We are team 4729 (EMU Robotics) based in the University of New South Wales (Sydney), Australia. Even though our team has been around since 2013, this will be our first year trying to maintain a blog on CD. Hopefully it is a useful resource for other teams! We know we are a week late but here is our blog for week two of build season:
Nuked all tank/differential drive code + any code that references it (since we are using swerve drive this year)
Flashed PhotonVision image onto the Raspberry Pi 5 which will be tested next week
Our strategy decision as a team has been to focus on AMP scoring and ONSTAGE points first before attempting anything else. With this in mind, we split into different groups focusing on drivebase, intake, the AMP mechanism and climb mechanism.
We built both intake and amp shooter prototypes out of 3mm laser cut plywood.
We discovered that the wooden prototypes are pretty flimsy (the CAD models were designed to benefit from the stiffness of the chassis/chassis mounting points) but realised the prototypes are for verifying the geometry of our CAD models.
Continued work of adjusting intake CAD to mount with fasteners instead of tabs.
Intake CAD cannot be finalised or manufactured in aluminium/properly mounted to the chassis until the interfacing point with the amp scoring mechanism is determined.
The laser cutters we have access to require us to learn how to use a software called RhinoCAD…
A decision was made to finalise the drivebase at 440mm (box tube) so that the chassis rail could be fabricated. A 3D-printed jig will be designed to aid in centre-punching mounting holes for swerve drive modules.
We confirmed the manufacturing processes that we have access to and should hence design for (laser cut wood vs. CNC aluminium vs. plasma cut aluminium).
For aluminium: plasma cutting and CNC have similar costing and expected wait times.
We decided that designs will be made for a 6mm CNC bit as that is compatible with laser or plasma cutting.
This is one of our whiteboards which finalized our strategic goals. The bottom circle is a list of our must have items and in the boxes were ideas for additional goals (along with some crude drawings).
We decided that if you can do speaker you do not necessarily need to do amp and vice versa. So, we decided to do amp and not worry about speaker which will give us more time to work on a climber.
Although week three was a tough one, we did have good news like getting three new mentors for our team and having multiple interested visitors this week. Here’s what we did in week three…
Driver training
is off to a slow but steady start with members revising how to safely perform the role of the technician on our test (mini) bot.
Created intake subsystem and added support to lock intake using a beam break sensor.
Testing of the Raspberry Pi 5 has been delayed to next week for two reasons:
The camera for the Raspberry Pi 5 still needs to be purchased
Swerve drive code has taken highest priority this week.
Swerve simulation has been debugged
We were trying to make a semi-accurate swerve sim by physically simulating each turning and driving motor on each swerve module, but this proved extremely difficult.
In the end, we stopped doing that and only used the controller inputs to simulate.
Vision simulation has been fixed
We were using an outdated version of photonlib which crashed the robot everytime we tried to instantiate a simulated camera - updating it fixed the problem and fixed vision sim.
Drivebase team
A 3D printed centre punch template was created to make the process more accurate.
Our frame perimeter has been finalised at ~96 inches (~244cm).
A decision was made to expand the M5 holes on the mounting brackets of our MAXswerve modules to M6 size.
Started assembling the climber before realising that we had wrong and missing parts (we were attempting to assemble two climber in a box kits).
We then took it apart and inventoried all components.
Fortunately, we did make some progress as we now have all of our holes ready, either marked out or already drilled into the aluminium tubes.
Problems and Mitigations
We faced two major setbacks this week:
As previously mentioned, both of our climber in a box kits were either missing parts or had incorrect parts. Currently, we are not even able to assemble one complete climber out of a combination of both of the kits.
We plan to mitigate this sort of scenario in the future by taking a thorough inventory of any kits we receive before we actually need to assemble them. (Instead of hoping that kits come complete/correct from the vendor)
We are using MAXswerve modules but with Falcon 500s instead of NEOs. We purchased shaft adapters for our Falcons expecting them to work without issue on our specific motors. Little did we know that there are three versions of the Falcon and version 3 brought a completely redesigned shaft geometry (incompatible with the shaft adapters we purchased). This set us back approximately two weeks but thankfully we were able to swap to a version 2 Falcon. (Thank you 6432!)
