Center pivot pen plotter

This center pivot pen plotter is an interesting take on the idea, and manages to somewhat simplify the fabrication when compared to a gantry-style built.

Normally we’d see a gantry that travels on two rails, with a print head that moves along its length. Here the gantry is anchored on just one side, with a chain driven system to rotate it along the plotting surface. The print head uses a fine-point felt-tipped marker. It still travels along the arm as you would expect, and can be tilted away from the paper for repositioning.

What was made easier in hardware ends up adding to software complexity. The benefit of a traditional system is that it uses X and Y coordinates to plot a design. The pivot of this mechanism means that as the print head moves further from the center of the machine, the distance between each pixel is magnified. But the clip after the break proves that this issue has been solved.

Comments

  1. bot says:

    just a polar-rectangular conversion, x=rcos(o) and y=rsin(o) where r is the distance from the marker to origin

  2. aweijnitz says:

    Very cool!
    Nice to see a different take on plotters.

    Reminds me of this, also very cool Arduino-based Polar Plotter.

    Arduino Based Polar Plotter

  3. Dr T says:

    I have been wondering about this technique/geometry for a 3D printer.

    Any suggestions as to why this would be a bad idea??

    • flink says:

      I think it’s a great avenue that’s begging for us all to begin tinkering!

    • Alex says:

      XY positional accuracy will decrease as the print head moves toward the end of the arm, given a fixed angular positioning resolution. This isn’t necessarily a showstopper, but it’s probably the biggest disadvantage.

      • Dax says:

        Only if you use open loop control. Closed loop systems can be arbitrarily accurate.

      • dbynoe says:

        There is definitely some difference in the movement depending on the extension of the arm. Right now my machine moves about 0.014 degrees per step. (at 1/16 micro-stepping). So on my machine that’s 0.17mm movement on the outside edge of the range, and 0.09 on the inside. Which for me is close enough as its less than the amount of wobble I get from the pen.

  4. Grovenstien says:

    The main issue I can see with this otherwise awesome design is one of rigidity. However with the right materials bearings etc there is no reason why this design could not make a good 3D printer. Ultimately it is only a couple of degrees of freedom from being a robotic arm.

    • flink says:

      It opens the way to use ever more complicated and heavier print head/extruder designs. Using arms that are sufficiently stiff and counterweighting, you are no longer hampered by weight and inertia of the print head…

      I wish work was over so I could spend the rest of the day on this idea.

      • Alex says:

        Inertia doesn’t go away with this design. In fact, with your counterweight you’ll have even greater inertia to overcome.

        • flink says:

          Really? Compared to the straight line movement on an XY table? The print head isn’t really swinging through an arc, which I could see making it really hard to control. I’m thinking about the weight of a three-filament extruders into a single, heavy hotend. That is a lot of weight to sling around in a long movement. If the head is making smaller movements and the weight of the arm is counterbalanced, then the inertia should be lower since the movement is smaller.

          Hmmm.. Maybe I mean momentum. I’m tired.

    • dbynoe says:

      Rigidity isn’t too bad on the arm, there is some play in the rotational bearing (it is meant for use as a heavy duty lazy susan) so when the pen pushes down it actually lifts the whole robot a smidge. I am using a commercial linear slide rated at about 50 pounds downward load. Its meant to be supported at both ends but seeing as its a formed steel channel its pretty much bombproof for this type of load.
      The main issue I could see with using it as a 3D printer is the amount of backlash I am getting on my rotational axis. The chain drive has a lot of play, which I get around by only driving it one direction, a cheap hack but effective till I redesign it. The chain was chosen mostly because it afforded a large gear reduction, let me do a big rotation, and was free, versus lots of money for large tooth count gear.
      The other thing I notice, even with the extra belt and reduction, the rotational motor is a lot more likely to slip and miss steps just because its moving the entire body of the robot, plus the arm, plus the motors, etc etc. And that is with running it at about a fifth of the acceleration of the linear axis. Not a huge issue, just put a larger motor but definitely a concern.

  5. Andrew says:

    Will it actually plot vectors? The demo videos show it simply rasterizing a bitmap. Why not send HPGL to it?

    • Beat707 says:

      I was wondering the same thing. ;-)

    • dbynoe says:

      Mostly, the code for writing a bitmap was easier.
      The motor drivers I am using default to receiving movement commands (ie move forward 200 steps), they are not currently linked so getting a synced movement is a pain in the butt unless I go to a step clock driving mode, which would mean rewriting the code.
      Combine that with the amount of backlash I am getting on the rotation axis and my open loop control system and a standard line by line raster was a more effective way to get this thing drawing.

  6. Jérôme Vuarand says:

    In the clip the polar coordinate problem is not solved. You can see that the lines are closer together near the top of the drawing (near the pole) than at the bottom. When you look at the overall logo at the end it’s clearly squeezed near the top, and the top filled arc looks darker than the bottom one.

    • Galane says:

      Yup. Not a good design for raster printing. Pen plotters of any type are very poor raster printers.

      They’re made for plotting vectors, which translates nicely to 3D printing and laser or knife cutting.

      Using a high resolution servo motor for the arm swing, a polar vector plotter can be as accurate anywhere in its print area as an X/Y plotter, but there has to be a significant amount of “excess” resolution to compensate for the increased sideways movement per “step” as the pen moves out towards the end of the arm.

      Software complexity goes up a lot to get the same precision throughout the plotter’s range.

    • dbynoe says:

      Yup, in that clip I was actually just feeding the X Y points straight to it. I had to get the thing running for our local makers faire.
      I since rewrote it to do the coordinate transformation, and its mostly accurate now.

  7. kabukicho2001 said says:

    kaganov said, solution: use a octocopter printer.

  8. Whatnot says:

    You are suppose to use vector systems for plotters not treat them like dotmatrix printers.

    Don’t feed it damn pixels.

  9. CRJEEA says:

    You could always print without compensation and then view it with a cylindrical mirror (:

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 96,656 other followers