Build Blog Update 1:
I’ve just been informed that “It was actually Fritos, not Doritos”.
Did we get the CAD done Saturday? Nope. Are we close? Hopefully.
Mechanism Overview
Intake
This mechanism is our furthest along. We have already done some testing…and found some bugs. Overall though, we think we are close.
We tried several different techniques for allowing the notes to center themselves. First we attempted surgical tubing which was a little less grippy than the silicone a lot of teams are using.
When that showed too grippy, we attempted to make it even less grippy with electrical tape, and then with gaffers tape. When that proved to still to be too grippy, we removed the surgical tubing entirely from the rollers and left raw aluminum - which finally did the trick. If other teams want to attempt this our rollers compress the note ~.125" This lets the notes go sideways, but still has plenty of grip to get them into the robot.
We are still using grip material (silicone) on our back roller to initially get the note off the ground, and of course we have our green compliant wheels which help initially grab the note.
Elevator
While maybe overkill, we opted for a 3 stage continuous elevator to be able to stay short, and maximize our height to 48" for a potential trap score. We opted to use a timing belt (mostly) in tube to try and keep things running smoothly. We have most of the parts cut for this, and will be assembling soon.
Shooter
This one was a packaging challenge. We have 4 things spinning independently, and had pretty severe dimensional constraints to get the angles we wanted. Mounting lots of NEOs here would have been painful, so we are very thankful for the packaging of the NEO Vortex/SPARK Flex.
We can adjust from 50 degrees to -25 degrees, and are driving everything with a laser cut section gear (made by Fabworks). Use code #frc461 for a discount.
We also embedded our MAVcoder into the pivot to be able to measure our angle easily. We believe we can connect it to the SPARK flex for simple wiring.
Climber
Full Assembly
It’s looking pretty…crazy. This has been a very fun packaging project, and has taken a lot of effort to get this far. Now…we build. Only a few weeks left until week 1 and our programmers would like a few minutes with the robot - probably.
Whoah, love Wolfpack’s design there!
Programming Update
Augmented Reality
Being passionate about reaching out to our community, we realized a year ago that we had a problem to solve. There are lots of demonstration areas that aren’t safe or suitable for showing off our robot. However, as the complexity of our robot grows, so does the challenge of letting others learn without movement.
Thus, we decided to take on the endeavor of integrating augmented reality into our robots to allow others to see visuals, CAD models, and video content easily. The above video is a demo of what we currently have developed. We chose the paid program 8th wall as it allowed for multiple developers. It also had tons of sample projects to work from.
We are excited to continue the development of this project, especially with its ties to this year’s theme, Technology Inspired Art!
Making Custom Stickers
(Written by Gabby, Digital Media/Imagery student, with the help of mentors)
If you visited our pit last season, you may have noticed we had some cool student-made stickers up for grabs. This year will be no different!
How do we make our stickers?
We designed the stickers using Canva and Adobe Photoshop. These initial designs and variations were presented to the team during a weekly design meeting. The designs received some feedback and then went up for a vote on which ones should be passed out in our pit during competitions.
Once the designs were finalized, we uploaded them to the Cricut Design Space app and used the newly released “Create Sticker” feature to convert them into print-then-cut objects. Designs were printed on glossy and holographic printable vinyl paper using an inkjet printer and then were cut out using a Cricut-brand machine.
Are the stickers durable?
Well, that depends on what you use them for. Using stickers on an everyday object, such as a computer or a water bottle, will fade over time. But, if you leave them in your room, on your wall, or any object you don’t carry around daily, they should last for years. The stickers could be made more durable with a laminate coating. The coating would give them extra protection from scratches and fading, but this would have increased the time and cost to produce the stickers.
Why do we make our stickers instead of ordering them?
We have found that making stickers is much more fun and sometimes more convenient than just ordering them online.
Sending out the sticker designs to get printed would be easier since it would save manufacturing time (and sometimes frustration). Many online stores offer this kind of service, but shipping the stickers will usually take a few days. Also, most online stores only offer a minimum quantity requirement and limited size restrictions. One advantage of making them ourselves is that we have complete control. We can customize the size, shape, and color. Sometimes, we can design and produce the stickers in just one regular team meeting. The quick turnaround time is convenient for making small-batch stickers for robot labeling and other projects. Overall, we find making the stickers fun because it is a hands-on experience.
Be cool man
Hard not to with these awesome stickers on our bot.
Save me a dragonfruit sticker and I’ll make sure to save you a next-gen 461 sticker!
2 Weeks Until Our First Competition.
CAD Update
We continue to add quality of life features, but the main components are about there. And…the climber is finally designed.
Climber
Some fun features to our climber:
Dual hooks allow us to grab different points on the chain. The lower hooks assist with a potential Trap mechanism (more testing is needed there). The higher hooks grab the chain higher to make harmony easier.
Shooter
…forgot the graphic in my last update. This mechanism is pretty sweet, so wanted to share
Manufacturing/Assembly update
The majority of the metal parts on our robot are now cut! We primarily use our Shopbot CNC router for this, but outsourced a few to Fabworks where it made sense.
We are well under way with getting the robot assembled. The drive-train is wired, and we have all major mechanisms mounted, except for the climber (it’s close). Found a few bugs, but overall it went together fairly well.
This is our first continuous elevator as a team, so it’s not too surprising that the stages aren’t going up in the right order (ugh. Friction). We will probably need some CF springs to bias things the way we’d prefer…but we’ll get there. This state is “good enough” for now while we continue with other priorities.
Is this a quality of life feature?
Systems Online
Well. Our new Kat is assembled and wired - except the climber. Even got some practice bumpers on, and they look pretty snazzy (these need some small adjustments before the inspectors will let us on the field). We have found a couple bugs we are addressing.
A big one is a severely under-powered shooting. We believe this is because of a poor hand-off from our feed to our shooter. We increased the speed of the feed t our shooter, and decreased the compression slightly on the release (and have a design ready to further decrease it to zero). Hopefully we find this improves the power.
We have also found our intake has a few jam points, and requires fairly high current just to run. We made a couple adjustments to gear it down, and to reduce friction, and we hope this improves.
But…we can drive fast, intake, and score a note, so that’s pretty cool.
I know this post was made a bit ago but shoutout to Twinkle for getting the AR project started! (It looks awesome by the way! Love the creepers)
Getting Close…
With only a couple days until Mishawaka, we are pushing hard to get a competitive robot on the field.
Onstage
We found some last minute issues with our climber. Assumptions on our CoM were not quite right, and the robot struggled to get all the way off the ground, even with the lowest hooks.
We thought through a process for using the back of the stage to help with the tilt. It is a bit more involved for the driver to be able to get the shooter over the chain…but it’s doable. Here is the result:
There is room for improvement, but we think this is going to be good-enough for our week 1 event.
Shooter Power
Our shooter power was a bit disappointing during our initial startup. We were finding (similar to 111) that the feed urethane belts were slowing down the shot dramatically. We were struggling to do a close-range shot.
To fix this, we reduced the diameter of the feed rollers closest to the shooter, from a 1.25" roller to a small crowned pulley. This created a “taper” on our belt run, so that we have about .5" compression at the beginning of travel, and have effectively 0" of compression at the end of travel…We also added a 3rd wheel to each end, to make sure we had consistent engagement on each note.
This gave us the power we were expecting, and we are beginning to run some cycles.
Amp
We are dialing in our set-point for the amp. We aren’t quite happy with it yet, but we think we will be before the event and with a little practice.
Auto
We haven’t given our programmers a ton of time…but they delivered something pretty good. Just need to work on ensuring consistency, GIF doesn’t show timing very well, but we are sitting ~10s with that auto. Plenty of room to go faster and get more eventually. Who knows? Maybe before Mishawaka.
Cannot wait to see you guys in 3 days!!! You look amazing as always.
Mishawaka Event - What Went Well?
We had a lot of fun at the Mishawaka Event over the past weekend. We ended up ranking 2 at the event and, based on our scouting, had one of the top teleop performances at the event. Our last minute Amp set-point adjustment was clutch and helped us amplify a lot of speaker-only bots in quals.
We accepted the #1 alliance’s invitation and were beaten by both alliance #4 and #6 who would ultimately make up the finalist and winning alliance.
Mishawaka Event - What Did We Learn?
This game is brutal. On our final two matches in the elimination rounds, we were primarily turned into a defense only robot. We were experiencing severe intake issues. We found destroyed bearings in the intake, and it was ultimately jammed. Looking at it after the event, we believe we took a heavy hit on the intake side, which caused our intake to buckle.
We also didn’t find ourselves climbing much at this event - certainly not as much as we expected. We climbed once in qualification matches to help with the Ensemble rank point, but the alliance was unsuccessful in getting the RP. We climbed twice in Eliminations while our intake was broken, since we felt that was a way we could contribute a few more points.
Our Mishawaka climber is pretty slow to align, so it was faster for us to get the last second speaker shot, and then park for the same number of points.
Changes/Updates
RIP Climber
Since our robot was within 1lb of the weight limit, we knew that adding material to make the robot more robust would be a challenge. To find the weight, we have decided to remove our climber - and not climb for the Columbus event. We are working on a plan to climb with our elevator which should be much lighter and likely faster.
We are a little worried that we cannot help with the Ensemble rank point, but hope to be able to make it up with lots of Melody rank points…and a more robust and consistent performance.
Making The Intake More Robust
So buckling…
Euler’s Buckling Formula:
So. to increase the force required to buckle, we can decrease the length of the “beam”, or increase the moment of inertia. We found a way to do both.
To increase the moment of inertia, we used 1/4" plate to reinforce the intake between both pieces of 1/8" 2x1 tubing. By doing this, we effectively tripled the thickness of intake in some key areas. We didn’t do a full analysis on exactly how much this increases the moment of inertia, but if we approximate it to be a rectangular beam, tripling the thickness with give us 27x increase to area moment of inertia.
To decrease the effective length, we have machined a new corner standoff for our MK4i swerve modules. This gives us a mount closer to the center of the buckled plate, which will offer us 4x the buckling resistance roughly.
The estimates are a bit bold…but quick napkin math says we increased our resistance to buckling by 108x. Hopefully that helps a little.
Autonomous
We continue to work on our autonomous routines. We are feeling pretty good about our 4 note autonomous (close). We are also working on a 3 note auto that grabs one piece from the center-line. We believe both of these will be useful for the upcoming event.
Final Improvements for the FIN Columbus Event
New Shooting Setpoints
Tall Shot
With some opaque 48" robots that block shots showing up at events, we thought it was appropriate to have a countermeasure and be able to shoot high from the podium.
ALSO…if there are any teams interested in becoming 48" opaque shot blocking robots, we have written instructions and free COTS supplies in our pit - and are ready to help. We believe it’s possible to add this feature to a kitbot while maintaining all existing functionality. Same is likely true for other architectures. This is intended to be a new feature…not your robot’s whole personality.
Shuttle Shot
We have also seen how at some events, it has become important to get lots of notes on our side quickly to be able to get that melody rank point. We added a set-point for this too.
Autos
We felt our auto was a bit lacking at the Mishawaka event, and have addressed that issue with two new routines we plan to utilize in most matches. We have tested the red and blue sides in our shop to help reduce any surprises.
4 Note Close
3 Note Far (Source Side)
We wish everyone good luck at their week 3 events!
Hey, I was curious if you guys were willing to post your advantage kit logs on your github or to this build thread? I’m really interested in seeing your camera and odometry estimates compared to the match video. We’ve noticed that we get a lot of error in vision pose estimates in the center of the field, which hurts our centerline autos.
Congrats on a great Columbus event and a well earned win! It was great to play with you! Not as great to play against you lol
Hiya! Disclaimer, I’m typing this on a phone, so I’m not going to be able to include as many pictures or files as I would like. Please ask for further clarification on anything you like and I can get back to it more after this weekends event.
We are definitely down to share our logging data, but since we attempt to keep logging from every match and practice backed up, it’s a lot of data to sift through. This, we will aim the share things based on people request of what they want to see.
‘’'Make vision more consistent"“”
Prior to this week, we found data from further than about the podium to be a little more noise than we liked, so we used limelights clipping feature. We clipped the bottom y portion so from a distance our robot couldn’t get a far reading of a tag. We are working on more ways to detect position from further, but that’s still a wip.
As for the log files, here is one from our Mishawaka competition. Since then we have began to log more data points, I just haven’t gotten around to pushing the data from Columbus. With the idea of putting them on our GitHub… it is such a huge amount of data that it is just too much to be putting it all on there. I can send you some more log files if you wish to see them.
Log_24-03-03_11-05-19_q57.zip (5.8 MB)