Gas station without pumps

2017 November 8

Ball dispenser

Filed under: Robotics — gasstationwithoutpumps @ 20:13
Tags: , , ,

For the mechanisms that deliver balls to the targets in the mechatronics projects, I’ll need to have one ball at a time.  I had considered using some sort of continuous feeding mechanism that fired one ball without getting jammed by the ones queued up waiting, but I wasn’t sure I could make something like that work, so I now envision using a servo motor to deliver one ball at a time from a chute or tube to whatever the launching or delivering mechanism is.

I made a mockup out of MDF using a scroll saw and drill press.  Once I get a chute design done, I can try it out—it might be good enough to be the final piece, if it works at all.  (There is no requirement that the parts of the robot be laser cut, though that is convenient when precise cutting is needed.)

Here is the ball dispenser, consisting of an L of MDF screwed to a servo arm. Here the arm is mounted on a Traxxas 2080 servo.

The idea is that the L-shaped part sits vertically in the middle of the channel that the balls roll down, with the inside of the L facing upstream. The first ball sits within the L. When a new ball is needed, the servo rotates to dump the ball it holds downstream, then returns to its initial position to let another ball roll into it.

The ball dispenser with a 40mm ping-pong ball for scale.

One possible problem I see with this design is that the return of the L to its initial position may get jammed on the next ball. If that is a problem, I can shape the back of the L not to allow balls in—essentially making it be a cut-out in a larger circle. I cut out that shape also, though my scroll-saw skills are not up to cutting it smoothly. If I end up using this design, I’ll probably want to recut it with the laser cutter—but I’ll probably end up writing SVG by hand and translating it to dxf using Inkscape rather than drawing it with SolidWorks or Inkscape. I think I can get the arcs and hole placements accurately done more quickly that way. (Why are the drawing tools so hard to use? Is it just me?) I probably will be required to do a SolidWorks version though, because we are expected to document all our mechanical designs in SolidWorks, and I might need to do that anyway to get the placement of the servomotor correct.

Here is version 2 of the ball dispenser showing just the MDF and the servo horn.

I might be able to cut away some of the circle away from the notch for the ball, once I figure out which parts of the dispenser are exposed to the ping-pong balls in the chute. It would be good to have the center of gravity at or close to the servomotor axis, so that the servo doesn’t have to fight gravity—that way I can get away with a low-cost, low-torque servo, which will save space and weight.

Here is version 2 of the ball dispenser with the micro-servo in place.

I used a Futaba 25-tooth spline servo and horn, because I have more of those than I do of the Hitec 24-tooth spline. I don’t know whether I have an 23-tooth spline servomotors. I don’t think so (though I did have one very tiny servo that seemed to have a completely non-standard spline). There is a list of the most common splines by manufacturer at, but there is a different list at

The servo horn is mounted to the MDF with 5mm M2 screws (the 4mm ones did not seem to be quite long enough). I drilled a 5/16″ hole (8mm) for the boss on the center of the servo horn, which allows the horn to sit (almost) flush with the MDF.

My next task will be to figure out how I’m going to put the ping-pong balls through the 4″ target holes for the AT-M6 targets. I mentioned in Starting blog for mechatronics project that the Adafruit 24V solenoids that I have won’t do a thing a 9.9V, so I’m now thinking of either batting the ball with flipper run by a servomotor. My fastest servo with a Futaba horn is a GWS S03-TXF-2BB, which claims 0.21s for 60° at 4.8V (with 69 or 5 of stall torque) and 0.17s for 60° at 6V (with 86.0 oz-in or 6.2 kg-cm). Oops, I take that back, the little Traxxas 2080 is faster (at least at 6V) with 0.11 s/60° but only 41.7 oz-in (3.0 kg-cm). I think that the servos (even the little ones) are geared for more torque than the gear motors I have.  (Note: some of the information is from the packaging for the servos, but most is from

I’m a little worried about aim with a flipper—even with a motionless ball, small changes in alignment might cause large changes in departure angle. I’ll probably have to build a mockup and see how repeatable the flipper is.

Another possibility is to feed the ping-pong ball between a couple of spinning wheels (like a pitching machine). The spinning wheels can be spun up with an ungeared motor before the ping-pong ball is delivered. This approach would be appropriate if I had to throw the ping-pong balls a long way, but the target is only going to be about 4″ away.

Still another possibility is not to throw the ping-pong ball at all, but to have a ramp that extends out to the target (perhaps mounted on a servo horn) and just roll the ball into the target. This may be the most reliable method, if I can arrange everything so that the ball dispenser delivers the ball high enough above the bottom of the target for rolling to work. A partly extended ramp with the track-wire sensor on the end could simplify the electronics for the sensor, because much less gain would be needed. I wonder whether I could also increase the passive gain by using a larger inductor and smaller capacitor—the less analog electronics I need to solder the sooner I’ll have the project done.

Leave a Comment »

No comments yet.

RSS feed for comments on this post. TrackBack URI

Leave a Reply

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

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

Google photo

You are commenting using your Google 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 )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: