My teams summer project is to design a swerve drive system to use as a base for next year. I have created a CAD model of what our design for the wheel housing. I have attached the CAD model.
the current design uses a bevel gear system to drive the wheel and a integrated sprocket for turning of the wheel housing. The rational method is going to be a turn table that is attached to the top of the housing. There is also a ring on the bottom for stabilization of the housing to stop the wheel housing from bending to the side.
Some of the issues with the design is the cost, my team is on a budget, of $135 per wheel housing being a bit much. see attachment for more details. Also weight and size are important to get down so if you see a area that can be cut out, we have access to a water jet, or know of a lighter component. It would be highly appreciated if you commented on it.
Hey, team 2158 is doing the same summer project! To cut down on weight and price, we’re making our modules out of 3/8" Lexan, if it helps at all. Aluminum should be comparable to both price and weight, though, so stick with what you’re team is comfortable doing. You’re right, the main expense is coming from the 90* gearbox. We haven’t really been able to find a way around that other than building a custom one, but since we don’t have any machining tools, that would end up being much less stable. If you feel confident enough, you could find some set-screw-on bevel gears and align them yourself, which should save you quite a bit of money. Good luck!
I have no problem with you reuploading the iges and for anyone else who is wondering i have no problem with you using the design for your teams own purposes.
and daltore that lexan thing sounds like a good idea. Ill do some research onto if that might be a good idea to use that in place of aluminum.
We used Lexan for our competition bot this last year (3-wheel independent swerve) because it was fairly cheap off of McMaster, but most of all, it can be cut with a band saw and drill press, as we don’t have a CNC machine capable of cutting metal (our mentor, Richard, has a home-made CNC machine that can cut wood and softer materials). It was a good learning experience, and it was quite fun. The main thing you’ll have to watch out for is side-load on the modules. This past year, there was so little friction between the wheels and the floor that we only had top support for the modules, and it ended up helping with breaking as acting a bit like ABS. However, on carpet with regular wheels, you’d DEFINITELY want a bottom support as well to take all of the lateral force. Actually, this is true for any material you make it out of, as they can all flex, bend, and snap eventually.
Another material I’ve seen used is sheet aluminum made into boxes. This provides a rigid structure (the cubic frame) that is very light weight, and doesn’t cost a whole lot. I forget which team did this, but searching “swerve module” on the Chief Delphi picture gallery will probably bring it up.
I looked over the design, and I really suggest some substantial changes.
The sideplates need to be beefier, 1/8" thick with that much removed for a module that tall… I obviously haven’t run the math on it, but I’m leaning towards scary.
The turn table scares me… a lot. They don’t handle sideload well (and even with the bottom support… there will be sideload) and I just would never trust them in a critical application. I highly, highly, reccomend a bushing or bearring for the module rotation.
The chain run, why does the bottom shaft need to rotate and have external sprockets? you should put the sprockets between the plates, make the very bottom shaft stationary (ideally built in such a way to rigidly attach the sideplates, adding strength) with a wheel that sits on bearrings and a sprocket bolted directly too it.
Good choice on the bevel gears, not sure what factors influenced it, but they’re the same gears 118/1625 (and we used on our crab) used.
You made me curious so I had to open up the cad file, but I share all the same thoughts Adam has stated. But FIRST, great design and keep creating awesome stuff. I really like the setup, it just needs some tweaking:
-The plates are way to skinny, you should run COSMOS or some FEA. If you need help, pm or email me. I personally would go with 1/4" thick plates, but you might be able to get away with 3/16th’s with less pocketing.
-As Adam also pointed out, why key the bottom shaft. A cleaner way to do it would to buy a bigger sprocket, just big enough so that the sprocket can fit an andymark bolt pattern on it. The wheel should have a bearing on it with a sprocket bolted to it. The first outer sprocket should be moved inside and I reduced the current hub by 3/8ths. Here is a link to the modified CAD file (IMAGE-1, btw hope you didn’t mind me modifying the CAD file).
-Also on the right side of the uploaded pic, I flipped the bearing so that the hub of the sprocket pushes up agianst the bearing. On the other side you are going to need collars to keep the bearings in the right spot (unless they are pressed in).
-As for the sprocket that turns the module, put a bearing instead of the turn table. Smaller turn tables tend to be weaker and a bearing can do the job just fine.
-Also make sure there is a spot to mount an encoder, you can either mount it on the module or on the gearbox that powers it. I have never made a crab drive and maybe I’m wrong about the encoder thing, but Adam or Aren should be able to ellaborate on the use of an encoder.
Distance encoders can be mounted at any point in the driveline, as long as the encoder can handle the rpm it’ll see there, in 2008 we had 1 on the chain the ran to the top of each module.
also 4" wheels make stuff alot nicer to fit, and as far as putting sprockets on the outside id only do it if i had hex shafts and snaprings. otherwise id run the bottom axle dead with the wheel on bearings (07 for us)
I love the critisim for the drive so far thanks every one who has posted.
Here is to answer some questions on why i did some things
the thickness of 1/8" is there because last year we built the prototype for this prototype and it used 1/4" plates. The 1/4" plate was complete overkill. It wieghs about 7lbs without the wheel and half the sprocket and bearings. It also could survive a nuclear bomb. But i like the suggestion for 3/16" plate. Ill see if we can use that
i like the fixed axle and interal sprockets concept. I will change the design to utilize that
I used the turntable because it was cheap. But after what you said i will chage to bushings and bearings for the turning method.
for the bevel gear choice i chose it because it is the cheapest set of bevel gears i could find with a 1/2" bore
Our crab drive that we built and used at comps this year had a problem with the 1/8 in. plates because at comps our pit crew had problems with the 1/8 in. plates bending during a match. Once realized, we quickly doubled up on them. The moral of this story is, keep in mind the stresses the robot is going to endure during a match. We (team 93’s design team) didn’t take into account those stresses and so we ended up replacing drive modules a lot.
It’s just something to think about.
After spending the extra time i had on my hands working on the model i completed the model with constraints and motion constraints then put four wheels on a weighted down cart, and applying the appropriate forces. Design simulation, part of autodesk inventor 2010, says that the plats on the side need to be thicker up to 3/16" should do it:D