What type of drive base to design depends a lot on your team’s resources, especially in terms of fabrication capabilities. A drivetrain that makes a lot of sense for one team may be very hard for another to construct. There are numerous examples of many styles of construction out there. Look at COTs options from VexPro, Andymark, and Team 221 for inspiration, and search terms like “West Coast Drive,” “Sheet Metal Drivetrain,” “Box Frame Chassis” “Tube Drivetrain” “Parallel Plate Construction,” and so on for numerous examples of custom work. Base your early work closely on inspriation from these examples, paying close attention to what exactly makes these designs effective, and then branch out from there.
Just a few specific things to look for and consider in your drivetrain design. I’m sure others will chime in with more. Here I’m assuming a basic 6WD drop center, some of this may not apply to other styles.
The drivetrain is the most important part of your robot. If it fails, your robot fails, plain and simple. Err on the side of caution when it comes to structure decisions. As you get more experienced, you’ll learn more about exactly where you can and cannot cut corners.
You’re generally going for complete rigidity in your drive frame, as well as leaving options to mount extra features to. Analyze how the individual joints of the frame interact, and make sure that it can’t bend or flex too badly. Common bending problems to look for include torsional rigidity
(twisting across the frame diagonal), parallelogramming (pivoting at the corners to go from a rectangle to a parallelogram), and impact resistance (how well will each of your individual frame members resist deformation when directly loaded) - Actually do the math when selecting your gearboxes and motors. Optimize your robot for the distances you plan to be travelling regularly.
Determine the materials and fabrication techniques that make the most sense for you to use in the construction of your drivetrain, and assess the limitations imposed upon you by these resources. Sheet metal sponsor? Build a sheet metal drivetrain. No lathe? Try to design around COTS axles as much as possible. And so on.
Live axle or dead axle? Cantilever wheels or dual-supported wheels? Each have their pros and cons, and the choice is often driven by construction technique.
Even if you go with a custom drivetrain, assess how you can use individual COTS parts effectively. Many custom drivetrains will use numerous small COTS parts throughout them to save on fabrication time, gearboxes being among the most common.
Make sure you have a plan for driving each of your wheels effectively. If you are using chain, make sure you have a solid tensioning system. If you are using belts, make sure you have a tensioning system, or that you design for exact Center/Center distance, and have the resources to hold those tolerances
Don’t forget about your bumpers. Please, please, please don’t forget about your bumpers. It’s much easier to integrate a bumper attachment system into your frame from the start, than to tack one on at the end. This applies tenfold if you go with cantilevered wheels. Some teams have even been using the bumper as structural members of their frame recently.
Consider the gearbox options available to you, and how you plan to integrate this to the drive. Direct driving a wheel saves weight, and can improve drive reliability by ensuring that a wheel per side will always be driven, but actually integrating a gearbox at the wheel can be a design challenge.
Based on how you worded your question, I guarantee that the thread below me is going to be filled with posts recommending you stick with COTs options. While a COTS chassis may well be the “best” decision for your team depending on where you’re at entering the build season, and while there are numerous solid options on the market (Versachassis, AM14U, Drive in a Day, C-base, SuperLight Chassis, etc.), that doesn’t mean that you should just abandon the idea of doing a custom drive. The offseason is a great time to practice and learn design, and practice is the only way you’ll step up to do one in-season (or, by getting forced to by design constraints, completely unprepared to do so…that was fun…). Assess design quality honestly and realistically compared to COTs options when selecting your competition drivetrain, the most reliable part of your robot, but by all means, work to improve your skills all that you can.
The diamond hatched bellypans you are seeing are made from raw aluminum plate, typically cut on a waterjet, and attached using rivets or screws. A bellypan which rigidly connects the corners and sides of a drivetrain together helps with torsional rigidity, parallelogramming, and gives you an easy to use, low down place to mount electronics. The diamond plates you are seeing are flashy and one of the most weight-efficient ways of doing it (1687 did one last year), but they’re also quite resource intensive (you need a dedicated manufacturing sponsor or the ability to cut it inhouse. The machine time on these plates is often better utilized elsewhere). Teams have used solid metal, plastic, fiberglass, and even plywood in the past for similar effect.
As Joe G suggested above that many subsequent posters would add: unless you progress very far with this before January, consider whether your precious design time is better spent on scoring mechanisms rather than a custom drivetrain. Last year the AM14U was hard to beat, and many designs that succeeded at beating it were also based on COTS platforms. During inspections, I heard quite a few teams express regret at having opted out of the KoP drivetrain in 2014.
All that said – drivetrain design is really fun, and can be quite rewarding for your team as well: IF YOU REMEMBER THAT THE MOST CRITICAL GOAL IS A ROBOT THAT NEVER MISSES A MATCH.
Joe’s suggestions above are all solid and worth reading over again several times. I especially like the one about doing the math on motors, gearboxes, wheels, etc. There are several good calculation tools out there to help you.
Another suggestion…which Richard hinted at…maybe you should get at least part of your team to work on some generic arm, roller, ball shooter, etc designs. We are pretty sure we’ll all get a useable drive base with the KOP (unless we opted out of it), but we’re also pretty sure we will get nothing for the part of the robot that actually plays the game. In my experience, the part of the robot that plays the game is the part that’s more challenging to design and build. And it’s where most of a team’s efforts should go.
As suggested about learning the math for drive base motors, gearboxes, etc. this same suggestion is very helpful for learning to design other mechanisms well.
For designing generics, it’s a good idea to consider how generic you want it to be-- a frame designed to be easily a WCD, swerve, or mecanum is going to be significantly more difficult that one designed to be a 4, 6, or 8 wheel WCD. Make sure you come into the project with a realistic assessment of what you can do in the time you have. Scope the project appropriately, especially if you’re working on this as a team offseason project. Even if you’re just doing it out of personal interest, it’s still a good idea to have a clearly defined set of goals before you start actually modelling anything.
For inspiration, I’d recommend looking at the AM 14U and the VexPro 2014 drive in a day-- both are designed to be heavily configurable by the end user.
One of the most crucial parts to creating a drivetrain is serviceability. You most likely will have problems with your drivetrain because nothing is perfect. Make sure that you can pull off a gearbox or a motor in case something happens.
We had so much power to each wheel last year (1 CIM and 1 Mini CIM per wheel (4 wheel mecanum)) that we ended up wearing the gears in the gearbox so much that only 1 CIM was driving a wheel. It took us a while to replace it (about 1.25 hours). Part of the problem was we didn’t use more than the standard andymark grease. (The funny part about this story is that since we had a PID on each wheel, we didn’t notice the loss of power until we heard the gearbox grinding)
Also, clever placement of items (in terms of making room for other attachments) can be a big help. We put our compressor inside a section of our frame that would have otherwise been unused.
Don’t forget to make some nice, rigid spots to attach end effectors/ game piece manipulators!
And last but not least, try to keep the CG as low as possible. We had 6 CIMs and 4 Mini CIMs all under a foot off the ground. It helped us not tip over, even with a relatively heavy ball throwing mechanism.
I would highly recommend looking at HOT 67’s tech notes, as well as other teams that have their drivetrains figured out. We based ours off 67’s and it worked out really well for us.
Something that you will mostly likely/should encounter this season is with the new control system. The new motor controllers(talon and victor) have your power, ground, and output are pre-attached inside the sealed unit. Those wires are minimally short so you will want them close to you PDB. Make sure you have room on your belly pan/electrical board for those two things to be relatively close. Things like gearboxes on a narrow chassis could become problematic.
This is all part of your extensive planning, which is the single most important part. Do not get ahead of yourself. Make sure you have a spot for each component required for your entire robot and its systems. Don’t just build an off-season drive train. If you can, experiment with other common FRC mechanisms(wheeled shooters, grippers, rollers, linkages, linear elevators,etc.)
Depending on what style drive you decide on, (tank, swerve, omni, mechanum) you may want to separate components of the drive train into modules for easy removal from the frame. Essentially, serviceability is key in good drive trains.
Last year our drive train was separated into pods. One left and one right attached only to the central frame for easy removal and servicing. It may or may not work with your robot, but I would recommend incorporating some easy disconnect devices, possibly captive nuts or similar and easy connectors for wires.
COTS drivetrain options are so good and so affordable nowadays that unless you (a) already have a tried-and-true design in your back pocket, or (b) decide you really need something out of the ordinary to execute your game strategy or robot concept, you are almost certainly better off going COTS (IMO).
All the same lessons that can be learned or taught with custom drive trains can also be learned or taught with custom manipulators and scoring mechanisms, and the latter is far more likely to set your team apart in the competition.
You would be wise to design for Serviceability and Manufacturability as many of the previous posters have said. It doesn’t matter if you have the best design if you cannot manufacture it correctly or it takes too long (and you don’t finish). It also doesn’t matter if you have the best design and when it breaks, you can’t fix it in time for the next match. I have seen a particular powerhouse team with deadly accurate scoring mechanisms sitting dead or only being able to play defense for multiple matches at multiple off-season events this year.
Think about how each part will be installed and what tools will be needed and how those tools will be used. Think about which parts you will need access to for servicing and what tools can be used. Are there parts that must have some exact dimension(s) to fit properly? If so, you may be setting yourself up for a lot of frustration.