Slides the gears to shift. Rather large, but a new approach on 3 speeds in the FIRST Robotics world.
Note: the shaft should be longer so it reaches that plate.
Slides the gears to shift. Rather large, but a new approach on 3 speeds in the FIRST Robotics world.
I think someone hit “start discussion” but forgot to post a reply. So, I’ll start it.
What do y’all think about this slightly different approach to three speeds? Although a little large, seems like it would be very simple to fabricate and not too heavy. And maybe you could place the motor(s) on top of the gearbox to compensate for the wideness.
PS. If anyone decides to further develop a similar design (or actually build it), I sure would like to see it! Thanks.
What are you planning to shift with? Pneumatics or another method?
Also, are you going to do anything to make sure that the gears mesh right away - make sure shaft 1 and shaft 2 rotate at the same speed, round off the teeth a bit (given that the gear is thick enough to do so), etc.?
One more question (probably with obvious answer, but I’m sleepy :p), the middle shaft isn’t just going to be hanging like that in first gear, right? It’s going to be extended through the plate?
Anyway, looks cool, I like.
The shaft would be exntended, no hanging in mid air.
I don’t actually plan to build this thing, but if I did here’s a few things I would do.
Actuation: Not really sure but I’m thinking pneumatics for speed and strength. Might try multipositioning but maybe just use two cylinders in series.
Gears: Might try 3/8" wide to save space, but 1/2" is probably a safer bet. 20 pitch seems to work well and be readily available. I suppose beveling the edges a little bit would aid in meshing, probably adding a couple thousandths in the c-c distance would as well. Other than that, I wouldn’t match speeds or anything, it should be fine
Any more comments/ideas?
First off, great idea Sanddrag.
I would actually like to experiment with this gearbox and see if I can design a complete drawing if you give me permission to do so. (I would have to pick up this project after the offseason that I am working on now.)
If I was to design this gearbox, I would make it much smaller as you did point out before. Maybe the output (sprockets) can be placed inside the gearbox. Now in order to do this design I would like inputs and ideas about how to shift (please share exactly how you would do it). Thanks in advance. 
I’m not exactly sure how to make it smaller without going to smaller face width gears. The thing that takes up all the room is the fact that you need to get all the way out of one gear before you move into the next one. For making up for the bigness, I had thought of putting the motors within the width of the gearbox (on top of it or something). Also, there is definitely room for sprockets inside.
To shift, I was envisioning some sort of rigid piece that couples the threaded end of one cylinder to the body of another cylinder. Maybe the end that screws on to the threaded cylinder rod can also have a larger diameter tube that slides freely over the body of the same clylinder for more support, less wobble, etc. Anyway, you would get three possible positions both extended, one extended one retracted, and both retracted. Of course this would end up being a long assembly as well. I’m just not sure how else to do it. I don’t really know anything about multipositioning but can it work? Maybe even lead screw but that doesn’t seem fast enough or robust enough.
And of course you can go ahead and design it all up fancy. I’d really like to see what somebody can do with it. 
57 uses a tranny like this but with just 2 gears.
You could make it smaller if you went with a rather counter intuitive shifting pattern. Call 1 the biggest gear on your shifter. If you arrange the gears on the shifter as 2-1-3 instead of 3-2-1 then you could have all 3 gears next to each other. Then your whole assembly would just need to be 7 times the width of a gear plus some tolerances. Downside of course is that you’d have to go through your high gear to get between your two low gears. Or low between two highs depending on which side of that is driving. Either way it’s bad news if you’re planning on using it like an automatic transmission in a car.
Also, beveling both meshing gears actually doesn’t help a bit. Beveling just one helps a little. Actually rounding or beveling the sides of each tooth helps the most and hurts your tooth face width the least. Of course it requires a good hand with a dremel and a lot of patience.
Shifting wise, we use a brass bushing with a shaft press fit inside of it that’s threaded on one end and attaches to a pneumatic. The two gears are welded together and free spinning on the shaft on some needle bearings. I can post a section view if that’s incredibly confusing.
This is a proven design, many early cars had transmissions like this. Of course this was long before synchro’s were introduced. A good example of a four speed of this design can be found in a book sold through Lindsay Publishing that has a collection of articles from the turn of the century. Included are the plans to build the entire car complete with two cylinder two stroke motor as well as steam turbines, sterling engine and a hand operated keyway mill.
Go here to see more.
Lots of teams use a tranny like this with only two gears. That’s where I got the idea. But my idea was to expand it to make it a 3 speed.
As for beveling the teeth, why would you dremel each one as opposed to chucking it up in a lathe and filing it or even making a light cut?
I think it will be more powerful than the three-speed in my car…
I’d have to make a drawing to show you what I’m talking about, I think. Lathing the gear would get you something looking exactly like if you’d done the same thing with a rod or other cylinder. I’m talking about beveling each individual edge of each tooth.
The point of this is to reduce the cross-section that has to mesh. If you bevel both gears, then it’s pretty must the exact same cross section that has to mesh and line up. If you bevel each edge like I’m talking about, then initially it’s like meshing two gears with much smaller teeth with lots of space between them.
Given the way we build gearboxes you might be correct. 
but not really, I think if a first robot used all the motors combined you’d still be under 5HP.
I know exactly what you mean now. That’s probably not a bad idea. I’ve seen a lot of meshing transmissions out there and they all seem to work okay. I wonder what teams like 254 or 60 do to their gearboxes to make them shift well. I’ve been looking at this one I wonder if they added c-c distance or beveled the gears or just left them.
I like mesh shifters because they are easier to construct, but what I don’t like is that in every single one I’ve seen there are metal shavings inside it like this. The gears themselves have always looked fine, but metal shavings in gearboxes make me uncomfortable.
Anyway, I thought about a dog shifting 3 speed, but determined that there is not good way to do it, where there is a pretty good way with mesh shifting.
Of course you’re going to have metal shavings in there–you’re smashing two metal objects together at high speeds, which is going to wear down the teeth and leave shavings in there.
They seem to run pretty well, regardless of what is or isn’t in em, so it doesn’t seem like a real issue
I have always wonder about this kind of shifting, how long can a gear box like this last?
I had this idea a while ago but I didn’t think it would work because of the gears hiting when shifting. What do you all think about using brass for one of the gears that meshes and steel for the other one. I think since brass is a softer metal it would start to take the shape needed to shift properly.
I know there was a team that use this method and had four speeds, i think it was team 33 but im not sure? someone please correct me if wrong.
Dave
When you build a robot for FIRST you really want the gearbox to last the season. There has been mesh gearbox made before and used in FIRST which lasted longer.
The best way to do this is by beveling the gear teeth edges so it would mesh properly. Personally I would 20 dp gears. If you do use brass, I fear that the teeth will sheer off.
Yes it was team 33 who has done the 4 speed back in 2004 and they used mesh shifting. It was very cool. Team 195 also used something like 33’s 2004 tranmission this past season.
Teams have been using this style of shifting for quite some time, and overall, when constructed properly, it seems to be extremely reliable. I suppose if you shift a whole lot, you may need to have extras of the meshing gears on hand, incase you happen to shear a tooth, or completely wear them down, but considering that a team who attends 4 events (2 regionals, nats, an offseason. Yes I know many teams go to more than this), with an average of 10 rounds per event, that’s only 80 minutes of drive time, which should be well within the lifespan of most things built for FIRST.
I would agree. It’s like one of those things that shouldn’t work but does. I will note that as SVR I know 254 was having a good bit of trouble with it (I think it was Friday evening), but I don’t know what it was caused by. So it wouldn’t be fair to say mesh shifters are inherently troublesome. I don’t think the design is nearly as robust as a constant mesh dog shifter, but there is no evidence to suggest that a mesh shifter is not robust enough for it’s use. I don’t like how mesh shifters work, but you can’t deny the fact that they do work, and seemingly pretty well.
I have a couple questions. Do mesh shifters prefer to be shifted under speed and load? I know the dog shifters just love it, but I’m thinking the mesh shifters might like a slower easier shift.
Can someone who has driven one enlighten us?
In our 2005 drive transmissions we added a little in the center-to-center distance (off the top of my head I think it was 0.0015", but I can check on that if you’d want) and chamfered the shifting gears.
Also, we didn’t care how fast we were running the transmissions when we shifted. Our problems on Friday at SVR weren’t related to the gears themselves. There were minor problems within the transmission related to snap-rings. We had no other mechanical issues after fixing that problem (which originated in Sacramento) for the rest of the season.