PDA

View Full Version : Tranny help


Thunder360
01-27-2003, 02:27 PM
Anyone have a good design for a 3 speed thats Pneumaticly shifted, I can only manage a 2 speed in the given space. And Im also having a down shifting problem its a lil rough, but thats expected from helical gears right? This is all for this years Cim's I have the pneumatic set up for a 3 speed but Im coming up short on space. So any help would be cool!

kmcclary
01-30-2003, 02:10 PM
Originally posted by Thunder360
Anyone have a good design for a 3 speed thats Pneumaticly shifted, I can only manage a 2 speed in the given space. And I'm also having a down shifting problem its a lil rough, but thats expected from helical gears right? This is all for this years Cim's I have the pneumatic set up for a 3 speed but I'm coming up short on space. So any help would be cool! MANY teams are finding the trouble and complexity of a shifting gearbox AT ALL may not be worth the time and MTBF reliability risk. It'a amazing at how much you can get done with a ONE speed robot. I'm not sure that you WANT to do THREE speeds, ESPECIALLY if it is NOT running by NOW. The KISS principle should rule at this point in the game, or you may not have ANY drive.

However, If you're REALLY set on doing that, consider combining a TWO speed gearbox with a "wheel switcher". That gives you three or even FOUR speeds.

A "wheel switcher" is any assembly where you drive two wheels off of the same shaft at different tangent speeds (via different sized wheels and/or different fixed gearing), but only ONE of them touches the ground at any one time. It switches speeds simply by using a piston to drop/raise one of the wheels to "take over", or rotates a triangular assembly where the three vertices are: Motor/gearbox, Wheel 1, Wheel 2. Just be sure you use a FAST, HARD throw piston for it so the time where both wheels touch the ground is tiny. It doesn't have to move FAR, just switch tires QUICKLY. :D

Here's a simple wheel switcher from last year's CDF Gallery by Team 401, which uses a DIFFERENT motor:
http://www.chiefdelphi.com/forums/pictures.php?s=&action=single&picid=2124&direction=DESC&sort=date&perrow=3&trows=4&quiet=verbose

The big advantages of wheel switchers versus gear switchers lie in that the FLOOR is the "switching part" NOT a mechanically complicated gearbox (which could be torn up by improper switching) so you have a SIMPLE FIXED transmission, and you inherently get "switch on the fly" so you don't have to stop the machine or otherwise unload a gearbox to change speeds.

BTW, I haven't seen it done yet, but given enough SPACE there's no reason why you couldn't make a THREE tire switcher with a pair of cylinders for the three speeds, and run a FIXED transmission. A motor gearbox runs a center, fixed tire. Two fixed length bars, both pivoting concentric to the fixed tire, run off to either side. Attach a wheel to the far end of each, driven off of it's own chain set from the fixed one. A vertical piston drives each end wheel down. Select which one is to be on the floor at any one time.

I don't know if this'll come through or not, but the side view of one end of the robot would look like this (ignore the ASCII Art spacing "dots"):

....Piston....................motor............... .......Piston2
.......V.....................gearbox.............. .........V
WHEEL_B===chain===WHEEL_A===chain===WHEEL_C

The three wheels are running at three different speeds. Wheel A is on the ground when B and C are raised. Push down on either end wheel assembly (ONE ONLY) to jack up the robot onto that wheel. Note the piston's upper end will have to pivot L-R. With the pivot being concentric to Wheel A, the fixed length chain and pair of sprockets for each side wheel each individually "orbit" Wheel A.


Note to save linear space you COULD put both the B and C wheels on the SAME side, adjacent to each other. The mounting mechanics might get trickier though to provide solid torque transfer to the frame. BTW... The orbit bar has to be tough, AND kept short, to prevent over torque at the pivot point.


My REAL advice though at this point is to get a ONE speed robot running. ANYTHING running at this point is a good thing. You should NOT still be in a design phase now for your transport chassis, but finalizing your payload package and practicing.

If you ARE intent on additional speeds NOW, try a simple wheel switcher. It can be added to any existing gearbox without much rework.


- Keith

Jeff Waegelin
01-31-2003, 09:10 AM
One of the things with a three-speed transmission is that it's almost better to build a four-speed transmission after a certain point. If you're using pneumatics, you have to have two shifters to get three speeds, and at that point, you might as well throw on a fourth speed. You'd need 4 sets of gears anyways, why not use that to your advantage. URL=http://www.howstuffworks.com/transmission.htm]HowStuffWorks[/URL] has a good explanation of four-speed manual transmissions work.

Keith is right, however. If you don't already have a gearbox designed, a one-speed drive, wheel shifter, or maybe a drill shifter (see the white papers) might be a better solution. Multiple speeds are nice, but don't spend too much time on it. We did that last year, and ended up assembling our robot in the last 3 days because we spent too much time on our transmission. Don't let your desire for a three-speed bot hinder your performance in other areas.

kmcclary
01-31-2003, 12:37 PM
Originally posted by Jeff Waegelin
If you don't already have a gearbox designed, a one-speed drive, wheel shifter, or maybe a drill shifter (see the white papers) might be a better solution. Multiple speeds are nice, but don't spend too much time on it. We did that last year, and ended up assembling our robot in the last 3 days because we spent too much time on our transmission. Don't let your desire for a three-speed bot hinder your performance in other areas. My old team also got caught in EXACTLY that transmission boondoggle last year. We had some GREAT college mentors (find some if you're near a college town), but we didn't watch one pair of them close enough and they took some students down a garden path of a too complex multigear drivetrain another team designed because it "was cool".

The problem was, they couldn't make it, and would NOT LET IT GO when the deadline passed. By the time we finally realized they were STILL on it (after repeated requests to drop it) we were finishing week FIVE with NO drivetrain in sight!
We YANKED them from it and SLAPPED together a SIMPLE drive (two drill motors in their cases bolted to a board, and a 3:1 gear set) in less than TWO DAYS. BTW, it worked FINE. In fact, we ended up locking it into "high" with the ring trick because RELIABILITY was more important than speed switching! (Those cheesy gearboxes tended to slip.) Turned out, after ALL of that hassle, ONE speed worked GREAT!

But as a result of that construction delay, we were almost DEAD. We had NO time to debug our payload, NO practice time, and basically shipped a non-functional robot to our first regional. We spent all of Thursday and half of Friday just getting it all RUNNING (we were a ten point "brick" for our first two matches <sigh>).

Once running, the lack of driver practice time started showing...

We didn't REALLY get the robot running as we wished until our SECOND regional, because of all of the OTHER things we had to debug AFTER having a running drivetrain. Needless to say, our stomach acid levels were pretty high from the stress, and no one could even WATCH other matches because we were stuck in the pits working. NOT much fun...

All of that could have been avoided if we'd simply made a simple mule in the first place.

Lessons learned:
- There MUST be DEADLINES for ALL phases of this project, and the number one construction deadline (after freezing the design) is to get that mule RUNNING.
- You MUST have a Plan B in place for each phase, AND the rime reserved to execute it.
- Once you've passed the deadline for Plan A, DROP IT, no matter HOW tempting it is to "just put a LITTLE bit more work" into it. You must know when to quit, and having hard deadlines enforces that.
- Complex drivetrains are SECONDARY to the schedule. Priority #1: Make it MOVE. THEN tweak it to make it move BETTER if you have the time.
- PRACTICE TIME MUST NOT BE SACRIFICED.

Don't forget, this year you also have the AUTONOMOUS period to debug! Depending on your strategy, you may need part or all of the PAYLOAD package running to even START debugging that phase, so you SHOULD be close to finishing your PAYLOAD construction by now.


IMHO, You're well beyond the "time to switch to Plan B" with your drivetrain. FIRST even stated that "you should have a drivetrain running by week two", and we're now finishing week FOUR.

My recommendation: Use WHATEVER is running now. If not running slap together the FIRST recommended drive and a pair of casters. Whatever you do, make it a hard goal that "before this weekend is out, we WILL have a running mule". Even ONE speed will get you beyond this point. Do NOT spend more time on the drivetrain. Instead, get to PAYLOAD and AUTONOMOUS MODE debugging asap.

Bottom line: IMO, my old team last year literally wasted the first regional's $5000, and didn't have much fun simply because we didn't stick to the build schedule and allowed tinkering to go on WAY too long. That's NOT happening this year with my new team.

THIS YEAR: We had a running drivetrain maneuvering on a "switch box" by the end of week TWO. NOW we're upgrading it with a better drive while the payload and electronics teams finish their work. However it WAS running then, AND we're set up to reinstate the original drive in less than a day as Plan B if pressed.

Trust me... If you're dead for the autonomous period, or spending all your time at your first regional in the pit fighting to simply get your payload to perform because you spent too much time on the drivetrain (or worse, you're still trying to get it to MOVE at the regional) you've just lost a LOT of the fun of the competition. Your performance will suffer without enough debug and practice time.

Good luck!

- Keith

Thunder360
02-04-2003, 03:12 PM
Well, I got it all built and it using the 2 chippy's and it runs like a dream, it has pin point up shifting, but it takes a lil more to down shift. I set it for 4 speeds insted of 3 like I was thinking. first gear is slow and very powerfull, and it works itself up to 4th gear and it will push our robot with the speed and torque about even with each other.

The only bad part is every time we shift gears we use air, and our compresser is always going. Thank god that only the drive train uses Pneumatics!!!

I also took some advice and built a second 1 speed drive train, that will bolt right into the existing holes, if my 4 speed happens to fail before, or at the compatition.


Thanks Alot, there will be pictures up soon