 |
| Apparently the robot belongs to Cormac. Wait, is that elevator supposed to be in the bag? |
11:59pm Eastern Time on Tuesday February 19th, our robot made its way into a giant plastic bag. Our 6 weeks to design and manufacture a competition robot for the 2019 FRC Game Destination: Deep Space, has come to a close. Our final days shared in difficult times and decisions, long nights/mornings, and a renewed excitement as we finally saw our complete robot shakedown and full-systems checks. We already have the next versions of many mechanisms and gearboxes imagined, CAD'ed and even cut. Our main goal for our first competition is reliability. Second will be adding speed, lightness, and functionality. We have a long season in front of us, with 3 district meets, then potential qualification for district championships and the Houston World Championship.
Making A Plan
Our robot was nowhere near complete this past Thursday. We had a three day weekend to mechanically finish the robot, software shakedown, and finish a million tiny detail tasks for electrical, pneumatics, mechanical, inspection readiness, and competition readiness. To make meaningful forward progress, we had to create a plan. Execution of a plan relies on two things: competent understanding of the plan, and the competency of the plan overall. A number of cliche and common phrases get uttered around this time:* Hope for the best, plan for the worst
* Take your time estimate and double it
* The last 10% of the work takes 90% of the time
And so on. I said a number of these over the past few days, and there's a very good reason for this: They are true. Here is an excerpt of the actual Slack post I wrote when we were discussing how to create our weekend plan: (if I were famous this might actually hold more weight...)
---
When building a plan, do not work from best case scenario and build down. start with worst case scenario and build up. If the lab lost power for the next 100 hours. what can we achieve. If we have mills but no router. what can we achieve, etc. My examples are obviously very unlikely, so maybe think in terms of: if we have no new parts by close of business today, what can we achieve? The second, more difficult part of the plan, is we cannot yet see what issues exist. each subsystem needs a <reasonable> shakedown period. If there's a critical flaw in the shooter, we need a time box for addressing it, and a fallback plan. If the issue is minor, then cool, that's an extra 30-60 minutes free.
Edit - adding more current context to consider:
we finished the elevator! ... there is a critical flaw. Caleb has put somewhere *3-4 hours* into a fix, and we still have not cut or replaced parts. That may take another 2-3 hours, so his shakedown window and critical fix time box is *6-7 hours*. That's just to correct a mechanical issue. We also need to make the brackets to host the wire guides. The string pot needs to be mounted. Limit switches need to be mounted. The bump stops for <if> we do our climb need to be mounted.
I'm am saying this purely as: this is the level of time and detail that EACH system needs. ... The realistic side of what I am thinking: If it is not mechanically complete on the robot by noon Saturday, it should be removed from our efforts.
---
 |
| One million little details |
Following this, most of our subsystem teams (and electrical, they're a real team too!) grabbed a whiteboard and starting writing their tasks and time estimates to be 'software hand-off' complete. This time of build (while unfortunately later in the season than we like) turns into the day of a thousand 5-minute tasks. Keeping detailed notes and staying organized is the only way to ensure that every system has been brought up correctly. All wires and pneumatic tubes routed and installed, shields cut and fit to protect certain parts of the bot, all the sensors and brackets and mounts and spacers and etc have been mounted and operate correctly. Finally around 8:30pm Sunday, we finished the required set of tasks to hand-off to software. (Note: we didn't finish absolutely everything, there's always something else...)
Our software team then executed on their plan. What order did they need to bring up systems? What tests did we need to run? What pre-requisites did we need to complete before running system checks? They too had a set plan for bringing all systems online. With basic systems checks complete, we started initial drive practice and cycle operations. Partially to get the drivers time on the new bot, secondly to keep testing mechanisms to do what they were built to do. And good news! We broke the robot. We uncovered some operational issues. We identified each issue, and now have a plan to move forward.
Wants And Needs
This post is getting long already so I'll try to describe the situation concisely. We wanted to score cargo and hatches at all levels and we wanted to climb. We needed to stay within the rules, and that means 125 pounds max. So, we had to toss one of our wants, and unfortunately that meant our HAB 3 climber got benched. We still have it, and we still have plans of lightening, but there's work to be done before we think about returning to the climber. |
| Our climber is still here, still together and waiting... |
This is a hugely important lesson to take away, and learning it early, and seeing it often is a good thing. Companies have succeeded or folded because they have understood when to cut losses, meet a more minimal set of priorities and make them great, as opposed to stretching thin and under-delivering on expectations. We made this decision very late in the season, and it will affect our "great" performance for at least our first or second district meets. The best part of all of this, is that this is still a high school program. (Albeit an expensive and time-consuming program, but let's not focus on that right now...) High school is a place for learning and application, and experience. This year, we had a wants list that we could not meet. So we've cut our losses and are sticking to our needs. And every student on the team will <hopefully> remember this lesson for next time.
Our Final Bot
Ha! Just kidding, as stated above, our bot will be different with every unbag. Our final bot won't exist until we finish our last match of the season. Our current bot however, meets most of the requirements specified at the start of the year. Our primary goal was to cycle game elements as fast as possible. (Humanly possible? Robotically possible?) At this point, I don't even know what our software team is fully capable of producing and how they will continue to amaze us. We have a HAB level 3 climber designed and 90% assembled, however we met our weight limit before fitting the climber. In a moment of difficulty, we had to cut our climb assembly and our losses, and focus on the game element scoring mechanisms. Our first version of the competition robot this year, code-named Andromeda (I think? Or 'Toss-Up'? Or 'TBD'? There's been some discussion on this still...), is a cargo ship and rocket bot. |
| The current final robot. |
Robot Basics:
Weight: 124.6 lbs
Chassis: aluminum monocoque
Drivetrain: 6 mini-CIM, 8 wheel west coast with shifters, traction wheels middle with omni ends
Ground Intake: piston deploy, 775 Pro positioning with encoder, bag motor for roller bar
Able to pick up all inflation sizes of cargo from the floor.
Elevator: 2 775 Pro driven, min and max position limit switches, encoder and string potentiometer positioning
Three stages plus carriage extends from ground intake height to all rocket scoring heights.
Shooter: 775 Pro driven
Holds ball firmly, hopefully won't drop the ball, or it will run a mile.
Hatch: Pneumatic deploy for collect, 3 pneumatic pistons for release
Pickup from loading station only, piston deployment forces the hatch against the rocket
We have completed a full shakedown (and broke the robot only a few times....) and practiced several cycles at home base. Looks like we are... mostly.... all systems go. There are always details to be resolved. For me, I have always loved seeing the cleanliness that shows in a 4911 robot. The electrical panel is well thought out, wire runs are clean, and the whole robot, to my eye, just looks good. We are looking forward to seeing all of you at Sundome, Glacier Peak, and Auburn.
Quote of the Day:
"This is not the robot we want. But this is a good robot. I firmly believe this robot can get us back to Einstein." - Head Coach
 |
| We're still making new parts, lighter, stronger, faster. |
No comments:
Post a Comment