Hi all, 6328 is happy to bring back the OA highlights thread which we ran in 2023 . As of writing this there are nearly 150 OA teams for the 2024 season, a big increase from last year! With all this in mind we want to help shout out the contributions that you may have missed out on. We are also continuing our collaboration with FIRST Updates Now to have accompanying video content which you can find here or on their YouTube channel !
Starting for the 3rd week of build season we are looking for feedback from the community for what posts you would like to see highlighted! Please fill out this form to submit those posts:
5 Likes
We are happy to present some posts that stood out to us in Weeks 1 and 2, look out for a YouTube video covering these highlights in the coming week!
Design Recap
We listed out some various robot types we have seen, sketched, talked about, etc. The slides are linked with their source for current year designs.
Thru-frame Intake
We also worked on ideas for the frame. After seeing 95’s intake plan we were pretty sure we wanted to do something similar if we could make it work. However, we also wanted to keep access for maintenance on our swerve modules and not change how the MK4is mounted that much. As we sketched it we realized we could probably intake the note through a gap in the frame. The concept below uses 2x1 and 1x1 to extend the belly pan 1 inch lower than a standard Mk4i mount. The front intake rollers (white and blue) could pass the note through the gap between two 1x1s.
Intake Test
We were able to mount version 0 of our intake on the robot with mixed results. The note did go in the robot some of the time so that is a success, but it has a lot of work to do before we are happy with the design. In the process of building this, we had a design mishap that led us to need two 35t 5mm belts that we didn’t have, we were able to print them from Carbon Fiber TPU and run the tests with the printed belts. We also tested using hot glue to hold in our roller hubs on ⅜” rounded hex dead axles.
Intake Prototype Testing
2 days ago, some videos were posted about our ground intake, so let’s take a deeper look at the final concept and how we came up with this.
Unlike last year, we decided that quality goes over quantity. This plus the fact that the notes wouldn’t arrive before Wednesday gave us the time to design and manufacture a decent first prototype.
Sunday started with a meeting in which we discussed what different concepts would work and which variables have to be taken into account while prototyping. We came up with the following concept:
The Shamper & Top/Bottom Shooter Prototype Update
We’ve been working hard on trying to understand the different parameters that influence shooting a note.
One thing we learned: Shooting from the wing is going to be very difficult to do consistent!
We’ll be updating you on two different prototyping concepts in this post.
16 Likes
Thank you for doing these highlights!
12 Likes
https://youtu.be/nrmOjhBQkgE Check out the weeks 1 & 2 CRESCENDO highlights from Open Alliance teams!
Presented by OA Team 6328 Mechanical Advantage
OA Teams: Submit your highlights to appear on the video: https://forms.gle/RVFccAfXcLViAHoR9
2 Likes
sjsands
January 28, 2024, 12:43am
5
For Week 3 we want to showcase the continued development of many teams’ prototypes and how they are being realized into more cohesive robot designs. With so many teams hitting their stride this is going to be a long post!
Climber
Currently, we are looking at using hooks on the end of a telescoping arm to achieve the climbing aspect of this game. We came up with four different hook designs, got them manufactured, and are currently in the process of testing them out. We did a quick test using our 2020 robot. While we were unable to do a full test because of the chain being too low we were able to confirm that the hook design with the slot is what we will move forward with as it grabs the chain securely.
Super Structure
Because we want an arm on a pivot, we had to make a structure to support the mechanism. Below is what we have come up with so far. This plate will mount on the outside of the chassis and run the entire length
As we combine these ideas, this is what we have come up with so far as we have started to throw everything together in one model which you can look at here .
“Final” Intake Design
FRC 2024 Self Centering Intake Rev 3 #FRC3255
Additionally we have decided to pursue the self centering intake design for our intake/transfer mechanism. We are now in the process of creating our Alphabot, and integrating all of our mechanisms together!
Robot Archetypes
During week 1, we also developed 3 distinct robot archetypes that we wanted to look into pursuing. Throughout all of our archetypes, we are aiming for at most a 26x26 chassis and must work around the PDP we are running off of
Shooter Prototype and CAD Updates
We finally got this shooter working well. Ended up with the ThriftyBot 4” Solid Urethane Wheels in pairs all running off their own falcon.
We’ve still got a lot to work on to see if we can get the desired spin-up time with just 2 falcons rather than 4, and how well this works when it’s sitting on top of a big arm, but we feel much better about the shooter than we did going into the weekend.
Team 88 SLS Wheels
We recently tested SLS Printed Wheels from Team 88 on our traction jig, alongside our usual array of wheels. The goal was to see how these wheels were compared in terms of grip. We’d like to thank Team 88 and Formlabs for providing us these tires for us to test.
The data from our tests show that the SLS Printed Wheels provide superior traction. They outperformed other materials with the highest average angle and coefficient of friction (CoF) measurements. Here are our findings:
Does more speed = more better?
We were thinking exactly the same. That is why, in the off season we have developed a tool to analyse the performance of our drivetrain. One of the features we wanted to simulate is the effect of the CoG height, for which we could not find a tool online. You can read all about it in our previous post . You can find the tool itself at frc-4481-team-rembrandts.github.io .
But is this accurate?
To find this out, we used the off season robot TR-XS. We took a slow motion video of it accelerating over a certain distance and noted the time every 0.5 meters. The results were put in a graph together with the results from the simulation. This can be seen in the graph below. The average error is just 4.6%, which is pretty accurate.
Shooting in the Trap
As previously posted, ZeroBot can score into the Trap with its shooter just about touching the Chain, and set to about 75 degrees from horizontal so that the Note just grazes the Trap Door near the hinge before hitting the top of the opening, and then bouncing back down and in.
We figured out that ZeroBot can score in the Trap around 60 percent of the time once lined up (8 out of 10 repeat attempts without moving).
With these learned we went back to the drawing board. The concept that bubbled up is a much more robust, well-built prototype. This higher fidelity prototype would have a few main features.
Be able to control the left and right shooter wheels independently
Use 3” wheels
Be able to have 1:1 ratios and 1:2 ratios easily swappable
Have shorter shaft lengths so there are less vibration issues
Use more solid materials to construct the prototype so we can reliably test over an extended period without worrying about wearing issues
With those, we designed this. 12 3” 40A stealth wheels from Andymark, powered by two Vortexs(big fan of the thru shaft bit, makes prototyping super easy), .5” compression, two sides independently controlled so we can adjust the speed, 1:1 and 1:2 ratios configurations available.
Intake Prototype
We left off the last post with a few different intake options, after some discussion we decided to further pursue the four roller intake, mainly for the ability to easily also score in the amp. After a little modification to the intake geometry, we drew some new side plates to cut to test and hard mount these to a chassis. Here’s a video of that below.
Indexer:
As much as a shooter matters, the indexer in charge of reliablying feeding and holding the game piece before it is shot, is equally important. We tried just a top-bottom feeder, but discovered that low friction rollers need too much compression to move the game piece. This causes the shot to be affected, as the roller still touches the note as it is shot.
Instead, we have chosen to go with belts for the feeding. This choice was based on the observations that as the price is shot, it is pulled into a more oval shape, separating it from the belts.
Shooter:
We tried some more 2in wheels since our last post, and the general consensus was as we expected: they aren’t good, and they eat pieces. We went back to 4in fairlanes, and with a reliable feed, are getting accuracy around ~80% from 25ft with our jank setup (it moves around a lot). While some teams are going with 3in wheels, 4in wheels package fine in our design, and we cannot think of a negative that outweighs our shooting accuracy.
Shooter V/Index, 25ft
What’s slightly novel is that the rollers are only on one side of the note. It creates a fairly smooth spin but at the cost of overall forward speed.
Shooting from the back of the wing, this video proved to be consistent during this run but in subsequent runs was less so.
We’re implementing closed loop velocity PID control to improve consistency. A lot of the consistency issue is from variations in feeding the note as many have pointed out.
If there is a team who is doing great things that we missed, please fill out the submission form with the things they are doing in Week 4!
16 Likes
https://youtu.be/WdgjeQTORko Check out the week 3 CRESCENDO highlights from Open Alliance teams!
Presented by OA Team 6328 Mechanical Advantage
OA Teams: Submit your highlights to appear on the video: https://forms.gle/RVFccAfXcLViAHoR9
4 Likes
https://youtu.be/W84nKjbLmSA Check out the week 4 CRESCENDO highlights from Open Alliance teams.
Presented by OA Team 6328 Mechanical Advantage
3 Likes
sjsands
February 12, 2024, 11:36pm
9
Shooter and collector update!
Starting with the collector, here is the roller configuration that we have been very happy with. The four rollers are passive and rotate on a fastener or standoff, nothing to fancy.
It’s All Coming Together
Got it all working together now - arm, shooter, amp scoring, intaking. And I swear that a software post about it is coming any day now - they’ve been a bit busy. Until then… How about some more testing videos!
We’re looking forward to hearing what teams are doing in Week 6! Please fill out the submission form to help us aggregate posts from the community!
6 Likes
https://youtu.be/PT6Hj0dP_K8 Check out the week 5 CRESCENDO highlights from Open Alliance teams.
Presented by OA Team 6328 Mechanical Advantage
3 Likes
sjsands
February 20, 2024, 1:08am
11
We started climber assembly tonight.
Before committing to cut CF tubes we agonized over arm positions for the climb, utilizing interference detection in OnShape’s section view analysis.
To fix that, today we attached a temporary fixture in order to find a good place to put a tube to support the note as it gets scored into the amp.
111 Amp Success
We plan to make this roller passive but on a bearing so it doesn’t add unnecessary friction.
On the other hand, our shooter pivot is working out splendidly. We have a ton of angular accuracy, and it’s super duper rigid. Using a linkage to pivot the shooter allows us to have a really accurate shooter once we get the “shooter” part consistent.
111 Shooter Pivot
Note simulation
A strong part of our codebase every year is being as close as we can to being able to run complete driver simulation, able to be run before the bot is built.
For this year, that means taking the output state of the shooter, and showing how the note will fly through the air
Moving shot
Alright! Theory explanation!
So consider a bot moving to hit a target, relative to the bot, the target is moving in the direction opposite to the bot velocity. The distance that the goal relative to the bot will travel depends on how long it takes to get to the target. Uh oh! Problem! Now that you compensated, it will take a different amount of time to reach the goal location, so you update based on the new timing estimate for a few iterations until its within reasonable error.
Auto-aiming: the Shooter LookUp Table:
One of the most important functions of the adjustable angle shooter this year is automatically determining shot angle and speed. This is important to reduce cycle times by allowing us to travel less distance per cycle, and allowing us to shoot even if our partners or a defender are against the subwoofer. To do this, we use Photon Vision to determine our distance from the goal, and then we check our lookup table.
To determine our Arm’s Angle and Shooter Speed we will be re-purposing code from Rembrandts 2022 robot Vision Look Up Table. We love the concept of using known positions since this provides a more robust and sturdy approach with less variables to handle. So what is a LookUp Table?
First Shootings
We got our first shooting today. Yesterday and today we continue to work on our arm and shooter programming. We solve our problems with the shooter after changing the spark max’s.
Today we tried the robot in the corridor to see the range of it. Here are some videos from today. In some of the videos, the note is not following a straight path. We think that it’s because of the weight distribution of the note. After it got damaged two times we put lots of tape to make it better. However, it changed the weight of that part. This causes the note to rotate almost 90 degrees. This only happens when that part comes to the side.
It’s Alive (For real this time)
Intake and storage changes are finally done and software team has been working on characterizing the mechanisms and swerve the last couple of days. Mechanically we are just working on some weight reduction on our intake.
Most of our code seems to be working correctly and we are just working on fine tuning shooting while moving and some interpolation tables.
Here are some videos of some robot testing and tuning, the final code should allow the robot to be faster than what’s shown in the current videos.
This is the last highlights post of the build season, thanks to everyone who submitted posts or tuned into our thread or YouTube!
5 Likes
https://youtu.be/M1bPthabRfY Check out the week 6 CRESCENDO highlights from Open Alliance teams | Presented by OA Team 6328 Mechanical Advantage
This is the last video/content piece from OA in 2024. Be sure to keep following OA team threads for more posts and content!
4 Likes
