A CIM motor is connected to a Nordex 10:1 worm gearbox through a spider coupling to accommodate slight misalignment.
Great idea!
If I were going to do this, I’d invert the mounting, so that the modules came in from the carpet side. The reason is that that way, the load carrying the robot is transmitted through compression load of chassis components, not tension in fasteners - more fail safe, and would put less load on the fasteners.
We used those same worm gearboxes in our arm this year, and we found they had a relatively low max output shaft torque rating. We sheared the worm shaft (or something like that) without putting what we considered to be excessive torque on the arm. And since they are non-backdrive-able, that could be a problem when doing hard direction changes with high CoF wheels. Make sure you take that into account when you design a robot around these.
Also if you’re doing a drop-center 6WD, each wheel would be driven by one CIM so at any point you would only have at most 4 CIMs of power going into the ground. In pushing matches, if you get pushed onto your back wheels, you will only have 2 CIMs of power. Maybe you should look into a belt/chain connection on the modules to link them and transfer the motors’ power equally between the wheels.
What is your goal with this? Flexibility or maintainability?
If we use this concept, it will be for ease of assembly & quick repairs at the competition. Using nutserts and an electric screwdriver I think we could swap out a drive module in <1 minute.
We also thought about inverting the assembly, but that would make the mounting screws more difficult to reach.
I’ve posted the CAD files (41 Mb, SolidWorks 2016) here, but this concept is not 100% done. We’ve built a previous version, but not this one. Just didn’t have time before the Kickoff. :o
[With apologies to Inigo Montoya] I don’a think that word means’a what you think it means.
Non-backdriveable means that torque applied to the output shaft will encounter sufficient friction that it will not turn the motor (with no voltage applied to the motor). It does NOT mean that the system cannot be driven in reverse from the motors.
While I have obviously not spent as much time with your design as you have, I don’t see this at all. As I’m picturing the fasteners, this would be a matter of switching the threads from the chassis to the module, with no net difficulty in assembly, other than the requirement to hold the module up as the machine screws are being threaded. If a few alignment tools were in place, even this could be accomplished with gravity.
I’m pretty sure I know what backdriveable means. According to a thread for one of the other pictures, the gearboxes are [\url] worm gearboxes. They appear to be from the same line as the worm gearboxes we used last year, where torque on the output shaft is not transferred to the input shaft. Therefore the ones we used last year were non-backdriveable. I was cautioning that if they are the same as the ones we used, then they probably are not a good option as a drivetrain gearbox.](https://catalog.nordex.com/pc2743_26612)
I guess I’m curious as to what scenario would require a complete removal of a drive module like this in less than 1 minutes.
Changing out wheels could certainly be done more quickly than removing the entire module, and if you’re blowing out motors/gearboxes often enough to require such quick repairs, then your problem isn’t lack of ease of accessibility, it’s more likely that the problem is that you’re using the wrong kind of gearbox/ratio/lubricant. IMHO, a good drive system is one you shouldn’t have to touch maintenance-wise (short of wheel/tread replacement) for the whole season.
Good points. Our pit crew chief complained that our 'bot was so reliable he didn’t have anything to do between matches except change the battery & check the fasteners for tightness.
One other feature of a right-angle drive is it leaves a lot of open space in the center of the chassis. This could be big advantage if we need to integrate some sort of floor pickup mechanism.
This point I will grant you, however I would think you would want to consider using bevel gears instead of worm gears. While back-drive resistance is nice, worm-gear efficiency can be as low as 50%, compared to bevel and/or spur gear efficiency of 93%-98%. Back-drive resistance can always be accomplished with brake-mode on your speed controllers.
Also, in situations where you expect significant stress (like a drive train involved in robot-robot pushing matches, or sudden shocks from big impacts), it’s actually preferable to have some give in the system. If something is going to give, I would prefer my motors to back-drive rather than breaking a gear. back-driving the motors does no harm, while breaking a gear ruins the rest of that match and forces a repair afterwards.
I hope this is a joke because isn’t this the dream? A reliable drive train?
I would be looking at iterations of your previous (low maintenance) drivetrain, not designing something that is less robust so your pit crew has something to do.
Worm gear efficiency is largely determined by the ratio. Yes, a high ratio (100:1) worm gear drive might be only 50% efficient, but at 10:1 they are 85-90%. Design World Article
Yes, I assumed our pit crew chief was joking. He loves to build & repair stuff, but he knows broken bots don’t win.
Worm gearboxes are not as “forgiving” as traditional spur gear drives, but we’ve had 3 years of failure-free use out of these Nordex drives. Interesting development: We can now back drive some of the older units a bit.
I don’t know if we’ll use these gearboxes in 2017. Depends on the game & what strategy the team decides on.
Great discussion points here on CD!
I’d be interested to know what size/kind of wheels you have used with the worm gearboxes and haven’t had an issue with backdriving. Especially in 2014, which was basically one big pushing match in many regionals, I am surprised you haven’t run into issues with the gearboxes breaking.
Also, the fact that some of your older gearboxes became backdriveable worries me. That means that either your worms or worm gears (or both) are wearing to the point that they have changed shape. I can only imagine that that means the gearboxes are less efficient. I wouldn’t be surprised if this effects auto modes as the gearboxes wear throughout the season.
True, which is why I said “as low as”, but even at the high-end of that efficiency curve, at 90% efficiency you still have a less efficient drive system than virtually any other machine on the field, and that efficiency curve drops off fast if you opt to gear any slower.
The gearing on a worm drive is kind of a catch 22, if you gear high you loose torque and stall your drive motors in a pushing match, but if you gear low you loose efficiency and probably still stall your drive motors in a pushing match since you’re fighting the other robot plus the added friction in the gearbox.
Not counting 2015 (we don’t talk about 2015 :rolleyes: ) our last two drive systems have been extremely reliable, so it’s definitely not just a dream. The only maintenance we’ve had to do was replace wheel tread about once per event due to wear. Needless to say our pit crew was also quite board most of the time as a result. Both robots are iterations of the same design, so, when possible, I definitely recommend iterating on what works. 
I don’t think that’s accurate. I’d say if we actually hooked up everyone’s robots to a dyno and tested their actual drivetrain efficiency, there would not be many with greater than 90% efficiency. I’m not saying worm gearboxes aren’t less efficient than spur gearboxes, but 90% efficiency is nothing to complain about. I’ve always been impressed with Category 5’s drivetrains in the past, so if they say it works well for them, I’m liable to believe them. Though I do agree with what someone said earlier (and maybe in another thread) about attaching the modules from the bottom, not the top of the frame rails.
There are 2 threads discussing this gearbox/chassis concept. I described our wheels in the other thread: https://www.chiefdelphi.com/forums/showthread.php?t=152912
Maybe “in theory” right angle worm gearboxes are not the best choice, but in reality they’ve worked very well for Team 3489, including helping us win the Motorola Quality Award at Smoky Mountain in 2014.
