Gear ratio for arm?

I have been doing some CAD recently and am making an rotating intake (it rotates from the vertical position to the horizontal position for the frame perimeter clearance). I’ve been using the JVN calculator and it’s spitting out ratios of around 700:1 on the 775pro.

Am I doing something wrong or do motorized arms typically have such high gear ratios?

If so what gearboxes can do the job?

If you made a intake that rotated down (motor powered). What gear ratio did you use?

1 Like

Would you mind screenshotting your JVN spreadsheet and tell us more about the mechanism that is rotating. (Length, Weight, etc.)

1 Like

I have certainly seen gear ratios for arms well over 100:1, but rarely that high. In addition to seconding @RSimpson for a screenshot or at least numbers, I would suggest trying two 775 pros to see what the effect is; if you are near the (power) limit of what a 775 pro can do with your current budget per motor, you’ll see a change of much more than a factor of two.

When more than about 100:1 is needed, most teams have used chains or belts or outsize gears for the final stage.

We had a rotating arm (with ~200 degrees of travel) on our robot in 2019. Here’s a quick clip of it in action for reference:

https://photos.app.goo.gl/by5JPrgHbJNu8pXKA

We ran the arm off of 1 775Pro with a 1:147 VP reduction (7-7-3) and a 15:44 chain-sprocket reduction after that. It was a fairly aggressive gearing and the motor got very warm after matches. We even burned out some motors during practice runs at our build space.

3 Likes

775 Pro free speed of about 18krpm, divide by 700 -> 0.4hz, so roughly 2s to move 360°, 1s to move 180°. That is extremely simplified math and assumptions, but the right sort of thing to do as a sanity check.

700:1 on a big arm with a super-fast motor is totally reasonable.

Selecting a motor that spins more slowly but with similar or superior power (Neo, Falcon, CIM) will reduce the gear reduction. Counter-balancing or spring-balancing the arm can also help tremendously. I’m a fan of gas springs or air cylinders in spring balancing because they tend to dampen the movement of the arm nicely, making controls tuning easier.

1 Like

700:1 isn’t crazy for a 775pro arm. This little guy has two 775’s at 430:1, and I can tell you that it’s faster than it should be, and torquier than it should be. A single 775pro at 700:1 sounds like it might be right in the sweet spot.

It also depends heavily on on how long the arm is and what kind of weight you’re lifting though, so we’d need to see some of those details to give “real” feedback.

I agree with the above. The numbers don’t seem outlandish, but we can’t be much help not knowing more details about the system you’re designing. Since the number indicate that this might be a beefy arm, I would either add another motor or go to something like a Falcon or NEO.

If you want to use a VersaPlanetary, make sure to check out the load ratings (https://docs.google.com/viewerng/viewer?url=http://link.vex.com/vexpro/pdf/VersaPlanetary-LoadRatings). Another option could be https://www.wcproducts.com/wcprotationss. AndyMark has some good options too.

Regardless which you choose, it always helps to have a chain/belt reduction outside the gearbox. This considerably increases your mechanisms ability to take impact loads which are never calculated really. I honestly wouldn’t ever design a mechanism that could interact with the field or another robot without a chain/belt stage somewhere.

I like to use this rule of thumb for designing arms, rather than a spreadsheet:

  • The actuator should have two or three times the stall torque necessary to hold the arm in place in the worst case.

Since an electric motor has a linear torque curve, it will be able to produce half of the stall torque at the midpoint of its curve, where the power is greatest. That’s the speed the motor will run when when your arm is fully loaded. Alternatively, the stall torque being twice the load should be enough to accelerate the arm at twice gravity.

Of course, you may want to gear it down even further to get less current use and better efficiency out of your motor. For a point of reference, our 2018 double jointed, ~30" long arm was powered by 775 pros, with a 400:1 gearbox at the shoulder (giving ~250 N-m or 2200 lbf-in of stall torque), and a 300:1 gearbox at the wrist.

As for which gearbox, you have more options than ever these days! I personally like the 100:1 Andymark 57 sport because it’s bomb proof. Pair it with a 3 or 4:1 chain reduction and you can prevent some of the shock load from reaching the gearbox. Mount it far enough away from the arm pivot and you can and use a chain turnbuckle to tension the chain and completely ignore C-C distance.

3 Likes

330:1 Works pretty for arms as long the biggest sprocket is yellow.

6 Likes

I suggest using my design spreadsheet to help with these calculations. I just posted an updated version today (though the mechanism calculator hasn’t changed at all).

If you know the length and weight of the arm you’re trying to move, it can help you find the ratio that will do that with the speed/current/voltage/whatever that you want. Plugging in guesses for length and weight shows that 700:1 should work fine, but you can probably get away with something lower if that’s easier.

To reduce the reduction needed, use a motor like a CIM or NEO that has higher torque at slower speeds. Our team’s arm last year had a heavy full width ball holding Intake utilising a 35:1 VP and a 4:1 sprocket reduction with the motor at 40% power. Very controllable and the NEO never got above ambient. What really helped was a gas shock to take most of the load, id suggest that if you can.

Hey,
Below is a screenshot of my CAD design as well as the Onshape link


I revisited my calculations on the gearbox and got a number around 350:1
JVN Calculation

So from what us said above I think I would go with a 100:1 initial reduction and then use sprockets or belts to get the rest. Or I could possible switch to a larger motor.

I highly doubt the intake would be 20lb, I just wanted to overshoot.

Here is my Onshape link
CAD File

1 Like

Yeah, I did some further calculations and brought it down to around 350:1

Pun appreciated, but we ran 300:1 on a VP then a 60:12 with #35 chain.

2 Likes

I shared it, I’m not sure how to calculate weight though, you can see the arm length on the bottom right of the first image (around 19 inches). To calculate the gear ratio, I overshoot and stated 20lb as the weight,

Looking good! Thanks for the update. FYI, that is definitely not a 20 lb intake, but I understand being conservative.

You can add materials to parts in a part studio. Just right click the part in the bottom. If you brought any parts in from MKCAD, it should already have a weight assigned. If you did and it doesn’t, just let me know. Once all or most of your parts have weights, you can select them in the assembly from the feature tree (not by clicking on the faces themselves which will only select a face not the full part). A little scale icon pops up in the bottom right hand corner which you can click on to get the weight. The next step is to figure out the arm length. This is not the distance from the pivot to the end of the arm rather the distance from the pivot to the center of gravity of the arm. Onshape will also give you this in the assembly, but you might have to do some math to get a number from the pivot rather than the origin of the assembly.

Is there a particular reason you are using brushless motors everywhere except for the arm?

1 Like

Weight is important, but also how that weight is distributed on your mechanism has some impact as well. For example, if you have an intake with a motor all the way out, you can move that motor closer to your pivot which reduces your moment.

For your kind of mechanism, I would recommend pneumatics if you’re already using them somewhere else, pivot as you’re doing if you’re not using them. Our general team rule is if we can find 3+ places to use pneumatics, we’ll add them. It’s become a lot less of a weight issue now, but it’s still irritating to manage.

Hard/soft stops can also be your friend if your gearing is set and you need a good way to support it. It doesn’t work nearly as well as doing PID on a motor to hold it up, but a hardstop at the end of the mechanism’s travel can often fix a lot of issues in a pinch. Just consider that you don’t want to be driving the motor into the hardstop. In 2019, we had a preset for our elevator that was 2-3 inches higher than our elevator could actually go, and we definitely stalled a couple motors before we figured out what the issue was.

As a frame of reference, I designed an intake mech similar to yours and weighed it at 4 pounds in CAD. Using a weight of 20 pounds is overkill(ish). I can’t really say exactly what yours weighs, but there are some certain ways you can reduce weight as well:

  • Switch solid aluminum to lightened polycarbonate/Lexan (half the weight, much better for high impact areas)
  • Switch bearings to TTB bushings or similar - most intakes aren’t handling a load that high to always need hex bearings (though your use case may need that)
  • Using captive shafts, retaining rings, or tapping end (shaft collars are heavy)
  • Using a pulley/gear reduction instead of using a VP/Sport. Usually custom reductions or custom gearboxes are lighter unless you’re gearing something down a lot.

Overall, gearing is one of those things that is a lot of rules of thumb. Design informs functionality, so making things lighter in the design process means you don’t have to gear down as much, and your mechanism can rotate faster.

1 Like

One more useful rule of thumb is adding springs to equilibrate the system, greatly reduces the load on the motor and helps to hold the arm when breaking