Gearing a Rotating Arm

[kitare102]

Our current robot design includes a dual-flywheel shooter on an angle-adjustable arm which we are currently trying to figure out how to rotate. For reference, the arm can be approximated by a mass of 3.5kg at .3m from the pivot. Our question is what would be a good way of gearing the arm such that it is responsive. To make another approximation, let's say "responsive" means traveling from horizontal (0 degree) to 45 degrees in .5 seconds, which should really be taken as a guideline over a restriction for the design.

The current model involves a versaplanetary (motor and stages undecided) followed by a sprocket stage which is necessary for the placement of the gearbox but can be made a reduction anywhere from 1:1 to around 4:1. Is a versaplanetary a good choice for this situation? If so, where to go from there and if not, what other options have teams used?

[asid61]

JVN calculator is my go-to for rotary stuff. Just plug in your ratios and it spits out speed.
For your application, I get a speed of 0.27 seconds to go 90* if I used 2 miniCIMs on 100:1 versaplanetaries (no extra reduction).

[pmangels17]

Just a heads up, that speed sounds like it could be quite fast if you are looking to control the arm accurately, and you might want to play around with it a bit. I don't know the specifics of your mechanism, but you could probably mess with the chain reduction part of it to get it to a quick controllable speed in any case, and that would probably only require the minimal maintenance of changing sprockets. This is by no means a must-do, but it might be something you'd want to keep in mind.

[JohnSchneider]

No arms please

[waterbott]

Originally Posted by JohnSchneider

No arms please

Out of curiousity why would you suggest against that? Our team is looking on doing something similar

[Joseph Smith]

Originally Posted by JohnSchneider

No arms please

The Second Amendment begs to disagree

[Jarren Harkema]

Originally Posted by waterbott

Out of curiousity why would you suggest against that? Our team is looking on doing something similar

Two teams wanting to work with severed arms? Wow... Where are you sourcing these spare body parts?

[Ginger Power]

Originally Posted by waterbott

Out of curiousity why would you suggest against that? Our team is looking on doing something similar

Their 2015 Robot could have something to do with it...

I think the teams in question are dicussing a very different kind of arm though. If I'm picturing it correctly, it's just a pivoting intake/shooter on a lever arm. A very simple concept as there is just 1 range of motion to worry about.

[Joseph Smith]

Back in 2011 when we had a long arm, we used gas springs on moment arms to balance the weight of the arm. If you do it right, the arm will stay at any position you put it without having to stall the motor to hold it there. A sprocket or gear bolted directly to the arm, driven by a miniCIM in a 100:1 planetary with a decent secondary gear/sprocket reduction should get the job done.

[D_Price]

You may want to tone down the speed if you are rotating the shooter up and down at those angles unless the "boulder" is firmly secured you could lose it when rotating.

Originally Posted by kitare102

Our current robot design includes a dual-flywheel shooter on an angle-adjustable arm which we are currently trying to figure out how to rotate. For reference, the arm can be approximated by a mass of 3.5kg at .3m from the pivot. Our question is what would be a good way of gearing the arm such that it is responsive. To make another approximation, let's say "responsive" means traveling from horizontal (0 degree) to 45 degrees in .5 seconds, which should really be taken as a guideline over a restriction for the design.

The current model involves a versaplanetary (motor and stages undecided) followed by a sprocket stage which is necessary for the placement of the gearbox but can be made a reduction anywhere from 1:1 to around 4:1. Is a versaplanetary a good choice for this situation? If so, where to go from there and if not, what other options have teams used?

[GeeTwo]

If you can swing that heavy an arm that fast, why not just remove the dual-flywheel shooter and its motors and throw the boulder a bit faster?

[Chris is me]

Arms can be tricky!

I would avoid driving a long lever arm with a live 1/2" hex axle, and I would avoid driving one directly off a VP, especially 100:1. The 10:1 VP stages are the least strong, and arms exert a lot of shock loading that can really do nasty things to the thinner gears of the VP.

How many positions does the arm need to be in? 2, 3, 4? If at all possible, using pneumatics to rotate the arm between a few positions will simplify control a great deal and is generally more durable.

If you must motorize a long lever arm, it's probably best to do the final reduction for the arm using chain, with the sprocket bolted to the arm to transmit torque. 973 RAMP has some good videos on how to design a simple durable arm power transmission system. The chain can deal with some of the shock loading the arm will face a bit better than a final gear stage could, as there is some flex in the chain and the sprockets are engaged at multiple teeth versus 1 or 2 teeth for a geared system.

Your gearbox will almost definitely backdrive, though you may be able to feed-forward the motor at a small voltage to prevent this. Only do this at low voltages or your motor will heat up quite fast.

[kitare102]

To answer a few questions and raise some more...

The arm is relatively short, and only rotates about one axis. The ball is contained by the arm as it moves, so the ball falling out isn't a concern.

Now, according to the chart here: http://content.vexrobotics.com/vexpr...e-20151221.pdf (page 25, miniCIM simple load ratings) a miniCIM is not recommended for a versaplanetary at a reduction of 100:1. How should I interpret this?

Thanks for the help so far, everyone.

[JamesCH95]

Originally Posted by asid61

JVN calculator is my go-to for rotary stuff. Just plug in your ratios and it spits out speed.
For your application, I get a speed of 0.27 seconds to go 90* if I used 2 miniCIMs on 100:1 versaplanetaries (no extra reduction).

1) JVN Design Calc is very good for this sort of thing

2) Read and understand your chosen transmission's applications manual there are a number of motor/gearing combinations that quickly exceed VexPro's recommendations (or bane-bots, or whatever). For example: running a mini-cim with a 2x10:1 vesaplanetary is counter-indicated, but running a mini-cim with a 9:1+4:1+3:1 gives you 108:1 reduction and is approved. I recall that when BB planetaries were first used widely in FRC there were a number of complaints about the transmissions failing, and it was because they simply weren't rated to deal with an RS775 geared down 256:1 (for example).
[moving on from addressing the above post]

3) Arms can usually stall to hold position, but you must make sure this current is acceptable (see JVN calc for this). I've designed arms that stalled RS775s to hold position with no issues, use your judgement here

4) If you counter-balance, and I think most arms should be counter-balanced, I suggest working out the spring geometry to counteract gravity. I.e. you want the most counter-balancing torque when the arm is horizontal, and the least amount (or zero) when it is vertical (meaning its CG is over the pivot point).

5) Some amount of damping on your counter-balance is fantastic for smooth control. We really like using air cylinders to counter-balance because they provide a constant restoring force, damping, and mechanical hard-stops all in one.

6) Back to actual arm gearing... my general practice is to pick the power I want the arms to have, which drives my motor decision. I gear as much as possible with a planetary transmission, staying inside the MFGs load guides, then I do I final sprocket reduction to the arm itself. Sprockets let me keep my drive motors down low in the robot, and generally help with arm packaging (many small details that I won't go into here). Chain is also a great place to couple on an encoder for arm control.

Examples of 95 robots' arm gearing:

2011: 2xRS775 + 64:1 bane bots + 14:72 sprocket reduction (IIRC), motors were mounted under the 'deck' of the robot, shoulder joint was near the height limit. Motors stalled to hold position. Some minor pneumatic counter-balance

2015: 2xCIM + 50:1 versa planetary + 16:64 sprocket reduction, pneumatic counter-balance kept arm neutral without game pieces, motors stalled to hold position. This robot needed over-balance protection in code since it could pick the arm+game piece up fast enough to flip itself over.

[JesseK]

Originally Posted by JohnSchneider

No arms please

Woah woah woah! You guys were the masters of arms last year!