Shop OBEX P1 Docs P2 Docs Learn Events
Somewhat OT: Optical Gratings — Parallax Forums

Somewhat OT: Optical Gratings

ArchiverArchiver Posts: 46,084
edited 2004-02-13 01:56 in General Discussion
Hello to all.

My project involves the detection of angular displacements as small
as 1/4th of a degree.

I will use simple optical gratings with five slits. Each slit will
have a width of 1/10th of a mm or perhaps less. When the slits are
aligned, I will have a window of 0.5 mm.

Which type of LED and photonic detector would be adequate? I will use
graphic arts photographic film for the gratings which I believe is
transparent for IR frequencies so IR is not adequate.

At this point I believe a bright white light LED might be adequate,
using a detector for visible light.

But there are several types of detectors. Perhaps someone here has
experience with them.

Thanks in advance.

Comments

  • ArchiverArchiver Posts: 46,084
    edited 2004-02-09 23:34
    --- In basicstamps@yahoogroups.com, "Pablo Roufogalis L."
    <proufo@c...> wrote:
    > Hello to all.
    >
    > My project involves the detection of angular displacements as small
    > as 1/4th of a degree.

    I meant 1/40th of a degree. That is the angular displacement of a
    nine-inch tonearm when it moves the width of a groove.

    My project is a tangential-tracking turntable.
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-10 01:06
    You mean like the Rabinow linear tonearm? I've actually played with one a
    quarter century or so ago. Or if not, then exactly what are you trying to
    detect and control?

    Mike Sokol
    mike@f...
    www.fitsandstarts.com

    " One should not increase, beyond what is necessary,
    the number of entities required to explain anything"...
    -William of Occam-


    Original Message
    From: "Pablo Roufogalis L." <proufo@c...>
    To: <basicstamps@yahoogroups.com>
    Sent: Monday, February 09, 2004 6:34 PM
    Subject: [noparse][[/noparse]basicstamps] Re: Somewhat OT: Optical Gratings


    > --- In basicstamps@yahoogroups.com, "Pablo Roufogalis L."
    > <proufo@c...> wrote:
    > > Hello to all.
    > >
    > > My project involves the detection of angular displacements as small
    > > as 1/4th of a degree.
    >
    > I meant 1/40th of a degree. That is the angular displacement of a
    > nine-inch tonearm when it moves the width of a groove.
    >
    > My project is a tangential-tracking turntable.
    >
    >
    >
    > To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    > from the same email address that you subscribed. Text in the Subject and
    Body of the message will be ignored.
    >
    > Yahoo! Groups Links
    >
    >
    >
    >
    >
    >
    >
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-10 16:27
    This is very difficult because even if you use a bright white LED and a
    photoresistor or photodiode it will clearly illuminate or show through the
    transparent art-work. however you will need to experiment with it, because
    i dont know how transparent it is.


    At 11:23 PM 2/9/2004 +0000, you wrote:
    >Hello to all.
    >
    >My project involves the detection of angular displacements as small
    >as 1/4th of a degree.
    >
    >I will use simple optical gratings with five slits. Each slit will
    >have a width of 1/10th of a mm or perhaps less. When the slits are
    >aligned, I will have a window of 0.5 mm.
    >
    >Which type of LED and photonic detector would be adequate? I will use
    >graphic arts photographic film for the gratings which I believe is
    >transparent for IR frequencies so IR is not adequate.
    >
    >At this point I believe a bright white light LED might be adequate,
    >using a detector for visible light.
    >
    >But there are several types of detectors. Perhaps someone here has
    >experience with them.
    >
    >Thanks in advance.
    >
    >
    >
    >To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    >from the same email address that you subscribed. Text in the Subject and
    >Body of the message will be ignored.
    >
    >Yahoo! Groups Links
    >
    >
    >
    >
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-10 22:22
    --- In basicstamps@yahoogroups.com, "Mike Sokol - Fits & Starts"
    <mike@f...> wrote:
    > You mean like the Rabinow linear tonearm? I've actually played with
    one a
    > quarter century or so ago. Or if not, then exactly what are you
    trying to
    > detect and control?

    Hello Mike.

    Coincidentially, I have described my project elsewhere as an updated
    Rabco.

    There are many linear-turntables. Mine will different because:

    1) It uses a regular pivoting tonearm, not an specially designed
    tonearm. In order to make it track tangentially, the armboard is
    mobile, controlled by a Stamp, a step-moptor controller, a step motor
    and a linear stage.

    2) Instead of waiting for the arm to develop an angular error and
    correct, it will keep a steady velocity and correct position and
    speed if required. I believe mass-market designs from the 70s act on
    error alone, including the Rabcos which used a mechanical sceme to
    move the arm.

    3) I aim to detect deviations of one groove. I believe mass designs
    use a much larger error window.

    BTW, Mr. Surround, my first iteration of the turntable will be for my
    CD-4 Quadraphonic setup.

    Best regards.
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-10 22:59
    --- In basicstamps@yahoogroups.com, Gary Denison <gdii@c...> wrote:
    > This is very difficult because even if you use a bright white LED
    and a
    > photoresistor or photodiode it will clearly illuminate or show
    through the
    > transparent art-work. however you will need to experiment with it,
    because
    > i dont know how transparent it is.

    Hello Gary and thanks for your comment.

    Graphics arts film is very dark. The figure of merit here is density,
    typically above 4.0. And there will be two layers of it.

    I'm sure it will require a lot of tweaking though.

    Would you use a photoresistor or a photodiode? What is easier to
    interface?

    Thanks again.
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-10 23:30
    Pablo,

    Most interesting....

    And yes, my main gig is consulting and teaching surround production
    techniques. As part of my surround seminars, I do discuss the original quad
    technologies from the 70's, including CD-4. I'm curious as to what CD-4
    records you have to play, and what cartridge and stylus are you using? Send
    me a picture when you've got it working and I'll put it in one of my
    PowerPoint slides.

    If anyone on this group would like an explanation of CD-4 and SQ quad
    records (which is what Pablo's project will decode), let me know and I'll
    post a quick overview. Tracking error was the Achilles heal of getting quad
    records to work properly, so Pablo's project has an excellent chance of
    working WAY better than the original turntables of the 70's... at least as
    far as quad records are concerned.

    Mike Sokol
    mike@f...
    www.fitsandstarts.com

    " One should not increase, beyond what is necessary,
    the number of entities required to explain anything"...
    -William of Occam-


    Original Message
    From: "Pablo Roufogalis L." <proufo@c...>
    To: <basicstamps@yahoogroups.com>
    Sent: Tuesday, February 10, 2004 5:22 PM
    Subject: [noparse][[/noparse]basicstamps] Re: Somewhat OT: Optical Gratings


    > --- In basicstamps@yahoogroups.com, "Mike Sokol - Fits & Starts"
    > <mike@f...> wrote:
    > > You mean like the Rabinow linear tonearm? I've actually played with
    > one a
    > > quarter century or so ago. Or if not, then exactly what are you
    > trying to
    > > detect and control?
    >
    > Hello Mike.
    >
    > Coincidentially, I have described my project elsewhere as an updated
    > Rabco.
    >
    > There are many linear-turntables. Mine will different because:
    >
    > 1) It uses a regular pivoting tonearm, not an specially designed
    > tonearm. In order to make it track tangentially, the armboard is
    > mobile, controlled by a Stamp, a step-moptor controller, a step motor
    > and a linear stage.
    >
    > 2) Instead of waiting for the arm to develop an angular error and
    > correct, it will keep a steady velocity and correct position and
    > speed if required. I believe mass-market designs from the 70s act on
    > error alone, including the Rabcos which used a mechanical sceme to
    > move the arm.
    >
    > 3) I aim to detect deviations of one groove. I believe mass designs
    > use a much larger error window.
    >
    > BTW, Mr. Surround, my first iteration of the turntable will be for my
    > CD-4 Quadraphonic setup.
    >
    > Best regards.
    >
    >
    >
    > To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    > from the same email address that you subscribed. Text in the Subject and
    Body of the message will be ignored.
    >
    > Yahoo! Groups Links
    >
    >
    >
    >
    >
    >
    >
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-11 00:13
    --- In basicstamps@yahoogroups.com, "Mike Sokol - Fits & Starts"
    <mike@f...> wrote:
    > Pablo,
    >
    > Most interesting....
    >

    >
    > If anyone on this group would like an explanation of CD-4 and SQ
    quad
    > records (which is what Pablo's project will decode), let me know
    and I'll
    > post a quick overview. Tracking error was the Achilles heal of
    getting quad
    > records to work properly, so Pablo's project has an excellent
    chance of
    > working WAY better than the original turntables of the 70's... at
    least as
    > far as quad records are concerned.

    Hello again.

    You'd be surprised how active the quaddie community is, even in 2004.

    Check the forums at http://www.quadraphonicquad.com/forums and also
    do a search for quad* at ebay. You'd be amazed.

    Would love to provide pictures when the time comes to that. But many
    at the forums can provide their current setups, up and running. Many
    have extensive collections of Quad recording, way better than mine.

    Will ask but you are invited to visit and join.

    BTW, whatcha doin'heah? ;-)

    Best regards.
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-11 05:01
    Pablo,

    I believe you can use an infrared source and detector for your
    application. Photographic film that has a density of 4.0 should block
    both visible and near-infrared light. You can set up a simple
    experiment to test this effect by using a matched infrared emitter and
    detector pair from Radio Shack (#276-142). Simply point the IR emitter
    at the detector, measure the signal, then block the path with the
    film. Be sure to measure the attenuation of a clear area on the film
    as well as the high density area. I would expect an extinction ratio
    (high density area signal/clear area signal) of at least 50:1. More is
    better.

    As for the ease of interfacing to a photoresistor or photodiode, I
    think the real question is speed. Photoresistors are slow,
    milliseconds. Photodiodes are fast, microseconds or nanoseconds. It
    seems like your application would benefit from using a photodiode, so
    the response time of the sensor is not an issue in your tone arm
    control loop.

    I would suggest that you do some preliminary experiments with the
    actual gratings and optical components in a test jig, to check that
    the on/off signal levels through the five tiny slits are sufficient.
    This will help reduce any surprises downstream.

    Dave


    >
    > Graphics arts film is very dark. The figure of merit here is density,
    > typically above 4.0. And there will be two layers of it.
    >
    > I'm sure it will require a lot of tweaking though.
    >
    > Would you use a photoresistor or a photodiode? What is easier to
    > interface?
    >
    > Thanks again.
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-11 12:04
    Hello Dave and thanks for your suggestions.

    Will test this for sure. Any idea on the size of the 'light
    acceptance window' of the RS detector?

    Thanks again.

    --- In basicstamps@yahoogroups.com, "barrowman11" <barrowman11@y...>
    wrote:
    > Pablo,
    >
    > I believe you can use an infrared source and detector for your
    > application. Photographic film that has a density of 4.0 should
    block
    > both visible and near-infrared light. You can set up a simple
    > experiment to test this effect by using a matched infrared emitter
    and
    > detector pair from Radio Shack (#276-142). Simply point the IR
    emitter
    > at the detector, measure the signal, then block the path with the
    > film. Be sure to measure the attenuation of a clear area on the film
    > as well as the high density area. I would expect an extinction ratio
    > (high density area signal/clear area signal) of at least 50:1. More
    is
    > better.
    >
    > As for the ease of interfacing to a photoresistor or photodiode, I
    > think the real question is speed. Photoresistors are slow,
    > milliseconds. Photodiodes are fast, microseconds or nanoseconds. It
    > seems like your application would benefit from using a photodiode,
    so
    > the response time of the sensor is not an issue in your tone arm
    > control loop.
    >
    > I would suggest that you do some preliminary experiments with the
    > actual gratings and optical components in a test jig, to check that
    > the on/off signal levels through the five tiny slits are sufficient.
    > This will help reduce any surprises downstream.
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-11 16:36
    Pablo,

    No, I do not know what the actual viewing angle is, but you can a
    rough measure by using the matching detector as an optical probe.

    Dave

    --- In basicstamps@yahoogroups.com, "Pablo Roufogalis L."
    <proufo@c...> wrote:
    > Hello Dave and thanks for your suggestions.
    >
    > Will test this for sure. Any idea on the size of the 'light
    > acceptance window' of the RS detector?
    >
    > Thanks again.
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-12 00:07
    Paublo,

    Hmmmm. Why bother with the angular displacement? Why not simply
    count revolutions of the turn table? Multiply counts x 0.04 to get
    total angular displacement.

    I worked with a shaft encoder once that used optical gratings and gave
    an 18-bit count. One LSB of the encoder was 9.88 minutes of arc.
    Your 0.04 degree resoulotion works out to 2.4 minutes of arc, so it
    should be possible to use gratings. However, the "gottcha" is the
    alignment and stability that you need. Also, one of the gratings
    should have one less line-per-inch than the oterher to get a stable
    Moirere pattern. Good luck! The encoder cost about $8000 and had to
    be sent back to the factory to have its light bulbs replaced at a cost
    of about $4000!

    I'm afraid that a stepper motor would create acoustic noise that would
    transmit to the tone arm. I'd suggest using a linear actuator.

    Russ


    --- In basicstamps@yahoogroups.com, "Pablo Roufogalis L."
    <proufo@c...> wrote:
    > Hello to all.
    >
    > My project involves the detection of angular displacements as small
    > as 1/4th of a degree.
    >
    > I will use simple optical gratings with five slits. Each slit will
    > have a width of 1/10th of a mm or perhaps less. When the slits are
    > aligned, I will have a window of 0.5 mm.
    >
    > Which type of LED and photonic detector would be adequate? I will use
    > graphic arts photographic film for the gratings which I believe is
    > transparent for IR frequencies so IR is not adequate.
    >
    > At this point I believe a bright white light LED might be adequate,
    > using a detector for visible light.
    >
    > But there are several types of detectors. Perhaps someone here has
    > experience with them.
    >
    > Thanks in advance.
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-12 00:33
    --- In basicstamps@yahoogroups.com, "russranshaw" <home.rr@c...>
    wrote:
    > Paublo,
    >
    > Hmmmm. Why bother with the angular displacement? Why not simply
    > count revolutions of the turn table? Multiply counts x 0.04 to get
    > total angular displacement.
    >
    > I worked with a shaft encoder once that used optical gratings and
    gave
    > an 18-bit count. One LSB of the encoder was 9.88 minutes of arc.
    > Your 0.04 degree resoulotion works out to 2.4 minutes of arc, so it
    > should be possible to use gratings. However, the "gottcha" is the
    > alignment and stability that you need. Also, one of the gratings
    > should have one less line-per-inch than the oterher to get a stable
    > Moirere pattern. Good luck! The encoder cost about $8000 and had
    to
    > be sent back to the factory to have its light bulbs replaced at a
    cost
    > of about $4000!
    >
    > I'm afraid that a stepper motor would create acoustic noise that
    would
    > transmit to the tone arm. I'd suggest using a linear actuator.
    >
    > Russ

    Hello Russ and thanks for your comments.

    It is not the best solution to count revolutions because the spacing
    of the grooves is not constant. The arm will lose tangentiality
    immediately, grossly in some cases.

    Yes! I saw the radial gratings one having one additional spoke in
    my "Mechanism for intermitent motion" book. I think I can do that
    with film but will try first with something simpler. I just need to
    detect small deviations within a very limited range. I agree it will
    be a challenge to align the gratings. One step at a time.

    Will use a linear actuator, with a very, very fine pitch. Check it at:

    http://www.sandbeads.com/siteimages/SS1.jpg

    and the pitch:

    http://www.sandbeads.com/siteimages/SS3.jpg

    I will use a lot of damping to diminish any chance of pulses going to
    the tonearm. Crossing my fingers, though.

    Thanks again for your comments.
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-12 01:41
    Hello to all.

    This is my current approach to the gratings. I increased the number
    of slits to ten, giving a window of 1 mm, somewhat spaced.

    http://www.sandbeads.com/siteimages/Gratings.gif

    This represents the two gratings. The bottom one will move over the
    top one (at this point in time).

    There are four sensors corresponding to five individual gratings
    pairs. The two extremes labeled 'Panic' will detect any gross
    tangential error, which should never happen except at the end of the
    record. In case any of those sensors activate, a servo will lift the
    tonearm.

    The one labeled 'Land between cuts' will move the tonearm inwards to
    rapidly keep tangentiality when the arm moves between cuts (songs).

    The comb-like slits detect angular deviation, inward or outward.
    Although the system should correct confirmed one-groove deviations
    immediately, these gratings are designed to keep state even with two-
    groove deviations. This may change if I need to increase the number
    of slits.

    Any comment is welcome.
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-12 01:46
    --- In basicstamps@yahoogroups.com, "Pablo Roufogalis L."
    <proufo@c...> wrote:

    > There are four sensors corresponding to five individual gratings
    > pairs. The two extremes labeled 'Panic' will detect any gross

    Sorry. There are FIVE sensors corresponding to five...
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-12 13:28
    FYI:

    Back when we used to cut records on a lathe, the engineer would try to vary
    the spacing between the grooves manually using the feed-rate control. It was
    pretty much by intuition, since you had to have more spacing for really loud
    passages to prevent groove-echo, but could cram the grooves closer together
    for soft audio passages. That way you could get the most audio length on a
    disc. There were a whole series of 45-RPM LP 10 inch singles for the DJ
    markets, where a single 5-minute song was spread out using really wide
    groove spacing and run at 45 rather than 33 1/3 RPM. Perhaps the first
    hi-rez LP's?

    Mike Sokol
    mike@f...
    www.fitsandstarts.com


    " One should not increase, beyond what is necessary,
    the number of entities required to explain anything"...
    -William of Occam-


    Original Message
    From: "Pablo Roufogalis L." <proufo@c...>
    To: <basicstamps@yahoogroups.com>
    Sent: Wednesday, February 11, 2004 7:33 PM
    Subject: [noparse][[/noparse]basicstamps] Re: Somewhat OT: Optical Gratings


    > --- In basicstamps@yahoogroups.com, "russranshaw" <home.rr@c...>
    > wrote:
    > > Paublo,
    > >
    > > Hmmmm. Why bother with the angular displacement? Why not simply
    > > count revolutions of the turn table? Multiply counts x 0.04 to get
    > > total angular displacement.
    > >
    > > I worked with a shaft encoder once that used optical gratings and
    > gave
    > > an 18-bit count. One LSB of the encoder was 9.88 minutes of arc.
    > > Your 0.04 degree resoulotion works out to 2.4 minutes of arc, so it
    > > should be possible to use gratings. However, the "gottcha" is the
    > > alignment and stability that you need. Also, one of the gratings
    > > should have one less line-per-inch than the oterher to get a stable
    > > Moirere pattern. Good luck! The encoder cost about $8000 and had
    > to
    > > be sent back to the factory to have its light bulbs replaced at a
    > cost
    > > of about $4000!
    > >
    > > I'm afraid that a stepper motor would create acoustic noise that
    > would
    > > transmit to the tone arm. I'd suggest using a linear actuator.
    > >
    > > Russ
    >
    > Hello Russ and thanks for your comments.
    >
    > It is not the best solution to count revolutions because the spacing
    > of the grooves is not constant. The arm will lose tangentiality
    > immediately, grossly in some cases.
    >
    > Yes! I saw the radial gratings one having one additional spoke in
    > my "Mechanism for intermitent motion" book. I think I can do that
    > with film but will try first with something simpler. I just need to
    > detect small deviations within a very limited range. I agree it will
    > be a challenge to align the gratings. One step at a time.
    >
    > Will use a linear actuator, with a very, very fine pitch. Check it at:
    >
    > http://www.sandbeads.com/siteimages/SS1.jpg
    >
    > and the pitch:
    >
    > http://www.sandbeads.com/siteimages/SS3.jpg
    >
    > I will use a lot of damping to diminish any chance of pulses going to
    > the tonearm. Crossing my fingers, though.
    >
    > Thanks again for your comments.
    >
    >
    >
    > To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    > from the same email address that you subscribed. Text in the Subject and
    Body of the message will be ignored.
    >
    > Yahoo! Groups Links
    >
    >
    >
    >
    >
    >
    >
  • ArchiverArchiver Posts: 46,084
    edited 2004-02-13 01:56
    --- In basicstamps@yahoogroups.com, "Mike Sokol - Fits & Starts"
    <mike@f...> wrote:
    > FYI:
    >
    > Back when we used to cut records on a lathe, the engineer would try
    to vary
    > the spacing between the grooves manually using the feed-rate
    control. It was
    > pretty much by intuition, since you had to have more spacing for
    really loud
    > passages to prevent groove-echo, but could cram the grooves closer
    together
    > for soft audio passages. That way you could get the most audio
    length on a
    > disc. There were a whole series of 45-RPM LP 10 inch singles for
    the DJ
    > markets, where a single 5-minute song was spread out using really
    wide
    > groove spacing and run at 45 rather than 33 1/3 RPM. Perhaps the
    first
    > hi-rez LP's?

    The (boosted) bass extension on those is astonishing, and the overall
    sound is much better. It does make a difference.

    There is a label that has reissued classic albums (one is Carole
    King's Tapestry) at 45 rpm. It takes six 12" sides for an Lp and are
    quite expensive.
Sign In or Register to comment.