthused about and the other group, well they started out very open, but they are now keeping their discoveries very private, because they are now selling machines, for laser direct imaging.

    As for reverse engineering one on my own... I just don't believe that I have the knowledge for that.

    @Don M
    How are you going to focus the beam? Blu Ray and other type players have a focus assembly much like a voice coil in a speaker.

    As mentioned earlier, my optics barrel can be adjust with a set screw. I am not saying that I will be successful, but I will be attempting to focus the laser through trial and error, using various measuring devices and an allen wrench :)
  • jmgjmg Posts: 15,140
    edited 2017-05-01 23:01
    idbruce wrote: »
    .... and the other group, well they started out very open, but they are now keeping their discoveries very private, because they are now selling machines, for laser direct imaging.
    That in itself is encouraging information, as it proves once you nail the 'secret sauce', it can be made to work...
    idbruce wrote: »
    As mentioned earlier, my optics barrel can be adjust with a set screw. I am not saying that I will be successful, but I will be attempting to focus the laser through trial and error, using various measuring devices and an allen wrench

    For your early proof-of-concept stages, I wonder about a laser fitting with teflon fingers to simply physically space to the film ?

  • jmg
    That in itself is encouraging information, as it proves once you nail the 'secret sauce', it can be made to work...

    Here is some information about the subject:

    https://blog.adafruit.com/2015/01/01/diyoupcb-is-an-open-source-pcb-printer-which-uses-a-blu-ray-pickup-3dthursday-3dprinting/

    https://youtube.com/watch?v=rDTOBbg9ZXU

    I now see that they are rastering as well.
    For your early proof-of-concept stages, I wonder about a laser fitting with teflon fingers to simply physically space to the film ?

    I really need a focusing and testing jig, before I start building the machine, but I will not get my laser driver for a while. I hate to twiddle my thumbs and wait for parts.
  • idbruceidbruce Posts: 6,197
    edited 2017-05-02 11:33
    In my latest research, I have found that cameras can be utilized to measure and focus laser spot size. Measurement is based upon size of the pixel within the cameras output image, or something similar to that, which makes almost perfect sense to me.

    I don't know if you guys remember, but several years ago, I unveiled one of my inventions, which was a manual pcb drilling apparatus. This manual pcb driller utilized a webcam for focusing and centering drill spot locations. I am attaching a photo of this driller, to provide a visual aid in understanding this discussion.

    Basically, I am now curious about using this same driller to help me size and focus my laser beam. By looking at the attached photo, you will see a 1/8" bronze bearing which guides 1/8" shank size pcb drill bits for drilling holes. Below that bearing is a webcam.

    It is now my intention to focus my laser through that bearing and down into the webcam. And yes I know I may need some serious filtering :) Anyhow, the laser beam should produce a spot size on the webcam output image, having specific pixel dimensions. By knowing the webcam image size, pixel size, and ppi, I believe I can roughly estimate, the laser spot size and adjust focus. This is all dependent upon whether I destroy my webcam or not, by shining a laser beam directly into it :)

    So what do you guys think?webcam_driller.jpg
    912 x 684 - 57K
  • kwinnkwinn Posts: 8,697
    It's a good idea, but shining the laser directly on the imaging chip is almost guaranteed to burn out pixels or the entire chip.

    Perhaps if you had the camera mounted beside the laser so it was focused on the dot the laser projected on the pcb. That would make damage to the camera far less likely, and you could then use the camera to control the dot size, focus, and positioning.
  • idbruceidbruce Posts: 6,197
    edited 2017-05-02 14:38
    Taking this discussion of the PCB driller, webcam, and laser focusing one step further, I now present a webcam image of the bottom of the bearing and a cropped image of the inside diameter of that same bearing. The actual bearing has an ID of 0.127in. and the cropped image of the inner diameter of the bearing is 184 pixels * 184 pixels.

    0.127 inch / 184 pixels = 0.00069021739130434789 inch pixel width
    Theoretical pixel size for 598 dpi = 0.00167391304347826 in. * 0.00167391304347826 in.
    0.00167391304347826 in dpi pixel width / 0.00069021739130434789 = 2.4251968503936991 laser dot size within the ID of the bearing

    To achieve the intended 598 dpi, my laser dot must be within 3 * 3 and 2 * 2 pixels, of the bearing ID of the cropped image. There is a tiny magenta colored dot within the cropped image that represents a 2 * 2 pixel area and a very close approximation of just how small the focused laser dot must be to achieve 598 dpi.

    bearing_bottom.jpg
    bearing_id_with_dot.jpg
    640 x 480 - 36K
    184 x 184 - 5K
  • kwinn

    It's a good idea, but shining the laser directly on the imaging chip is almost guaranteed to burn out pixels or the entire chip.

    Perhaps if you had the camera mounted beside the laser so it was focused on the dot the laser projected on the pcb. That would make damage to the camera far less likely, and you could then use the camera to control the dot size, focus, and positioning.

    I did say
    And yes I know I may need some serious filtering

    I think with the right filtering, it might be okay. Or perhaps shining it onto a semi-translucent material, such as a cigarette rolling paper or film used for photo masks.
  • Okay... I think I may have the answer.... Suppose I diffuse the laser through 12~14 welding shade.... I think that might do the trick.
  • Will the laser not fire at reduced power?
  • idbruceidbruce Posts: 6,197
    edited 2017-05-02 16:23
    potatohead
    Will the laser not fire at reduced power?

    I am not really sure what it will do just yet. My driver is coming from China and I must wait on delivery before I can do any serious testing.

    Until the laser driver arrives, I am just trying to develop some sort of focusing scheme to quickly get me close to where I want to be.

    I would imagine that it will still fire at about 5mw, but then again, I really wouldn't want to risk damaging my PCB driller, so I am thinking that it would be wise to at least try to filter some of the brightness out of the beam.

    Besides.... Having a variety of welding shades for the old helmet is never a bad idea :)
  • The filter is thick enough to affect your focus point.

    It's small, but so is your intended focus.

Sign In or Register to comment.