Difference between revisions of "Talk:DEWBOT VII Finger Lakes Regional"

Revision as of 21:03, 8 March 2011

What Worked

• Teleop scoring was okay. We could score 3 logo pieces fairly reliably. 4 max. Practice should improve this further. We believe a roller claw would also speed up tube acquisition (scoring speed is solid).
• Pit work - fortunately there wasn't much to do, but the fixes that did happen were relatively smooth for a first-week regional.
• Drive team - it takes some doing, but we can definitely be on our game. More practice will also help this more.

What Didn't

• Minibot - general bumping around and vibrations during match caused the spreader to come out and motors turn on (and burn out).
• Minibot - easily damaged
• Pivot drive repeatedly jammed (drive jam, not steering). Problem appears to originate with the wheel axle collar, some of which cannot be tightened securely. When the collar moves, the wheel shifts on the axle. The wheel shift causes the chain to jam on the 9T sprocket, which is not chamfered. Sometimes the jam is momentary. Sometimes not. One such jam led us to change a pivot on Thursday, another unfixed one effectively disabled us in the semifinals. A few more likely froze us at different times in qualifications, once getting us stuck in an opponent's lane causing a penalty. Since the jam stalls a CIM motor, these jams run down our batteries extremely quickly.
• Drivetrain lag - likely due to early camera interference with the FMS, the drivetrain lagged severely until the data load was decreased.
• Belt-driven arm - The belt shredded during the quarterfinals. We replaced the belt and it shredded again in the first semifinal match. We need a better mechanism. However, the switching procedure is pretty efficient (nice, we should keep up practices like this).
• IR sensors were apparently not working throughout the qualifying matches. This appears to have been a cRIO problem. They seem to work now.
• IR sensor - one was shorted to the frame, complicating inspection. These need to be insulated before mounting on metal.
• Autonomous line following - while the sensors themselves appeared to always correctly register the line, the autonomous code did not, usually causing us to veer off the line near the wall and hang the ubertube on top of the next scoring grid post (not on the horizontal pipe, but the vertical). This appears to have been a correction speed problem, likely dependent on the different carpet. It seemed to be fixed during Eliminations, but IR sensor problems and an off-course alliance member left this inconclusive. We did score once in autonomous (match 2) and lined up correctly (lateral and height) once more (match 1). All other times we failed.
• Outboard autonomous code - screws extended out along the low wall prevented us from running our outside-peg autonomous that requires sliding along the wall. We were unable to square against the wall at a (non-zero) distance with the IR sensors because of issue above.
• Video feed from the camera - we were originally running this through wireless via Internet Explorer, but had to disable the network connection and IE in order to work with the FMS. There are ways to make the feed work, though we were unable to implement them in time for competition.
• Pit work (yes, it's on both lists) - especially if the pivot axle collar is responsible for all our drive problems, we're missing some pretty big things on the checklist. We've also got pit processes to practice and quick-fixes to understand, though we're doing well so far.

What Broke

• Minibot
• Pivots - see jams above
• Arm drive - shredded two belts

What Next?

An Action Plan in preparation for Philadelphia is proposed, together with a new Minibot Specifications breakdown.

Priorities

Project owners in parentheses
Proposed in a 8-March email:
"The Finger Lakes Regional was a great experience, but also a great learning experience. We have a lot of work and practice to do before the Philadelphia Regional on 7-9 April. We've got four weeks.

To keep exeryone directed and to prevent overworking people, I would like to meet three evenings per week, but to split the evenings by function:

• Driving Practice - Tuesday (starting next week, not tonight)
  
• Mechanical - Wednesday (includes Minibot)
  
• Programming - Thursday
  

I would encourage some cross-over between groups, but also want to avoid idle hands. Let's organize the cross-over ahead of time, rather than just waiting to see who shows up. Please contact me is you want to be involved.

The Minibot building can take place on any of these evenings, as long as it does not require the robot. If you need to interact with the robot, please make arrangements with the drive or programming teams ahead of time. If you want, I can facilitate this.

I sent out an Action Plan yesterday, but this did not include priorities. I propose the following priorities:
A - Must do/have

1. Must fix the drive-train chain jam problem
2. Must reinstall spare parts on deux
3. Must, must, must have a reliable minibot & minibot deployment system
4. Must have reliable straight-line autonomous (Programming team, Julie)
5. Must have a different (durable) arm drive mechanism
6. Driver practice - Hang 4 logo pieces & deploy minibot
7. Test and clearly label batteries - Battery deployment plan
8. Diamond plate v. plywood test and/or install diamond plate at target (Programming team, Julie)

B - Really ought to have if we are serious

1. Reliable autonomous targeting outboard targets (or at least one of these) (Programming team, Julie)
2. A fast minibot and minibot deployment system
3. Anchor & cover sides of minibot post base
4. Resolve camera issues (Programming team, Julie)
5. An operating Pivot display
6. Ergonomic mods to human player station
7. Updated Spare parts & Tools & Pit to-so lists (necessary but not resource-intense, Siri)

C - Nice to have

1. Battery Caddy
2. Practice with a set-up similar to competition FMS (Programming team, Julie)

D - Some other time

The world's best roller claw (an excellent summer project candidate)

If we need, we will add one or two Saturday meetings before Philadelphia. Let's discuss this week.

Best regards,
Clem McKown
Head Mentor - FRC Team 1640"

Improvement Discussion

Claw

Scope

Unreliable gripping

• difficult to tell quickly if tube is in
• gripping tubes with side of claw is common and doesn't hold for long
• gripping tubes at different angles is just as common and unreliable
• even if we don't lose the tubes, wrong gripping configuration makes it difficult to score on upper pegs (less common)
• Lower Priority: can't get back & low in the frame perimeter (haven't broken it yet)

Specs (Acceptance Tests)

• no dropped inflated tubes (-5pts/ea)
• no failed pickups of inflated tubes (-3pts/ea)
• all picked up tubes can be scored on top row (-2pts/ea)
• minimum average pickup time from claw down to tube secure and above bumper perimeter. (-1pt/second)

Acceptance Test: top score after 5 realistic matches (with OPFOR)

Concepts

Practice (accepted):

• Will definitely help, but there are inherent design obstacles to acquisition

Camera (accepted):

• Will definitely help, but there are inherent design obstacles to acquisition. We need to get this working, though.

Change Maxillae Shape (needs testing):

• We often lose tubes off the edge. Needs testing. This is primarily a trade off between wider contact area and higher pressure.

Roller Claw (for post-St. Louis):

• Pros: fast, secure, low-accuracy-required tube acquisition and fast, flawless release. Possibility of rotating tubes. Many of the best robots (perennially) are roller claws. From a drive team perspective, we wouldn't have lost/failed to pickup a single one of our inflated tubes we with a good roller claw.
• Cons: Heavy, drastically increases moment arm, complicated (especially when trying to move weight closer to center of rotation), not easy to implement, not particularly cheap. From a drive team perspective, we would have lost the same number of tubes with a bad roller claw and maybe 2/3 with an average one.

Arm

Scope

• Primary: belts failed twice
• Secondary: slow movement
• Tertiary: "bounce" in tuning

Specifications (Acceptance Tests)

• doesn't fail at competition (20 matches as chief scorer with OPFOR)
• 2 minute pit switch (marginal 4 minutes)
• within 115% of status quo belt weight
• within BOM budget (prioritized below minibot deployment and above claw)
• gear down (speed up in software)
• reduce backdriving/ease PID

Concepts

#35 chain

• Pros: easy, quick fix, light weight
• Cons:

Helical gear

• Pros:
• Cons:

Worm gear

• Pros: prevents backdriving
• Cons: prevents backdriving (can't use power to remove robot from field, can't have arm up when removing)

180 Degree Encoder (<360):

• Pros:
• Cons:

Potentiometer

• Pros:
• Cons: