Connect with us

high resolution optical encoder patterns.

Discussion in 'Hobby Electronics' started by Cory Seligman, Sep 25, 2003.

Scroll to continue with content
  1. Hi all.

    I've successfully made a linear optical encoder with a stroke of about
    25cm using a flattened steel rod with a nice shiny surface, and gluing
    on a laser printed transparency with a pattern of stripes. This works
    find with a 1mm pattern, and a homemade detector. The detector is an
    IR LED, a small lense and a Honeywell HLC2701 encoder detector
    chip/opto thingy.

    No worries, except it's fiddly. The lense and detector and pattern
    have to be adjusted just right, and it's hard to put together.

    So I got some Agilent HEDR-8100 detector chips. they have the lenses
    built in and do all the hard work for you. The only problem is that
    they're 150 lines per inch, which is more than I can do reliably on a
    laser printer. So I'm looking for alternatives.

    Does anyone know of somewhere I can get stripes like that printed
    directly on the steel? Or etched?

    Any thoughts on producing a clear strip with black stripes
    photographically?

    Or maybe a replacement reflective optical encoder detector that works
    on a 1mm stripe spacing? As far as I can tell, the Agilent one is the
    only device of its type.

    Thanks in advance,
    Cory
     
  2. Rod Speed

    Rod Speed Guest

    Or use an optical mouse that doesnt need
    anything special on what the sensor passes over.
     
  3. Paul Burke

    Paul Burke Guest

    If you get hold of some old HP Deskjet printers, they have a nice little
    linear encoder (a transparent plastic strip) for the head positioning
    system. It's about the width of an A4 sheet and probably has a
    resolution of 300DPI! I expect they used to produce the stuff in long
    lengths, so maybe HP have it in their catalog somewhere.

    Paul Burke
     
  4. David

    David Guest

    So I got some Agilent HEDR-8100 detector chips. they have the
    lenses
    In terms of getting a better printed resolution trying asking a
    graphic designer studio or a printer if they can produce a
    photographic bromide image or a photo-plot.

    BTW, why does the 150 lpi limit your application? You only need
    to count the transitions, and as long as you know how many pulses
    there are per distance unit you can scale it. You might need to
    make your lines nice and sharp and solid, but unless I am missing
    something it should still see them and encode accordingly. The
    HEDR-8000 does 75 lpi, would that be easier?
     

  5. Have you tried using a high quality ink jet type printer with ink jet
    transparency? Even the local Walmart will sell Lexmark printers that claim
    up to 2400dpi or up to 11ppm capability and only cost $35. 2400dpi aught to
    be way more than adequate for the job. You have to use the right type of
    transparency though (a laser transparancy or the wrong type of ink jet
    transparency will yeild terrible results).

    Using my ordinary Epson ink jet printer I regularly develop etched copper
    PCBs with line sizes sometimes as small as say 5 mils wide. I haven't
    extensively tested smaller features because I haven't really had much need
    for it, but I think it is certainly possible given todays ink jet
    technology. If you aren't going to be using any photoresist and copper
    etching I should imagine the quality would be a fair amount better as well
    (with just the plain transparency).
     
  6. nospam

    nospam Guest

    Create your pattern with a PCB layout package, export to Gerber and get
    your local PCB house to plot a sheet of it.
     
  7. They can be created on an imagesetter from a Postscript (or PDF) file.
    I just paid $60 for two oversize pages on litho film, price seems to
    be going up as more printing goes to direct-to-plate. But that would
    be a heck of a lot of strips.

    Best regards,
    Spehro Pefhany
     
  8. Tim Shoppa

    Tim Shoppa Guest

    At the very fine end, ruled diffraction gratings are available, but
    they're in the hundreds or low thousands of lines per mm.

    Ronchi targets are very high precision and have the resolution you
    desire, but mounting them to your device may present some difficulties.

    Getting a pattern put onto lith film is something your local
    repro shop can do for you, and probably something you can deal with
    in mounting.

    You told us you want a resolution of a few hundred lines per inch, but
    you didn't tell us how precisely they must be spaced. Others have
    suggested using consumer-quality printers to do this, and those are
    probably capable of giving the resolution you desire, but they are
    very abysmal in terms of absolute precision.

    Tim.
     
  9. Zak

    Zak Guest

    A litho house can make this for you - even from your PC produced file.

    They usually work with Postscript to produce a 2400 DPI output on silver
    halide film. Huge sized are posible, but 30 cm wide wide is the beginning.

    It shouldn't be expensive either.


    Thomas
     
  10. If you place a transparent filter on top, also printed with 1mm stripes,
    you don't need a very well adjusted sensor. When the rod with the stripes
    move behind the filter, even you can clearly see that it moves. Most
    optical encoders work like that, makes them get away with pretty coarse leds
    and fotodiodes.

    Situation 1:
    # # # # # # # # # # # # # # # # # ROD
    # # # # # FILTER

    ^^^ looks like 100% black from this point


    Situation 1, rod moved a bit:
    # # # # # # # # # # # # # # # # # # ROD
    # # # # # FILTER

    ^^^ looks like 50% black from this point


    Place two filters with the stripes 90 degrees apart for A & B signals.
    The effect is named after Moire, it's inventor, I think.
     
  11. Thanks for the ideas. I have a suspicion that the optical devices I'm
    using already exploit the Moire effect, since they appear to only work
    at 150 lpi. This would indicate that they already have a 150lpi screen
    internally.

    The optical mouse sensor idea is something I already thought of, but
    the specs for some of the modules they use show fairly lousy
    repeatability, which I guess amkes sense. I need pretty good (0.5 mm
    or so) repeatability.

    I have yet to have any success with ink jet printer on transparencies.
    The ink is either too think, blobbing up and obscuring the gaps, or
    not dark enough to block the reflection. Perhaps I need better
    stationary.

    In any case, I've ordered the lower resolution (75 lpi) encoders. I've
    printed out some 75 lpi stripes, and checked them under a microscope.
    They look clear and distinct, so I'm just going to go with that.

    thanks for the responses though.

    Cory
     
  12. John Eaton

    John Eaton Guest

    : I have yet to have any success with ink jet printer on transparencies.
    : The ink is either too think, blobbing up and obscuring the gaps, or
    : not dark enough to block the reflection. Perhaps I need better
    : stationary.

    : In any case, I've ordered the lower resolution (75 lpi) encoders. I've
    : printed out some 75 lpi stripes, and checked them under a microscope.
    : They look clear and distinct, so I'm just going to go with that.

    : thanks for the responses though.


    Find an old hp ink jet that someone is tossing and salvage the encoder strip.


    1/150 dpi.


    John Eaton
     
  13. Rod Speed

    Rod Speed Guest

    Makes no sense at all, you dont see any problem like that using a good one.
    Should be completely trivial with a decent optical mouse
    and the most you might want to do is make sure that what
    you are running it on isnt dead smooth like glass etc.
    Yeah, thats completely trivial to do properly too. Basically
    just the right combination of transparent film and ink.
     
  14. hey, yeah. You're right. They have good repeatability. I'll give em a
    go for the next version of the prototype. As it is now, I have the 75
    lpi encoders and they work fine with laser printed stripes.
    Yeah. It's machined steel, and an optical mouse picks it up fine. I do
    have pretty limited space though, and it'd be a tight squeeze with the
    sensors used in the mouses I've looked at so far. I'm sure they make
    tiny ones though....
    Yeah, I assumed so. But the combination I had was lousy.

    Anyhow, it's all moot now, since the 75 lpi stuff all works fine, and
    the laser printed stripes are nice and clear when glued to the steel
    rod.

    thanks anyhow,
    cory
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-