2468 worked on an off-season chassis design to address maintenance issues we encountered over the last couple of seasons maintaining the sprockets and chain.
The same project has served as a platform of our programming team to improve the control of the chassis, the AUTO functions and vision. We’re exploring the use of RealSense cameras.
You can find the CAD and programming files on the Technical Resources section of our team website.
Just curious, what maintenance issues did you have and how did you fix them in your off-season project?
In 2014 and 2017 we used chain in tube as inspired by 118/team221.com. Over the season, we found inspecting and maintains the sprockets and chain a challenge. One of the goals of this project was to investigate maintenance port effectiveness (strength, ease of access, etc.).
Did you have any issues with chain in tube? Which sprockets did you use?
Our students worked on this project for the Fall semester.
They wanted to address a few issues we had
- Chain in a tube design maintenance
- Learn best practices in design, manufacture, wiring and programming
- Improve on previous skills
- Implement new code development
- Build an 8 Wheel West Coast Drive since we never have before
- Learn to use new to use Vertical Mill
- Manufacture in house bellypan
Here was our reasoning and plan of action for the project. In the end, we actually eliminated some items from the plan. I have deleted those from the plan below. Those will be investigated in future projects.
The WCD Mule Project Proposal
To create an 8-wheel WCD drive train in the in-school offseason that will:
- Serve as the potentialbasis for the 2018 year chassis design.
- Provide the software team opportunity to develop code for a chassis prior to the kick off.
Design a chain-in-tube west coast style chassis (CITWCD) with access ports for chain and sprocket maintenance (without having to disassemble the robot, drill out any rivets, remove any gussets, etc.)
Evaluate the acceleration/speed and voltage/current using mini CIM vs CIM motors (3)
Provide a real-world design challenge for training and practice
- Develop better encoder wiring practices
- Develop component placement on belly pan
- Provide a real-world wiring challenge for training and practice
- Develop auto-shifting software
- Ramping software (to manage current spikes and avoid wheel spin during acceleration)
- Velocity control with Position feedback
- Navigation library for constructing AUTONOMOUS sequences
- Addition of vision to the navigation library
- Provide a real-world programming challenge for training and practice
- Must publish mechanical designs and software prior to 2018 kick off event to allow use of the design and software developed during the offseason in the 2018 season.
- Cannot use any parts manufactured during the offseason on the competition robot. In other words WE CANNOT TAKE PARTS OFF THE PRACTICE ROBOT AS SPARES.
Students developed access holes in the bottom of the tubing rails for the sprockets. https://d2t1xqejof9utc.cloudfront.net/screenshots/pics/f0b4c20ee7396c0c268eb6f933b528bc/large.JPG
You can find the CAD for the MULE Project here: https://grabcad.com/library/frc-2468-mule-design-project-1
Software team researched and experimented with shifting automatically. Thanks to FRC 33 for the resources they published in regards to their designs. In the end, the programmers were not pleased with the results.
Software was successful on ramping, velocity control/feedback, navigation control and Intel RealSense Camera implementation. Quote from one of the programming students;
A game changer in FRC vision, the RealSense R200 contains a high-speed RGB camera and an integrated depth sensor. Using a laser, the camera projects a matrix of light onto its surface to obtain depth measurements. The depth sensor provides real distance measurements, rather than using the traditional proportions and math used in FRC vision logic, essentially providing a R,G,B and Depth value associated with each pixel. This off season, we used the existing LabVIEW wrappers for previous RealSense camera versions and the API to develop wrappers for the RealSense R200 such that it is compatible for LabVIEW development.
Programming code for Mule Project: https://github.com/FRC2468/MulePublic
This was the first offseason project we have taken on. It was a beneficial project for our team and something we will continue to do in the future.
We had some issues last year with chain in the tube but we believe it had to do with use steel mandrel rivets. From time to time, we had to remove a side rail from the robot to remove the chain and sprockets. To do this required removal of the entire side rail from the bellypan (drilling out lots of rivets).
We wanted to be able to check and maintain the sprockets without the removal of side rails or the bellypan.
We used the 221 LLC double sprockets (no longer in manufacture). This is the second year we have ran chain in the tube with the 221 LLC sprockets. We are happy with the results as long as the game allows #25 chain. In Stronghold, we ran #35 chain and we did not do a chain in tube design that year.
Heavy sprocket wear with the team221 sprockets (both alloys). In both years, our design team didn’t factor in a method to conveniently inspect the sprocket wear. Moreover, replacement involved considerable rivet removal (time sink).
It’s quite nice when the design includes feature for inspection, lubrication, and eventual replacement. Chain/sprocket replacement time was greatly reduced.