Our Prototype Mecanum Drive train and setup using KitBot chassis.
I must say great setup there, making a complicated drive system as simple as possible. I must say that these mecanum drive systems seem to be all the rage these days, at least from the number of prototypes i have seen. But I am still puzzled by the perceived advantages of them… The first myth I can see is the idea that they are more maneuverable than a basic tank or 6-wheel drive. The fact is that the majority of the successful robots using a mecanum or holonomic drive systems all have low COG. If you have ever seen a good 6-wheel or tank drive system with a low COG which can turn on a dime. I would argue they are far more controllable and more effective (traction!). Have a look at some of the 2002 (or was it 2003) videos just posted here, and you will see some low bots (SPAM 108) which were blazing fast, could turn on a dime, and seemed controllable. Our robot in 2005 was our first year using a 6-wheel design, because of design faults with our arm we ended up being a defensive bot (in the most offensive oriented game so far :ahh: ), that robot didn’t score much but was fast and maneuverable for defense.
Another fact is that having a high COG with a mecanum or holonomic drive system will cause horrors to the controllability of the robot, because all 4 wheels need to be on the ground (with relatively equal pressure on them) to work properly. I have heard teams argue that these systems allow them to be more maneuverable and therefore better offensively. But in the era of FIRST where the majority the scoring is up high, this will cause havoc for teams trying to use these types of drive systems. I have yet to see a strafing drive system (other than crab, I like!) that can out perform a tank drive or 6-wheel system while still having an offensive robot (presumably with a high COG).
I applaud you guys for designing a simple implementation of a mecanum drive system. Maybe my reasoning is totally off base here… feel free to prove me wrong. But I feel that the advantages are not there to use one of these systems for one of the games, thou I would like to prototype one myself.
Are those little orange things encoders? You might want to reconsider their placement.
Have you ever seen a 6-wheel bot go sideways? That, in a nutshell, is the advantage of a mecanum or holonomic system. Easier realigning after someone knocks you offcourse, sliding around a defending robot, and outswerving a 6-wheel are some of the advantages here. And, there are ways of making them have better traction. Roller angles and treading can have an effect.
Could you explain how the CG has affects on driving a “normal” drivetrain as compared to an “omni” one, other than that a slight tip can throw off one wheel?
Why the angled gear boxes?
One word, strafing. The ability to move sideways is the inherent advantage to the “omni-directional” drivetrains (such as mecanum). That is especially advantageous in maneuvering through narrow spaces or when quickly correcting direction or orientation. The most cited con to omni-directional drivetrains is their weaker pushing power (which is true), but they are often rather difficult to be pushed as well (if they are attempting to resist the opposing motion).
As for the CoG, the rest of the bot should have little dependance on what drivetrain you impliment, and thus the CoG shouldn’t be terribly effected depending on your drivetrain. But even with higher scoring games, you can still have a low CoG, especially if the robot design doesn’t recquire you to elevate any components until your in scoring range (or near it).
I’m also wondering why you angeled the gearboxes.
I’m guessing that they angled the gearboxes so that the chain can be run straight from the gearbox to the wheel. The way their gearbox spacers are placed, the chain would rub along them if the gearbox was horizontal.
My question has to do with the width of the gearbox. There’s an awful lot of unused real estate there. I understand that the gearbox is mounted on the kit chassis, which requires that spacing. But, the CIM could be reversed 180 deg, and be mounted on an “intermediate” plate between the two outside plates. Then, by turning the entire assembly 180 deg, the encoder is inboard, the motor is inboard, and you keep your spacing and loose all the excess width. Also, this places the chain run on the inside of the wheel, which allows you to move all your wheels out another inch or so, effectively widening your wheelbase by two inches. This makes a “more square” drivetrain, which has been argued to be an improvement for Mechanum systems.
Overall, I like the integration into the kit chassis. I’m just not sure why you didn’t use the kit gearboxes as well.
BEN
I also enjoy seeing how folks work with the kit frame–it’s cheap as free, has zero lead time, and handles just about anything you could want to do. (Alright, fine, live axles take a bit more work.)
I’ve got three possible theories on the absence of the KoP gearbox:
-
It may just be cheaper for them to build four ridiculously-simple gearboxes than to order two more from IFI. (AndyMarks would be another option, but then you’ve got two different sets of spare parts.)
-
I don’t quite know how you’d run the encoder on the kit gearbox.
-
Perhaps there’s some weight advantage to it?
Yep encoders. This design is a prototype that we want to use in our CannonBot which is surrounded by plywood, so we figure the encoders are safe. Reasons for doing this design…
- An improved drive system for our CannonBot that weighs around 160 lbs.
- Mecanum drive system presents a challenge for the students to figure out during the preseason. During the build season, we may stick with a 6 wheel design like we’ve done for the past 3 years.
- Give our students practice designing a transmission. Yes, there’s a lot of wasted space in there, but it will easily drop into the kit chassis.
- The angled transmissions were a way to try a new method of tensioning chain. Usually in the past there are tensioning slots like in the Kit tranny. In this design, there is a pivot on one side and on the other is a shaft in the transmission and another shaft in the chassis. In each shaft are holes, clearance in one and threaded in the other. Put some bolts through the holes and tighten to tension the chain.
http://www.chiefdelphi.com/media/img/df9/df905e22d221a812054fcb8f6eb1d7f7_l.jpg - Give our students practice in fabrication of the transmission. Currently they have just finished writing the G-Code and are in the process of CNCing their first transmission plate.
- Give us more experience using Hex shafting and Hex broaching.
- Give us more experience programming. We will run the robot off a Vex controller.
Thanks for all the feedback and I’ll post some pics of it finished in the upcoming weeks.
So our students get more practice designing and fabricating them.
Awesome design!!! looks very easy to fabricate and robust. You guys need to post more pics on the website, obviously a lot is going on back home!
What are the two slots near the encoder form?
Hey stranger!! Website?? I wish! A lot of noobs, not enough getting done!! Anyhow, we’re still working on the CNC plates. The CNC is a bit different than the other one. The coolant system is cool, but messy, so we’re still working through that. We should have all the kinks out by tomorrow and have our first plate done. Just seven more to go, but it should go faster, now that we know what we’re doing. The slots, on the side next to the encoder, are for access to the mounting screws for the CIM.