We haven’t tested it on the V2 balls yet because we’re planning to change the gear and test again. Once that’s done, we’ll test with the V2 balls.
Climb Update
In the past two days, we’ve manufactured, assembled, and tested our deep climb concept. The design features two arms that rotate against the lower plate in the cage, similar to Penn State Ri3D. Achieving the right center of mass has been challenging, requiring us to experiment with different weight placements to both match the real robot’s drivetrain weight and achieve balance.
The system CAD:
Some tests:
Key conclusions from the tests:
The gear ratio needs to be increased (currently 1:120 on each Falcon motor).
The arm geometry is not yet optimal, during most tests, one arm would get stuck in the cage hole and fail to reach a low enough position.
The assembly was not rigid enough, and some parts bent (some issues were due to assembly and CAD errors, leading us to improvise).
The center of mass must be extremely precise—any imbalance causes the robot to tip.
CAD Update
Yesterday, we completed the manufacturing of the nearly final coaxial rotating intake. The next version will have only minor changes and will be the one used on the robot. This intake version is fully functional and will be used for software testing and further tests.
The system CAD:
We are also experimenting with some custom TPU intake wheels so stay tuned for an update on that
We are also making progress on the elevator geometry and are currently working on the cscade drive.
Programming Update
The control code for the systems, including the basic logic for intake and angle changes, is finished and will be tested today with the new prototype. We are also working on our autonomous alignment and have seen some promising results.
Today, we will have our drive team meeting to finalize robot control methods. Expect another update soon!
11 Likes
This is very impressive. How do you do that so quickly? Are there any special processes in your software development you use?
2 Likes
Nothing out of the ordinary. Before we start writing the code, we create a documentation document with all the details and plans (I can send it if you desire), which makes the actual coding process much faster.
In the offseason, we developed a state machine framework that is surprisingly simple, adding states and functionality is easy and fast. We will probably explain this a bit more once we get into the software in more detail.
3 Likes
I would love to hear more about the documentation your team has written (mainly related to state machines) in a full post!
Successful Deep Climb
After our first deep climb prototype didn’t work out, we were eager to try something new. One major issue with the previous design was its sensitivity to the center of mass. To fix this, we swapped the two smaller levers for one wider one to increase the contact area with the cage. We also increased the gear ratio from 120 to 200.
Check out some test videos below:
Today, we’ll focus on improving the lever geometry to make it more stable.
We also have another deep climb concept already CADed, today we will discuss if we want to test it as well.
CAD
We finished assembling and started testing the final coaxial rotation intake prototype . We’ve only done intake tests so far, but they look promising. We’ll keep testing today.
Here’s a video from the intake test:
The elevator is almost ready, just the gearbox and mounting left to complete. We’ll post a more detailed update once it’s finished.
CNC
We had an issue with our CNC where the Z-axis wouldn’t home. After troubleshooting, we think the problem is with the CNC controller. We’re working on getting a new one, which should fix the problem.
Programming
Yesterday, the driver team met to discuss robot controls and how the driver will control the bot. Today, we’ll break down the software plan and start writing the code. I’ll post a more detailed update once we have more progress.
8 Likes
Yes, please. We now start writing our code and that would be helpful.
2 Likes
This is our code docs and plans, still work in progress so maybe not up to date and you may see us working on it.
Feel free to google translate this doc.
2 Likes
Thanks a lot!
If you have any questions feel free to ask!
2 Likes
What is “sorting” state? How does the elevator know which height it should go to?
As demonstrated in this clip
After intake, we will probably do a littel back and forth to center the coral. This action is the sorting state (if I am correct, it’s not yet in the triggers in the doc, so it may be a bit confusing).
The operator will choose on the keyboard. I still haven’t written the docs for it as well.
1 Like