Here's a look at the latest iteration of the 6WD module I've been working on for our team in case the mecanum drive is inappropriate for the new game.
The goal is make something that is small, light, easy to manufacture and, if at all possible, pretty.
The gearbox contains a 12:40 first stage reduction using gears. That is followed by a second stage of 25 chain and sprockets at 12:60. Overall ratio is 1:16.67 – approximately equal to the kitbot arrangement.
Wheels are 5.75" OD. Custom omniwheels at the corners allow for fantastic maneuverability and otherwise are fun to look at. You could just as easily get away with a single pair of omniwheels and four regular wheels. All axles are in line and are made from 1/2" hex stock. Dead axles may be substituted if desired, but the hex stock allows for easy addition of encoders for feedback.
The chain runs from the gearbox to the rightmost omniwheel which is fixed in place. From there, the middle and far wheels are each connected by a single chain. The middle and far wheels ride in bearing blocks which fix to the main module via slots to allow for chain tensioning.
i love how compact it is with the cims mostly in side of the assembly not sticking out in the middle of the bot and how the chain runs work. you’ve made good use of the allowed space in there keeping it compact and it looks like the whole thing would weight 10lbs. i like it
Another beautiful design from the Xbot team, I also love how compact it is. I have a few questions, why do you use omni wheels for the front and back?? (it will help you turn better, but the traditional 6-wheel design turns just fine) Also, what are the extra two sprockets on the front omni wheel for?
We ran omniwheels on the corners last season (http://www.chiefdelphi.com/media/photos/23578) with good success. It’s true that there’s a bit less resistance to being spun around on the field, but that wasn’t a challenge we faced because our robot didn’t rely on being in a single spot to score accurately. I don’t think this method is any better or worse than lowering the center wheel or using a single pair of omniwheels – it’s just different. My strategy and, subsequently, my design assumes that if you’ve gotten yourself into a pushing match during a FIRST match, you’ve done something wrong.
The large sprocket on the rightmost wheel is driven by the chain coming from the gearbox. The other two sprockets on the wheel each drive one of the two remaining wheels. This’ll allow us to move each wheel independently of the others to properly tension the chain.
I like! Reminds me of a design I did… But anywhoo, about what speed do you get out of it? Also, with how rough the games seem to be getting, aren’t you worried it’s a little flimsy?
If welded / screwed together properly it should not break all that easily. I Like that design and have another question. The Drive System seems to be well organized and hopefully built but what about the many other modules this robot will have, will there be enough space inside the robot to do a motor transplant w/o having to move anything out of the way? Is the CIMs easy to get to? Will other modules be easily able to be gotten to? If not will all the other modules be modular so that they can be taken out quickly and easily with little to no hassle?
Looks nice and compact! I’ve always loved these double plate designs, something we’ve yet to integrate into one of our robots.
As I understand it, there are two chain runs that go over the motors? How much clearance do you have b/w the chain and the motors? Are you concerned about a) tensioning the longest run b) having the chain (even properly tensioned) slapping against the motors?
The hex shaft is nice as well. How are you adapting the sprockets to the shaft, as it looks as though the sprockets are either AM’s, or kit sprockets. Neither of which come with a bore you could use.
Finally, I’m calculating about 6.5ft/sec, right around the kitbot speed. Have you found this to be a good balance b/w speed and torque?
This is based heavily on our drive from last season. It looks considerably different because we now have the capability to plasma cut much larger parts, but the design is informed by what we learned from our effort last season. The single largest problem we had with that drive was borne from my stupid, aesthetically-driven decision to expose the ends of each module and extend them beyond the frame. As a result of that design decision, we sometimes had the modules bent inward at their ends when they were hit by other robots. The frame this will sit in will resolve that issue. I’m otherwise not concerned at all about its resiliency. I think most teams in FIRST overbuild their machines to an absurd degree, but sometimes that’s the only method they have available to them to compensate for a lack of engineering support.
I didn’t explain the gearboxes in too much depth in the caption. They are wholly self-contained and can be quickly removed from the outside face of each module. Replacing a motor, which we’ve never had to do, shouldn’t take more than a few minutes.
If they’re properly supported by the rest of the robot frame, these are a great way of compartmentalizing the function and assembly of the robots. We’ve found that it’s much easier and goes much faster to have two teams of students assembling the individual modules than to have them all huddled around a single frame. It also allows other folks to do their thing with the frame without getting in the way.
Of course, there’s also a pretty big advantage to repairability and that sort of thing as well. In the short time we’ve done this, we’ve not had to replace an entire module, but I’m sure that we’d need to as soon as we abandon this design
As I understand it, there are two chain runs that go over the motors? How much clearance do you have b/w the chain and the motors? Are you concerned about a) tensioning the longest run b) having the chain (even properly tensioned) slapping against the motors?
Admittedly, I’m always bad at accomodating chain. I like things to be snug and compact and chain doesn’t always cooperate. If it turns out that the chain is banging against the motor casing, I imagine we’ll make a sleeve or something to minimize the noise and losses out of delrin.
The hex shaft is nice as well. How are you adapting the sprockets to the shaft, as it looks as though the sprockets are either AM’s, or kit sprockets. Neither of which come with a bore you could use.
Since we’re building our own omniwheels, they’ll have a hex bore in them and we’ll just bolt the sprockets right onto the wheels. The center wheel, if we again use a Skyway wheel, will have 3/16" aluminum “hubs” that’ll be hex bored and they’ll tie that whole thing together. It means that the wheels are tapered toward the middle, but I don’t think that’s a bad thing.
Finally, I’m calculating about 6.5ft/sec, right around the kitbot speed. Have you found this to be a good balance b/w speed and torque?
BEN
That seems about right. Feedback from our drivers last year suggested that the kitbot arrangement was even a bit too fast for their liking. I’m hesitant to slow it down too much, though, as I think there’s some range where the machines are too slow to outrun defensive machines, but too weak to fight their way through them.
Madison, we are quite impressed with the idea of removable six wheel modules, and have begun to think about the design of one.
My question is, how to you intend to attach the modules to your frame? I’ve considered welding angle onto the outside frame, angle onto the modules, bending tabs, etc., but I’m not completely pleased with these methods of attachment. We intend to use 1x1 tube, in the kitbot arrangment style- perhaps even the kitframe, and I run into problems intersecting gearboxes and motors with necessary frame members, just in my head.
As Andrew said, We have been developing a similar drive train, but are wondering how much each of your modules will weigh completed. As is, A completed frame with our currently designed module will weigh about 35-40 lbs, with each module weighing 15 -20 lbs. Is this comparable to yours?
Also, our current design differs from yours in that we will attach the modules directly together using aluminum tube, and that will form our primary Frame. Because of this, our modules had to be somewhat more… Robust, and are going to be made of a thicker plate, probably .20-.25 inch. As soon as we have some renderings we will post them, and thanks for your help in advance
I think I’d go with more standoffs and thin plate rather than thick plate, but, as always, it all depends on a lot of things that aren’t easy to predict and sometimes aren’t easy to test for.
That sounds pretty close to the weight calculated by Solidworks. My model is missing bolts, chain and little things like that and comes in at about 10 lbs. per module and 25 lbs. for the whole chassis.
I was planning on seating this design inside a frame of 3x1" aluminum channel, somehow using the spacers and bolts on each end in the scheme to attach it to the frame. I didn’t develop it much further than that because I began in on yet another 6WD setup that’s integrated wholly into the frame.
I’m looking forward to seeing your renderings. I’ll likely be putting up some new renderings of the latest 6WD arrangement we’ve designed later today.
You get more contact area between the drive wheels and the carpet when you use more wheels.
Some will argue that the reduced weight on each wheel makes it no more effective than a four-wheel system. Others will counter with the idea that the carpet fibers bend and thus have a maximum static friction force they will support for a given area.
