Difference between revisions of "Talk:DEWBOT VII Philadelphia Regional"

Revision as of 03:42, 17 April 2011

Preparation - To Do List for Thursday

Add part locations (probably the spare parts box), hardware and tools, etc.
Can we estimate how long each of these changes will take, or have a time to shoot for? --Julie

What did we say, Clem? Pulley-Chain 30min; Deployment 20min; Compressor, Brace, Hangar 10min/ea; Camera & Pivots 5-15min/ea? Thursday-time, that is. - Siri

Is the BOM up to date for all these (minus compressor), or would you like a hand? - Siri

Prioritization

Best to practice with ASAP (in order)

Pulley to Chain Drive
Mount sonar (test autonomous)
Mount camera (test on FMS)
Turn around deployment
Competition Claw & Torsion Brace (nice but not necessary)

Must be done before inspection

Switch Compressors
Mount Hangar
Pivot check (ensure no inspection-worthy changes)

Technically ok post-inspection

Full pivot check
Test all sensors
Tune line-followers
Tune arm PID (but would also be nice for early practice)
Test Fangers (inspected separately)

Switch from Pulley to Chain Drive

Remove arm drive motor assembly from arm column, cutting the belt
Unbolt arm angle sensor mount; remove arm sensor
Remove elastic counterbalance
Unbolt lower bar pillow blocks
Unbolt lower bar at triangle
Remove lower bar
Unbolt (2) ¼"-20 bolts securing 56T HTD5 pulley; remove pulley from lower bar
Replace pulley with 42T sprocket/hub assembly; align key with keyway perpendicular to the ¼" mounting holes
Insert lower bar axle back into pillow blocks
Rebolt pillow blocks to columns
Rebolt lower bar end to triangle
Restring elastic counterbalance (5 passes); secure elastic using wire-ties out of the sprocket path
Remove arm drive motor from mounting plate & bearing block
Remove Remove 20T HTD5 pulley from arm drive gearbox shaft
Install ½" ID x 0.32" PVC spacer on arm drive gearbox shaft
Install 9T #35 sprocket (McMaster-Carr part # 6280K311 - bored thru to ½" with 1/8" keyway) with sprocket end towards the motor (hub end towards the shaft end); use a 3/32" key
Install (1 or) 2 ½" Nylon Washers on the arm drive gearbox shaft outboard of the 9T sprocket
Place the bearing block on the shaft end
Drape the prepared 59-link #35 chain loosely on the 9T sprocket
Bolt the arm drive motor/gearbox/bearing block assembly to the arm motor mount plate using (6) 10-32 x 0.50" FHCS
Loosely secure the Arm Motor Mount to the column using (7) 10-32 x 2.00" FHCS & nylocks
Secure the 59-link chain around the 42T sprocket with a master link; master link clip facing away from the lower bar
Pull down on Arm Motor assembly to tighten chain (rotating arm to take out slack on both sides)
Remount arm sensor mount
Support the lower bar parallel to grade (use a spirit gauge)
Reinstall & Calibrate encoder
Check/set preset heights (floor, low, mid, high--HOW?)
(Tape measure + Ben or mark a couple PVC banner stand pipes. I'll document the heights Saturday. Don't know how/how crowded the practice field will be. - Siri)
As long as the encoder is calibrated where level is 90 degrees, there shouldn't be much change to the arm angles for the preset heights. --Julie
(So run a practice match and then see? Works for me. I'd still like those heights for reference though. Do we have them or should I measure? - Siri)
We've never had heights, just arm angles. --Julie
Tune PID
Attach rear chain guard
C-clamp side chain guard, drill new hole & bolt on

Put on Competition Claw

Time: <5min

Unscrew jam nuts at linkages & unplug pneumatics
Screw in Competition Claw's jaw nuts at linkages & plug in pneumatics
Test floor height and wrist & jaw actuation.
Any time: put on longer start position Velcro. (But always take off the bungee cord before the match!)

Switch Compressors

On standby pending Q&A
Suggest we pre-mount the compressor on new cross-bars at DRC. In mounting this to the robot, use one set of existing holes; use the holes in the mount bar as a template for drilling the two new holes needed. - Clem
That was my plan (hence "new bars with compressor"), but I don't know how rigid the orientation is. I remember with the Thomas's old standoffs, squaring the mounting was kind of annoying. Looking at the diagrams we should be fine, but if it's in Saturday I'll try it. - Siri

Unscrew compressor mount bars
Unplug pneumatic hose, disconnect wires and remove compressor
Bolt in new front/rear mount bar (with compressor attached) and clamp rear/front bar in place. (I'll pick which when I finish CADing the new assembly (pending GDC cutting us a break) - Siri)
Drill (2) new mount holes for bar that's shifting. (Do we need to tap those holes? - Siri) (Are existing holes tapped? - Clem) (I'll check Saturday - Siri)
Screw in new bars with compressor & plug in pneumatic hose & connect wires (Do the wires disconnect compressor-side on both and/or do you know how to do this? - Siri)
Test

Turn around Minibot Deployment

Unscrew Bident & staff and pull out
Unscrew back plate (with cylinder). Ensure pneumatic hoses are labeled.
Unscrew Standoffs (keep rollers on them)
Put jig in place & drill new Standoff & back plate holes
Screw Standoffs & back plate in place. Ensure back plate is within frame perimeter. Ensure rear-most rollers are off-centered towards the robot front (as far from the Bident as possible, or it may go outside of the frame perimeter).
Screw the new Bident & staff into the cylinder. Ensure Bident is in frame perimeter.
Wire tie the pneumatic tubes so they can't be pinned by the retracting cylinder.

Mount Hangar

Put on top edge clamp (4 bolts)
C-clamp bottom angle in place. Ensure Hangar is in frame perimeter.
Drill bottom angle mount holes in frame perimeter & bolt (3 bolts)

Install Torsion Brace

Install top brace clamp (8 bolts)
C-clamp bottom plates in place & drill H holes in frame (4 total)
Bolt bottom plates in place (4 bolts)

Mount Camera

Drill mount holes (2 holes, on reverse deployment jig)
Mount camera & plug in (wire tie wires)
Mount iron sight (needs jig)
Test/calibrate
Remove old camera mount

Mount sonar sensor

Remove and unwire one front Sharp IR sensor
Mount and wire one sonar sensor
Test
Update BOM accordingly

Check pivots

Check nylock nuts for nylon engagement. If they do not engage the nylon, remove and loctite.
Fix broken/disengaged shaft collar. Check other collars for engagement on the shaft and tightness. Add plastic spacers (same as Duex) and wire tie.
Check pulleys. Raise if necessary, add spacers and wire tie.
Check chain & belt tightness and wheel turning. Loosen CIM if drive is too stiff.
Check sensor stand offs. Are they bent or showing signs of damage?
Check treads. Make decision when to change. Check for bolt stripping throughout and determine how long change would take.
Modify chain guard to fit with hangar.
Test (Optional: color/tape 1 triangle on black pulleys for visual confirmation)

Test all sensors

Time: 5 minutes

Run Testboard VI to check that all sensors can be read (this includes digital encoders).

Good call. Time? (All your guys know how to do this, right?) - Siri We should get the VI on your computer as well. --Julie

Adjust sensitivity of line-following sensors

On real field if possible, ensure that photoelectric sensors indicate no signal (green light) on carpet and signal (yellow light) on grey tape.

Test Fangers

After set of fangers pass inspection, test each one manually on the field poles. Make any fine tuning adjustments. Rank fangers for performance.

Post-Mortem

Okay. We won. Why a post-mortem? To learn and to do things better next time.

What worked?

Perserverence
Flexibility
Fanger and deployment
Robot robustness & reliability
Pivot drive (when mojo available)
The pit
Arm changes since Finger Lakes

What could have been better?

Autonomous - Needs calibration - eliminate oversampling
There is no switch to turn Autonomous on and off
Noise reduction
Lower hanger scoring than at Finger Lakes
Need better driving & operation
Drive team critique must go through the coach
Stronger grip on claw
It would be very beneficial to be able to lift the claw (for defensive play)
Better claw needed

Action Plan

Mechanical

Rebuild deux (Clem McKown lead)
Build Claw II (Ben Kellom lead) - better grip - less floppy - tuck in?? - sensor?
Design/build a faster Minibot (Mighty Mouse) (Gary Deaver lead) - to use same launch pad - lighter - faster - legal
Toe-in test for autonomous
Speed up deployment (Clem McKown lead) - split deployment cylinder extension and retraction to two solenoids to allow depressurization of the retraction side prior to actual deployment

Programming

Reliable autonomous - all modes
Off position using current switches
Eliminate dead time during switch between crab & snake modes
Is there a benefit to keeping the compressor off during autonomous?

Drive Team

Directed, coached (Clem McKown or Siri Maley) Driver Practice - as much as possible!
Defensive opposition-force robot with coached (Mike Rizzo), trained driver

Execution on Action Plan

Accelerated Minibot Deployment

A fast Minibot is of limited value of the deployment is slow. In the original Bident pneumatic system, extension and retraction were accomplished via a single, single-acting solenoid. Pretty standard stuff. But slow, requiring 1.58s (σ = 0.08s) to fully deploy (all measurements taken on deux with 5 samples)

An alternative pneumatic arrangement was tested using two solenoids decoupling extension & retraction. Prior to actual deplayment, the retraction side of the deployment cylinder would be vented (solenoid activated). Then the extension side pressurized for deployment (solenoid acticated). Finally, deactivating both solenoids retracts the deployment cylinder. Deployment time dropped to 0.84s (σ = 0.1s). A good start, but the need to vent retraction-side air (even though no longer under pressure) through the solenoid with its small Cv slows extension and makes the movement uneven.

A third arrangement was tested. Returning to a single solenoid, the retraction side of the cylinder is open to vent. This output on the solenoid is plugged. Only the extension side is powered. Faster deployment at 0.50s (σ = 0.1s). A normally-closed (NC) manual valve on the solenoid extension-side vent prevents Bident extension prior to deployment (due to the vacuum which any extension would create). The Bident must be retracted manually by opening the NC valve and pushing in the Bident.

Original pneumatic system, used at Rochester & Philadelphia
Interum design utilizing decoupled extension & retraction solenoids
Final schematic, requiring manual retraction

Claw II

Mighty Mouse

On April 16 mighty mouse was constructed and at 10:35 PM was launched for the first time. It's fast. Direct drive 118 style with a weight of 2Lbs 1 OZ and 3/8" tubing on the fangs. More testing will be needed on Sunday April 17. The down side to direct drive is stall and the motors smoke. Durability of the cantilevered drive will be found.

Drive control software

Autonomous Software

@@ Line 172: / Line 172: @@
 ===Claw II===
 ===Mighty Mouse===
+On April 16 mighty mouse was constructed and at 10:35 PM was launched for the first time. It's fast. Direct drive 118 style with a weight of 2Lbs 1 OZ and 3/8" tubing on the fangs. More testing will be needed on Sunday April 17. The down side to direct drive is stall and the motors smoke. Durability of the cantilevered drive will be found.
 ===Drive control software===
 ===Autonomous Software===