Weight can be a factor depending on how heavy each module ends up. The amount of motors used is another problem many teams that run swerve may run into. Typically you will end up using 8 PDP slots to run a swerve that leaves only 8 slots for all of the other functions on your robot.
Honestly, the mechanical production time and maintenance isn’t really much more then a standard custom skid steer chassis in today’s FRC. This was more true 5 years ago. With 3D printing and the amount of COTS solutions available (AndyMark, Swerve Drive Specialties, WCP) mechanically it is not that complicated or difficult to pull off. It isn’t even much heavier then a skid steer.
The biggest thing really is programming and then just driver familiarity. For these reasons I would still suggest taking time in the offseason to build one and develop the code for it, and to give drivers a chance to get used to it. That said, imo this should be done for any drive train really, even skid steer.
Generally expensive (encoders, motors, motor controllers, gears, plates if a COTS module)
Takes up room on the corners
Draws quite a bit of power and clogs up the PDP.
Heavy (unless you’re a big powerhouse)
May be a waste of resources (time, money) better put to a competitive mechanism on your robot.
Careful not to underestimate programming complexity - the 90 90 rule applies here, big time. You don’t want to end up uncontrollable or dead on the field cause your programming team didn’t have time to fine tune PID or get the encoders reliable.
Really well actually. It helped that we had a mentor who was good at programming. It wasn’t the programming that was the problem though, just the encoders that we spent a good amount of time trying to figure out.
In 2018 we didn’t have problems with the swerve. This year, the only problems we had were electrical problems. We had an encoder wire detach one match and the can bus fail the other 2 matches. So like others said, a lot can go wrong.
A huge one is weight, the Swerve Drive Specialties Mk2 weighs in at 4.9 pounds, where the AndyMark weighs in at 9.4 pounds. The Swerve Drive Specialties is also shorter, coming in at 8.26" tall, and the AndyMark coming in at 12.04".
The SDS MK2 modules are heavily based on (if not exact replicas of) 2910’s swerve modules. They’re a lot more mechanically robust than AM’s Swerve and Steer modules. They are also designed to include sensors, so they don’t have the same problems that the S&S absolute encoders have and they have built in drive encoders rather than using the finicky AM CIMcoders.
Just to echo what others have said and nothing against AndyMark but I don’t see much reason to ever buy the S&S modules at this point. The WCP module is also heavily baed off of 1323s design, although there is a lot less documentation available so far for it (missing CAD etc still). I expect R.C. will have this before the 2020 season starts.
You’re not actually buying it from 2910, you’re buying it from Swerve Drive Specialties, which is a company owned and operated by one of 2910’s mentors. As long as that company meets all of the requirements for a vendor laid out in §10.1 (which I think they do) then it’s a COTS part.
I think that 3707 used a modified Andy Mark swerve module for their 2018 robot and probably did the same this year, so there might be SOME benefit. I’m unsure of what it would be though because we made ours custom and most top swerve teams do as well.
The fabrication and assembly od Swerve modules is much more time consuming than any other kind of drivetrain. A Swerve module contains many more parts, many of which require fabrication and assembly geometric, dimensioning and tolerances that are much tighter than your average Tank Drive kit.
We are in the process of building our first Swerve Drivetrain, using a Swerve Module we have developed internally. Fabrication and assembly during the build season will monopolize a good portion of the Build Team and that will definitely play a role in whether we decide to use the modules in competition.