Even with an "easy" schedule, rankings can turn on a dime. A wire wiggling loose, a bolt head snagging, or a haywire auto can turn a probable win into a hard loss quickly and remorselessly. Our pit crew, drive team, and strategy group were top-notch through qualifications though, and got us to the end of Friday with an undefeated record and a 4RP average (a first for us at Champs). That gave us a lot to talk about when we got to...

This picklist was the hardest of the year for us. There were several teams in strong contention as a first pick, and the first pick decision alone took over 2 hours. Our eventual decision was that 4-time division champions 3476 Code Orange was the perfect partner for us. They were the second highest scoring team available, but did it shelf loading rather than floor loading which would keep us out of each other's way to maximize scoring. We called them to our hotel conference room for a joint picklist meeting, and spent a further 3 hours working with them to build a picklist.

The 1678 picklist process is as follows:

1. DNPs. The first thing we do is create a list of teams we can't see ourselves playing with. Reasons for being DNPd at champs can range from low performance to consistent penalties to egregious unsportsmanlike conduct. The goal is to leave around 40 teams on the remaining list. This year we DNP'd 39 teams, leaving 38 remaining on the board.
2. Autosort. All of our scouting data feeds into a spreadsheet with a metric called "pickability". This is our best calculation of the number of points we expect a team to contribute in a match, based on scouting data. We sort the non-DNP'd teams by that list as our first pass.
3. Pairwise Comparison. Once we have our auto-sorted list, we compare each team to the one directly below them and decide if they should stay where they are, or swap. Each time two teams swap places in the list, we compare the newly "promoted" team to the team above them before returning to where we were. It can be slow, but it results in a well-considered list.
4. Review. Once the pairwise comparisons are made, we slowly read down the list team-by-team, making sure that our decisions make sense.
5. (Watchlist). This step is not relevant for Champs, but at a Regional we would make a list of teams whose performance we want to take a closer look at, or pairs of teams who we struggled to make a decision about.

From this list, we managed to pick up 461 and 59 onto our alliance, both of which were strong reliable scorers who had been putting up great performances all weekend.

One of our discussion topics in the first pick meeting was that by picking anyone other than 1114, we were going to leave a strong alliance 2 in the mix that we would have to face eventually. Sure enough 1114 picked 3005, 836, and 2337 for a truly terrifying powerhouse of an alliance. In Match 11 we faced off against them and they won the match by a single penalty, 205 to 209. This relegated us to the lower bracket for the second time this year.

After fighting our way up through the lower bracket, we again found ourselves across the glass from alliance 2. The first match went our way, but in Finals 2 we took a hard hit against 1114 immediately out of auto, snapping our intake. We were able to shelf load cones from the right side of the intake but the scoring disadvantage was too big, and we took a loss to alliance 2 again. The pit crew had a frantic few minutes to fix the intake, but it came apart again in Finals 3, forcing us to play defense for the back half of the match. That defense worked, and we took home a miraculous division win after 3005 suffered a drivetrain failure during a pushing match. Winning with a broken robot against a broken opponent isn't our preferred way to win, but we'll take it.

Once on Einstein, things didn't go our way. After repairing our intake it was Code Orange's turn to break in our match against the eventual champions 1323, 4414, and 4096. We managed to win our first match in the lower bracket with Code O out of commission for repairs and both 59 and 461 on the field which was a good validation of our picklist. Another mechanical failure in our last match of the season gave us a final loss against the eventual finalists, 125, 5460, and 870.

All-in-all, this has been a rollercoaster of a season which culminated in an unprecedented ninth consecutive appearance on Einstein field. We made new friends, lost some matches, won some banners, and kept building up our program and experience. We're all ready to decompress before doing it all agian.

259 days until kickoff.

Despite the good win at SVR, the work never stops. Brendan's to-do list posted to slack after SVR looks like this:

The forks are doing their jobs well, but the Emma and Yi are finding it unnecessarily difficult to guide the forks under partners because of how long they are. We're going to shorten the forks to help maneuverability. It also opens the option for us to start auto with the arm up at its scoring angle to get the first points on the board faster. 

On the Tuesday after SVR, HQ announced an unprecedented change to the rules. Once the grid has been filled, nodes can be double-scored for an additional 3 points each. To my knowledge, a rules change this big has never been made before. It's also being made with just about no time to adapt strategies or practice.

From our limited testing time, stacking cones onto mid/high nodes seems very reliable. Double-scoring cubes in the hybrid nodes seems fastest, but it's very easy to bump a cube out when putting another one in.

​After a day of testing and talking, we think this change is beneficial to us. 

Division assignment  is always one of the most exciting parts of the year. On the Thursday before Champs, we got assigned to Galileo. There is a ton of depth in the division, which means we should be able to hit elims with a strong alliance from any seed. The two teams standing out at the top are 2910 and 3005, but there are a ton of reliably great teams who could come out swinging hard this weekend like 67, 1114, ​1690, 3467, 3476, 4481, and 6800.

Looking across at some other divisions, 254 and 2056 are in a Archimedes together along with some incredibly strong teams, and 1323 and 4414 are together in Hopper.

​The match schedule came out on the Wednesday before Champs, and ours is pretty nice. Statbotics gives our schedule a rating of 0.18, meaning that out of 1000 randomly generated schedules, only 18% of them were easier than the one we got. After our brutal match schedule last year (0.97 on Statbotics), we'll happily take this one. Our hardest match appears to be qm15 where we face 2910 and 4451, but we're backed up by 67 in that match making it a bit of a tossup.

During quals both us and 254 lost a single match each, where 4255 went undefeated from some impressive gameplay (not an easy task!). The rankings at the top of the board were in flux up until the very last few matches, where 254 managed to secure a spot in first. There was a lot of excited and nervous murmuring leading up to that point, as 4255 ranking first would likely end up with us facing 254 across the glass at some point in elims (a scary proposition for anyone).

Once 254 ranked first, they extended us an invite along with Team 100 (The Wildhats). There was a fun moment in Match 1 where Team 4010 on the blue alliance accidentally bumped a cube across the line and we happened to bump it into a hybrid node. If we hadn't then scored our last auto piece on top of it, we would have had 8 auto pieces, theoretically letting us score 194 of the 193 maximum possible points. It didn't work out like that, but we can dream.

From there we were able to make a clean run through eliminations, subbing in 2489 (Insomniacs) for the first match of finals when 100 ran into some robot troubles. Wildhats were able to hit the field for finals 2 though, where we set a score of 190, just one auto mobility short of 193.

SVR showed us what we already knew about this game: hitting max points will not just be possible at Champs, it will be necessary for success. We need to be able to do it reliably, cleanly, and quickly. This year will be won with reliability, efficient driving, and strong scouting.

I know I haven't said much about software on this blog, but I assure you they were working this whole time!

At this point in the season, the Software Robot team was working on tweaks and optimizations to make autos more reliable and subroutines faster.

"Quick Scrape"
The fork climb has three states: "pre-fork" where the forks hover a few inches above the charge station, "scrape" to press the forks down onto the charge station and slide under a partner, and "curl" to complete the climb. The move from pre-fork to scrape was a little sluggish at Sac, so it got a little more aggressive for SVR.

We also made a clean 3 piece auto for SVR.

Finally, Zoe got us an auto that we think is going to be very valuable at champs. We expect that robots with 2 piece autos on the bump side will be fairly common at champs, but 3 piece autos over the bump will be rare. This 2 piece over the bridge will make it easier to get max auto points with more partners.

Rohan and the Hardware Electrical subteam made the following changes before SVR. All-in-all a pretty short list.
​
Major Things:

• One of the addressable LEDs on a fork sheared off while practicing forking, the issue was found to be that when we replaced a fork, we placed the leds around 15 degrees too close inwards, which caused the issue. Fork has been replaced, and made sure to be mounted properly.
• Intake motor lost can during driver practice. Issue was found to be the CAN entering the PDH, one of the wires was loose and only held on through hot glue. I put it back in properly, and fully pull tested all of them before re-hotgluing it.

Minor Things:

• All electrical-related issues found during sac were given a more permanant resolution than the temporary fixes that were given there. This includes:
• LED splitter made functional
• Extra sheathing added to points which were in moderate danger of being cut
• RSL moved a few inches to prevent hitting the arm

2910 has the highly critical long-intake maneuver as part of their 3-piece clean side auto, so we need to catch up. Here's a high speed video of the Tumbler doing the same.

It's about 700ms from extension to acquisition, probably improvable down to 500ms. We can definitely drive faster than that though. Probably not an SVR play, but you never know.

Sacramento went well! The robot had no major mechanical failures, although we found a lot of smaller issues in preventative maintenance in the pit. We went the whole event undefeated, and in eliminations followed a tried-and-true strategy: get picked by 254.

Full disclosure, we were very thorough about tracking issues at Sac, but less so at Hueneme and SVR. This list is what we aspire to do every time.

We always appreciate a team who can do a lot with a little, and simplicity is always one of our goals as a design team. With this in mind, we loved getting to play opposite 8793, the Pumpkin Bots (and wish we could have played with them, at least just once!) 8793 built a simple cube running robot on the KOP chassis. It was comically fast and skillfully driven. With just a roller claw on a pivot, they were outscoring many much more complex machines, and really stood out to us from the stands.

Here's a Pumpkin Bots match for your enjoyment.

Our to-do list after Hueneme is as follows:
Hardware

1. New swerve modules to swap & reinspect old ones (maybe big bearings dying) (Emma, Henry, Sabrina, Soren) 
   1. brand new modules to build
   2. taking guts from comp to put in new modules
   3. C- becomes comp
   4. remember to wiggle test!
2.  Reinspect all falcons
3.  Use JJ tech on gears, buy shim stock from McMaster to shim gears & reduce backlash, results inconclusive as to feasibility
   1. put shim tape on gear not on the shaft 
4.  Strap down battery (Michael)
5.  Maybe partial belly pan in the back to replace current battery holder (Michael)
6.  Get shavings and whatever out of the robot to stop the rain stick, 
7.  Add led strips to inside edges of forks, clear heat shrink (Jack)
8. Better radio mount (Jackson)
   1. move spot

Arm

1.  Replace the bearings on the arm pivot gearbox with bushings (Brendan)
2. Look at making arm gearbox more serviceable (Brendan)
3.  Get max planetary 3:1 stages to put on arm (Mike) 
4. Inspect chain tension post 3:1’s being mounted 
   1. Maybe add channels for arm to sit in when all the way down (Vamsi) 

Elevator

1.  Add reinforcement strut part way up elevator for string routing / anti-twist, position is in Sketch 247, need guidance bearings/bushings as high as possible, mounted to the plate, probably a 3d printed part or smth, needs to be basically tangent to the bolt for the pulley, bottom strut is ok low 

Intake

1. Remove exposed bolt heads so we on don't have to tape over everything
2.  Replace CF on intake with 2x1 or 1x1 polycarbonate square tubing to help with twist (CF might be bad, replace with poly)
3. Less holes close to edges

Wrist

1. Go to 50:1 gearbox on wrist (needs to be uber-light, maybe buy lightweight gear from WCP) (Taye)
2. Need gearbox to have hole for echain mounting 
3. Fix wrist wobble somehow (add reinforcing tube thx Taye)

The gear ratios on both the arm and wrist are higher, meaning they're much torquier. No more failed fork climbs.

The new intake looks both stronger and more rigid, with aluminum square crossbeams instead of round CF (which snapped at Port Hueneme).

We also got lights on the forks, which looks amazing.

We have some new autos, including this 2.5 + balance.

Here's what drive practice looks like heading into Sac. 15 game pieces means we can conceivably solo the link RP. We apparently had a 17 piece match, but nobody thought to take a video.

The first thing we saw at the event was 973 rocking the Epsilon design that we wanted to pivot to, and it's actuations were some of the fastest I've ever seen in FRC. That put us on edge from hour 1 of the event.

However, as we got into our routine, things started looking up. Our schedule was about as fair as it gets, and we managed an undefeated run through quals landing in a safe 1st rank spot. Heading out of quals, our only real complaint was that the intake was very floppy. It was wiggling around for the entirety of every match, but it held together with help from the best pit crew in the world.

We managed an undefeated run through quals and a 1st seed position, but we were not the best robot of the event. 4414 was consistently outperforming us. During quals though, they tipped other robots a few times while making high speed runs across the field, which drew multiple red cards and dropped themselves to 9th place.

When alliance selection came, we invited 4414 onto our alliance and made a strategy note that they should avoid tipping opponents during our run through elims. During Match 1 of elims however, 4414 and 696 gave an accidental 1-2 punch to 3647, drawing another red card and dropping us down to the gauntlet of the lower bracket. Our run up through the lower bracket went smoothly, with us and High Tide managing some smooth cycles, and 696 managing their match flow very cleanly.

Finals 2 went even worse. The fork climb failed us again, causing us to invalidate the engages of both 4414 and 696 and leaving us out of the community zone without even earning park points. A 16 point loss put us on the back foot for a third time in this eliminations round.

In finals 3 we finally found our footing again, pulling out a win, and leaving everything down to a Finals 4 tiebreaker.

Finals 4 was, to say the least, stressful. The bridge got wedged up partway through the match, and we played most of the way through before the match was foghorned and we had to play the same match again. A fifth​ finals match. On the final final finals match, the auto failed entirely but the drivers focused up and pulled out a narrow 7 point win, with the fork climb barely making it in time. It was the most excited I have ever seen 1678 for a regional banner.

This event validated our design, and put to sleep our worries about the Epsilon design. It also confirmed for us that traffic in the scoring and pickup areas along with a capped max score makes this a hard game to be exceptional at. The underpowered climb needs to be fixed, too. We're going to have to bring the reduction back up and deal with the risk that the robot can damage itself.

All in all, this was event to remember. 4414 and 696 were great partners. 4481, 6036, and 3859 were scary opponents. We came home with a banner and a to-do list.

​17 days until Sacramento.

With the wrist unable to lift a game piece, the wrist pulley blocking the middle of the intake from being used, and the pulleys on the wrist power transfer falling off frequently, things are looking grim. Citrus has never been this far from a competitive robot heading into an event.

In order to open up the middle of the intake, we're biting the bullet and moving the intake motor into the wrist plate instead of back on the carriage with the wrist motor. It's going to make our CG higher and the intake heavier, but at this point we're out of options.

The rope is having some issues staying on track and spooling nicely, and the elevator isn't all that rigid side-to-side. We need to come up with some fixes.

To fix the play in the elevator, we added some reinforcement plates to the front and bottom of the base and first stages of the elevator.
​
Author's note: The screenshots below are from post-season. I think the bottom plates might have gone on after Hueneme?​

To fix the rope rubbing, we moved from a bolt into the drum to a high strength rivet instead, which sits flush and doesn't snag on anything.

Comp's wiring is... okay, but Beta's wiring is truly bad. We made the decision to fully re-wire the machine. While the elevator and intake changes were being made, Rohan and the hardware electrical team measured and pre-cut every wire on the robot. As soon as the new intake was on, they stripped and replaced every single wire.

Every. Single. Wire.

The new wiring is the nicest that has ever been on a 1678 robot. Turns out that electrical is an iterative process as well, who would have thought?

The night of 2/28 was the only time in the 2023 season that the Citrus mentors had to get their hands dirty building. With the robot still incomplete by the end of the meeting, we couldn't justifiably keep the kids in the shop all night on a Tuesday. The coaches stepped in to make up some time. We had a small punchlist of items, including re-riveting the bellypan on with more high strength rivets, finishing installation of the elevator end-plates, tapping some holes in Falcon shafts, and tying down some last few wires. People often wonder how much hands-on work the mentors are doing on "good teams," and this was it for us this season.

With Beta's beautiful new wiring, we decided that Comp would sit out for the first event, and Beta would play at CAPH instead. Is a comp that doesn't play at competition still "comp"? Who knows. Regardless, Beta is finally ripping some fast cycles. Here is a 12 game piece practice match (counting auto).

4414 posted a dope reveal video and we're stealing their format. This took like 30 tries, but we eventually got it clean enough to post. Beta was packed up immediately after filming, and we headed down south to Ventura. There are some serious powerhouses at this event, and we've only been competition ready for two days. Confidence is low, tension is high, but all we can do from here is our best.

On Saturday, Comp and the drive team went into San Jose to practice driving on 254's full field while the hardware teams finished the R2B2 rebuild.

The practice session went well, and below are Mike's notes as they were posted in Slack:

• When the battery gets really low, the robot struggles to keep the arm up for L3 scoring. We might need to further adjust the gear ratio, but we can play around this for Port.
• We had one falcon power wire come out of the PDH. There is no strain relief so it just popped out. We need to fix this.
• We had multiple bad solder and strip joints in the CAN loop. The CAN issues are brutal and hurt big time in practice. We must get better and inspect all cables
• The e-chain mount on the arm broke. This needs an AL backplate
• One of the steering falcon shafts is coming loose. We should inspect and consider adding a output shaft support bearing like poofs do. Poofs also said they swap all their falcon nose bearings out.
• when switching swerve tread, one of the wheels was almost impossible to remove because of too much loctite.
• Bumpers are still too low.
• The rivets in the bellypan sheared. Let’s use high strengths.
• Intake power belts are still falling off/slipping and results in low/no power to the intake rollers. This is our most common issue.
• Grabbing from the shelf was working better than grabbing from the ramp. Both are challenging for our robot
• Cubes on the ground are proving tough for the intake to hold onto. We hope the newest intake is better in this regard.
• Intake skew is still a problem. We hope the new intake fixes this too.
• Software sequences are slow and could be sped up in places.
• The wrist chain was loose but never really skipped, which was good.
• Elevator DOF was solid
• Arm DOF was fine except for a few times the flathead holding the elevator drum backed out causing chain rub. There is potential gear tilt again that we should inspect.
• The 1/8” inner intake plate snapped. We need a spare cut.
• The swerve twitches at the start of auto, sometimes causing a cone miss. We need to look into ways to mitigate the twitching.

R2B2 was mechanically and electrically complete by about 5pm today, finally ready for drive testing.

Immediately after this, we found that the wrist doesn't have enough torque to lift the intake with a game piece in it.

 Beta is doing better than comp though, since Comp's wrist had fully snapped in half.

5 days until competition.

Something you sometimes hear mentors in FRC say is that robotics is a third space. Your first space is your home, your second space is your school or workplace, and your third space is where you have fun or pursue your hobbies. Your priorities in life should follow the same order. Your first priority in life should be your health and wellbeing, followed by your responsibilities as a student, employee, or parent, and finally in third is your hobbies.

With that in mind, this is the point in the season where work started to overtake FRC in my priorities list, and I stopped having time to be in the shop daily, and didn't have time to dedicate to this blog.

Following this, each post will be a summary of each event and how we spent time preparing for the next one. They are being written well after the conclusion of the season, using the pictures and information available to me in Slack. Hopefully next season we won't have an interruption like this in the build blog, but I hope you find the information here interesting and valuable nonetheless. Thanks for reading.

Regards,
​Matt