Turns out we have to resort to plan B, we have to abandon our forklift due to weight and time issues.
Luckily we have prepared for this to happen, and have built an arm incase the forklift doesn’t work out, or if we want to try a different manipulator (both built from 80/20, can be switched in under 5 minutes).
Now, I was thinking about using two FP motors with the KOP gearboxes. BUT, I was also thinking at the same time that it might be overkill, and we are also starting to face space constraints. Now, we HAD 2 42mm 64:1 gearboxes for the FP motors, but we wound up returning them because we decided not to use them before we even opened the box. Seeing as this is the 6th week, and since we operate off of purchase orders, we can’t sit around waiting for another order from BaneBots to come in.
This brings me to the 64:1 36mm gearboxes given to us in the kit. Could two power an arm that weighs about 15lbs max? We can reduce it even further with an 11 tooth sprocket on the gearbox, going to a 30 tooth sprocket on the arm. I think it’ll have more than enough torque to lift it, I just want some input from somebody who has a little bit more experience with them, I’ve never ran them before.
Adding more reduction isn’t a problem. We’ve got room for almost any size sprocket on the arm end. Possibly a 60 tooth aluminum sprocket from IFI, making it 11:60.
For that size arm I believe that you are not going to get enough reduction out of the 64:1 gearbox’s. I would take a look at the Globe Motors and Kayang motors, they both have more than enough tourqe to power an arm when properly sprocketed.
The problem that you would most likely have is that the 36mm gearmotor (or the globe) will not support the arm when not powered, the arm would fall down of it’s own weight. We have a similar design, and purchased a gas spring from McMaster-Carr, which seems to do a decent job of supporting the weight of the arm. We are using a BB 125:1 gearmotor, and the arm weighs about 5 lbs total, and is about 5 ft long at it’s longest. We have an 8:1 chain reduction at the arm pivot.
You might be able to use one or two of the window motors, with quite a bit of chain reduction, and have it support the arm without backdriving the motor, but look carefully at the torque it will take to lift the arm, and try to pick a motor or support the arm with some type of spring so the motor is only loaded to about 1/4 of it’s stall torque (as found in the Guidelines, Tips, and Good Practices manual)
I agree: The BaneBots trans is great, but it can be backdriven. You need to use a motor that is not backdriven, as squirrel mentioned. Some sprocket reduction - a tradeoff between power/durability and movement speed - and you should be good to go.
We did exactly that in our rookie year robot, a 12 tooth sprocket on the motor, about 48 tooth on the arm, using a window motor and the van door motor. The van door motor did the work, the window motor was there only because it did not backdrive - it acted as a brake of sorts.
In another thread, Dr. Joe said that the 36mm may fail with repeated use with as little as 100in-lbs of torque. You want to use two, so you safely have 200in-lbs at your disposal. A 5ft arm that weighs 15lbs and is uniformly distributed weighs 0.25 lb/in. Integrate from 0-60, and you get 450in-lbs. You’ll need to add more to that, because your weight is not uniformly distributed (due to your gripper at the end), and you’ll have to add 60*(tube weight) to account for actually manipulating a game piece. Roughly estimating, I’m thinking that the 11:30 reduction is cutting it way too close to be safely operating your arm. A bigger sprocket at the top (the IFI 60 or 72) may work, but as others have mentioned, you still have the backdriving problem. My team is taking our chances with that, but you may not want to.
So if we had 11:60 gearing on two of the motors, it could power the arm without a problem.
About the backdriving problem - would it be a bad idea to regulate this by applying current to the motors to keep it from backdriving, like the Victor’s brake mode?
OR who thinks I should just abandon the idea of using the 36mm gearboxes, and just suck it up and find somewhere to mount the fisher price gearboxes?
5’ long arm 2 banebot 36 mm motors, However we we are using 125:1 gearboxes with a futher 6:1 on chain.
This was done for speed tho, and torque was not the factor.
It lifts the arm no problem, back drive was a small issue but we made some modifications and now it stays where ever we put it.
If your wondering about torque, these motors can handle it. The gearboxes are something to worry about potentially, but we’ve been running it for about a day and not seen any of the problems that other people have been seeing. Its run enough for me to be happy with it. ( Maybe 200 cycles up and down? )
The arm weighs ~7.5 lbs with tube in tow. Its displaced over the distance of 5’ b/c the pvc is like .75 lbs a foot or something, but the gripper ways ~ 2 lbs and the tube ~ 1lb.
To use a Victor to keep the motor still, you would need to use a feedback loop and PID control - simply because the force on the arm pivot varies according to load on the arm, and angle of the arm. No load straight up, but max load sticking out paralllel to the floor.
The Victor brake mode will not do anything for you in this case - it is dynamic braking, and its force is linearly proportional to motor speed. Zero speed, zero braking effect. Search on Victor Brake to find the thread that explains this in detial, I think it’s from a few years ago.
We are at the same weight, just about. The arm is actually going to be made of 80/20 1515 , I just modeled it as normal aluminum square tubing for sake of simplcity in seeing my design. The one 5’ piece of 1515 weighs 6.7165lbs. (based on the 1.3433lbs. per foot given in the catalog), and our claw, by my guess, weighs maybe 2lbs, also made from the 1515 extrusion - so realistically, we are looking at ~9lbs. I just said 15lbs. as an overestimate - tomorrow I’ll actually weigh everything out so I can have an accurate sense of how much this thing actually weighs.
I’ll probably just put an order in tomorrow for a 2 72 tooth sprockets from IFI, and have them next day aired, just to play it on the safe side, 11:72 should be more than enough to lift this thing. I think we have some 60 tooth steel sprockets laying around our storage room somewhere, and we also have the 3/8" bore 11 tooth sprockets for the motors themselves, so I’ll put those together real quick with those other sprockets and see if it works.
Ok, we’ve made one change, the motors and everything are fine and the backdrive was negotiable but we went to Home Depot and bought a gas spring strut typically reserved for cars. We mounted it aprox. 5" down the arm that way at the most extreme of cases the arm was nearly neutrally boyant. The motors are then able to keep the arm completely still just by putting in the pwm jumper for the " break " ( also for testing purposes shorting the motor does the same thing )
The mounting for the gas shock and all weighed in about 1.2 lbs and will greatly increase the life and functionality of our arm. Infact the gas shock was so affective it negates the need for our second motor, however we left it on because we like high safety factors. If your worried about motors wearing out or backdrive this is a wonderful solution thats overlooked and under appreciated. If you need some pictures of what I’m talking about I’ll get someone to take some and send them your way.
This is true…I really like the 1515 due to easiness of mounting stuff with t-nuts, but you are right, using normal alumimum would take down the weight significantly. In fact I think we have some square tube just lying around in the storage room as well…looks like I’ll have to start drilling mounting holes again…
If teams are worried about arms backdriving, stop.
If you have feedback on the arm (a potentiometer is the easiest) a simple PID loop will hold the arm steady.
If the PID loop is perfectly tuned, it may cause the motors to stall as it tries to maintain a perfect angle. If you’re using the FPs (which I’m assuming you are) then you can add a slight oscillation. The oscillation can be so minimal it has no effect on scoring, and it is easy to do; Just have the programmers turn up the I slightly. In fact, anyone who has programmed a PID loop can attest to this, it’s easier to make it oscillate slightly than stay perfectly still.
Read Matt Krass’s paper on PID and you will have a much better description.
As to falling arms, this is a non issue if you do the right thing and counter balance your arm.
NOTE TO EVERY TEAM WITH AN ARM:
DO YOURSELF A FAVOR – COUNTER BALANCE YOUR ARM WITH SOME LATEX TUBE.
Seriously. This is the single best thing you can do to make your arm more controllable – More significant than increasing your gear ratio. Really. Design in a way to wrap 20 or 30 loops of latex tube on the back side of your arm pivot and you will be much much happier. Here is a rule of thumb for you: plan on about 7lbs of force per strand of latex tube – deciding how many loops you need is as easy as dividing by 7.
Example: Trying to hold up an arm that weighs 15lbs and has its CG at 30"? That means you need 450in-lbs (=15x30) of counter balance. If you can get a counter balance bar 6" behind the arm pivot, that means you need to pull on that bar with 75lbs (=450/6). To get 75lbs of force from latex tubing you’ll need 11 strands of latex (=75/7) or 5-6 loops.
As to the 36mm gearbox, I have made a prediction about them failing with 100in-lbs of torque with repeated cycling. I predicted that the 64:1 gearboxes will fail giving situations where they see stall loads and dynamic loading. While I stick by this, the number of cycles is quite high (many hundreds in my very abusive testing).
Bottom line: Given what I know now and given good counterbalance and given that you don’t make a habit of ramming your arm at high speeds into the floor or the end stops, I think that the BB 36mm gearbox could be a good choice. I would use as big of a ratio as I could in going from the gearbox to the arm (say a 96T on the arm from IFI and an 8T on the BB gearbox for a total ratio from motor to arm of 768 – that would make your arm’s freespeed be something like 20 RPM – fast but not too fast. Using 85% per planetary stage and 95% for the chain you would have 125N-m (1100in-lbs) of torque at your arm. That would give you 22lbs at 48" – Not too shabby. BUT don’t forget that you should not design your mechanism to routinely stall these motors or they will become copper slag heaps. You should really only routinely use about 20-25% of the stall torque of these motors – this would mean that you could only use about 5-7lbs out at the end of the arm – not all that great – but you could use 2 BB gearboxes and be back in business OR you could go to the 125:1 gearboxes (but I don’t really like the high ratios so you have to pick your poison).
If your still nervous about the impact load replace a couple of sections of chain with spectra string (one on each side of the drive sprocket) and you’d effectively have an infinite life mechanism.
Today I replaced the 1515 with 1.25" square aluminum tubing, taking down the weight by more than 50%. We had two 60t sprockets laying around our shop so we put them on, making the ratio 11/60 - I didn’t get a chance to run it yet - tomorrow we’ll mount the grabber and fire it up and see how it works, and see if I’ll need to make a Grainger run to get some smaller sprockets for the motors.
Thanks for your help Joe. We are running two of the 36mm 64:1 BaneBots gearboxes with the pillow blocks to power this thing.
We ordered two 10 tooth 3/8" inner bore sprockets from mcmaster for ours. Then ran them to a sponsors machine shop and begged them to place a 1/8" broach in them for us. The 6:1 is a pretty good ratio, but if your still worried about weight and lifting power put some serious consideration into using a gas strut, if only one sided arm and correct placement it can keep the arm nearly neutrally bouyant its a great great addition to any arm.