for fun and boredom

[mjt902]

A friend and I are on team 902. We're wanting to spend the rest of the school year working on a sweeeet transmission for our robot next year. We want maybe a 5 speed stick shift with the steering wheel, gas (not actual gas) and brake pedal system. I think i can figure out the gas and brake, the steering will come to me eventually, and even the control system for an automatic/manual stick shift doesn't seem that far out of my scope. However, the delimma comes when we actually want to build the clutch system for the shifting.

The teams that 'shifted' in this year's competition had to stop and then manually shift into another gear, thereby bypassing the necessity for any clutch system.

What we want to do is more like a car, but we simply don't have the knowledge, nor the resourses to do so. We would like to make the decision as to whether we want to have lots of torque (like our last robot) or be dangerously fast (if the game should require a fast bot). There's a friend who can get me the gears, so the clutch system is all that stands in our way.

The building process as not stated yet, so if anyone has any wild or even retarded ideas, please share. I've got oodles of time on my hands and can't wait to start another robotics project.

Lastly, I wanna say: BIANCA!!

226 (and i know there are more I just can't think of them) can shift on the fly

[mjt902]

thanks eko!!! i'm so happy to hear this. i'm even happier to see that its tema 226, our good friends who experienced a wild ride with us (902) in cleveland. if there is anybody that could help us with this quest on your team, i would be in your debt if you could inquire about it.

just for a thought- were you the volunteer at western mich that i talked to a few times, once right before our tragic radio cable failure in finals??

oh well... much thanks

[Lloyd Burns]

If you want to make an air-operated brake, make sure the power-off default is on only if you won't be moving too fast at 2:00 :-P

Kincardine (781) had a braking system system independent of its motors, this year.

[Scott Duhaime]

you can shift on the fly with out a clutch. this year we used an air cylinder to changed gears by engaging them with a 3 tooth dog. This is by far the easiest way to do on the fly shifting.

Cars use clutchs because it wouldn't make sense to turn off the engine every time you want to change gears (changing gears on the fly does not mean you want to keep throwing power into the system as you shift to hard on the shifting mechanism)

Some good pics on our 2003 drive

http://www.chiefdelphi.com/forums/pi...gle&picid=3611

http://www.chiefdelphi.com/forums/pi...gle&picid=3601

http://www.chiefdelphi.com/forums/pi...gle&picid=3707

also look in the white papers for team 45's 6 motor drive from 2002 that what we based ours from.

Anything else you want to know about our drive let me know

[Jnadke]

From what it sounds like, you want it to drive like a car like our current year's robot. You will need an Ackerman steering and a differential to allow the robot to move correctly. For more information and pictures, go here. Ackerman steering works much like a car steering, it points the front wheels in a direction and then the rear wheels drive it forward or backward, following the direction of the front wheels. The robot will generally move in an arc. A differential is required because the rear wheels are so far apart that the robot actually follows 2 arcs. Obviously, one of the arcs will be longer than the other, so one wheel has to make more rotations than the other. A differential allows this (mechanically), without having to skid wheels across the ground.

The positive side about the differential is that the motors are coupled together to make 1 output, since the rear wheels operate at one speed. For this reason, you'll want to use the CIM motors for your drive system (the drills have timing, they spin faster in one direction than the other). The singe output means you'll only need 1 gearbox, cutting the amount of parts you need in half, but also the parts of the gearbox/differential will need to be twice as durable, since they need to cope with the stress of 2 CIM motors instead of 1. I recommend no less than 3/8" shafts and keyways. Then, use 2 pneumatic cylinders to create a 4 speed on-the-fly shifting gearbox (2 reductions per cylinder). Each cylinder will shift between 2 gear choices. As each engages, you'll have a different gear ratio, giving you 4 different speeds. If you space the reductions enough, you'll have a very low speed, a medium-low speed, a medium-high speed, and a high speed. There are many whitepapers around here on how a 2 speed shifting gearbox works. Just copy it twice and do the math for the reductions. If you have a 2:1 reduction (gear up) and a 1:2 reduction (gear down), copied twice, the possibilities are 4:1, 1:1, and 1:4. Essentially you have a 3 speed using 2 shifting stages. Now the trick is to change those so they don't match, such that you have 4 different reductions. Now you have a 4 speed gearbox.

You don't need a clutch because you are using an electric motor and not an engine. A engine can't afford to stall, because it is a combustion process and must be continually running to continue the process. If the process stops when you are engaging gears, you'll either shear some gears from the massive torque, sieze up your engine, or bust a piston. In either case, it won't be pretty. Electric motors don't have that problem because you can turn them on and off in an instant, and nothing seriously major goes wrong if they stall.

The on-the-fly pneumatic shifting works by brute force. As the robot is moving, the pneumatic cylinder forces the gear off the current gear that it is locked with (they slide apart), and slides it along the shaft until it hits the 2nd gear. These cylinders have anywhere from 30-180lbs of force (30 is plenty). As the opposing gear rotates (becasue the robot is moving) eventually the gears will mesh and the cylinder will force the gear to slide into position. Usually the gear slides on a keyway, so that no matter where the gear is positioned, the entire shaft rotates. I suggest you hit the whitepapers section of the website if you want to read more on this and see some designs. I recommend the document on a 2 speed gearbox designed by 116 I believe. It's the easiest to build and the easiest to understand. The design (above) by 226 would be a good (and easy) one to implement as well. I can already think of a good picture of how it would look and work.

If you don't understand how gears/gearboxes work, then you have a lot of catching up to do.


This task is daunting for even a veteran team to pull off. If you are truly serious about it, do some research. Anyone can do it, it's all about the amount of time you have. If you are truly serious about it, then I guess we'll see you next year with the design. There's no doubt about it you'd win a technology award if you can pull it off.

[ZACH P.]

A place I know of to get such a differential is www.sciplus.com just go to drives, and then wheels.

[Jnadke]

Originally posted by ZACH P.
A place I know of to get such a differential is www.sciplus.com just go to drives, and then wheels.

The bad thing is that differential is only 21" long (axle). Unless you want to make your robot narrow, you'll want something longer.

It's not too hard to create a differential if you understand anything about gears and have the machining tools available. Just look at the page I provided. I gave some very basic instructions on how it can be done.

For gear ratios on the shifting system do it like this. Caculate a raw gear ratio (including wheels) such that your robot would move at 8 FPS. Then create a 2:1 and 1:2 option on the first shifter, and a 1:1 and 1:2 option on the 2nd shifter. You'll have options of 2:1, 1:1, 1:2, and 1:4, creating speeds of 16FPS, 8FPS, 4FPS, and 2FPS. I can see the advantage of this, because you can tailor your speed (and torque since they are proportional) to what competition you have.

You'll probabaly want to find the wheels first. Find something with major traction on the carpet, or something that is modifyable so that you can change the traction material. You'll need the wheels because you need them in calculating the overall gear ratios for calculating your torque and speed.

[Adam Y.]

The on-the-fly pneumatic shifting works by brute force. As the robot is moving, the pneumatic cylinder forces the gear off the current gear that it is locked with (they slide apart), and slides it along the shaft until it hits the 2nd gear. These cylinders have anywhere from 30-180lbs of force (30 is plenty). As the opposing gear rotates (becasue the robot is moving) eventually the gears will mesh and the cylinder will force the gear to slide into position.

I have a question? It sounds like you would shred the gears if you try and force them to mesh as one is moving.

[Clark Gilbert]

The cool thing about the type of shift on the fly transmission that my team designed and that Robotic Scott is talking about doesn't mess with actually shifting the gears.

This gearbox works best when shifting “on the fly.” It had no problems shifting from low to high gear, even while pushing 180-pound goals around the field. Due to the geometry of the dogs (details 11 & 12), this assembly actually prefers to shift while moving. --TechnoKat White Paper

http://www.chiefdelphi.com/pics/bin/...59100_0293.jpg

If you look at the top of the picture you can see what we call "dogs" between the gears. The dog in the middle is attached to the shaft which is moved by the pneumatic cylinder. The cool thing is that the dog is made so it will engage every time. It would all make much more sense if we had video.

[Will Hanashiro]

team 322 has been using shift-on-the-fly now for a couple years. in 2002 we had a three speed transmission and this year we had a two speed transmission using six motors. Both years we used a planetary gear box system. we also used the pneumatic air cylinders to shift, and we also found that forcing the gears to shift caused teeth to break and eventually cause them to go bald. we found that a simple solution to this is to program a split second no power delay to the motors. if you shift while there is no powers to the motors, then the gears will hold out. therefore, stopping is not needed, and the program stops the motors but the momentum will keep the robot moving.

[Jnadke]

Originally posted by Clark Gilbert
The cool thing about the type of shift on the fly transmission that my team designed and that Robotic Scott is talking about doesn't mess with actually shifting the gears.



http://www.chiefdelphi.com/pics/bin/...59100_0293.jpg

If you look at the top of the picture you can see what we call "dogs" between the gears. The dog in the middle is attached to the shaft which is moved by the pneumatic cylinder. The cool thing is that the dog is made so it will engage every time. It would all make much more sense if we had video.

Yeah, I was looking at this, it looks pretty neat actually. What do you use to allow the "dog" to interface with the shaft?

This system would be the easiest to create a 4 speed with, because you'd have the motors going to the 1st shifting stage, then I'd put a 1:1 gear ratio on the outside of the 1st shifting stage to another shaft. This shaft would have the 2nd shifting stage which is geared to the final output shaft.