pic: what do you think?

[Tom Bottiglieri]

Originally Posted by JVN

Seems like after all that hassle, it would just be easier to do a 4-speed similar to 33, or 968. (I'm especially partial to 968, but it is more difficult to manufacture than 33's).

Then again, I've found no need for more than 2 speeds.
You can get along fine with just 2, more is overkill.

John

Well If you go for 2 speeds, your probably going to have a manual tranny set up where the driver switches speeds as needed. Going more than 2 implies you are going to have an automatic that switches accoring to speed and load. (See: Team 33)

Either one works, but a 4 speed gives a hungry programmer more to work on

[RogerR]

before you take my advice, realize that i'm not an expert (yet... )

it would apearr that there isn't enough space between the gears that are being shifted. i would say you need at least double the face width of the stationary
gears, otherwise you would end up engaging two different gearsets in mid-shift.

then again, i didn't see anything saying that this drawing was to scale, so this might be a non issue.

[sburro]

Originally Posted by RogerR

before you take my advice, realize that i'm not an expert (yet... )

it would apearr that there isn't enough space between the gears that are being shifted. i would say you need at least double the face width of the stationary
gears, otherwise you would end up engaging two different gearsets in mid-shift.

then again, i didn't see anything saying that this drawing was to scale, so this might be a non issue.

you are right, it is not to scale, but as aI said, a work in progress. I figured a team would use at most three gears, a "speed gear", a "tourque gear" and a "driving Gear" for normal use. our past robots just used the drill motors. I have been in First for one yaer now and I cannot stop thinking about it .
The gears that are on the piston are attached to the same rod.

[Paul H]

Originally Posted by Greg Needel

i think that this might be problematic when it comes to meshing the gears....this is the reason that most teams use shifting "dogs"

it still looks like a sold design but will only be effective if you stop to shift

Not at all. It actually works better when it is moving. We played the whole season on our transmission which follows the same principles. Just get the gears hardened and leave a little gap between the gears being meshed and unmeshed. I don't recall how much ours is, but I can find out if you want.

Very nice transmission, it looks a lot like the one we build, did you ever see it? The nice thing is that it is much lighter than using a shifting dog, and requires much less resources. Instead of one piston, it uses two, so there is not multi positioning of pneumatics. Maybe you could use a servo instead for the multi positioning/shifting. The whitepaper on ours will be here some time soon, check it out when it gets posted.

[Kevin Sevcik]

We used a similar design for a simple 2 speed tranny this year. We also had no problems at all shifting on the fly, but we ground the outsides of the gears to points so that the would mesh better. I will say that you need to be very careful about the size, pitch, strength, etc of your gears, as it's easier than you think to strip them.

Also, I'm also questioning the need of 3 gears, when you only really need a pushing and driving gear. I'm doubting that a mid-range would really help the acceleration of your robot THAT much, considering how short the field is, and how hard it is to make a turn at full speed.

That being said, I was thinking of a way to accurately shift one of these cylinders half way. I was thinking of using a split system of pneumatics with 60 psi and ~30 psi. shifting full distance would involve flooding one side with 60 psi. to shift half way, you lower the pressure on the extended side to a little more than 30 psi, then flood the other side with 60psi. Then P1V1=P2V2 does all the complicated positioning for you. The valve system would be complicated and/or impossible. I haven't looked into that yet. but it should work in theory, atleast.

[Andy Baker]

Originally Posted by Paul H

Not at all. It actually works better when it is moving. We played the whole season on our transmission which follows the same principles. Just get the gears hardened and leave a little gap between the gears being meshed and unmeshed. I don't recall how much ours is, but I can find out if you want.

Jim told me that it was 0.015" + center distance between the gears. Shifting works like a charm with this beauty.

It was a pleasure to be next to the Killer Bees all weekend. Many people came by to see the 4-speed shifter. Heck, even I was demonstrating it to people by Saturday!

Andy B.

[Ryan F.]

Are there any pictures/specifications about the killker bees transmission. Also, I was wondering if someone would explain how the "dog" shifters work in a transmission. I've been looking at specs..such as 45's newest gearbox, but have had no luck figuring it out myself.

Sorry if this is a little off topic

[Veselin Kolev]

Here's what I think.
If you want to just go for the design, meaning make it even though you probably will never use it, awsome. I think you're just going for a challenge, which is great. As for the mechanical part of it...

Sure, it looks like it'll work and all, if you get the right gear ratios, etc... but I myself have 2 "major"issues with this form of 3 speed. You notice that the spacing on the 3rd gear is a lot larger than for the first 2 gears. That is necesary to clear the first two from engaging, but it eats up space. Second, you are literally using the gear teeth as a shifting medium. Sure, cars do it too, but cars have syncros. Sure, other teams do it a lot, but I look at their gears by the end of the season and they are worn out a lot. For robotics purposes, sure it works. For anything else, I would look for a different method.
If you get a flexible coupler on the output shaft of this transmission so it takes the forces of shifting, your gears will probably be fine. Just look up "lovejoy" couplers online if you want an example of one.

If you're interested in slick transmissions, check out the white paper on the Technokats 2003 tranny. It's a two speed, but, as we've seen from other teams, it can be modified to be much more.

Other than that, this concept can work easily. Cool idea!

[Phil 33]

Are there any pictures/specifications about the killker bees transmission.

The design is being drawn up in CAD, and the code to make the transmission automated is being perfected. Before next season however, we will have CAD drawings, code, design details, etc.

i think that this might be problematic when it comes to meshing the gears

We had almost no problems with this. As Andy said, take the spot where the gears fit together perfectly, and move the center of the shafts apart 0.0015'' and its enough to allow the gears to mesh nicely. I believe we also originally filed the edges of the gears a bit to make them mesh a little better. Also, be sure to heat treat the steel gears to harden them which will reduce grinding. Making the gears a bit wider so they have more contact surface area also helps some.

[Paul H]

Originally Posted by rforystek

Are there any pictures/specifications about the killker bees transmission.

http://www.chiefdelphi.com/forums/sh...slickest+drive

That's the thread about it, and in it, Raul posted a picture of it that can be found here: http://www.chiefdelphi.com/forums/at...achmentid=1962

[kmcclary]

Originally Posted by JVN

How will you get the pneumatic to go to 3 different positions?

Easy...

1) Stack two identical short throw pistons along the same axis, on something resembling a drawer slide. (One piston's shaft pushes the next piston's back end.) If placed on the same side as in the drawing, in-in = 1, in-out (or out-in) = 2, out-out = 3.

2) A variant of (1): Make one piston's length twice the other. You now can switch between four gears. However, switching between gears 2 and 3 without hitting 1 or 4 MAY be difficult.

3) Another way to couple two cylinders is the "square Y" configuration. Place them parallel to each other. Add a simple flat bar as a linkage between the two clevises (clevi???). The two cylinders and the link bar forms the "U" part of the "Y". The gear throw rod is then attached somewhere in the middle of the crossbar with another clevis, forming the lower tail of the "Y". By varying the stroke length ratios, and where you connect along the center crossbar, you could even have a similar situation as case (2). BTW, You'll either need to allow the cylinders to swivel from the back end, or one side of the crossbar needs to be a slot, to make this work. Otherwise, it'll bind up and probably bend a cylinder shaft. This method lends to better packaging: "Fold it over", one cylinder on top of the gearbox, one on the bottom. (Now it is a "W" linkage...)

You really should check out Team 33 (Killer Bees). They have a very tight "binary" FOUR speed "autoshift" gearbox this year, made out of only two plates, shafts, chunks from only a couple of different gear sausages, and two cylinders. VERY simple. Theirs had the two pistons in separate places. By placing an encoder disk on the wheel, the software could even autoshift it for them at predetermined speed points. They said it keeps the motors running at near peak efficiency over a LARGE speed range. IMHO, VERY cool. They said once the parts were cut and the plates drilled, it only took ONE student ONE night to assemble the entire thing.

They may be putting up a Whitepaper on it soon. (One can only hope!)

- Keith

[Madison]

I realize that this is simply a sketch, and maybe I am misunderstanding something still, but it seems that each of the three stages in this transmission gears the output up rather than down.

The final ratios would be 2:1, 6:1, and 18:1. That seems a bit backward.

[sburro]

I do not know a lot about gear boxes, as of yet, so this may be true, but I ask you to remember that the picture is just for lookes, the ratios, size ect may not be the real ones.