Arm rotation speed?

Is a rotational speed of 39 degrees/second, no load, and 33 degrees/second loaded reasonable for our Cargo arm to be easily controllable by the driver and/or the software? Rotating through 90 degrees would take between 2 to 3 seconds.

We used the JVN Calculator yesterday evening to choose a motor and gearbox combination to actuate our Cargo arm. We would be installing one of the AMT-103V encoders (up to 2048 PPR resolution) from the FIRST Choice program on the shaft of the arm.

kind of slow, by our standards…but yeah, it should work. Winning matches takes a fast robot, though.

1 Like

We are currently at 99 degrees per second and 96.4 loaded works very well and it’s pretty fast for our standards.

This rotational speed really depends on the rotational inertia of the arm… The lighter you arm is (and mass closer to center of rotation) the faster you can get away with.

Polling other teams on this matter is difficult without specs of the arm.

There are indeed a lot of variables.

A second for full travel is usually controllable.

And it makes things better if you have at least 10 times as much torque as needed, with the motor running at 80% of free speed.

Have fun!

MrForbes,

Could you please elaborate on the 10 times the torque and running at 80% of the free speed requirement?

Thanks for the feedback. The calculated values are with a single CIM driving a VersaPlanetary with a 22 tooth sprocket going to a 60 tooth sprocket. The 60 tooth sprocket was the biggest available. We figure if we can accomodate the 60 and need it, we are okay. We can drop VP stages or go to a smaller final sprocket to speed up the arm.

The arm is about 23 inches long. I think it has around 25 to 30 pounds on the end of it. The girls left some weird number in the spreadsheet for the weight.

:grimacing:

Back-of-the-napkin-math says you better invest some hard earned moolah in some extra versa-planetary stages for when one inevitably goes nuclear due to that loading on that arm if you move it too fast.

Perhaps look into seeing if you can reduce the effective loading on the arm by making it “weightless” with surgical tubing, torsion springs, gas shock/springs, etc. You may be able to get away with a smaller motor on there with this technique.

2 Likes

Blockquote
Thanks for the feedback. The calculated values are with a single CIM driving a VersaPlanetary with a 22 tooth sprocket going to a 60 tooth sprocket. The 60 tooth sprocket was the biggest available. We figure if we can accomodate the 60 and need it, we are okay. We can drop VP stages or go to a smaller final sprocket to speed up the arm.
The arm is about 23 inches long. I think it has around 25 to 30 pounds on the end of it. The girls left some weird number in the spreadsheet for the weight.

Hi philso,

Are you sure the rotational speed numbers in your first post are correct?
For this arm load and length plus sprockets JVN calculator shows VP reduction of 300:1

This is way outside VP load rating for CIM motor.

1 Like

I suggest trying the calculation using the Mechanism tab on my design spreadsheet. It does the same stuff JVN’s calculator does, with a number of additional helpful features.

If you plug in one CIM with an arm of 23" and a load of 25 lbs, you will see that turning 33 deg/s means you’re working on an inefficient point on the motor curve, either slower than max power or higher than max efficiency. This is bad for the motor, your battery consumption, and your arm will be unnecessarily slow.

The fastest an arm under these conditions can turn under load (when the motor is working at max power) is 0.66 rps (240 deg/s), using a ratio of 67.4:1. But that draws 54 Amps, which is more than you want for an arm. If you lower that to 25 Amps, you get a loaded speed of 0.45 rps (160 deg/s) with a 155:1 ratio. You can get pretty close to that using a 63:1 VP with a 22:54 sprocket reduction.

That VP ratio is not recommended for use with a CIM motor. Look below for a better ratio combination.

Thanks again for your feedback.

The VP would have a 22 tooth sprocket on it’s output shaft and a chain going to a 60 tooth sprocket. If I have done my math/physics correctly, the VP shaft should see about 1/3 of the torque at the arm pivot. I forgot to mention that the VP has a 300:1 reduction. The overall reduction is 818:1.

I am estimating that I will have around 150 ft. lb. of torque at the arm pivot. The 60:22 sprocket pair will apply around 55 ft. lb. of torque to the output shaft of the VP. It looks like Table 5-3 of the “VersaPlanetary Load Ratings Guide” says this is a No-Go.

It looks like we will have to go back to the drawing board on the weekend and choose different (stronger) motors and gearboxes and aim for a faster travel time, say 1~2 seconds for 90 degrees of travel.

Unless I’m misunderstanding your problem, you don’t actually need stronger motors or gearboxes. You just need to lower your total reduction and you’ll get a faster arm that doesn’t break the VersaPlanetary.

A 45:1 VP with a 15:54 #35 sprocket reduction* should give you an arm that rotates 160 deg/s (90 deg in 0.56 sec), with 230 ft-lbs of torque, draws 24 Amps, stays within the VP load ratings guide, and uses COTS sprockets. That seems to meet all of your requirements.


*or a 18:60 25 sprocket reduction should give about the same specs

Also don’t forget that just because the motor can go that fast, there is no requirement that your programmers do move it that fast if you are worried about control – you can slow it down in code.

1 Like

And to add to that - if you are worried about breaking the versa planetary, consider current limiting the arm to reduce the torque. This is going to be a very common issue this year - arm loads can be huge. Consider when you use that arm to place a hatch cover and you have to “ram” the hatch cover on.

We will most likely add a bicycle disk brake to lock the arm in place once it is in the position we want.

I have plugged in the recommendations that you all have been making and they make perfect sense. I want the girls to do it themselves so they can see why it should be done rather than “Mr. So said we must do this”. Their initial design was a single stage clamping gearbox with a ratio of 5:1 to 7:1 that my gut feeling said had no hope of working. We didn’t have time on Monday evening to check the VP torque rating so I will go over that with them too.

I am also worried about the shock loads and will probably apply a margin of at least 2 to 4 X which is similar to MrForbes recommendation of 10 X. I don’t want what happened to the team I was with in 2016 to happen again.

Look at the efficiency curves - they typically peak around 80% or 90% of free speed. This means that this is where the greatest fraction of the energy from the battery is going into moving stuff, and therefore you are heating the motor and wires the least. Also, when operating at over 50% of free speed, if you encounter a higher load than expected, the system slows down and gives more power - usually solving the problem. If you’re operating under 50% and encounter a higher load than expected, the system slows down, and gives LESS power, making it slow down even more, until it eventually stalls.

1 Like

That explains the 80%thing. The 10 times more powerful thing is about making the robot a beast. Look at some of the arms on 330 robots in past years. Look at 842s elevator last year. Fun

1 Like

Understood. The team I worked with in 2017 ended up with a climbing winch that had all sorts of margin and just worked, match after match, with no need for maintenance.

1 Like

I see.
Beast is good.

I will use the opportunity to ask a few somewhat related questions.

I see many of the robots from past years using 2 large sprockets to turn the arm.
Like 330 from 2013 where sprockets are on opposite sides of the arm or 233 from 2011 where both sprockets are on the same side.

Would single sprocket be sufficient if one is using #35 chain?

Second question:
Anyone knows how 233 solved the problem of square tubes moving against each other in their 2011 telescopic arm?
I understand how their chains and rigging worked, just not sure what they have used for bearing block/slides/bushings?

First question: It depends on your design. Using sprockets on both sides can help ensure that both sides of your arm rotate simultaneously. Depending on a number of factors (arm width, arm weight, arm material properties, live/dead axle, etc) that may or may not be essential.
From a purely chain load standpoint, a single #35 is likely to be sufficient for normal loading (note that I said likely, not guaranteed). If you have a particularly long arm and are planning on rapid stops, shock loading could present some degree of concern for a chain. Having a second chain not only reduces load, but also provides redundancy against failure.
Third, build quality and especially chain alignment matters. How accurately you tension and align your chain will impact the odds of your chain failing or skipping off the sprocket.

Second question:
233 used custom bearing blocks with roller bearings
254s 2013 telescoping arm used skids made of delrin or teflon (IIRC) as fixed bearings

1 Like