Designing a Chassis/Drivetrain

Hello all.

Our team used aluminum extrusion last year, since it was our rookie year, but this year we’re looking to move towards custom parts if we acquire a machining sponsor that we plan to soon approach.

I’m reasonably able at CAD (SolidWorks) but I know little about how to approach a design for a chassis.

I wanted to get some feedback from CD in regards to what kind of chassis is optimal, what thickness of sheet metal to use, which metals, using bearings in the sheet metal, supports between walls, axles, gearboxes built into the chassis, bellypans, etc.

I really am completely new to this, so any guidance in terms of how to design and construct a chassis would be helpful.

(I’m particularly interested in what designs you use to eliminate weight from the sheet metal used in the chassis - I see people use ellipses, circles, rectangles, and triangles as cuts in the chassis walls to reduce weight).

Just curious - Have you seen 1114’s Kitbot on Steroids? (www.Simbotics.org/kitbot).

If you haven’t, go look at it now.

If you have, tell me this: What does your custom chassis have that the KBoS doesn’t? In what way is it better? I’m not trying to tell you not to innovate a new chassis. Going custom is fun. I’ve been there, done that. But when it comes down to things, your custom chassis has to be REALLY GOOD (I’m talking 254 down to a science good) to be a better choice than the kitbot on steroids.

Too many times I’ve seen rookies or even veteran teams decide that they want to go custom, and then completely fail come competitions because they have a half-functional drivetrain.

Well, how good are you with the stress analysis tools with SW?

FRC teams use a variety of chassis designs. There probably isn’t one optimal design (no design is ever optimal in all aspects, only optimal for a particular application). 2473 has used a mix of custom welded chassis and the kit base chassis. I’ve heard of other teams using sheet metal chassis. The base function of the chassis is to support the drivetrain, control system, and manipulators. Now there are a bunch of ways to organize this, and that should be the decision of your team. Such ways include discrete modules that bolt onto the chassis, fully integrated systems, etc. Each has their own advantages and disadvantages, but in each scheme, the chassis is the backbone and support for everything.

In other words, the chassis must be strong and rigid enough to support the entire weight of the robot as it is moving and getting banged around. Note too that the robot must be stable as it corners (robots don’t get up after tipping), so the center of gravity for the robot should be fairly low. Eliminating weight in the chassis may not be the best idea, but again that is dependent upon your chosen design criteria.

I would suggest coming up with a couple concepts at this point. Look at the different drivetrains that are possible, and evaluate the strengths of each of those. Since you guys are in your second year, I imagine you don’t have many experienced student members, so I’d stick to a simple drivetrain (6wd or 4wd). Then, look at some ways of mounting such a drivetrain. The kit chassis is okay. It’s not particularly rigid, but it’s fairly modular and will take a fair beating. Welded aluminum chassis give some more room for customization, but they do require a design and an understanding of the manufacturing process. To my understanding, sheet metal chassis are completely custom, but require a full design prior to manufacture. When you do design these chassis, remember your factors of safety, and do as much analysis and prototyping as possible. Once you commit to a particular design, you may not have time to redesign it if it fails.

Yes, I’m familiar with it.

To answer your question: nothing.

Firstly, I hope you realize I’m talking about simply the metal frame. I’m not saying we’re going to engineer some magical new way to drive the robot. I’m not going to be designing some ball drive or a magical swerve system. I’m just interested in designing a custom metal frame - not redesigning a new way to drive.

I basically want to design things such as:
http://www.chiefdelphi.com/media/photos/37992
http://www.chiefdelphi.com/media/photos/37987
http://www.chiefdelphi.com/media/photos/37190
http://www.chiefdelphi.com/media/photos/36978
http://www.chiefdelphi.com/media/photos/34397

Maybe you understood this, but it seemed like you got the impression that I was trying to pilot an entirely new way of driving.

Thanks for your reply - it was very comprehensive.

I recognize that the way we design our chassis will vary with the game or our goal - no doubt about that - but are there generally any concepts or rules of thumb that should be applied to most designs? In my above post I linked to a few that I found on CD.

I’m not really familiar with SolidWorks stress tests or even the SolidWorks Simulations program, so I can’t say much for that.

Sorry if I came off the wrong way, but I mean to say that even though tank drive is tank drive, and it’s all the same old song and dance, building the kitbot on steroids is generally a smarter decision for most teams because it provides the same benefits a normal tank drive chassis does (and all of the ones you posted), and it has something none of them do - You can build it in a day. That’s the big hitter right here. Custom chassis can take a week or more to be sent off, machined, brought back, and put together. And I know you know that time is something very valuable in the FRC 6 week build season. The goal of a drivetrain/chassis is to be strong, efficient, and to effectively move the robot from point A to point B. The way you score is via your manipulator (Why am I saying this? You know it already.). I’d rather spend 6 weeks working on perfecting a manipulator that is better than everyone else’s than spending a whole week or more making a chassis that does the same thing everyone else’s does.

Tl;dr - You save lots of time by going KBoS instead of a custom chassis design, both drivetrains do the same thing, and both are equally as effective.

Keep it simple. The chassis is not the most complex part of the robot, but it is the most critical. Your choice of drivetrain will drive the chassis design, followed by your choice of control system mounting and manipulator mounting. If you don’t have a set of guys familiar with stress analysis, I suggest you stick with the Kit base. There’s nothing wrong with it, you just need to be a bit creative at times. If you go for a welded aluminum chassis, overengineer it for now until you better understand the forces on the chassis. We’ve used 1" and 1.5" extruded U channel from McMaster (1/8th inch wall) and haven’t had a problem with it, aside from the obvious (or not so) problems from welding. I’ve never worked with sheet metal, but I don’t suggest you try that just yet until you find some people who can work with you through the design of that.

Well, I understand your concern and what you’re saying.

However, the way our team is structured and the way we plan to approach this, I don’t think it would be an issue.

We’re basically going to build our chassis out of extrusion as we did last year, starting on day 1, and then begin prototyping, testing, etc, with that chassis. We’re not going to lose any time waiting for a custom chassis to come in. We will, however, try to make this quick extrusion chassis similar to what we plan to design in terms of, say, the dimensions of the opening, placement of manipulator, space for controls and electronics, etc.

Then, while most of mech team is working on that chassis and approaching the challenge, manipulators, collectors, etc, our design team will be CADing up our custom chassis, which, as mentioned, should be similar in dimension, purpose, and usability as the test chassis we’re using.

That way, if we get that chassis machined, there’s little headway in terms of moving from our test chassis to the custom one - it should be similar enough that we need to make just minor adjustments to our designs to have them ready to put on it. Furthermore, we’ll have in our minds the entire testing time that we are having a custom chassis on the way and we need to plan ahead for it.

Basically, the building of a custom chassis shouldn’t really detract from our efforts in terms of building, say, the perfect manipulator, like you say.

To add on to the post I just made…

As I said, we’d have the test chassis that’d be ready to go on the competition field and work with our manipulator, collector, etc. It’s there. It’ll work. It’ll be there if the custom chassis fails.

At the same time, I much rather have some students working on a custom chassis, getting that CAD experience, learning more about sheet metal fabrication, learning to do stress tests, and have our team start to delve into that kind of design rather than stick to the same old thing because it works.

Don’t get me wrong, I’m not saying we should not use the KBoS because it’s “used by everyone” or “boring”; I’m merely saying that I think it’s beneficial to students, and really, in the vision of FIRST to have students learn more about other chassis designs, drive systems, etc, especially if it doesn’t hurt our teams standing (considering we have our backup extrusion chassis)

I am not disagreeing with your decision, you have a point. However, 6 weeks is generally, and I mean this respectfully, not a good time to be developing technologies for your team. Now I don’t know the skillset of your team, and perhaps you do have the skills and knowledge to develop the chassis. I just wouldn’t trust my team to do it.

Even if those students that would be working on the custom chassis wouldn’t be doing other stuff?

We aren’t taking students from their current roles and switching them into chassis fabrication. Instead, since we’ve grown in size as a team, we’re trying to create more opportunities for students to get involved in order to have everyone engaged, and this happens to be one of them.

If it causes no human capital strain on our team, and no excessive financial strain, AND we have a mentor who has experience with sheet metal fabrication, don’t you think it’s reasonable to attempt?

As I said, I don’t know your team. If you feel you can, by all means go for it. I just wouldn’t trust my team to do it, and I’ll bet our teams are very different.

Good luck with it, though.