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

New member, surplus ER-III 'bot!

Expand Messages
  • Alan
    Hi to the list, I just bought a surplus Scorbot ER-III, and I m working on getting it working. Just a few questions, if I may. Can the Scorbase (or other)
    Message 1 of 3 , Jul 1, 2011
    • 0 Attachment

      Hi to the list,

       

      I just bought a surplus Scorbot ER-III, and I’m working on getting it working.

       

      Just a few questions, if I may. 

       

      Can the Scorbase (or other) software drive more than one axis at a time (coordinate moves)?  I get the impression that the controller does closed axis control for each of the joints, but can it control (PID loop?) more than one axis at a time?

       

      Has anyone written another control program for the ‘bot?  Or built a multi-axis hardware controller to run the arm? 

       

      I didn’t get a teach pendant, and am doubtful of finding one.  The pendant seems to be a simple RS-232 interface, with buttons to generate commands.  Is the protocol documented somewhere?   I suspect it’s similar to the control program protocol. 

       

      I’ve seen the arm move a little now, not all of servos, but I’m getting closer (jaw servo left), and then on to running Scorbase.

       

      I wrote a CNC controller program for my CNC’d Sherline mill and lathe, and I’m familiar with IK (Inverse Kinematics) from my hexapod robot control programs.  I’d like to add IK to my CNC program, and be able to run Gcode.  This would allow me to plan moves with a CAD/CAM program. 

       

      Comments?

       

      Alan  KM6VV

      Central coast, CA

       

       

    • Keith Mc
      Hi Alan, Welcome to the group! ... I too now have SEVERAL of the ER-III arms, in various stages of repair, but only one working controller (and no pendant
      Message 2 of 3 , Jul 1, 2011
      • 0 Attachment
        Hi Alan,
        Welcome to the group!

        On Jul 1, 2011, Alan <KM6VV@...> wrote:
        > Hi to the list,
        > I just bought a surplus Scorbot ER-III, and I'm working on getting it working.
        > [...]
        > Has anyone written another control program for the bot? Or built a multi-axis
        > hardware controller to run the arm? [...]
        > KM6VV

        I too now have SEVERAL of the ER-III arms, in various stages of repair, but only
        one working controller (and no pendant currently). I too am just getting the
        arm up, so I can't tell you much about the controller yet. (Been busy with
        many many other projects, but I'm trying to get back to this.)

        FYI, The ER-III is driven from a number of low voltage Pittman (~12V) DC PM gearmotors.
        The encoders are simply low res quadrature encoders implemented on the high side
        of the gearboxes (raw motor side) via a pair of LEDs and phototransistors peeking
        through either a 3-vane or 6-vane star disk.

        To get you started: I generated and posted the DD-50 cable pinout in the FILES
        area, as an XLS file, listing all of the signal names. You can quickly see there
        are patterns to the connector.

        I wish to generate a table of the various ER-III motor and encoder specs:
        - Motor: Pittman p/n, rated voltage, inductance, free speed, free current, stall torque, stall current
        - gearbox ratio for each axis
        - the encoder count per rev, for each axis,
        - each segment's base/arm lengths, for I.K. purposes
        - etc.
        so if there is ANYONE that wishes to help with that, it'd be appreciated!

        I also still need an inverse kinematic model for this robot arm (link lengths, etc.),
        as I too wish to attempt driving it from something like EMC2 or Mach3.

        BTW... I also have the linear axis, so for I.K. purposes, I'd like to consider this a
        SEVEN motor arm, vs a SIX motor arm. (The base slides back and forth on another axis...)
        We can consider the basic arm a "frozen seventh axis" version.

        At the recent CNC Workshop in Michigan, I brought in my ER-III, and talked with
        both Artsoft (Mach3 creators) AND some EMC programmers about it.

        Unfortunately, Mach3 can currently only support 6 motors. Mach4 WILL EVENTUALLY
        "support multiple instances of 6 motors", but that's a ways off, so it doesn't help for now.

        OTOH, EMC2 already supports a *9* motor machine now. Some programmers there
        were also VERY interested in my arm and said that IF we CAN create the inverse
        kinematic model for it, one of the EMC2 programmers might be willing to take a crack
        at writing a HAL model (Hardware Abstraction Layer) for the arm!

        So I feel our best bet for CNC control of the ER-III is attempting an EMC2 model for it.

        As to drivers - Given EMC2, you can use any number of multichannel bridge boards
        (12V at a few amps is rather trivial...). It is not the cheapest solution, but using
        established "known compatible" EMC2 hardware axis controls would certainly
        speed up development of an EMC2 driven ER-III arm.

        There's also the possibility I may attempt a direct drive from a Propeller MPU, but that
        is for later, after I get it running in some form. A single Propeller chip could potentially
        run the entire arm, but we'd be getting close to pin saturation so if I also wish displays
        et all, we may wish to mux some encoders, or make a second "slave" Propeller chip
        drive the motors while the first one listens to the encoders and runs the show.

        However, motor drivers for THAT could be MUCH simpler, and CHEAPER than std CNC
        axis drives. Ex: We can create opto isolated H bridges and drive it with an old PC-AT supply
        FYI, I'm one of the principals in the Y3MD (cheap motor driver) project:
        ... http://www.mirobotclub.org/mediawiki/index.php?title=Y3MD
        Seven channels of that circuit is one way to go. As one set of three WAS less than $100
        (it's an old design, so I'd have to look at newer part availability NOW), I expect the whole
        seven motor interface could probably be done for under $200 in parts.

        BUT, since this is a fixed installation (vs on a mobile robot), I was thinking about
        attempting to make a HALF bridge variant of that, for this arm, and use a DUAL (bipolar)
        power supply. THAT cuts the bridge circuitry in HALF, to only two MOSFETs per motor.
        A "Half H-bridge" version would be simply a pullup MOSFET with an opto (and resistors),
        against a pulldown MOSFET and its opto. Putting 0-100% PWM on one opto now drives
        it one way, whereas sending PWM to its complement would reverse it.

        A "raw interface" 7-axis version of this design could potentially be made for about $100,
        (plus two power supplies, and micro to drive it).

        > I wrote a CNC controller program for my CNC'd Sherline mill and lathe, and
        > I'm familiar with IK (Inverse Kinematics) from my hexapod robot control programs.
        > I'd like to add IK to my CNC program, and be able to run Gcode. This would allow
        > me to plan moves with a CAD/CAM program.

        Did you use EMC2 for the Sherline? Have you written an EMC2 HAL before?

        Do you think you can help with a EMC2 compatible inverse kinematic model for the ER-III?

        - Keith Mc. - in SE Michigan
      • KM6VV
        Hi Keith! Good to hear that there is still some activity on this list! I got my ER-III talking to the controller today, and all the servos move! More comments
        Message 3 of 3 , Jul 2, 2011
        • 0 Attachment
          Hi Keith!

          Good to hear that there is still some activity on this list!

          I got my ER-III talking to the controller today, and all the servos move!

          More comments below:

          On 7/1/2011 5:54 PM, Keith Mc wrote:
          > Hi Alan,
          > Welcome to the group!
          >
          > On Jul 1, 2011, Alan<KM6VV@...> wrote:
          >> Hi to the list,
          >> I just bought a surplus Scorbot ER-III, and I'm working on getting it working.
          >> [...]
          >> Has anyone written another control program for the bot? Or built a multi-axis
          >> hardware controller to run the arm? [...]
          >> KM6VV
          >
          > I too now have SEVERAL of the ER-III arms, in various stages of repair, but only
          > one working controller (and no pendant currently). I too am just getting the
          > arm up, so I can't tell you much about the controller yet. (Been busy with
          > many many other projects, but I'm trying to get back to this.)
          >
          I don't have a pendant either, but I'll bet if it does just a basic
          keypad and sends values via RS-232 to the ScoreBase program, then it
          might not be too hard to mimic.

          > FYI, The ER-III is driven from a number of low voltage Pittman (~12V) DC PM gearmotors.
          > The encoders are simply low res quadrature encoders implemented on the high side
          > of the gearboxes (raw motor side) via a pair of LEDs and phototransistors peeking
          > through either a 3-vane or 6-vane star disk.

          A few more connectors would make it easier to check out each servo,
          encoder and limit switch.

          How do you set up the cams for the limit switches? I've got a drawing
          of the "ruler", but no instructions or drawing showing what is measured
          (if it's in the manual, I couldn't find it).

          > To get you started: I generated and posted the DD-50 cable pinout in the FILES
          > area, as an XLS file, listing all of the signal names. You can quickly see there
          > are patterns to the connector.

          Found the pin-outs, thanks! Priced the connectors, $22 each! I'd like
          to build up a driver box using RoboClaw drivers.
          >
          > I wish to generate a table of the various ER-III motor and encoder specs:
          > - Motor: Pittman p/n, rated voltage, inductance, free speed, free current, stall torque, stall current
          > - gearbox ratio for each axis
          > - the encoder count per rev, for each axis,
          > - each segment's base/arm lengths, for I.K. purposes
          > - etc.
          > so if there is ANYONE that wishes to help with that, it'd be appreciated!

          I haven't seen any specs to speak of.


          > I also still need an inverse kinematic model for this robot arm (link lengths, etc.),
          > as I too wish to attempt driving it from something like EMC2 or Mach3.

          Does EMC have a "model" to work from? I've done legs for hexapods, and
          there are IK models for arms on the Lynxmotion website, that would be a
          starting place. I did just realize something from watching the arm that
          I probably already knew but hadn't applied: As the arm "rotates"
          around, the pulleys on the shoulder and elbow joints add motion to the
          distal joints. I haven't thought it through yet. Otherwise, a little
          "cosine law" does most of work of the calcs.

          >
          > BTW... I also have the linear axis, so for I.K. purposes, I'd like to consider this a
          > SEVEN motor arm, vs a SIX motor arm. (The base slides back and forth on another axis...)
          > We can consider the basic arm a "frozen seventh axis" version.

          IK gets really hard after about 3DOF. Too many solutions. A
          combination of FK and IK will help.
          >
          > At the recent CNC Workshop in Michigan, I brought in my ER-III, and talked with
          > both Artsoft (Mach3 creators) AND some EMC programmers about it.
          >

          I have Mach3 as well, and have exchanged ideas with the authors. I
          normally run my own CNC controller program on my CNC'd Sherline mill and
          lathe. I might be able to modify that.

          > Unfortunately, Mach3 can currently only support 6 motors. Mach4 WILL EVENTUALLY
          > "support multiple instances of 6 motors", but that's a ways off, so it doesn't help for now.
          >
          > OTOH, EMC2 already supports a *9* motor machine now. Some programmers there
          > were also VERY interested in my arm and said that IF we CAN create the inverse
          > kinematic model for it, one of the EMC2 programmers might be willing to take a crack
          > at writing a HAL model (Hardware Abstraction Layer) for the arm!
          >

          That would be great! I do have EMC on a dual boot machine, but I seldom
          run it. More fun to run my own creation!

          > So I feel our best bet for CNC control of the ER-III is attempting an EMC2 model for it.

          Yeah, I can see that!

          > As to drivers - Given EMC2, you can use any number of multichannel bridge boards
          > (12V at a few amps is rather trivial...). It is not the cheapest solution, but using
          > established "known compatible" EMC2 hardware axis controls would certainly
          > speed up development of an EMC2 driven ER-III arm.

          I have some in mind. I'd even consider swapping out the encoders for
          better ones, and running a step/dir "servo" encoder, like Gecko's
          (compatible). However I have some cheaper dual servo driver modules
          that will do the job.

          >
          > There's also the possibility I may attempt a direct drive from a Propeller MPU, but that
          > is for later, after I get it running in some form. A single Propeller chip could potentially
          > run the entire arm, but we'd be getting close to pin saturation so if I also wish displays
          > et all, we may wish to mux some encoders, or make a second "slave" Propeller chip
          > drive the motors while the first one listens to the encoders and runs the show.

          I have a few Arduino UNO boards in projects (for work and a 'bot) that
          come to mind. OR more pins on a ChipKIT UNO (PIC32) board. Some of the
          dsPICs have encoder interfaces built in.
          >
          > However, motor drivers for THAT could be MUCH simpler, and CHEAPER than std CNC
          > axis drives. Ex: We can create opto isolated H bridges and drive it with an old PC-AT supply
          > FYI, I'm one of the principals in the Y3MD (cheap motor driver) project:
          > ... http://www.mirobotclub.org/mediawiki/index.php?title=Y3MD
          > Seven channels of that circuit is one way to go. As one set of three WAS less than $100
          > (it's an old design, so I'd have to look at newer part availability NOW), I expect the whole
          > seven motor interface could probably be done for under $200 in parts.

          That could work!

          > BUT, since this is a fixed installation (vs on a mobile robot), I was thinking about
          > attempting to make a HALF bridge variant of that, for this arm, and use a DUAL (bipolar)
          > power supply. THAT cuts the bridge circuitry in HALF, to only two MOSFETs per motor.
          > A "Half H-bridge" version would be simply a pullup MOSFET with an opto (and resistors),
          > against a pulldown MOSFET and its opto. Putting 0-100% PWM on one opto now drives
          > it one way, whereas sending PWM to its complement would reverse it.

          Good update to the driver design! They've already proven that a simple
          pair of transistors and a bipolar power supply can do the job.

          > A "raw interface" 7-axis version of this design could potentially be made for about $100,
          > (plus two power supplies, and micro to drive it).
          >
          >> I wrote a CNC controller program for my CNC'd Sherline mill and lathe, and
          >> I'm familiar with IK (Inverse Kinematics) from my hexapod robot control programs.
          >> I'd like to add IK to my CNC program, and be able to run Gcode. This would allow
          >> me to plan moves with a CAD/CAM program.
          >
          > Did you use EMC2 for the Sherline? Have you written an EMC2 HAL before?

          Haven't written a HAL, but I know the concept.
          >
          > Do you think you can help with a EMC2 compatible inverse kinematic model for the ER-III?

          > - Keith Mc. - in SE Michigan

          Let me study the IK that is on the LM website. It will take some
          thought about the implications of the "pulleys" adding motion components
          to the distal axis. Interesting stuff!

          Alan KM6VV
          Central Coast, CA
        Your message has been successfully submitted and would be delivered to recipients shortly.