now that we are getting to the last week of regionals and the championship is coming up I thought it would be good to start a thread of simple tricks and fixes or improvements people have come up with
or things you have seen at regionals that teams can do to improve their bots in the pits, or things teams can do to improve their performance on the field.
(Obviously) I have a few in mind:
ball herders. if you are trying to herd balls with arms or plows or panels that swing out, and find the balls jam and wedge, or if you have a problem with your bot riding up on balls, we found a simple solution. Spray the surface with silicon spray, then wipe it down good with a clean rag. The surface will not be wet or oily, and the silicon will not transfer to the balls and make then slippery, but it will greatly reduce the coeffecient of friction between the balls and your bot, and the balls will slip instead of wedging and jamming. You really have to try this to see the difference. its unbelieveable!
auton mode switches. We had switches on our bot to setup auton mode to turn left or right, and to disable it, and our team had an aweful time getting them right, because they were out of site on the bot and you had to reach under to set them. You can put switches on the OI input pins of an unused joystick port, and read them in the code - the code will hold the last value it saw before auton mode was enabled - its much easier to have a switch at the driver station clearly labeled AUTON L & R and to look at it and see L is Left and R is Right
auton mode is 1/8th of the match - have your bot do SOMETHING in auton mode. if nothing else have it dead recon forward about 20 feet to the center of the field, or to whereever you point it at the start of the match (other side is going for the release ball? you could at least get in their way) or push your mobile goal to the other side of the field? creating a sequence for auton mode is easy - set a loop counter (auton_count) = 0 on power up
and do something like this
while (auton_mode)
{
…
if (auton_count <120) // 40 = roughly 1 second, 120 = 3 S…
{
pwm1=pwm2=254; //drive motors on
auton_count++;
}
else pwm1=pwm2=127; //drive motors off
…
go after the release ball in auton mode - teams have put paint stirring sticks on servos that twist out and the bot simply runs along the wall- when it get to the release ball the stick knocks the ball off
use the IR sensors: its incredibly simple - if you connect them to digital inputs 1 and 2, then the two flags: INTCON3bits.INT2IF and INTCON3bits.INT3IF will be set when the sensors see the beacon - you DONT have to do anything with interrupt to use these - they are already there in the hardware waiting to be used.
all you have to do is something like
if (INTCON3bits.INT2IF ==1) //sensor on rc_dig_in01 sees the beacon
{
//do what you need to when the beacon is seen
INTCON3bits.INT2IF=0; clear the flag for next time
}
to make this more robust I would wait till you see the flag set two or three times in a row (clearing it each time you see it) then you wont get mislead by a glitch or a slight reflection - it will be a good solid IR signal hitting the sensor - to use the sensor house it in a light tight box, like the black hobbie boxes from radio shack, with a hole or slit cut in one end to let the IR in from that direction only.
if you have a simple frame bot, put an arm on it. make an A frame on the top of your bot with 2" angle alum from homedepot, have the top be near the 5foot height limit, then put the van door motor sticking out one side and put a 4.5foot lenght of PVC pipe on it - what good is this? turn the motor on so the pipe spins upward and you can fling the 2X ball out of your mobile goal - then you can push the goal under the opponents ball drop and you will get at lest 20 or 30 points when balls fall in it
or you can knock the 2X ball off the goals after your opponent has capped them - it would take about 2 hours to add this to a frame-bot in the pits and wire the victor up to control the motor
bring a camcorder and put it up in the stands - have someone film your matches with the camera zoomed in close on your bot the whole match- later you can review the tape with the drive team or the pit crew - if something didnt go right in the match you got it all right there on tape, and you can review it over and over when you are not pumped full of adrenilin, and you can measure time and estimate distances if your auton mode is not working right - we did this at buckeye and it was EXTREEMLY useful - and a fringe benefit is you end up with lots of video of you bot in action to take home - do it for every match- its well worth the effort.
Have a reliable method of lining up autonomus if you need it. We tried shall I say “dead reconing” and it worked most of the time, but sometimes failed us.
Check the PWM’s on the Spikes. They have a nasty ability of comming out right as you’re about to hang.
…just a small bit of the infinite things you could do
Confirm that all your electrical connections are tight.
Make sure the nuts holding the positive and negative 6 gauge is secured tightly to their distrobution blocks or circut breakers. Also check the screws on the new distribution blocks to make sure those wires are secured as well.
Check to make sure the (3 conductor wire, german cable, PWM wire, whatever you call it) is pushed ALL the way into the Spikes and Victors (these loosen over time.
I second that. Geez! We discovered that after shipping to the Annapolis regional, several screws/connectors became lose somehow and started giving us intermettint results (and as it was the result of two different things happening simultaneously, it was really fun to track down!) So, even if it worked at another regional, do extensive systems checks, and check everything visually too. Especially screws on speed controllers stuck in the depths of the robot (the problems always happen on the one speed controller that had to be put out of the way due to space … some sort of physics rule I think).
Also, just a general tip if you’re using pots … check your values before a match. A pot slipped on us a few times, and seemed to work in a systems check, but going up in a match it didn’t perform correctly because of a bad pot value … and afterwards, I always check all three pots on our robot before a match (I have a nifty debug pot on the oi to change between modes so I don’t have to upload new code). It’s horrible losing a match because of something so simple, and something that shouldn’t happen in the first place, so just keep a heads up for it.
Well, 12 hours and counting till we leave for Columbia! W00t!
one thing i noticed and will bring up to my mentors is that when we would do “safe move” auto and get out of the way, the other team would ram the goal into our side and the front omni wheel axles would bend and the chain would fall off. have a “safe” auto… go infront of the ramp instead of the side so when a team goes staright at the goal and hits it it wont hit your robot and mess it up as much. :]
edit: even though we still had 3 wheel drive we still made it to the top and hung every matcht hat the chain fell off, also we had screwed it and pinned it 3 times to the frame, what a mess i have in the pits in atlanta. :rolleyes:
even easier then hot glue is just a bit of silicon gives u the same effect that is easier to apply and switch out we learned the hard way until 237 came to our rescue at NJ
Also making a check list for after or pre-match things is a great way keep these little problems away
For example, at every event they go to, everyone on team 93 gathers together to watch their matches. Everyone is there, cheering like mad. And at the end, no matter how they did in the match, they gather 'round where the robots come off of the field and cheer for not only 93’s drivers and their partners, but for the opposing teams as well. It’s neat, fun, and people remember you for it!
Really, no matter how good or bad your robot is, if the whole team is there, watching your matches, cheering like crazy and just having fun everything will be better. From what I’ve seen, drive teams that have a lot of support tend to do better. Even if that’s not the case for your team, at least you’ll be having fun! Who knows, maybe you’ll even get a spirit award out of it!
that has helped me out so many times, when i went through the list i found out that our tranny chain was off and that one of our motors didnt work… it helps a lot to have what tools you need to do the check list on the check list too.
if your robot has appendages, make sure you have a perimeter quardend off so noone walks into a flying arm.
At BAE, we had a pit stall that was right next to a walkway and people kept walking past our bot…so make sure you have a perimeter. also, keep un-designated pit personell out of the pit. it helps with traffic, and keeps the pandamonium to a tolerable level.
This sint’ so much about the bot performance but about the team. One thing that every team should do is meeet outside of FIRST and robotics, juts go out on your own with a few of your teamates to just relax and get to know each other. Granted you mihgt already go with your friends but invite others on the team that you dont really get to know outside of robotics so it gives you a chance to learn and understand each other better on the outside so during the stress of build you can understand where someone is coming from and how they typically act so things dont get out of control or blown out of proportion. It is simple but very effective.
I’ll say something… I think many, many rookie teams should sit down and honestly tell themselves what their robot can do and how long it will realistically take to do so. Then write it down, memorize it, and leave yourself enough time in a match to do what you plan. At GLR and on the VCU and Portland webcasts, I saw many teams that spent time herding balls, blocking people, etc, and then lost a match cause it took them longer to hang or cap than they expected. Honesty with yourself and your alliance mate can only help you do better.
Heck, thanks to my poorly designed transmission, our robot was stuck in low gear for all of GLR. But we knew that we’d never get high gear to work in that timeframe, so we went with that and just crawled to the center of the field and hung. We did remarkably well and actually got picked for an alliance.
We wrote a checklist on the side of our gearbox. =)
-check that you’re on the right autonomous code
-are the breakers in?!
-is the radio connected?! (at BOTH ends of the cable)
-is the robot on?
And yes, we did miss out on matches for each of those reasons. Luckily those were only practice rounds. We’d perfected the whole checklist system by the time we got to the qualifying rounds.
In reference to what Ken said about ball herders: We had a similar problem. We drilled holes into the lexan plates that functioned as our “wings” and wove pneumatic tubing through those holes. The balls primarily came in contact with the tubing and not the lexan, which had a lower coefficient of friction.
Wow, we didn’t really test the robot out before we got to the regionals, we had a whole truck load of problems. So we needed alot of help. So I think we may want to pick up the pace next year. Instead of slacking off. I’m still agonizing over the problem with too little people.
If you are using something cylindrical for herding balls, we suggest using PVC rollers. Cut the PVC a little bigger than the rod that’s herding (and try to have another in the front of the robot), and just slip them on. Ken’s solution is more universally applicable but if you can use them, PVC works wonders. We’ve never ridden up on a ball and managed to herd really well in Sacramento.
i would say that it is important to define roles…
know who will be driving and hping the matches.
know the scouts and who to go to for strategy
know who is supposed to be in the pits (repair and judge’s team)
also…it is always good to attempt to stay with one strategy as long as possible throughout the competition. this lets the drivers become better at knowing robot timing issues and capabilities.
this is something I was thinking about yesterday. We are done for the year except for expo matches, but I still want to try it on our bot.
in control systems there is a concept called ‘jogging’ - its when you normally move something fairly quickly but sometimes you want to bump it just slightly to one side or the other - and some control systems have a JOG button that does this
on our bots the Victor speed controls reduce the speed of the motors by chopping or pulseing the 12VDC to the motor on and off - when you want the bot to move slowly it puts out a short pulse of 12V, then nothing for a few mS, then a pulse…
it works well mostof the time - but when you want to nudge the bot just a slight bit forward, or to the left, then sometimes its not enough or its too much
you could implement a jog function in your code by counting SW loops - for example, then the jog button is pushed, turn both PWM drive motors on full for one SW loop, then turn them off - or turn them on full for one SW loop, then turn them off for 37 loops, full on for one, off for 37…
this would be like taking a hammer and tapping your bot once per second - when you just need to nudge it over to line up with a 2X ball, or the ball corral, or to get exactly under the bar to place your hook
if being full on for one SW loop jogs the bot too far, you could try half power, or 1/4 power
I think this will work really well and cant wait to try it on our bot - if anyone else trys it let us know how it works for you.
(Though a pain to get off…)
we learned the hard way until 237 came to our rescue at NJ 
