Learn about the GrabCAD Platform
Get to know GrabCAD as an open software platform for Additive Manufacturing
Visit our new homepage
For most of our preliminary test versions of the drive system, we used several Tetrix robotics parts for the internal components of the Swerve drive. While the tetrix parts that we used would likely be adequate for an FRC version of this drive system, they lack much of the strength that other FRC robot parts normally have. With this in mind, we are working on creating a more heavy duty version of our drive system that focuses on stronger internal parts that can handle stronger drive motors. The question is, what type of hardware do people suggest for the internal mechanisms of this design, and what are the benefits/drawbacks?
I have several questions and comments on your design.
1. Have you though about possibly moving the second chain run from its current position to in between the forks? This would allow you to do dead axle and move your bearings into the wheel rather then in the forks. doing this would make them stronger and stiffer.
2. Using belts rather then chain for the first chain run might be better as the pulleys can be 3d printed unlike the sprockets. This would allow you to quickly change the gear ratio by swapping pulleys. Also a chain spinning at 5000 rpm wont be as efficient as a belt. The only thing is you will need to make the cim able to slide to adjust belt tension.
3. Currently your drive ratio looks very close to 1:1. You will most likely need to either swap to a bigger sprocket or belts to get the ratio down to something around 6:1
I really like the idea of a 3d printed swerve but I am not sure it would be robust enough for actual competition.
We are currently working on the next version of our design, which will have the chain moved to between the wheel and the housing. The main reason that we had the chain exposed before is that this is just a prototype version that we are using to test programming.
We opted to use chain as opposed to belts in the design, as we have had problems with belts slipping on earlier prototypes during testing, and chain typically has less backlash than belts when they are properly tensioned.
Also, CIM motors only produce an RPM of 5000 when they are completely unloaded. With our calculations/test data, we have found that the fastest that they will likely spin is around 3000 RPM.
Why would chain be less efficient than belts?
Currently, this model has a drive ratio of about 4:1. As most swerve modules that have a gearing of 6:1 use 4 inch wheels and we use 3 inch wheels, our swerve will produce a similar strength/speed output to most other swerve drives.
Finally, we are printing our modules out of Nylon, and destructive testing has shown that each wheel can take a little over 400 pounds of pressure and 250 pounds of lateral force before failure. Even if this system was printed in PLA, each wheel should be able to handle over 100 pounds of force. Considering this, the bearing that we use in our module will most likely fail before any of the 3D printed parts do.
Update, we have completed much of the testing phase of our prototypes, and we are finalizing the designs of our competition ready swerve modules. We are currently working on 5 designs that will each have different strengths and weaknesses, so that we can choose the design that best fits the strategy that we choose for competition.
We have also decided to ditch the annoying round axles and bushings in favor of 1/2 inch thunder hex. This will eliminate the only serious problem that we experienced during our testing, being that tetrix parts come loose very easily, so we had to constantly maintain our modules between drives.
The latest prototype that we have began assembling, files will be released soon, utilizes hex axle and vex bevel gears, meaning that we will no longer have to have our bevel gears laser sintered, which is a big bonus.
We will be releasing more designs soon, so stay tuned.
We will release our code before the start of the season, but it currently needs significant work done to eliminate bugs, so we do not feel comfortable releasing it at this time. We do, however, have several google spreadsheets that demonstrate all of the math necessary to code swerve drive. If you would like access to these spreadsheets, send me an email at bartlethowarde@jacksonpec.org.
