Are your drivetrains too practical? Are you not going sideways enough? Over one CAD binge session lasting one long Tuesday night, I created a monster. A being so cursed that the world should have never seen it. Behold…
Made possible by the new 5.5w motors a full four module drivetrain would only use 66w, leaving 22w for other subsystems. Other features include 300rpm driving with 2.75 in wheels and a FRC style coaxial design allowing for full 360 degree motion on the steering axis.
Real talk, obviously this wouldn’t be a great idea to use on the field. I just wanted to see if I could make a coaxial swerve with Vex Components. That brings me to the name. Why Project Malcolm? Well, one of my favorite lines in cinema history was uttered by the character Dr. Ian Malcolm in Jurassic Park. That line of course being…
…which basically sums up my thoughts after finishing the majority of the CAD. Speaking of cad, here’s a link to the Onshape document featuring the individual module and very quickly thrown together drivebase. The module has some issues still; some fundamental, some minor but I think its together enough for you guys to get the picture.
Wrapping up, this was a fun thing to try and design. Making the V5 ecosystem of parts (figuratively) bend to my will was a cool challenge. Whether or not the results were worth it is something I don’t really want to think about. Thankfully, this thing doesn’t exist in the real world… yet.
Unnecessary dramatic writing on my part aside, I was aware of other VRC swerves but wanted something a bit different to what they have. For all of these, the motor controlling the drive of the wheel is rotating with the wheels. VRC slip rings don’t exist (at least to my knowledge), so it takes a lot of work to not have motor wires destroy themselves. Project Malcolm doesn’t have the drive motor rotating with the wheels so the module can turn whenever and however it wants without risk of damaging anything. It’s one more step to being actually viable, but VRC swerve is in a canyon of impracticality so deep that my semi-major overhaul doesn’t really matter anyway. It was fun to work on though.
although it may be slightly larger, it might be better than a mechinum drive because it is easier to gear and would only take an extra motor if you gear the turning together.
Another design that might be valuable might be something like This (where the swerve is central), but instead of a differential, it is coaxial. This would get rid of the space issue because the wheels are central and wouldn’t get in the way of other components, and someone could pump 33w per wheel, so 66w(+5.5w for steering) swerve could be possible. The only issue would be the barrier, but this could be used for a later season when the field is flat.
The OP and all of these examples devote a swivel motor at each wheel, when all wheels need to align to function correctly. It would be quite easy to have the swivel motor on one wheel, and chain or linkage to rotate all 4 with the one motor. Probably more readily consistent to align the wheels mechanically than with sensors.
Of course, all of these are mostly for the fun of it.
Like this? Welcome to my seventh grade lockdown project I had 4 built, but have since scrapped all but one, too keep as a memento. Certainly not the best design, due to the awkward gearing to power the wheels, but still fully functional. Don’t have videos of it driving, since the programming was not a challenge I wanted to tackle; I didn’t know all of the requisite math. But it does work! Unfortunately, the 8 motors required for swerve to function makes it completely illegal in VIQC, but it is a fun idea to think about! Feel free to ask design questions, if you want.
Connecting the modules with chain like that is called Crab Drive in the FRC world and was swerve’s predecessor. It would cut down on wattage, but since each module has a lot of scrub when turning I would be worried it might not be enough power. Crab Drive also limits the maneuverability, and modern swerve code actually does have modules acting independently on occasion.
Still Ready to see 2 swerve VIQC modules, and then 4 omni wheels. Making that use only 4 motors, and using the pneumatics, I think that is perfectly doable and is actually what I am planning on if my org exists this year. (Check out 21037A if that team still exists (Selfish Promo)) By worlds, I hope at least 1 team has this.
Fun cad project I did a few years ago. This was before the addition of new gears, so I’m sure the height could be shrunken down quite a bit, making it more viable.
Swerve with mecanum wheels on it. Interesting choice, now the software needs to make sure the wheels are oriented correctly for it to be able to translate.
I had designed a swerve drive myself a while back. More of a semi-swerve than a swerve drive, it would function roughly like a swerve drive but be more realistic to construct. Heres some pics:
First, having dual traction wheels on the same axle will create significant resistance against turning (changing the direction that pod is facing). One or both wheels will have to slip. This could be addressed by adding a differential, just like it’s done in cars.
As for the second example, you have to consider that the tractive force contributed by each wheel is limited by the weight supported by that wheel. If the weight of the robot is distributed unevenly then some of the motors will not be able to contribute their full power towards the driving task.
In omni-wheel based tank drives, where motors on each side are mechanically interconnected by gears or chain, if one of the wheels loses contact with the ground then the power from all motors on that side of the drivetrain naturally flows to the wheel(s) that still have the contact, so no power potential is wasted.
