2024 Off-Season Project: Black Cherry
This off-season, we built Black Cherry (a 1678 clone robot (link to their CAD) with various Windupifications™) as a practice project for the 2025 season. The goal was to experiment with the design practices of top-tier teams, refine our processes, and learn from the inevitable challenges that come with building one of the most complex robots to touch the Crescendo field.
Here’s a breakdown of the project and the lessons learned along the way:
The Why:
Summary
There was a combination of reasons that led to us deciding to build an offseason robot the most notable being:
- Nocturne yearns for retirement. Our 2024 robot Nocturne had struggled to stay on the field since the end of the New England Championship. We used our alpha robot less than we did in 2023 due to the number of last-minute changes and wiring issues with that robot. We’ve put Nocturne through a little over 110 matches, and with the nature of the 2024 game, it was only a matter of time before it became very challenging to keep it surviving through the rest of the season.
- 2025 Build Practice: We feel we are in a good position going into 2025 but still have some areas where we lack expertise, and it’s always great to get hands-on experience with the hardware and software of top-tier teams that you wouldn’t otherwise.
As for why we chose to replicate 1678 specifically:
- Building an Elevator: Our team has not built an elevator since 2018, and that was not the most successful Windup machine
.
We have successfully avoided building an elevator since then, but we really needed to learn how to do linear motion right, especially in the age of COTS elevator solutions. - Over bumper intake: In past years, we’ve built a few over bumper intakes. The 2019 intake was before integrated encoders became the norm (used limit switches and an external encoder for control), and the 2020-2022 intakes were all pneumatically actuated. We’d never run a “simple” motor driven pivot using the integrated encoders on modern brushless motors.
We know 2024 was the year of under bumper as the intake meta, but we saw this as a learning opportunity for future seasons given the prevalence of over bumper intake in most games. - New skills to develop: The 1678 CAD used round tubing with tube nuts, a rack-and-pinion for the shooter pivot, a new bumper mounting system to try, etc. Really from the ground up their design was full of small details and ambitious packing we’ve never dared try in-season before.
What we Changed:
We made a few changes to the mechanical design of the robot:
Roller inserts
We elected to use the 6328-style rollers for the Intake and Amp systems.
Elevator changes
The one system we redesigned from the ground up to use more COTS products and try to limit the number of unique machined parts Citrus used was the elevator. We used several parts from The Thrifty Bot elevator kit, WCP elevator bearing blocks, slightly different sprockets, and used a common hex shaft at the top to support the upper set of sprockets. We somehow overlooked that this hex shaft prevented us from running the energy chain the way 1678 did, preventing us from fitting under the stage by ½”:
We re-designed to add some printed parts rather than custom plates to support the top of the elevator:
Compared to the original Citrus CAD:
Shooter changes
We swapped the lanturn gear that 1678 uses on their rack and pinion for some “normal” spur gears. We didn’t love the toothpick plates they use to support the outside of the pivot so we thought that doubling up on the rack/pinion might help give the support they’re getting from the toothpick plates.
Citrus Version:
Our version:
We also added a torsion spring pushing up on the shooter from the Y-split to try to reduce slop in the shooter pivot.
Y-split rollers
Throughout the season, 1678 switched what wheels they were using on their Y-split rollers. We tweaked it from what they had in CAD, adding some SpinTakes. We found this to be more consistent for us during testing.
LaserCAN
The release of Grapple Robotics LaserCan interested us and we decided to use it on the Y-split instead of traditional beam breaks. So far the only issues we have had with them are note dust getting stuck in the centering wedges that they are looking through. Our solution has been to turn down the sensor zone and clean them out frequently.
Something we love about these sensors is the live feedback they provide when powered with a flashing amber light indicating if the sensor is tripped and increases as objects get closer.
Great feature and fantastic way to check your systems when disabled. Check them out from The ThriftyBot!
CANopy
Custom Swerve CAN PCB adapter boards (CANopy). More on these in the future.
The Nubbin'
We had issues with the surgical tubing running from the Y-split to the intake being fragile. Eventually, someone thought of the idea of putting a simple “tongue” on the front Y-split roller (similar to what we used on our 2022 intakes). The tongue uses pulleys driven off the roller and an elastic ziptied from the tongue to a structural hex shaft. We are very happy with how this turned out after our initial tests. To start the match, it rotates back so the gray roller is tucked behind the main intake roller until we deploy the intake.
The nubbin is spring-loaded with surgical tubing pulling the hex standoff at the back toward standoff here marked in yellow:
Centering Wedges
3DP Y-split centering wedges: Instead of bending polycarbonate plates to create centering wedges like 1678 did, we designed a 3D print that sits against the Y-split plate and makes use of captured nuts for fastening.
Other changes
- Lightened Belly Pan (for wire management)
- Spray Painted Citrus Bot! (if you won’t do it @Michael_Corsetto then we’ll paint it for you)
Challenges:
This robot is extremely compact, which served us several novel challenges, especially for our electrical team. There are 6 independent degrees of freedom to work around. Some of these don’t move during a match, but pivot out of the way to access other areas of the robot.
Challenges
-
We initially did not pass the Isolated Frame requirement in R611. Our Orange Pi mounting was a little too compact with the screws contacting the frame of the robot. Thankfully this was discovered during a robot inspection at the Governor’s Cup but it made for quite a surprise.
-
We miscounted our PDH slots! Originally we planned to use the Mini Power Module to power sensors on our robot. Citrus did not… and we learned why. In total there are 20 Krakens on this robot and on wiring the last mechanism we realized our 20th slot was filled. We changed over to using two VRMs to power devices.
-
Because the robot is just so compact, the RIO and PDP are stacked on top of each other, both of which are underneath the Y-split:
The Citrus Y-split is built to flip out to give access to the electronics, but that still leaves the PDP very difficult to access. To solve this, we designed and printed a custom 3dp hinge so that we could flip the RIO out to do quick fixes:
Failures at the offseasons
At the NH Governor’s Cup and River Rage, we experienced:
-
Three forms of Intake failures
-
Bent intake plates. Typically the intake still works fine if it’s bent, but taking a hard hit causes greater impacts to other components.
-
Pulleys breaking has been the pain point of the intake. At the Governor’s Cup, the pulleys in the 2” leading polycarbonate roller broke repeatedly. We had two issues with the first being the flanges of the pulleys breaking easily. We suspected when taking hits and the intake was flexing the belts were putting a high degree of strain on the flanges until they failed. The second was a 3D printed spacer to retain the need bearings had an OD slightly too big so it could contact the printed pulley spinning next to it. We’d see it melt into the pulley or start spinning itself and melt into the lexan creating a live axle that could cause the bolts holding the roller on to back out. For River Rage, we moved to use a flanged bearing from Amazon, made the pulley wider to allow the belts to walk, increased the flange thickness, and adjusted our printer settings.
-
At River Rage we saw an increased failure in small pulleys with the ⅜” Thunderhex bearings being too large to leave enough material support. For our upcoming events we are moving from PA6-CF20 to TPU and transitioning the bearing on the pulley side to a bushing so we can increase material support.
-
-
Because of all the intake failures, by River Rage, we had 2 spares built up. From our third qualification match through Match 11 of playoffs, a new intake was put on for each and every match. The last over-bumper intake we did in 2022 was a lengthy swap we couldn’t always get done in time.
The Citrus intake is designed to be swapped quickly with just 6 bolts.
We timed a few of our swaps as sub-5 minutes: 2 minutes to remove/install the intake and 3 minutes for wiring. We highly recommend teams check out their CAD for how they’ve designed the mechanism to be removable quickly (good given we’ve had to swap it so frequently). Overall, very happy with what we’ve learned through adapting 1678’s intake and it’s been a big confidence boost that despite so many failures, every time we get the robot back on the field with a working intake in minutes.
-
Our LaserCANs lost power in the playoffs at River Rage from a cut power wire.
-
In some intense matches, the shooter can go askew and skip teeth in our rack. When this happens we can still Feed, Amp, and Trap but it is limiting.
The Results!
Governor's Cup
- At the Governor’s Cup, we ranked 5th and were the 1st overall pick by our friends 5813 Morpheus and picked 7674 RaiderBots to round out our alliance. We lost in the first round of the lower bracket and fought our way to the Finals tiebreaker to win our third Governor’s Cup in a row!
- Recap video here.
- Recap video here.
River Rage
- At River Rage, we introduced four new drivers to our qualifying rounds and all did a great job taking Black Cherry for a spin. We ranked 20th, were the first overall pick by our friends 6328 Mechanical Advantage, and selected 4905 Andromeda One and 8724 Mayhem from the Great Pumpkin as the 3rd and 4th members of the alliance. We then went undefeated in the playoffs!
We’ll be taking Black Cherry this weekend to Battle of the Bay and Girls Behind the Glass up in Alton, NH for the final matches of our Crescendo season!
Would We Do This Again?
No… just kidding.
For the 2024-2025 team we have no regrets going so outside our comfort zone. This project was a huge learning experience for the students and mentors. We saw younger students step up and have safe space to learn when cloning simple mechanisms while our more experienced students were challenged with redesigning entire systems from scratch. We tried a plethora of new techniques in the most ambitious robot we have built in 15 years.
The project also built up more respect for 1678 ) and how easy they make it look). It was also a good reminder of why machines this complex are not for us right now.
We’ve found great success the past few seasons focusing on “W8” and simplicity in terms of what we’re confident pulling off in season. This experience with new mechanisms and styles of construction will stay with us, however, and maaaaaaaybe we’ll have an additional degree of freedom in our future designs (this is NOT a license to design a turret next season @burbozo3467)
For other teams considering an offseason robot project we highly recommend it. In 2026 we may not go as complex as a 1678 machine, but a simpler concept will likely become our new offseason training method.
In the coming weeks expect a few more updates on software, project management, and other aspects of the project we may have left out.
