gear shifting

[Jack Jones]

Originally posted by Stephen Kowski
Any ideas are appreciated!

Thanks
JSK

OK - I specialize in "Any Ideas"

Welcome to the club when it comes to shifting problems. The trannys on CD8 have had the engineers grinding their teeth since day one.

In my humble ( newbe parent/research-engineer, vis-a-vis FIRST-mentor/real-engineer ) opinion, you won't get that transmission to shift reliably without some major enhancements such as the forked slider collar that PMGRACER mentioned. Even with that, I don't know if I'd trust it to shift me to victory in the finals at the NATs. With a 3.5:1 ratio change, the odds are of it aligning are slim. Seems to me that, rather than using brute force, some sort of synchronization is needed. But simply filing lead-ins into the plastic ring gear will not work, at least not for long - as they will surely blunt.

FWIW from a freshman's parent, take a look here < http://www.dewalt.com/us/articles/ar...ordless&ID=695 > Maybe the ALL METAL gear set would hook up, or maybe (if it's kosher) a DeWalt/Bosch switch would do the trick. (How about that 3 speed tranny??) Simple and FREE pre-test would be to visit local tool store and check the DeWalt shifter?

I have some "pie-in-the-sky" ideas on how to synchronize. But it would involve considerable bandwidth and a whole bunch of hand waving; contact me directly if you're interested.

[JAlpert]

Team 469 shifted the drills with the servo in 2001. The problem is that when you first apply power to the servo, it pulls the collar to the neuteral position. The collar will hardly ever actually shift into the desired gear on the first try. I believe our program called for two separate power bursts to the servo - the first to disengage the collar, and the second to mesh the other gear. Current must also be applied to the motors in between the two servo bursts in order to promote meshing. You may even want to do this 3 times. Our servo shifting routine actually ran independent of the main body of the program, so the loop times would be smaller and the duration of power to the servo was more consistent. It was pretty cool becuase everything else shut down while the shifting occured.

The shift was accurate 5 times out of 6 when the robot was at a dead stop, and 1 of 2 on the fly. Shifting on the fly with those drill trannys sounds VERY ugly, however.

Good luck with your drivetrain!

Jeff Alpert
Team 469

[PMGRACER]

A new idea I have been toying with for shifting is totally inspired by a local Transmission builder-Liberty's Gears- in Taylor Mi. They build custom Drag Racing Trannys that you can shift effortlessly under full throttle accelleration. Their design has a synchronized 1st gear, the rest are unsynchro'd. The way it works is very simple. The "slider" has only 12 teeth with a .500 gap in between each tooth. On the face of the gear upon which the slider grabs, there are 12 matching teeth with the same gap. The front edge of the teeth have a lead in angle cut into them to help "grab" the slider. I had this modification done to my Ford Toploader that I ran in my Camaro 10 years ago. They call it "Pro Shifting" and let me tell you, I could shift the trans under full power @ 7800 rpm!! I'm sure if I apply that same technology to our planetary, we'll be shifting even smoother than we do now. With the amount of play that is inherent in that design, having the slider only partially engaging in low should go away. Sometimes ours will not go fully into low because there is a neutral position between the 2 planet sections abd the slider is "floating" momentarily. This has to be, since you cannot be simultaniously in high and low. Pneumatics, to me are a better way to shift. It requires less code, and is virtually trouble free. I suppose you could shift usig a Globe motor. You'd have plenty of power, but still need more code than pneumatics. Just my 2 sense...

[Stephen Kowski]

Well thank You all for your suggestions we finally figured out how to consistently shift this year's gearbox. took us a little over a day of work, but it will pay off at nationals

Thanks Again!

[Eric Bareiss]

We use the drill shifted using one central servo. To fix the meshing problems we took a great idea from the Cheesy Poofs' 2001 bot.

Our programmer rigged us up a little program so that whenever you switch gears, it stops the robot and moves back and forth really fast for a cycle or two while it switches so the teeth don't grind.

I didn't write the program but it works really well. The guy who wrote it goes by wakezero on CD. PM and i'm sure he would send you the program.