Lately several members of the team have been discussing the possibility of doing a wooden drivebase with a live axle system (both things that we’ve never done). This is one rendition of the concept that I came up with.
For this particular version I was trying to cut manufacturing down to a minimum. Ideally it could be built with a drill press and a bandsaw (or even a hand drill and a hacksaw), but there are some details that still need to be worked out.
The siderails are wood, the endrails are 3" fiberglass channel which we happen to have lying around, and the base is a sheet of aluminum. Each wheel module consists of a banebots wheel on a keyed shaft with flanged bearings, and since each one just needs a hole in the wood, this concept could work for 6 or 8 wheel drive. (again, there are still a few details to work out, but it looks doable)
I would guess weight to be about 45 lbs for this particular base, not bad at all.
I would be a little concerned about frame-torsion. Your design looks like you intend to screw into the wood, and over time the twisting of the frame may well loosen it up, especially since you’re screwing into end-grain. I’d consider L-brackets with through-bolts to avoid that.
Also, unless there is a pressing reason for 8 wheels - I would avoid it. Unless you drop the middle 2 sets, you’ll have trouble turning, and making 8 wheels the correct heights is really difficult. There’s a reason stools have 3 legs rather than 4 (wood warps and twists over time). You can avoid that in part by using plywood, but that has problems all it’s own if you plan on screwing into the end of it.
I love the simplicity of this design. I too would suggest that the wood be made from plywood.
Another suggestion… if you make the 4 interior wheels have a slightly bigger diameter, that will work too. What I would do is put treads on the 4 interior wheels so that they have a diameter that is 0.2" larger than the outside wheels. This may be easier than varying axle heights.
If all 8 wheels are driven, wouldn’t that cause the outer wheels to fight the inner wheels since their smaller circumferences cause slower linear translation than the middle wheels? Is it too minor to worry about?
I like it in general. It looks very similar to my design I posted. I too would switch to plywood, particularly a good 1088 meranti or baltic birch plywood as it is much stronger and more stable than a solid chunk of wood. Also u should look into epoxying the wood to the fiberglass plus screwing. If you do that you will have an indestructible joint.
Another option is to use omnis on the 4 corners–that way, all eight wheels can contact the ground for stability, but still be able to turn smoothly…and I second the idea of adding L-brackets or some additional support to the 4 primary corner joints.
Question for anyone on the aluminum belly pan: what is the minimum sheet thickness one would use on such an application?
yes use L-brackets the dont add to too much weight for all there worth
and if the ends are fiberglass why not just glass it all its not very hard and that way you would have all the great performance of wood and the added strength of the fibergalss but it would still be able to give a little allowing for a smoother ride then a metal frame:cool:
the only problem with fiberglass is its kinda messy to work with and if it gets on a tool or something you might have to throw it away
and I second what andy said about the inner wheels
Even if it weren’t too small of a difference to actually make a difference, there is a wee bit of theoretical benefit to a minor mismatch in speed. If the corner wheels are any less grippy than the specially-treaded middle ones, they’ll be slipping slightly. The way friction works, they’ll thus slide sideways more easily than if they were in solid lock-step with the rest of the wheels.
But it’s not an issue. Andy Baker’s suggestion is a proven one. The TechnoKats 2008 robot uses six-wheel drive with in-line axles and extra tread on the center wheels, and it drives great.
Good points, I forgot to cad in the top corner brackets that would reinforce those joints. It could be constructed in a very similar way to our 2008 drivetrain which was able to go through three competitions with no problems. I am not sure whether wood screws would be enough or if through bolts would be needed, that’s something that may need to be tested.
Oh, and the middle four wheels are all dropped 1/8" on this model.
Last year we used a .050 or .060 sheet, I believe… not sure on the actual thickness (squirrel might know?). It held up great and provided a solid mounting surface for all of the ‘guts’. Adds about five pounds to the weight, but it compensates for additional bracing and a sturdy electronics board.
We’re a little hesitant to take on a project like that… like I said this particular model was designed with manufacturing taken into consideration. That means that we cut a length of wood, drill some holes, and we have a side rail. The fiberglass ends come as a pultruded channel already, so we just hack off the correct length and bolt it on. (We have this problem of being lazy and trying to avoid too much labor )
Some of you also mentioned using plywood, and I’m curious as to how it would best be implemented. One team member suggested laminating several sheets of plywood together to make the siderails…
One more thing I forgot to mention before (and why I like this design so much): Since it is constructed with siderails that require just a proper sized hole for a wheel module to fit, we could really stick wheels wherever we want. So, if it turned out that we didn’t want an eight wheel drive robot after all, we could just drill another hole in the center of each side and we’d have a functioning six wheel drive robot. The difference between a 6wd and an 8wd bot with this frame is just two extra wheels and chains.
I’ve been discussing this with some of the students and Steve, here are a few more comments. The plan for the sides was to use a hardwood such as oak. I’m sure we’ll get the “that’s way too heavy!” comments…that’s ok…last year our robot with it’s way-too-heavy drive base was still 10 pounds lighter than it could have been, and it never wanted to fall over. Hardwood has the advantage of being dense, so it won’t let the bearings open up their holes too much, and it should hold screws better than a lighter wood (although putting the screws in would be more of a challenge). We used .060" 3003 aluminum for the belly pan and corner braces last year, it worked fine, it’s pretty easy to rivet to the fiberglass channel to make a strong structure.
One interesting point about this design is that it uses three quite different materials for the chassis structure, and each material is used in that place where it is most effective.
I do like the way that you are using multiple materials and your use of fiberglass has actually gotten me thinking about incorporating it possibly. I do think that fiberglassing the side rails and bonding the whole thing together with epoxy would be your best bet. Your frame would be indestructible. Also for all the naysayers, oak is much less dense than aluminum and stronger as well for the weight so go for it. I hope you guys get a chance to prototype this and post it up on here. Maybe we should see who has a prototype frame done first :yikes:
I think the wood will work fine for the overall stress of the base, but I worry about how about the stresses at the axels and how you are attaching the bearings. I could see big issues with however you attach the bearings for your cantilevered wheels (can someone rationally explain to me why everyone likes to have cantilevered wheels??? seriously…). I see the wood warping at the bearings (even hardwood), unless the bearings are very securely supported. For this reason it just makes more sense to me to go with two 1/8" plates on both sides of the wheels with cross-bracing.
The 2 major reasons for cantilevering wheels is to ease wheel replacement, and to maximize your robot’s footprint. Many teams are able to change a wheel on their cantilevered system in a matter of seconds by removing a snap ring such as 254.