Loading ...
Sorry, an error occurred while loading the content.

Re: [SeattleRobotics] Re: Piezo touch sensors

Expand Messages
  • Steve Greenfield
    ... I think the original poster was attempting to convert an existing LCD graphic display into a touch sensitive screen. As for using this for a robot sensor,
    Message 1 of 18 , Jul 31 10:17 PM
    • 0 Attachment
      --- RoboVac <robovac@...> wrote:
      >
      > One problem with all this is that while it works very good at
      > detecting
      > people it fails to detect other things like chairs, walls, etc.
      >

      I think the original poster was attempting to convert an existing
      LCD graphic display into a touch sensitive screen.

      As for using this for a robot sensor, I'd make it one of many. We
      have more than one "sensor", why expect our robots to get along
      with one sense? Some things aren't picked up well by IR but may be
      by ultrasonics, etc.

      Steve Greenfield

      __________________________________________________
      Do You Yahoo!?
      Yahoo! Health - Feel better, live better
      http://health.yahoo.com
    • Steve Greenfield
      ... Hype is part of marketing. The key behind their chip is that it is just that- a chip. IE, I don t have to put together a collection of chips and parts to
      Message 2 of 18 , Aug 2, 2002
      • 0 Attachment
        --- Alan King <alan@...> wrote:
        > Steve Greenfield wrote:
        >
        > >Alternative: use the Q-Prox matrix scanning capacitive sensors
        > with
        > >a matrix of very thin wires. I mean thin, I'm thinking take
        > apart
        > >
        >
        > Heh I love it when people tout products like they're
        > important,
        > especially when I could likely bust their patent.. :) I guess
        > I'll
        > share since I'm not using it right now anyway and I'm working on
        > other things..

        Hype is part of marketing. The key behind their chip is that it is
        just that- a chip. IE, I don't have to put together a collection of
        chips and parts to do the same thing, automatic drift and
        contaminant compensation, scanning the matrix, 3 different types of
        output interface, etc. The Pentium is just a collection of
        transistors and resistors, after all, the key is that it is put
        together in one chip.

        Yes, we all know that capacitive sensing is as old as tuned
        circuits. They've just really simplified it. I'm hoping I can use
        their 64 key capacitive matrix for some projects I have in mind,
        it'll make it a lot easier.

        So what other things are you working on?

        =====
        Steve Greenfield // Digital photo scanning, retouching,
        Polymorph Digital Photography // and photomorphing to your specs.
        253/318-2473 voice // We use the best little computer in
        polymorph@... // the world, the Amiga!
        http://www.polyphoto.com/ // Based in Tacoma, WA, USA

        __________________________________________________
        Do You Yahoo!?
        Yahoo! Health - Feel better, live better
        http://health.yahoo.com
      • warrenhb
        Your circuit looks something like the one I had in mind for use with three piezo disks used as pressure sensors to triangulate the position of touch on a touch
        Message 3 of 18 , Aug 12, 2002
        • 0 Attachment
          Your circuit looks something like the one I had in mind for use
          with three piezo disks used as pressure sensors to triangulate
          the position of touch on a touch screen.

          Your use of r2 leads me to conclude that you're connecting this
          directly to a digital input. That means you're using the threshold
          of the digital input as your sense threshold. As long as you're
          using an op-amp, you could use it in a comparator circuit instead
          of a voltage follower, then use a couple of resistors or a pot to
          adjust the threshold.

          For the touch screen, I would want to measure bipolar analog
          pressure, so I would not use r2, and would connect the
          output to an A/D converter input. To get bipolar sensitivity,
          I would add a resistor in series with the inverting input, plus
          resistors from the inverting input to ground and V+. These
          three resistors could be used to control the scale and zero
          offset.

          For a robot bumper touch sensor, I would think the voltage
          doubler circuit I mentioned in my first message would be
          preferable to your circuit, since it would respond to the peak-
          to-peak voltage developed by the transducer, and so wouldn't
          depend upon the direction it was deflected, only that it was
          deflected. Here's what I had in mind:

          diode
          ___________>|_________________to A/D, comparator
          | | | | or CMOS digital input
          piezo | | c1 r1
          | |_/<_ |_||__ |
          | zener | || | |
          G G G G

          A single ended voltage doubler circuit would normally have
          a capacitor in series with the voltage source, but the piezo
          transducer is itself a capacitor, so I've been experimenting
          with leaving out the initial capacitor. The zener voltage sets
          the positive voltage limit and drains off negative charge so
          that when the transducer returns to neutral deflection it will
          develop a corresponding positive charge. Positive deflection
          should cause an additive voltage which will appear on c1.
          Together, r1 and c1 set the time constant.

          The output could be connected to an A/D converter input to
          measure how hard the robot was touched, or to a comparator
          with an adjustable threshold, but it's probably sufficient to
          connect the output to a CMOS digital input and just use the
          fixed input switching threshold to set the threshold.

          Looks like a fun contest! It would select for a robot that can
          cover the most carpet in a room without getting stuck, as well
          as being able to pick up rice. Maybe SRS could try it. It would
          be intermediate between our sumo and fire fighting in some
          ways, and simpler than the sumo in others. It would require
          less effort than our line following maze.

          Looks like your shroud is the slotted sheet metal cylinder.
          Do you use more than one sensor? Looks like two on either
          side could tell you whether you bumped on the left or right
          or both.

          The piezo cable listed on the Digikey page looks pretty pricy.
          Vetco in Bellevue has the piezo transducers for $.85 each, and
          I was able to buy 10 for $6.80.

          Doug Bell

          ----- Original Message -----
          From: "krowell98" <rowell@...>
          To: "warrenhb" <warrenhb@...>
          Sent: Friday, August 02, 2002 9:14 AM
          Subject: Re: Piezo touch sensors


          | I've used the folowing piezo sensor circuit with some success.
          |
          | The 1m resistor slows down the discharge time of the piezo and
          | drains a characteristic voltage buildup in piezo's.
          |
          | The zener pair discharges the majority of the dangerous voltages
          | above 5 volts.
          |
          | The op-amp is used as a follower and repeats whatever input it gets,
          | safely repeating it in normal ttl voltages to the controller.
          |
          | The last resistor r2 pulls down the controller pin.
          |
          | op-amp
          | ___________
          | ________________________ | |
          | | | | | |_-_|\ |
          | piezo r1 | | | \____|______to controller
          | |________| >//< |___+_| / |
          | | zener | |/ r2
          | G pair G |
          | shunt G
          |
          |
          | The Piezo sensor worked well for me in a vacuum contest in Atlanta.
          | See: http://www.botlanta.org/rally2002/vac_photos.html
          |
          | Kluge II didn't pick up alot of rice, but he didn't get stuck.
          |
          | My biggest problem with them is the mounting scheme. I used a shroud
          | on the front of the bot to provide a skin on which to push. The
          | piezo was mounted between the shroud and the frame.
          |
          | |<
          | shroud|
          | ----->| frame
          | |
          | |<
          |
          | Good flexing can give good signals, but poor mounting can demolish
          | the piezo by bending too far. Having positive contact with the
          | shroud provides "pressure" sensing even if the contact is not a
          | discernable bump. "say if it sidles up along the wall and pressure
          | builds gradually"
          |
          | I think there is good merit to the concept of mounting the piezo
          | from one end and weighting the other. the end mass can flex the
          | material in a bump. But, can it detect the "pressure" effect that
          | doesnt include an abrupt inertial change. I suspect not.
          |
          | I particularly like the weighted end version because it could
          | dispense with the shroud. I will try it out. It would certainly be
          | easier to mount.
          |
          | I think it might be best to use some other detection scheme in
          | conjunction with the piezo to overcome the "pressure" sensing blind
          | spot. The optical mouse sensor has been on my mind for that. Just
          | comparing wheather the bot is moving when it's supposed to be could
          | get it out of any jam that the Piezo might miss.
          |
          | I've also been curious if the piezo cable material might easier to
          | mount for shroud pressure sensing. See:
          | http://info.digikey.com/T022/V5/0936.pdf
          |
          | this material is used for sensing traffic at intersections, as well
          | as vibration sensing to make microphones for acoustic guitars. I'm
          | currently trying to aquire some of this cable to experiment with.
          |
          | Keith Rowell
          |
          |
          | --- In SeattleRobotics@y..., "warrenhb" <warrenhb@e...> wrote:
          | > I was inspired by an article in the book Robot Builders Bonanza
          | > and by having many times passed by the bin of piezo transducers
          | > at Vetco wondering what I could do with them. The article tells
          | > how to use similar piezo transducers, usually used as alarm
          | > speakers in smoke alarms, etc., as touch sensors.
          | >
          | > Returning to Vetco and looking at the sensors more closely, I
          | > noticed that they didn't seem to have places to solder contacts.
          | > I talked to the proprietors - they pointed out that there is a thin
          | > metal coating on the piezo button to which you can solder a wire.
          | > I also found one transducer in the bin to which wires had been
          | > soldered.
          | >
          | > The price on the bin was 85 cents each. I asked for a deal on 10
          | > transducers, and received 10 for the price of 8.
          | >
          | > Trying it myself, I found that if you're quick, it isn't hard to
          | solder
          | > a wire to the thin metal coating. If you're not quick, your iron
          | can
          | > melt away the coating, in which case you can try again on another
          | > spot. Solder a second wire onto the brass plate, either on the same
          | > side or the opposite side.
          | >
          | > Connecting a crystal earphone directly to the transducer, I found
          | > that dragging the transducer over any surface produced enough
          | > sound to be heard in the earphone.
          | >
          | > The article warns that with sufficient deflection of the piezo
          | > material, an arbitrarily large voltage may be produced, so when
          | > connecting the sensor to a microcontroller, you should protect
          | > the input with a Zener diode.
          | >
          | > I reasoned that in addition to this, it may work to add a standard
          | > half-wave voltage doubler circuit (using two capacitors and two
          | > diodes) which would output a positive voltage corresponding to
          | > the peak-to-peak voltage envelope from the transducer, so that
          | > the microcontroller would always receive a positive transition no
          | > matter how the transducer was tweaked by whatever touched it.
          | >
          | > I've been investigating this at NSRG (North Seattle Robotics Group)
          | > meetings. So far, I've figured out that since the transducer is
          | basically
          | > a capacitor, the first capacitor of the voltage doubler circuit
          | isn't
          | > needed. So, my current circuit consists of a zener diode across the
          | > transducer, then a diode to a capacitor. I've been trying
          | different sizes
          | > of capacitors and parallel load resistors to get the time constant
          | right.
          | >
          | > I also investigated the arbitrarily large voltage capability at
          | the last
          | > Weird Science meeting, by connecting the transducer to a neon bulb
          | > and striking the transducer on a table - this generated enough
          | voltage
          | > to light the neon bulb, which requires at least 90 volts!
          | >
          | > At the SRS meeting someone mentioned having once built a touch
          | > screen by mounting an LCD screen on three such transducers, and
          | > deriving the touched position from the voltages generated by the
          | > three transducers from the pressure of a touch. He said this
          | requires
          | > amplifiers that can measure the voltage generated by the
          | transducers
          | > without draining the voltage off, i.e., a very high impedance
          | input.
          | >
          | > At the last Weird Science meeting, someone gave me a bunch of
          | > circuit boards from smoke alarms - each of these should have an
          | > electrometer on it, which is just such a very high input impedance
          | > amplifier. I plan to try using these to make such a touch panel, or
          | > perhaps a touch dome on a robot.
          | >
          | > Doug Bell
          | >
          | > ----- Original Message -----
          | > From: "sdk6772" <sdk6772@y...>
          | > To: "warrenhb" <warrenhb@e...>
          | > Sent: Monday, July 29, 2002 2:23 PM
          | > Subject: Re: Piezo touch sensors
          | >
          | >
          | > | Hi Doug,
          | > |
          | > | Do you have any good technical details or info you can send me
          | on the
          | > | project you discussed at the last meeting? I'm trying to write
          | up
          | > | some notes for the meeting minutes but wasn't there to see
          | things
          | > | first hand.
          | > |
          | > | Thanks,
          | > |
          | > | Steve Kaehler - SRS Secretary
          | > |
          | > |
          |
          |
        Your message has been successfully submitted and would be delivered to recipients shortly.