Team 1640 2010 Summer Program

From DEW Robotics
Revision as of 13:38, 13 July 2010 by Siri (talk | contribs) (Possessor Consumables Replacement Processes: moved)

Jump to: navigation, search

Objectives

Objectives for the 2010 Summer Program are focused on our desire to perform well at IRI. Towards this end, we are:

  1. Performing basic maintenance on the robot, especially the drive-train to compensate for the wear and tear of our competitions and demonstrations;
  2. Developing and implementing a possessor which really works;
  3. Getting the control code right & tested, including the angle-drive autonomous; and
  4. Adding a mechanism to raise the mirror.

Necessary Repairs & Maintenance & Drive-Train

Basic maintenance & upgrade work started 2-June and comprised:

  • Pull all pivots. Add spacers between driven sprocket lower 1" bearing. Check condition and refurbish as needed. All treads will need to be changed (they're beat). I will order more treads.
  • Replace Steering Jaguars with Victors and test.
On Wednesday, 2-June, we followed the above plan, without finishing.
  • Pivots removed
  • (4) 0.520" PVC Spacers cut from 1" PVC pipe (we've got miles) - thanks, Douglas & John
  • General cleaning & tread replacement. Steering sprockets pulled from pivots.
  • Drilled holes for Jaguar > Victor switch. We made a template first, so that holes were drilled in the correct spots.
Generally, pivots are in good condition excluding tread wear and carpet fiber accumulation on the transfer axle. No "broken" pivots as at PARC. All steering sprockets are 0.52" above lower bearing (same as model dimension - 0.521") except Pivot #3 at 0.33" (thereby our steering chain problem). Pivot #3 steering sprocket set screws were loose (not quite finger tight).
The Pivot Replacement Process was documented.
On Wednesday, 16-June, we continued:
  • Molly & Garrison joined us - welcome!
  • Steering Jaguars replaced with Victors. All drive Jaguars are black.
  • Pivots 1 & 4 reinstalled. 2 & 3 still in service.
  • PVC spacers installed.
Work was completed on 19-June, just in time for the Upper Uwchlan Block Party.

Possessor

We've got a adequate possessor for demonstrations. At PARC, Monty Madness and (BR)2, we were dissatisfied with our possessor's performance. While it seemed to perform well enough at the STEM Defined demonstration, in the heat of competition, it tends to either fail to possess the ball, or to cause or allow the robot to drive over the ball. We really cannot effectively herd balls into the goal either (because of a tendency to drive over them). This is an area which needs work for IRI. Options seem to be:

  1. A vacuum possessor (like Team 25's)
  2. A Possessor with a low roller
  3. Adapt the possessor to grab the ball by pinching it (without lifting it off the floor) rather than just applying a backwards spin

Vacuum Possessor

Preliminary analysis in Inventor indicates that it is feasible to accommodate a vacuum-type possessor similar to Team 25's in the DEWBOT VI chassis. The driver would have to accurately center the ball. Analysis and design were not pursued past this point.

Low roller Possessor

We've observed that robots with effective possessors often have passive rollers at or below the ball centerpoint. Installing such a low roller on DEWBOT VI without a major redesign is not trivial. Key challenges are:
  1. Avoiding interference with the kicker
  2. Avoiding interference with the pivots
  3. Maintaining our ability to cross bumps without difficulty. A low-position roller intrinsically is in the way when climbing a bump

Trirol Possessor Design

A possessor was designed having an articulated, low, 3rd roller.
  • The low roller is a ½" steel rod (like the middle roller)
  • The low roller is set in needle (roller) bearings for free rotation
  • The low roller articulates by pivoting forward and up out of the way for bump crossing.
  • Two 3/4" diameter x 4" stroke pneumatic cylinders provide the actuation. Both run from a single solenoid. Restrictor valves will need to be installed on the solenoid vents.
  • A hard stop is provided for the possessor in the lowered, working position (the design does not rely on the pneumatic cylinders to provide this stop). This hard stop is necessarily quite close to the pivot axis.
  • While rotating between the deployed and retracted states, the low possessor assembly does not break the frame perimeter.
  • It is proposed that the low possessor will automatically retract when the operator relaxes the kicker to cross a bump; and automatically deploy when the kicker is armed; normal kicking will not change the possessor state.
  • In addition, it is likely that the combination of reversing the driven possessor and retracting the lower possessor roller would be a useful, low-velocity means to positively eject balls into the goal from the goal's ramp (without having to drive the robot entirely into the goal).
  • The kicker is modified by having its lower portions cut away so that it clears over the 3rd roller.

As designed, this is a back-rolling possessor, not a pinching possessor. It could in principle be converted to a pinching possessor if the low roller bearings were removed or disabled.

View the detailed drawings.

Prototype Testing

3rd roll prototype installed on DEWBOT VI
At Jon Davis's suggestion, a fixed 3rd roll prototype was installed and tested on 30-June. Key observations:
  • Possession performance appears unaffected by the 3rd roller - poor-to-mediocre
  • On the other hand, it is now impossible to override the ball
  • and herding now works very well, possessor on or off
Overall, some small improvement, but not at all what we were hoping for.
Adding friction tape to the bottom roller resulted in no observable improvement.

Hi/Lo Roller Pinching Possessor test

We were planning to replace the middle steel roller with fiberglass, but before installing the fiberglass rod, we tested the possessor without the middle roller. This provided a phenomenal improvement in possessing performance. This Hi/Lo 2-roller possessor really works! It hangs on to the ball. The Hi/Lo 2-roller possessor is definitely a pincher. The bottom roller needs to be high-friction and needs to not roll easily (or at all). The bottom roller also needs to be stiff (so steel is better than fiberglass). No ball overdriving problem. The possessor will lose the ball if the ball touches the Ball Dam Gusset on either side. Transverse driving is therefore a problem if taken too far, but it can also serve to center the ball. We tested filling in the top roll spiral gaps with friction tape, but this had no obvious effect. Existing spirals are definitely not beneficial (they do not center the ball), but don't seem to hurt performance either.

Back to the Drawing Board (3 & 4-July)

The elimination of the middle roller makes access for actuation of the low roller simpler. Narrowing the top roller is no longer necessary. A design with this change was started, then abandoned. The possessor went through two major redesigns in two days.
Our tests make it clear that a great deal of backwards force is applied on the low roller. The Trirol's low roller mount was not designed with this sort of force in mind. We therefore need a more robust approach for the articulated low roller.
In light of this need for robustness, we started investigating vertical retraction (linear motion) design. Robust linear motion 80/20 roller wheel assemblies (80/20 part # 2751) were gallantly liberated from DEWBOT IV. Mounted on the existing 80/20 kicker spring mounts, a quick prototype test indicated promise. CAD design and physical prototyping and testing ran together in parallel. The revised design is simpler, more easily executed and more robust. Things are still tight and require careful design and fabrication. Key points:
  • The low roller is no longer intended to roll. No bearings provided.
  • The (2) 3/4" diam x 4" stroke pneumatic cylinders will still be employed (they have already shipped, so this is good)
  • The actuators will act in the direction of linear motion. Actuators will be hard-mounted (not pivoting).
  • Actuator mounts are moderately complex and must be welded to the chassis frame
  • The low roller is further forward and, when lowered, expected to occupy a lower position than the Trirol low roller
  • The height of the low roller will be mechanically adjustable
  • When retracted, the low roller mechanism will be entirely within the ball dam gusset envelope and therefore protected from impact with the bump
  • The kicker width will be reduced from 15" to 13". The cutting of the lower extremities of the possessor for clearance will be less extreme than for the Trirol design
  • The pinching design puts a lot of stress on the possessor drive motor and the polycord drive belts. We destroyed the original motor in testing.
Regarding the last of these points (the excessive load on the possessor motor & drive):
  • Our first thought was to limit current to the motor via a current-based feedback loop (a control/electronic solution)
  • but an alternative approach would be to reduce the friction coefficient or the normal load between the possessor drive belt and either the motor pulley or the possessor tube, turning this into a poor man's clutch. This would continue to apply pinching force while at the same time avoiding motor stall, motor burn-out, excessive battery drain, and polycord melting. for example:
  • reducing polycord tension by slightly increasing the belt length would reduce normal loads; while
  • a different surface on either the pulley or possessor tube could reduce the friction coefficient;
  • this approach could allow us to keep unloaded possessor velocity high, provide force for pinching; preserve the motors and drive parts; save battery power; and limit force on the low roller.
The Possessor is effectively narrower.
Additionally, benchmarking of other successful bottom rollers (which neither roll nor facilitate rolling) revealed an inordinate number of teams using 1/2" angle. Prototyping (with aluminum) seems to reveal why. The ball tends to roll (and thus fall out) much less than with even the misnamed bottom roller bar, and holds somewhat better in snake and much better in strafing. (Strafing is a beneficial aspect of our drivetrain could prove very helpful on defense against a certain looper bot at IRI.) The system isn't yet to death-grip standard, but I (Siri) wouldn't be overly surprised if changing the roller(s) surface(s) and/or speed (recommended at 2x robot speed) got us quite close.
We appear to be having sporadic DS-robot communication malfunctions, both during bootup and randomly mid-test. It'd be great to determine the root cause(s).

Hi/Lo Possessor Installation & Testing

Tempus fugit!. Only a week and a bit before IRI! Time to finish our work and test it all out.
Almost like a week taken from build season.

10-July Update

Mother nature played a little joke on 10-July. Morning rain was so intense that it flooded the robot shop via the outside door (at the bottom of a ramp). Fortunately, we arrived at the start of the flood and were ably to stay ahead of the water, so no damage was done. We pulled over 30 gallons of water off the shop floor. Opportunity was taken to organize the shop.

11-July Update

We have a Vulcan death grip (and it doesn't just convince the Romulans). The ball goes in. The robot goes forwards, backwards, sideways, slantways, longways, backways, and spinways. The ball stays in--approximately 2.75" but not more than 3", and not off the floor. It passes the ruler test, the paper test, and (most importantly) the Rizzo-the-Ref test. The only way to release the ball is to turn off/run backwards the possessor or kick. We don't even need to reduce the torque, though we can, since the vertical IR ball sensors work. The death grip doesn't visibly affect the kick range or arc (both of which are quite nice). The only way to pick up another ball is to get rid of the first--no way to possess two a once since the first on stop the top roller.
There is a catch, though. The system has quite a few consumables:

  • Lo Bar friction tape: The Lo Bar is completely wrapped in spiraled friction tape, roughly 1/2" shift per turn. When this wears out (maybe every ~10 matches), the possessor goes from a death grip to more of a standard nerve pinch and continues to deteriorate from there. The kicker needs to be back (with the cocking arms up for safety) to remove and replace this. Also, don't use the Fuzzy Soccer Ball for practice--it turns the roller white.
  • Possessor motor pulley friction tape: We've added friction tape to the possessor motor to make the driving band slip on the Hi Roller rather than the pulley. This prolongs their life considerably. (But only in a relative sense, they're still consumables.) It needs to be wrapped to tighten as the possessor spins inwards: counterclockwise as viewed from the outside of nearest long side, or away as viewed from the front. This should last us maybe 5 matches (a little more than the polycord, but less than the Lo Bar tape). We've like to use 1/2" friction tape for this so it doesn't fold up the sides of the pulley. Either way, we'll need more friction tape--lots.
  • The possessor bands: (2) 15-11/32" (?) polycord bands, or a 6% reduction between the physical geometry and the band length. These don't take as much abuse now that we've added friction tape to the roller, but they've still got a pretty short lifetime. We're thinking 4 matches or so. We've got a forearm full of spares (18) for IRI and the replacement process isn't horribly painful.

The Lo Bar cylinder mounts also limit bumper wingnut access. It's pretty annoying. (But not DEWBOT IV level.) We'll live though; the drive team has put their hands on a diet.


12-July Update

Driving Practice. Well, sort of.

Problems Encountered

Possessor Failure
Early in practice, the possessor began to fail to capture soccer balls. Unlike prior Hi/Lo failures, it did not backspin the ball, but simply failed to move it at all. Standard examination revealed minimally worn bands that barely slipped on the Hi Bar, stripped pulley friction tape, and entirely intact Lo Bar tape. In the name of expediency, the bands were replaced along with the pulley friction tape. In retrospect, I (Siri) will aim to address this more scientifically. In all likelihood we could have retained the original bands. At least one of these changes fixed the issue. We should keep an eye on this: failures of this type will be harder to diagnose during competition than ball backspin. Videographers need to be trained in what to observe and tape in these respects.
IR Sensor Failure
Largely simultaneous with the possessor failure, the dashboard's binary (ball/no ball)IR ball sensor readout failed. Originally it would display only "no ball" regardless of possession, later it began to flicker rather randomly. Anecdotally, the two values flickered together (despite supposedly reading separate sensors) and the value readouts corresponded to the binary. The issue eventually disappeared, though (again anecdotally) not exactly simultaneous with the possessor fix. The drive team should be briefed on how to (non-anecdotally) diagnose, and more importantly use, this. Tangentially, "Normal [Kicking] Mode" on the dashboard is fully manual (no IR sensor requirement). Override mode requires a positive IR ball reading.
Immobile Wheel 2
Soon after the possessor and IR failures, Wheel 2's steering froze. The motor was quite hot, though oddly not as hot as the other rear motor, which had yet to freeze. Entering calibrate mode and adjusting the pot had no effect. The wheel could be turned by hand easily, implying no mechanical impedance. Presumably it thermaled out, though after letting it cool it reheated faster than window motors normally do. We applied Freeze-it (of which we do indeed have a full can, though could use another), which worked acceptably but not profoundly. We could run into trouble in a finals situation. It eventually cooled and was usable for approximately 2 matches, until we stopped to address the below mentioned issues. Wheel 2 also had problems yesterday: drifting due to the Victor and immobility due to a loose PWM. These issues are in all likelihood unrelated, though it seems Murphy's favorite number is actually 2.
cRio Modules Loose
After a rather violent bump maneuver (cause unclear), we realized the wiretie holding in the cRio modules had been removed. The time of this is unknown, though it would have been replaced after welding the Lo Bar cylinder mounts. Module locations are now labeled and re-wire tied. The order is (as seen from the back):
Slot 1: Red board with 13 & 18 wires
Slot 2-4: open
Slot 5: main white cable
Slot 6: open
Slot 7: Red board with 12 & 3 wires
Slot 8: Red with 11 & 27 wires
Battery Box Break
Perhaps also due to aforementioned bump maneuver, the battery box has started to crack in the bottom left bend. We have one spare and will have a cutting template for tomorrow. Another alternative is to use galvanized strap to brace the bottom on the box. This has been rejected in the past, but maybe it's time to reconsider.
Communication Dropouts
Throughout practice, drivers reported brief (~1s) communication dropouts. This is most evident in the stopped possessor (which loses any ball in possession). This may appear as a possessor motor issue, but simultaneous driving has shown this is not (at least) always the case. Cause unknown. It's rather annoying. It appears not to occur in competition, though.
Kick Angles
Kick angles proved rather temperamental throughout practice, though especially at the beginning. Having the possessor on tends to increase the kick angle to the normally desired range, though occasionally 'something' will happen/it will catch 'oddly', and have virtually no arc. This warrants further examination.

To do:

  • Print xbox controller and possessor joystick control maps
  • Add relax kicker and Lo Bar manual extend/retract to the possessor joystick
  • Practice driving
  • We need to get used to using the new possessor. We shouldn't have to dawdle picking up balls anymore, but drivers are understandably hesitant.
  • Keeping moving. Successful (especially swerve) robots rarely stop. This is appears to have multiple causes. Firstly, it's a technical control issue: wheels really don't always cooperate, especially on short movements. Secondly, drivers' lack controlled and consistent experience, especially with defenders (which DEWBOT V is excited to remedy). Thirdly, it's a strategic problem, which requires big-picture coaching with eyes on the field (not necessarily DEWBOT, though lack of practice makes this more complex).
Fast & accurate aiming. Especially for the home zone, this is extremely helpful if not excepted. We can move the ball now; we should do this.
  • Practice possessor consumables changing.
  • Add thumbscrews to the front of the polycarbonate lid.
  • Determine battery box plan. Make a spare and/or switch the box or add the galvanized strap(s).

Mirror

Looks really good, but we should:

  • Verify that it stays within the frame perimeter.
  • It would be good to have a means of raising it during competition. A servo might be adequate.
  • A working possessor, or a sensor to kick automatically would be a benefit now that we can find the hidden balls.

3-July Update

  • The mirror was moved back 1" to keep it within the frame perimeter (Jen).
  • Initial prototyping and CAD took place for a servo raise/lower mechanism (DJ).

4-July Update

Much effort - little progress. Servos do not appear to have the necessary torque to raise the mirror.

Plan B would be to use a 3/4" pneumatic actuator.

7?-July Update

It works! With a little help from Foster and his VEXing, the servo pushes the mirror up with ease. It does not bring the mirror down. That's ok (in fact, it was in the functional specs), we'd rather have it up. A second latex tube has been added to the other side to prevent torquing.

11-July Update

We've loosened the second latex tube to prevent the mirror from re-extending outside the frame perimeter. We've also added a simple latch (a deliberately damaged wire tie to allowed unlatching) to hold the mirror down during maintenance. The mirror is not raised during autonomous, though it can start up. Retrospectively, the servo placement makes bumper wingnut and lifting handhold access a bit more complicated. (The non-possessor motor side would have been better.) We'll live, though. We've also realized that we've gained some primo advertising real estate on the back of the mirror, and that we can add a polycarbonate back shield for the same reason. This is good, especially since we'll have to cover up some of the current sponsor logo space with our autonomous tape guides. Plus, this new space is vertical.

Mirror photo gallery

Programming Version Control

Implement and use version control to enhance communication and collaboration.

Programmer Jobs

  1. Review the final code specs (below).
  2. Communicate and distribute assignments. Ensure team knows what you are working on and when your draft will be done.
  3. Indicate on meeting work plans what you will be doing during the meeting. Check off items (and upload your code) when you complete them.
  4. Use the Mercurial version control. Ensure versions are kept up-to-date and well labeled. Store code locally before competition (don't rely on internet at competition venues).
  5. Write, debug, and comment (describe how it works, possible issues/changes, etc) on your code.
  6. Ask questions as needed.
  7. Meet your deadlines and/or inform the team of issues.
  8. Inform the design team/meeting management what you need to test your code and how long you think it will take.
  9. Consolidate individual assignments into a single code body. Upload and test this as well.

Current Mercurial Assignment

  1. Create a local working copy of the repository on your own computer.
  2. Create a text file in the "wall" directory announcing your presence. (Hint: You will need to let Mercurial know about the new file by "add"ing it.)
  3. Get your new file to the repository.

Photo Gallery


Catch up on other team information at FRC Team 1640