For the past several months, team 6635 has been hard at work developing a series of 3D printed swerve drives for possible competition use. We currently have four completed designs, each being designed to fill a different role in a competition.
All of our swerve drives are 3D printed in either Nylon, Polycarbonate, HIPS, and ABS (for low stress parts)
V1, We designed this swerve drive to fill the role of a practice bot swerve drive. For our practice bot, we decided that we wanted a lightweight swerve drive with as open of a construction as possible. This open construction also allowed us to use a single piece of aluminum or poly carbonate as a mounting plate, greatly simplifying the manufacturing process. https://grabcad.com/library/frc-team-6635-off-season-swerve-drive-v1-1
V2, This version is nearly identical to V1, except the motors in this version have been moved to the top of the module. This makes it necessary to include a second mounting plate for the motors, which detracts from the easy construction and easy maintenance of V1. The primary reason for this change was to increase the ground clearance of the module, allowing for better terrain navigation. https://grabcad.com/library/frc-team-6635-off-season-swerve-drive-v2-1
V3, The third version of swerve drive that we produced during the off season is the first heavily specialized swerve drive we have designed. We designed this swerve drive to fill the role of a high durability, high torque pushing robot. To do this, we decided to put the bevel gear and the wheel on the same axle, with the wheel at an offset. This allowed us to make a much shorter, almost stubby, module that can much better handle large amounts of lateral force. https://grabcad.com/library/frc-team-6635-off-season-swerve-drive-v3-1
V4, This is the fourth and final off season swerve drive designed by team 6635. This swerve drive is a less common design, which uses a vertical off center axle to transmit the power from the motor to the wheel. This is by far the least tested of our designs, as we have only created mockups and have not yet done any powered testing. https://grabcad.com/library/frc-team-6635-off-season-swerve-drive-v4-1
We would greatly appreciate any feedback or questions that anyone has regarding our designs, as we are still rather new to swerve drive and easily could have missed some important details.
I can’t speak to the 3D printing aspect, other than to say you will probably need some advanced materials if you want it to hold up to the rigors of an FRC season. You also need to consider the differences between designing for classical machining and for 3D printed parts. Some features are missing what appear to be critical fillets to keep the parts from shearing.
v4 definitely looks like the most reasonable option. That being said, it has a few “features” that I’m not very convinced are the best options. In no particular order:
I’m not sure the Neverest will be strong enough as a steering motor. You might consider a more powerful BB550 to make sure you have the power when needed, then you can lower the motor power by decreasing the max voltage.
The steering motor is only attached by a clamp, which itself is cantilevered off the side of the module. If the module hits a bump, I could imagine the motor turning in the mount, throwing off the control loop.
The smaller vertical axle only appears to have space for one bearing.
The main plate has bearing retaining features, which I don’t see how they are going to be made. You really should add bearings, axles, fasteners, and everything else before even considering actually making it.
The wheel is just a cylinder with a 3/8" hex bore. I assume this will be replaced with a bearing bore wheel and hub, but you really should model that.
The large sprocket is floating in midair. I assume it will be attached to the module, but you should probably put that in the module.
Overall, I’d say these are good starts but probably not ready for competition use. Keep working on it, and we’ll be happy to give more advice on future versions
We have noticed some of the things that you have mentioned during testing, and we are working on mending them. We have been experiencing some minor issues with power from the neverest motors when going over rough terrain, but it hasn’t been that detrimental to the functionality of the module. We are planning on swapping one of our modules over to a BB550 or bag motor to improve the strength of the steering loop. For the time being, we are also going to start trying to use the mounting holes on the face of the neverest, but we are having some difficulties properly tensioning our chain with that setup.
The wheels that we are using are the 3" banebots T81 wheels, which use a hex axle hub in the center, so we didn’t feel the need to include an accurate wheel design at this time.
Currently, V4 is not a completed design, we will add more bearings into it as we continue to work on it.
There are a few parts missing in these designs, but we are planning on adding them soon.
As of right now, we are planning on attempting to use swerve this year if Deep Space is a good game for it, but we are also going to build a WCD bot with a nearly identical chassis design so that we won’t be stuck with swerve if we can’t get it working as well as we would like. We do have our code working well, and we have had several prototypes driving without any major issues, so I don’t think we will have too much of a problem, but it is always good to have a backup plan.
We will continue to work on testing and modifications, and I will make sure to post updates as our designs continue to evolve.
We print our modules in mostly Nylon and Poly Carbonate, so we are printing materials that are significantly stronger than your average PLA. We have also done some destructive testing, and we have found that the wheel housing section for V1 and V2 can support up to 400 pounds in any direction. Actually, the thrust bearing that we were using would tend to fail before the printed parts would.
During testing, we have noticed almost no wear and tear on the modules, but we do plan on replacing all of the printed parts on each module with brand new parts before each competition. After all, each module only has 15 dollars worth of printed parts.
We have added some fillets into the designs, as the ones currently on grabcad are slightly out of date. I plan on adding the updated designs later today.
Honestly, I strongly suggest you don’t try swerve in-season unless you already have it ready at a competitive level before the season starts. The time and resources you will spend developing swerve to a workable point will only take away from the parts of the robot that actually score (i.e. the manipulators). Even more so if you try to build it in addition to a WCD, and have to make the superstructure work with either structure. All of this is a good recipe for a less-than-working robot.
Of course, you’re welcome to disregard this advice and prove me wrong . But if this were my team this is what I would recommend
I respect your opinion, but we have already had multiple swerve drives running relatively well, albeit with some minor bugs to work out (most of which we already have solved).
It is also worth noting that we have nearly 40 people on our team not including mentors, so we normally have half of the team not doing anything at any given point in time. A big part of why we are going to build the WCD is to help get more of the team involved in actually doing something, even if we don’t end up using it for competition.
I do agree that we will have to be very careful if we want the superstructure of our robot to be compatible with both chassis, so we plan on designing our robots systems so that the most important systems are on the front or back of the bot, so that we can still use them effectively even if we do not have any omni-directional capabilities.
