Friday, March 28, 2014

3/6/2014-3/8/2014 Week 2: Hub City Regional, Big Test #1



This is our first regional.  After having scares on the way from El Paso to Carlsbad involving gas and not being any gas stations for about 150 miles or so, we were ready to play with robots, and see other designs/approaches to problems.  Also, it was time to make the several fixes (and rebuild half of the robot) needed, as the tweaks are a large part of the detailed design phase.

Shooter

Shooter requirements for the robot:
- Winch back consistently and effectively
-Shoot consistently
-Have a continuous hit range (make the peak of the shot only reach the top of the goal)
-Keep the ball on the shooter for shots

Bonus for shooter:
-Hit from 20 feet back

Unfortunately, we did not get to test the shooter much on Thursday, since almost the entire day was spent trying to make the changes to the robot.  Since the field for practice matches closed at 4:30, which was when we finished, we did not get to play around with the shooter on the field, but got to begin to play with it on the small practice area.  While we did get to play around on the practice field, we did realize the black tubing was far too weak for the energy output compared to the energy input/the friction within the bands made the bands not return as quickly as the other thinner walled bands.  Before we could really replace the bands, it was about time to go.

Friday morning was a bit more interesting, as there was the driver's meeting and everything else.  After the night before, we went ahead and changed the surgical tubing out from the shooter and went back to the thinner walled stuff.  This worked extremely well, but now the winch had trouble winching back, and would stall just before latching.  After looking at it, we realized the churro we were using was bent, which definitely is not a good thing.  After seeing this, we went ahead and pulled our extra hex shafts from the practice robot (which we brought along for some reason as a just in case we needed parts), since they are a lot sturdier. After burning a decent amount of time with this, we went ahead and tried it again, but again realized it was having trouble.  Because matches were happening, we had to just decrease the amount of tension and shoot from much closer.
After a while, we realized the way it was wrapped for the competition bot was different from the way the strap was wrapped for the practice bot.  After seeing this, we realized the strap was constantly pulling on the middle from the sides of the catapult, rather than pulling straight down on the sides from the sides.  Because it was pulling on the middle from the sides, there was a quick angle change, meaning the force was no longer radially perpendicular to the rotation of the axle, so that the two motors would have to work harder and harder by a factor inversely proportional to the sine of the angle between the strap (spanning from the side to the middle of the catapult) and the axle.  This was an easy fix, since it was just changing the way the strap ran on the catapult, so that the winch was pulling on the sides rather than the middle, so that the sine of the angle between the strap and the axle would stay nearly perpendicular.  Once we changed that, the winch worked perfectly again.
While the winch still worked well, we also remembered how important the radius of the "spool" is, as that was the original reason we couldn't use the yellow strap.  Since the radius is more important than the angle of the strap with the axle (although both are important), and it is harder to pull back the further back the surgical tubing is pulled (or it is easier to pull back if the tubing hasn't been pulled back much), it doesn't matter so much what the angle is between strap and bar when the catapult is beginning to be pulled back, but rather toward the end, when there is the most upward force.  Because of this, we went ahead and mounted the strap to the axle a little further in, decreasing the radius, but keeping the angle between the strap and the bar nearly perpendicular as the catapult was beginning to be latched.  This allowed for smoother operation with the winch, and, finally, the winch problems are solved.

Finally, after seven matches or so, we had the shooter winching back and shooting consistently.  Saturday was a similar day to the end of Friday, in that we just made sure the latch was okay and continued to shoot consistently, although from much closer up.  By the eliminations, we finally realized we had to re-tension the surgical tubing, and get it back up to the power it was supposed to be, and, since the winch was now working.

Requirements met by end of Regional/first test:
-Shooter winches back consistently and effectively
-Shooter shoots consistently
-Had continuous hit range (did not miss shots due to ball being too high)

Bonus trait of shooter by end of Regional/first test:
-Can now hit from 20 feet back after re-tensioning

The ball still stays on the shooter while shooting, but can be knocked forward when nudged, or when the robot abruptly stops.  Also, while the shooter still works, the surgical tubing seems to be way overstretched. After looking up how much surgical tubing is supposed to be stretched/the most efficient amount, we found it to be about 150% the original amount, so that would be something we may possibly change.

Intake

Requirements for the intake:
-Pick up the ball quickly
-Spit out the ball quickly
-Actuate/move outside of the frame perimeter
-Be sturdy/able to take hits from other defending robots

Bonus for the intake:
-Keep the ball held on the shooter
-Pick up from a wide range

The intake worked well, except for the second and third matches.  With the physicality of this year’s game, we did not realize how semi-ineffective the rivets on the intake were.  After the second match, we realized a large portion of the rivets were entirely knocked out and broken, which caused the intake to flop around during the match.  Unfortunately, we had matches close to each other, so we ended up getting to replace some of the rivets with bolts, but not all of them, so some more of the rivets holding it all together were knocked out.  Also. the chain slipped a couple of times, but, luckily, it was in the same matches as when the rest of the intake wasn't working.

In terms of the worries of picking up over inflated balls, as almost all of the balls were inflated to pressure rather than size, the intake did just fine, and the motors did not stall at all.  At most there were two times where the robot had small amounts of trouble picking the ball up, and just needed a bump, either from a wall or another robot to pop the ball the rest of the way in.

For the shooter, we will need to add something extra to the intake to try to keep the ball somewhat contained.

Requirements met after regional/first test:
-Can quickly pick up balls from the ground
-Can quickly spit balls out onto the ground
-Extends out of frame perimeter to get balls
-Loads ball directly onto shooter

Bonus traits for the intake met:
-Can pick up smoothly from the sides due to the side rollers

Need to work on something to keep the ball contained and stable on the shooter, so the ball stays even if the robot is hit.  Also may need a second motor, so that if the chain slips on one, there is a back up to still drive the intake.

Drive Train

While the drive train works, we need to make the robot have a higher top speed/be faster, based on the style of play.  Also, due to what happened, we found we need to increase the friction of the wheels, because, often times, the wheels would just start slipping.  Overall though, pretty much it was solid, except for those two things.  In terms of the wide, square, or narrow debate, the robot was just fine in terms of stability, and tipped over forward a total of once (immediately got back up by putting the intake down.  For maneuvering between robots, everything would most likely have been easier if the bot was square, because it would have a slightly smaller footprint in the width department, but would still be easy to turn and have a large enough intake.

Catcher


Thursday, we went ahead and added the catcher, as well as the yellow straps on the catcher.  Pretty much, it did everything we could ever ask for.  As a human player, I was able to throw to the robot from between 12 and 15 feet away, which made it easy when defense was beginning to be played against us.  As the goal was just to make the human player job easier and allow the human player to have a larger range for error, the catcher was doing exactly what it was built for (and, it made the top half of the robot look like the top half of a lawn chair).

Regional Summary

In terms of requirements set and completed for our final product, every requirement set was completed and done by the end of the regional, and it is just a matter of adding the things that would be nice for us to have, but the final product is definitely finished and working as it is supposed to work.

Overall, we ended up as 29th seed at the end of qualifications and ranked 12th in OPR for the regional, including the first 5 matches where we were not working.  Once we were working, we were one of the obvious top teams, which is why we were picked second overall, meaning we were at worst the fourth best robot at the regional, and, pretty much showed, we were at least one of the top three robots out of 45 there.

While we ended up getting eliminated in the first round, partially due to the alliance partners (easily defended and had trouble with mechanisms when defense played on them), we did with the Quality Award.  The Quality Award, as described in the FIRST Administration Manual for this year's game, "celebrates machine robustness in concept and fabrication".  This, of course, is a huge award for us in terms of showing that the engineering design process has worked very well for us, and, for the sake of the engineering design process, is one of the more important awards in judging the build season as a whole.

In terms of things to do once we get back, to improve performance for the Arizona Regional, we need to change the gearing so it goes faster, change the wheels, figure out a ball retaining device, and (preferably) figure out a new way to mount the surgical tubing (and add lights, although it probably would more so add a cool factor rather than enhance performance).  While the latch has been working alright, it would also be nice to see if there is a different latch we could make that would not wear down as quickly, although it is okay for now.


Approximate times of working Thursday: 8:30 AM - 8:00 PM

Approximate times of working Friday: 8:00 AM - 7:00 PM

Approximate times of working Saturday: 9:00 AM - 1:30 PM

Wednesday, March 26, 2014

3/2/2014 Week 2: Preparation for Hub City Regional

Today was the last meeting day before heading off for the Hub City Regional.  We went ahead and made a list of what was needed in terms of changes made on the practice robot since the competition robot was bagged.

Shooter

While the winch was almost entirely changed yesterday, we went ahead and got to test the newly built winch, as it now had the two CIM motors as well as the new yellow strap.  As expected, everything worked as needed, and the winch operated smoothly.  Again, from the way it was wrapped, everything went as it was supposed to, and it worked until the battery would dip below 12 when inactive.

Things to do with the shooter once we get to Lubbock
-Change out winch motors
-Rewire winch motors
-Replace Kevlar rope with yellow strap
-Attach the surgical tubing

To summarize the above, replace everything on the shooter except for the structure/shooter itself (replace the actuators)

Intake

Today, as before, the intake worked decently, and, while the intake was not really tested much with the larger overinflated balls, we found it shouldn't be that much of a problem due to the gearing/the intake motors should not stall at all (nice to have it 25:1).

Things to do with the intake once we get to Lubbock
-Change the gearing on the intake motor to the 25:1

Drive Base

As we had to take out the motors from the gearboxes to put in the winch, we had to take those mini CIMs previously on the winch and put them into the gearboxes for the drive base.  This happened yesterday, and, as before, not much difference was made in the driving as a whole, just not quite the same amount of acceleration.

Things to do with the drive base once we get to Lubbock
-Change out the CIMs and mini CIMs with the winch

Catcher

Today was the first day in a while of getting back to the catcher in a while.  After seeing what many teams did, we figured all we needed was some sort of extension upwards, and angle it downward to bounce the ball into the robot.  Before, we tried playing around with plywood, but that seemed a little odd to mount at an angle that we would have wanted, so we went to what we used before, conduit.  Pretty much, all we had to do was bend the conduit to be mounted on the tower bars supporting the top bar (the bar that holds the surgical tubing), so it looked like a large rectangle on top of the robot.  We went ahead and mounted it to the robot using hose clamps, since it wouldn't be hit too hard, and angled it slightly downward, and added strap so that the ball couldn't go through.  Overall, it worked extremely well, and bounced the ball straight down into the robot, which is exactly what is needed.  While it may not be what was thought of in the beginning of the year, it was useful at least to me as a human player, where I could just hit the target/catcher, and know it was going into the robot.

Things to do with the catcher once we get to Lubbock
-Mount it on the robot

Approximate times of working: 12:00 PM - 6:00 PM

Friday, March 14, 2014

Competition Season and Regionals Week 1: Alamo, Inland Empire, Southfield, and GTR West

While the title only has half the events that happened this week, those are the major ones with some of the national powerhouses and interesting happenings.

Alamo (San Antonio, Texas): 118, 148, 624, 2468 (seeing at Hub City)

From Alamo, most of what was shown are that 3 robot cycles are possible, and a much more plausible way to go than originally thought (this was found when watching matches with primarily 118).  Also, we found that defense was a big factor for teams that had slightly lower shooters (148), and quick turns will definitely be needed to maneuver around defense, rather than pushing, since the clear shot is needed.

Inland Empire (Grand Terrace, California): 1678, 399, 1828 (seeing at Arizona)

For Inland Empire, 1678 was definitely the team we would like to be.  Although they would get continually hit, it was about quickly maneuvering, lining up to shoot, and getting the ball out as fast as possible, and, just in general, trying to go around than through.  Also, one of the main strategies seen with them is the over truss to human player, then getting the ball from the human player and scoring, allowing them to do quick 20 point cycles and not have to rely on other robots if other bots on the same alliance were built primarily to be defensive.

Southfield (Southfield, Michigan): 33

Main thing here was watching how great the Bees are (Team 33, the Killer Bees), but watching how teams would respond to them.  Pretty much, I saw a new strategy not previously seen or heavily talked about, which was double team the powerhouse, and one of the other bots run back and forth doing the 20 point cycles (video can be found here).  This will definitely be helpful for anything we may run into at either Hub City or Arizona.

GTR West (Toronto, Ontario, Canada): 610, 1310, 1241

This regional was particularly interesting due to the teams attending.  610 and 1241 are the current defending world champions, so it was nice to see what they had planned.  Overall, the general strategy taken by these teams were similar to 1678, but, rather than doing everything on their own, they would have a different bot positioned next to the human player, and that different bot would be the one that scores.  This opened things up to the 30 point cycles, and does not burn much extra time in the process.

Palmetto (Myrtle Beach, South Carolina)

Pretty much, don't violate G40!

Central Illinois (Pekin, Illinois): 1986

This regional showed the physicality of the game, as can be seen in the last finals match, that people really hit each other hard, and there are not any fouls called if bots are hit hard

2/23/2014; 2/26/2014; 2/28/2014; 3/1/2014 Week 1: Detailed Design Continuation

2/23/2014

Today, we went ahead and started making clamps for the winch axle so we can prevent the strap from ripping out.  General idea is to secure the strap by wrapping it around the axle, then adding clamps with bolts that both hold the clamp and the strap, and have the clamps squeeze the strap and into the axle.  Hopefully, we will be able to add these to the axle Wednesday and start to test the winch again.  Other than this, not much else happened in terms other design things.  For now, the winch is now the only mechanism that has not reached or exceeded the standards and requirements set for it thus far, so it is not as far along in the detailed design process as the other mechanisms.

Approximate times of working: 12:00 PM - 6:00 PM

2/26/2014

Today was a heavy programming and making the other things for the bot day (finishing bumpers, making extra parts, finishing making conveniences for the regionals like battery cart, etecetera).  While we did get to attach the stuff for the winch, we did not get to really test it extensively.  We did try to winch back a couple times, but, when it wouldn't work, we saw that we were missing a screw that tightens a lock collar, which caused the bearing to pop out, so the shaft wouldn't really turn.  By the time we found that, we didn't have time to change it out, so we couldn't really test the part we wanted to.

Approximate times of working: 3:30 PM - 7:30 PM

2/28/2014

Today was another day without changing too much, although the winch was further tested.  Unfortunately, the clamps were not going to be the solution, because what they add to the radius of the axle is far too much, and it stalls out the two mini CIMs.  Because of this, we went ahead and figured that possibly still using the backpack strap could work, as long as there is a heavy duty safety strap/hard stop of some sort, to prevent the backpack strap from breaking due to overextension.  Also, we cut down the clamp a bit, and realized we didn't have to have it encircle the entire axle, but just part of the axle, which definitely helps with the radius of the axle.  Other than working on the winch strap, lights were worked on, but that both doesn't change the overall design and is only to help the drivers.

Approximate times of working: 3:30 PM - 7:30 PM

3/1/2014

Today was Maketopolis, a public mini Maker Faire, hosted by Xerocraft (where we're building).  We got to demo the robot in public, but also got to heavily test the winch and the robot as a whole.  Overall, it worked extremely well, and the shooter was extremely consistent, as well as intake and so on.  Everything worked well, until the strap broke again, but, rather than while firing and unwinding, while winching back.  This showed us we absolutely need to use the yellow strap.

After the strap broke, we went ahead and decided to look at the stall torque of the mini CIMs, and found that it was at about 200 inch pounds, which, by far, is not high enough for what we need, and that is why it stalls with the yellow strap.  Stall torque is the amount of torque needed to stall a motor, and inch pounds is one of the units, which means how many pounds of force the motor can take from one inch away from the axle (foot pounds is more commonly used, but, since we are dealing with smaller distances and motors, it's a lot easier to use inch pounds).  From this, we knew we couldn't stay with mini CIMs there unless we really changed the gearing, which would be complicated and a bit of a process.

Because of this, we figured we could switch the mini CIMs with the CIM motors on the drive train.  When we looked at the stall torque for the two CIMs, we saw it was 360 inch pounds, which, as can be seen, is much better than the mini CIMs.  Also, this would only make us lose about 12% power in the drive train, so, overall, it seemed like it was gaining a lot for not giving too much.  By the end of the day, we were able to change the motors out and reconnect everything, but not test it, as it took the last hour or two of the day.  Hopefully, this ends up working, and all that is needed is testing it tomorrow.


Approximate times of working: 8:00 AM - 4:00 PM

Monday, March 10, 2014

2/19/2014 and 2/21/2014: Week 0 of Competition Season/Week 7 of Build Season

2/19/2014

Over today, not really anyone came, since most of us were working until at least 9:30 PM on the robot last night, and had school, so people were mostly wiped.  As I was expecting no one to really show up and not much to happen, I ended up leaving early at 6:30 PM.  Today kids went ahead and built a battery cart, and we began to talk a bit about the catcher, although not much.  Also, the reveal video got lots of work done on it, and should be ready by tomorrow.

2/21/2014

Today the release video officially was released, and is right here.  As shown, the shot is nice and flat, but we definitely can increase the shooter power still to make sure autonomous goes in.

In terms of things that happened today, the unfortunate happened, and we continue to struggle with finding a permanent fix for the winch.  Today the kevlar rope ended up snapping, so we went out and just got the heaviest duty stuff we could find finally, which is in a way a double strap with nylon in it rated at 4000 pounds, which better hold up when we end up installing it.  Also, we ended up deciding we needed to use some sort of clamp for the strap to make sure it wouldn't rip off entirely if it ended up not being to stay together near the holes for some reason.  Pretty much, as before, the main thing still bothering us is the winch, and everything else is pretty a okay.

End of Build Season: Comparison and EDP Post Build Season Conclusion

As compared to the past years, we went at a pace more than twice as fast.  While in past years we have only bagged everything but the main mechanism, which needed work, this year we bagged a fully working robot that could play the game, and built two different practice robots that were both functional, in addition to anodizing one of the robots.  Overall, the Engineering Design Process was worth it, as it was a lot more organized than past years, and is not too heavily affected by time, which has an effect inversely proportional to the preparation of the people within the group.  Because we were fairly prepared this year, we did not really have that much of a problem with the time restraint, and, as we have been getting more and more prepared, we have gotten further and further in the process as a whole of building a bot.

In terms of my hypothesis about the Engineering Design Process, it is basically the same as my current conclusion of the Engineering Design Process.

In terms of everything else that will be happening, and seeing other robots and so on, most things will just be confirmations of our conceptual design process.  All that is left is to get the results for the performance of the product, and that will conclude my project.

2/18/2014 Day 45: Bag Day!!!

Today was the final day of the build season, and, finally, we got to bag a fully working robot.  While we normally don't meet on Tuesdays, as it is the last day of the build season, we went ahead and finished what we needed to finish since this past Sunday.  Also, we began watching some of the videos that were beginning to come out, mainly those of teams 610 (defending world champs) and 11 (a really good and nationally known team) (click on the numbers to go to their release videos).

Shooter

Today we went ahead and added the kevlar rope to the winch, and, as shown with the practice bot, was effective.  Also, we went ahead and added the surgical tubing, so, by the end of the day, we bagged a robot that had already made shots, which is a huge improvement from past years.

Intake

Today we went ahead and only had to add the pool noodles onto the shooter as the thing that helps retain the ball, and, as it worked on the practice robot, it worked well on the competition robot.

Drive Train

As we finally had a fully driving robot, we were able to play with it, so we got to drive it around and play with it.  As expected, the moving of the battery to the back of the robot was huge, and the robot was no longer rocking back and forth at all.

Overall, the competition bot did everything it was supposed to do, and there was not really anything more that we could ask from it. The robot is pictured below.



Catcher

After looking at the 11 robot, and from our experience at Duel in the Desert, where someone had built a full-out catcher, we figured all that was really needed was the ability to catch/receive from a human player, and all we had to build was something that would settle the ball downward into the robot.  As of now, the general ideas are just attaching what 11 had, or adding an angle to that so it bounces it downward more.

Approximate times of working: 3:30 PM - 11:00 PM