1
\$\begingroup\$

Basically our project brief is to design a turn table, an object weighing around 10kg which will be placed off centre (but allowing around 20kg in case) to rotate in 0.5 degree increments, pausing for a few seconds to have a photo taken and continue through to 180 degrees. We're currently using Arduino.

We thought of having the turn table on a lazy Susan bearing to distribute weight. From research we've figured we want a stepper motor with high torque and possibly a worm shaft to prevent the table from continuing motion on the bearing once stopped. Our concern is the specific increments of movement as most stepper motors move in 1.8 degree steps. We've looked at micro-stepper motors but we haven't found one which can carry the load desired/ hasn't got enough torque. We're not super experienced in this but we'd like to play around with a prototype given reasonable suggestions.

Could anyone provide any advice about how we could set our system up for precision and efficiency?

\$\endgroup\$
6
  • 2
    \$\begingroup\$ How about a timing belt reduction setup? It will also tolerate inevitable small misalignments in your mechanical setup. \$\endgroup\$ Commented Nov 27, 2014 at 3:14
  • 1
    \$\begingroup\$ I was going to suggest a gearbox, but yeah, that could work. \$\endgroup\$ Commented Nov 27, 2014 at 3:25
  • 1
    \$\begingroup\$ Unfortunately recommendations for specific products are off-topic. \$\endgroup\$
    – tcrosley
    Commented Nov 27, 2014 at 4:36
  • 2
    \$\begingroup\$ A worm turns once per revolution of the motor. 1.8 deg motors are 200 steps per revolution. Pick a drive gear with the nice integer ratio like 180 teeth and use 100 steps or 60 teeth and run 300 steps, etc. The belt and cog pulleys gives better motor position and the M4 size is all over eBay to work with 5mm (typical stepper) and 8mm (typical rod from printer/scanner) shaft. Often sold as a couple pulleys and 2 meters of belt. \$\endgroup\$ Commented Nov 27, 2014 at 7:34
  • 1
    \$\begingroup\$ One thing I learned from programming the robots that assemble airplanes is that for precise position control you always need to have a way to actually measure the position of the object you want to move, as for so small movements one step of the motor relates to anything but that exact amount of movement of the object. \$\endgroup\$
    – PlasmaHH
    Commented Nov 27, 2014 at 10:22

1 Answer 1

5
\$\begingroup\$

Belts or a gearbox are a good way to improve your resolution, but rather than relying on the stepper and a potentially sloppy coupling I suggest you include a rotary encoder (absolute or incremental, optical, hall, whatever). Couple that with a closed-loop position control algorithm (PID is easy enough in an MCU, or even an Arduino) and you can start to make some credible guarantees about turntable orientation.

Apart from the sloppiness inherent in soft couplings like belts, a good reason to do this is that steppers can jump steps, particularly if they're heavily loaded and having to fight a lot of inertia. Unless you have a way of detecting this, you have no way of knowing that it has happened, let alone compensating for it. otoh a closed-loop position controller will (if properly written & configured) compensate for any mechanical slop, in either direction.

\$\endgroup\$
2
  • \$\begingroup\$ A timing belt isn't what I would consider a soft coupling. \$\endgroup\$ Commented Nov 27, 2014 at 6:29
  • 1
    \$\begingroup\$ @IgnacioVazquez-Abrams: I disagree, at least for certain applications. Exactly this was a huge problem for me when programming a part of the A380 assembly plant (where the hardware subcontractor lied to me, saying they had a better setup). Those things get deformed, it is just a question whether its within your tolerance or not \$\endgroup\$
    – PlasmaHH
    Commented Nov 27, 2014 at 10:41

Not the answer you're looking for? Browse other questions tagged or ask your own question.