DEWBOT VI IRI Problems Encountered

From DEW Robotics
Revision as of 18:40, 9 August 2010 by Siri (talk | contribs) (spelling)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

We ran into several problems (some major, others less so) at IRI 2010. On the up side, there were 78 teams at the competition, which means we often had over an hour in between matches. On the rather less elevated side, we often had plenty work to do in that time.

Losing Communication (Persistent)
This pretty much sucks. Though we've determined that we can connect the DS to the robot through the wireless in the pit, we sometimes (though thankfully not even more often) flicker the connection on and off on the competition field FMS. It's even worse on the practice field, when we have to tether. Part of this can be attributed to a loose Ethernet connection due to a bad port on the Classmate. We need to repair or replace that--sooner rather than later (though we'll have time enough after IRI). Secondary issue: we don't know what the diagnostic error messages mean. Ben to look up as full a list as possible Friday night.
To Do: Find out what's wrong (ok, more of a goal than a plan). Send the Classmate in for repair/replacement. (Foster to build a pigtail to be used to release stress on the Classmate)
Contact: Ben, Foster for technical. Carly, Sasha, Clem, Siri for observed instances.
Possessor Eating the Ball (Persistent)
Ball pulled in too far by possessor
The possessor repeatedly pulled the game balls into the front of the robot over the lower possessor roller. This breaks two rules (G44 & G46) and generates penalties if observed by a Referee (which it was). G44 (Ball Carrying) generates two penalty points (we were called on this once). G46 (Ball penetration restriction) generates one penalty point and earns a Yellow Card if not remedied promptly. We are unable to "eat" our practice ball, so the field balls and our practice balls are clearly different. The field team relates that the field balls were fully inflated, but have a "padded" feel to the cover. The kicker will not work with the ball completely pulled in, but reversing the possessor will expel the ball--usually. Raising the Lo Bar roughly 1/16" did not produce observably different results in competition. (Granted, things like this are difficult to test from 10 feet away at standing level behind a polycarbonate shield during a late qualification match while watching the rest of the field and communicating with alliance coaches with a ref standing over the robot. I admit I'm pretty well biased towards the "avoid penalties" call. - Siri)
To Do: Procure accurate game pieces and test. (Ordered - Delivery date 27 July - Foster) Can a secure back-stop be added? The Lo Bar height may (also) need adjustment.
Contact: Siri, Carly, Sasha for observed instances. Clem and Siri for previous testing. Clem, Ben Kellom & Gary Deaver for mechanical development.
IR Sensors & Autonomous Kicking (Persistent but Understood)
The IR sensors have come loose and lost their angles, thus yielding entirely inconsistent (but mostly just "no ball") readings. We started to fix this on Friday but were sidetracked when we realized the extent of our communication problem (issue 1, above). This caused autonomous kicking to work infrequently at best. When it did kick, it came too late and always caused the ball to miss the goal. We scored zero points in autonomous. This is unfortunate, our angled autonomous (especially near zone) is really quite cool.
To Do: Reset the IR sensor angles, then carefully measure and fabricate plastic blocks to secure them. Checking these also needs to be added to the pit checklist.
Contact: Clem, Siri, Carly, and Sasha for observed instances. Siri and Matt for technical/mounting.
Auto-Fire (Matches 7-15+?, Fixed Itself?)
To everyone's continued bewilderment, the kicker fired on it's own in teleop (though ironically, not often in autonomous) for our first several matches. It's not in the code. It's not in the specs. It doesn't happen on the practice field. Fortunately, it's sometimes a pretty smart auto-fire, kicking only when Carly is in line with the goal. This is actually because she lines up slightly away from the ball and then inches forward--apparently almost exactly the correct distance. Unfortunately, the operative word is "almost". Even at its best it actually kicks just short of contact, making the arcs rather anemic. Oddly though, in Match 15(?), "almost" turned into "nowhere near". We're still not really sure what happened, but we seemed to regain control by at least Match 57. Be nice to know, though.
To Do: Take the kick off the lower xbox joystick button, or practice with it that way. Investigate other issues and/or causes.
Contacts: Carly, Sasha, Clem, Siri, Ben. Paul also consulted with understandably little enlightenment.
Hoboken Death Grip (Fixed, More to Do)
We worn down the standard consumables on the Hi/Lo Possessor, making our death grip rather less grippy (deathy?). We switch everything pretty easily, though the Grip-All rubber band didn't work out so well. We do have issues finding the correct replacement friction tape, but we should have enough to make it through.
To Do: Get the right friction tape. Work on replacement procedures and making items less consumable.
Contacts: Clem, Siri, Doug, Matt, Sasha, Andrew.
Lo Bar Actuation (Largely Understood)
One on the cylinders on Lo Possessor Bar was bent. It took us a bit to discover, but we did replace it. The left stop also fell out once and over-loosened once more. We also managed to twist the upper rollers and slide the corner gusset stops. This is probably what caused the Lo Bar to slip and made us eat a ball in Match 67. We added stops to the outsides of both rolling 80/20 struts, which seems to have improved things. Oddly, we also managed to completely rub down the friction tape on the bottom side of the Lo Bar. This is interesting, considering it's physically impossible for the bottom of the bar to hit flat ground before a significant portion of the 80/20 struts bottoms (like 3/16"-ish).
To Do: Fix the sloppy geometry that's inhibiting reliable operations. Specifically, get the 80/20 rails squared-up and the cylinders running parallel to them.
Fix the mounting hole/angle for the right (viewed from the front) cylinder.
Add polycarbonate spacers between the roller mounts to keep them square.
Investigate adding an aluminum angle backstop to the bottom of the moving 80/20 struts to help keep things square.
Also, Add tapped endcaps to the moving struts to hold the corner gusset stops in place. This wouldn't secure any adjustments that lower the stops (raise the Lo Bar), but would at least keep them in the general vicinity--better than some situations we've been in. Finally, figure out how the heck we managed to rub off the friction tape on the bottom side of the Lo Bar (and/or determine that whatever did cause it won't present further problems).
Contacts: Clem, Siri, Doug, Matt.
Jammed Kicker (Match 32, Understood)
We could not get connection in the pit (this was before we learned the wireless trick) and thus had to start with only 80psig upstream pressure(the original minimum we were comfortable with). Unfortunately, that minimum was decided pre-Lo Possessor Bar. Come kicker relaxation, the much larger diameter kicker cocking cylinders start creeping back while the Lo Bar shoots up, and you have yourself a jam. Also unfortunately, we've removed the Lo Bar extend/retract override from our competition code (...but not our final specs).
To Do: Re-add manual Lo Bar retraction and extension to the competition code. Ensure the rest of the competition code does follow the final specs, and that the entire programming team understands the specs and code.
Contact: Sasha, Siri, Carly for observations. Ben for fix.
Kicker Bands (Friday Night, Fixed)
These came loose early in the day, but luckily that's easily fixable. They (namely the forth band) also proved abnormally difficult to replace Friday night. We managed though.
To Do: Teach the rest of the pit crew the wire tie trick. Determine why the fourth band was so stubborn and/or that both the symptom and cause are no longer issues.
Contacts: Clem, Siri, Doug, Andrew, Matt, Sasha, Rita.

Other pages about the DEWBOT VI Indiana Robotics Invitational

DEWBOT VI IRI Photo Gallery
DEWBOT VI IRI Scouting
DEWBOT VI IRI Matches
DEWBOT VI IRI Problems Encountered
DEWBOT VI IRI Trip Reports
DEWBOT VI IRI Post-Mortem