Wednesday, February 22, 2012

Build Log 46: Last Day

Build Season is over


It looks pretty close to the CAD Model



Bag and Tagged at 11:16pm

Headed to Dallas West on 29th-31st of March and Lone Star on April 5th-7th.


Monday, February 20, 2012

Build Log 45: The end is near

We're coming to close of the 2012 build season only 1 day left and we have to put the robot in the bag.

We spent the day doing final changes to the competition bot, we fixed a few collection issues we noticed during the scrimmage this weekend and we also made a new camera mount.

The collector is nice a speedy now, it's 1.4 secs from the ground to the shooter. We have had it faster but the torque it has right now prevents some of our jamming issues.

The drive train is working great with the addition of the counter weight. It can turn on a dime with ease.

The programming team is still working on several automation issues and we are beginning to design our custom control panel.  Both these tasks will carry over to the pre-competition season.

Tomorrow we will put the decals on the robot and demo it on two different occasions. We should have some more photos and videos up after that.

- Allen Gregory

"Individual commitment to a group effort - that is what makes a team work, a company work, a society work, a civilization work." - Vince Lombardi

Sunday, February 19, 2012

Build Log 44: 2nd Scrimmage Day

Today was the 2nd day of the Houston Scrimmage. We had a really good turn out, today. 624, 441, 3335, and 2585 all showed up for some amount of time. We are excited for the Houston Robots this year, they are coming along very nicely. Lone Star is going to be fun.

Today was almost entirely drivers practice and testing Ultraviolet (the robot). We were able to get the barrier crossing down a bit smoother, we can now controllably ride on the back wheels over the barrier providing a smoother crossing.

We had a few minor issues here and there but nothing that would have kept us out of a match at competition so that is a good sign. We're also learning what we can fix to prevent those issues in the future.

We took some time working on shooting from the other side of the barrier and it was pretty consistent (> 30%) for not having any form of auto aiming this weekend. It might be a viable strategy to act as the feeder/long ranger shooter.

Here is a sneak peak at something we filmed for our release video.

FRC#3847: From the 2nd Floor from Spectrum 3847 on Vimeo.


We did some touch up painting when we got back to our lab and we have tomorrow to work through a couple fixes that we came up with over the weekend but overall she's ready to rumble (pun intended). Also we have a very long list of projects to get done between now and the Dallas West Regional.

- Allen Gregory

“People are smarter than you think. Give them a chance to prove themselves.” - Timothy Ferriss

Saturday, February 18, 2012

Build Log 44: Only 3 Days Left

We spent the day at the Houston Scrimmage with teams 2587, 1429, 2585, and 441. It was great to see the other teams robots and see how ares works on something closer to the real field.

We had our share of problems today but the last few weeks of tweaking have been making us better and better. When we were working well, the collection system was feeding well and the shooter was great. We still need to work on our control panel and automation but mechanically we are looking pretty good.

The bridge device is our biggest problem, we aren't using pneumatics and a Jaguar just to run that motor seems like overkill but it's looking like a real possibility. The spike issues we had today were some of the strangest problems we have had all season. Not really sure what's going on but there doesn't seem to be a clear solution.

We did have some fun playing on the bridge with the Discobots. This was staged but it's interesting that this could even work.

We'll have some more video up tomorrow of actual shooting and driving.

- Allen Gregory

"I know you've heard it a thousand times before. But it's true -- hard work pays off. If you want to be good, you have to practice, practice, practice. If you don't love something, then don't do it." - Ray Bradbury

Friday, February 17, 2012

Build Log 43: First Day of Scrimmage

We spent the morning making more fixes and working on the practice bot, we still have a few ways to improve but it's coming along. We fixed some collector issues where we had some jamming.

The robots counter weights got some improvements. The IR sensors and elevator automation is coming along nicely.

The second half of the day was working on setting up the scrimmage at Rice. We should have a few local Houston teams joining us and the Discobots tomorrow for some live game action.

At the end of the night we had an issue with the #25 chain on our collector so we are going to be moving to #35, which we should have done along time ago.

- Allen Gregory

“What you do every day matters more than what you do every once in a while.” - Gretchen Rubin

Thursday, February 16, 2012

Build Log 41: More tweaks

We kept working on the copy bot, we have a lot of the mechanical done but we are still missing a few parts to make it work like the real robot. Another AndyMark order and we should be home free for a second robot.

The real robot is now not shooting strait consistently again. A couple changes in the wrong direction and now we have to fix it again. It appears that the hood and the compression bar are both not strait now. They should be relatively easy fixes but we still have to make them work. Hopefully before the scrimmage on Saturday.

We did some more testing and we noticed that the battery voltage alters the shots far more than we though even with velocity control. We think it's due to the torque required to maintain speed through the ball. Since the voltage is less thus the torque is less than the ball slows the wheel more than at full voltage.

We worked on getting the PID controller setup for the hood and now have accurate angle control.

We also worked on adjusting the center of gravity and we have found that we can make it turn properly by adding weight to the front. It's still not perfect and it isn't how it's actually going to be for competition.

We're still missing decals and netting from the side of the robot as well. Bumpers will be finished this weekend.

- Allen Gregory


"Slow down and remember this: Most things make no difference. Being busy is a form of mental laziness—lazy thinking and indiscriminate action." -
Timothy Ferriss

Wednesday, February 15, 2012

Build Log 40: Crunch Time

We're coming down to the home stretch, we only have 6 days until we have to put her in the bag.

We're making our last efforts on the competition robot before we have to bag it We're adding the body work that will prevent balls from entering the robot where we don't want them too. We're also fixing a few minor issues here and there as well.

The scrimmage is coming up this weekend and we need to be ready to use the field time to the best of our abilities.

The practice robot is coming along nicely though it seems like it's going to be difficult to get it to match perfectly. It will be close enough for our testing purposes but tuning on the Thursday of competition is going to be very important.

- Allen Gregory

"I do not think there is any other quality so essential to success of any kind as the quality of perseverance. It overcomes almost everything, even nature." - John D. Rockefeller

Tuesday, February 14, 2012

Build Day 39: Mostly Practice Bot Parts

We're gearing up for the scrimmage this weekend. So that means the robot is handed over to the programming and electrical teams, while the mechanical folks work on replicating all of our parts. We are on pace to have a functional practice bot the week after we bag the real one.

We weighed the robot at the end of the night and we are right at 100lbs which is what we were going for. This means we will have about 20lbs to play with to adjust are CG for turning. We will be messing with the weight a lot this weekend to get it to a point where we can steer perfectly.

The electrical team is cleaning up all the wiring and making sure we won't have any electrical problems during competition.

Programming is working on more of our control loops and making sure our sensor data as clean as possible.

We should be able to do some final testing before the scrimmage on Thursday night and Friday Morning.

- Allen Gregory

"I don't care if my team wins. I care if they try their hardest to win." - pfreivald









Monday, February 13, 2012

Build Blog 38: More Programming

We worked out a fix for our bridge device but we still haven't tested it.

We have tip bars in place in case we fall over so we don't damage our shooter motors or signal light.

We made a setup to be able to test ball velocity and exit angle from the robot, we should be able to use it at the scrummage to get good data from our robot. We know what the shooter wheels are doing but we don't know how much of that velocity is getting imparted on the ball and at what angle compared to our compression bar.

We have been doing wire management for the last few days and it's coming along pretty well, it should be about finished tomorrow.

The programming team made great advances. We now have the beagle bone taking data from our IMU and sending it to the cRIO. This will allow us to know our pitch, roll and yaw. The Beagle bone will also be doing the vision processing in the near future. We worked on a couple other control loops today including our drive base brake. The robot now has a setting where it will prevent the drive wheels from turning. We haven't been able to test it with another robot but we break traction with the tile floor before the wheels ever spin. This should give a nice stable platform to shoot from.

The copy bot is coming along nicely, getting it wired is going to take some time but the mechanical systems should be about finished this week

We still have some work to do on the bumpers and our body work. It's nice that we have time to figure out where our decals go before competition.

- Allen Gregory

"Truth is ever to be found in simplicity, and not in the multiplicity and confusion of things."  - Sir Isaac Newton


Sunday, February 12, 2012

Build 37: Body Work, Robot #2, and Code

Mechanically the robot is very solid. We have some small modifications but the collector and shooter assembly are working really well. We reinforced the collector motor mount and that is taking a lot of play out of the collector chains. We are using #25 chain right now but if we have any problems we can switch to #35 very quickly. We have to do a bit of CG management to pull our weight forward but that shouldn't be to difficult. We still need a simple solution for the bridge manipulator that will work reliably.

We did some body work on the robot so that we can put logos and sponsors on. and to protect certain areas from flying balls. We still need to add our diver's mesh to the sides to do ball protection. We also still have to make our decals for our logo and sponsors.

The programming team is working on getting our beagle bone to talk with the cRIO, we should have more progress on that tomorrow. We still have a lot of automation system to get running and integrated but we're confident we can get it done. If we went to competition with the code that we have now, we would be very competitive but if we get all the automation working we will be even better.

The practice robot is actually starting to take shape we have most of the structure is finshed we just need to make all the small parts and mount the mechanisms. Shouldn't take that long hopefully we can have it ready for bag day to be able to make the change.

- Allen Gregory

"Teamwork, dedication and friendship have a lot of play in all factors of life, including a robot contest." - Damian Eveland











Saturday, February 11, 2012

Build Day 36: Half Court Shot

We fixed some of the issues with the elevator by removing things, which is always a plus. The ball gather got a bit of an improvement by adding guide plates to the bottom to avoid messing up the balls. We are still having a bit of a problem with the polycord slipping out of it's grove but we now know what's causing the problem.

The beagle bone development is coming along, we have it sending data to the cRIO. We still have a long way to go on the vision system. The rest of the programming is coming along well but there are a bunch of things to do once the vision and IMU systems are up and running.

We tested the shooter improvements with flying colors, we were far more consistent this time around. We are still shooting a bit left on occasion but much less noticeable than before. We were able to set the wheel speed and make the same shot consistently.

We took the robot to half court to see what it could do on the 10 foot rim and it didn't have any problems. In fact this was our first shot, with just sort of estiating on speed and angle.



FRC#3847: Half Court Shot from Spectrum 3847 on Vimeo.


We are continuing to work on the copy bot, we have the drive base, and most of the collector bars attached. We also are making all the practice rollers for the collector and elevator. Still awhile before it is functional.

Tomorrow we will design something to push down the bridge.

- Allen Gregory

"Almost all quality improvement comes via simplification of design, manufacturing... layout, processes, and procedures." - Tom Peters

Friday, February 10, 2012

Build Day 35: More Progress

We did a lot to the robot today. We made new polycord belts to replace the yellow ones that we previously had. The new ones are clear. We mounted the new collector and elevator motors and they are working perfectly. The new mount is much more stable and we were even able to briefly tested the sensor feedback on the collection system and it's looking good.

We are continuing to make spare parts for the practice robot and we should be able to start assembly this weekend.

We also worked out some issues with the shooter velocity control. We are now getting a much more consistent velocity from the shooter motor which should make testing far better. In case anyone is interested we did this by implementing a moving mean function that removes some of the erroneous values we were getting back from the sensor.

We haven't posted a CAD picture in a while because we have not been updating it much, but we did more this week.


Now you can see the side panels that will be incorporated into the robot to prevent the balls from entering certain areas. The final product will look similar to this.

- Allen Gregory

"What happens is not as important as how you react to what happens." - Ellen Glasgow

Thursday, February 9, 2012

Build Day 34: Banebots Came In

We got our shipment of Banebots motors and gearboxes today. 5 x P60 16:1 Gearboxes and 4 x 550 motors. This will give us plenty of spares for the two that we plan to use to run our collector and elevator. We finished making the mounts for them as well but we don't have them mounted because we are planning to change out the polycord that we have been testing with and mounting them would get in the way.

We also made the new and improved shooter hood. This one looks a lot better, we'll have to wait and see if it makes any drastic improvements to the hood. You can see the new hood below. Also you can see the bridge device on the robot as well.


We did more code testing, we are getting good signals back from the shooter wheel velocity and the angle of the hood. We need to do some tuning but those look like they will be working well. We also have valid data back from our IR sensors along the collector/elevator.




We weighed the robot tonight and we are at a very happy 96.65lbs. After body work (side panels and things) we should be right at 105-110 that give us 10 lbs of weight to pull our CG a little more forward on the robot to help improve it's turning characteristics. We have a 3/8" thick plate of aluminum that was donated to us by Trident Metal. It should be easy enough for us to machine to fit where we need the weight.
.
Here is a quick video of us testing the short range of robot yesterday.


3 in a Row from Spectrum 3847 on Vimeo.

- Allen Gregory

"Hard work is the price we must pay for success. I think you can accomplish anything if you're willing to pay the price." - Vince Lombardi

Wednesday, February 8, 2012

Build Day 33: Testing the "fixes"

Our fixes didn't exactly work out as planned. Our max range decreased and our accuracy was worse but this is why we test early so we know what works and what doesn't. We found out that how we guide the ball into the shooter is far more important to consistency and velocity than we had thought. We will begin building a redesigned hood to feed the ball to our compression bar tomorrow.

We actually did some out door testing in the court yard today because our normal room was being used for another event. We took out our middle height goal and were able to shoot on it for the first time. Even our close range consistency is having problems.

We spent a lot of time working on the PID control for the Shooter velocity and for the angle control of the hood. Both of which are making progress. Getting all of the code functional is becoming a higher and higher priority as the build dead line approaches. We're still behind on the vision processing and several of the other programming features.

We also began making the motor mounts for the Banebots P60 Gearboxes that will be running our collector and our elevator, we have been using a 775 with a CIM-U-LATOR and a CIM Motor but those will finally be changed once we get our latest Banebots order.

- Allen Gregory

“Slow down and remember this: Most things make no difference. Being busy is a form of mental laziness—lazy thinking and indiscriminate action.” - Timothy Ferriss



Tuesday, February 7, 2012

Build Day 32: Fixing more stuff

We were still in the fixing stage today but tomorrow we should be ready to go test our changes.

Today we finished putting the shooter back together, we're still working out the hood design to solve the problems that we noticed over the weekend. We also spring loaded our shooter bar so that the play that is inherent in the motor is taken out by the spring. This way our shoots are more consistent. We also fixed the problem of the polycord coming out of it's groove by fixing the one misaligned tread.

We were also working on sensor mounting, the potentiometer mount that we worked out was not keeping accurate angles so we moved to a much more secure mount. Getting accurate angle reading will be crucial to making accurate automated shots. We also mounted the limit switches for our bridge device so we can easily move it up and down with out problems.

We successfully tested the IR range sensors that will be controlling our collector and elevator. We are now able to get accurate distance measurements from them and stop the collector when the ball is at the sensor. We still need to test the algorithms that will keep us from jamming and will keep track of how many balls are in the collector.

We're also working on the practice robot, we are much closer to being able to assemble it.

- Allen Gregory

"Everyone you meet is better than you at something." - David Cain

Monday, February 6, 2012

Build Day 31: Making it better

We were off yesterday for the super bowl and the second rookie build day in Houston. We had a smaller turnout but we got to provide a lot of information and to demo our robot.

Today we started on the long list of improvements we made during Saturday's testing. We have to get the shooter mechanically consistent before we can expect the controls team to make accurate shots. We're making small changes such as making sure the ball aligns properly to the shooter wheels and changing the spacing of the shooter wheels. We also rebuilt the shooter CIM-sim gearbox to make sure there are no metal shaving on the gears and to fix the airflow issue with the Fisher Price motors.

We also are working on integrating all of the sensors that are on this robot. There are sensors for every motor and as of right now non of the programs have been tested yet. Some of them are simple and should work right away other are far harder and will require a huge amount of testing and tweaking.

We also are cleaning up some of the electrical system. We're making custom PWM cables for each Jaguar so there isn't to much excess wire that we have to hide. We're doing the same with all the sensor wires.

- Allen Gregory

"Mistakes are the best way to learn. Don’t be afraid to make them. Try not to repeat the same ones too often." - Leo Babauta


Saturday, February 4, 2012

Build Day 29: Tweaking everything

We were working through the list of improvements that we saw from last night's testing. We moved the polycord guides and that seemed to fix the jamming problem. We also riveted the guide rails and that made the shooter more accurate. We mounted the bridge device motor but not the device it self, that will come this week. The shooter hood needs to be refined to allow for more consistent shooting, we are planning to spring the hood forward so that the slack of the mount is taken up by the surgical tubing or other spring system.

We think the overheating issue with the Fisher Price motors is due to the CIM-sim blocking the air intake of the motors. We have  a couple solutions in mind and we'll make those changes when we make a few other shooter improvements on Monday.

We mounted the camera to the front of the robot to mostly just to make sure the mount will be sturdy enough with all the vibrations and impacts.

We also mounted more our sensors today. The potentiometers form Digi-Key came in last night so we made the mount for that so it can measure the angle of the shooter compression bar. We also mounted our IR range finders in the elevator these will let us know where the ball is on the elevator. We are using them similar to line break sensors but they provide a bit more feedback because you can actually detect the curvature of the ball.

We also tested our robot on carpet for the first time. It drives better than it does on the gym floor but it's still not very good, we have to adjust our center of gravity to make our pivoting much smoother. We also did barrier testing for the first time on carpet and with the shooter on. The robot climbs the barrier with out issue. It was a little scary the first few times but the robot really doesn't want to tip. In fact if we go at a little higher speed the front wheels hit and the robot ends up driving on its back wheels until they hit the barrier. Completely skipping the middle wheels, that was the design but it seems to work.

We also tested ball collection and driving over balls. Collection seems to work well but we have a problem driving over balls. We haven't tested with bumpers but we may change our bumper to design to take advantage of this handy rule clarification by the GDC (http://www.chiefdelphi.com/forums/showthread.php?t=100575). Bumpers at varied heights could help with several aspects of the competition.

Here is a quick preview of the robot in action, more to follow.


FRC#3847: Bank Shot + Barrier from Spectrum 3847 on Vimeo.


- Allen Gregory

"I don't know if I practiced more than anybody, but I sure practiced enough. I still wonder if somebody -- somewhere -- was practicing more than me." - Bird, Larry

Friday, February 3, 2012

Build Day 28: Making Free Throws

We had the robot out at the gym making free throws tonight, we'll post video this weekend. The automation isn't there yet. We were allowing everyone to attempt shoots in a very open loop control mode but we made about 4 of the 40 free throws. The shooter was actually pretty consistent with it's in line accuracy we just weren't controlling speed and angle consistently. We need to make a manual control scheme that is able to set wheel speed more accurately and see how precise it is. The shooter angle adjuster was working very well for such a simple device. We also think we hit the thermal cutoff on the Fisher Price motors once during our testing. We're going to need to deal with the heat that these guys generate running at a constant high speed.

We did notice a few minor collector bugs but nothing that isn't fixable in the next few weeks. The variation in the balls is even a challenge on our collector some of them compress far easier than others and that was causing a bit of jamming. We also had one of our polycord bands hop out of it's groove but it's only when we reverse the collector. One of our treads is made a bit off center so it allows it to climb out. This should be easily resolved as well.

The drive base is having a few issues because our weight is distributed to far back in our robot. It's extremely low so we won't tip even going over the barrier but turning is a bit bumpy even with the six wheel drive configuration. We should have enough left over weight that we can play with our CG to will get the feel that we are looking for.

We made more progress on our bridge device but because the Van Door Motor uses metric bolts we weren't able to mount it tonight.

Today was also our chairman's review day. We have 13 days till the award submissions are due and this being our second year it's the first year we get to do an actual chairman's submission. The chairman's submission is interesting because you have to  decide which parts of your team to highlight.

Everything is looking good for 16 days of practice, programming and iteration with our competition robot. "Perfect practice make perfect", so we are going to need to get access to field somewhere in the Houston area so we can try this thing out on an actual court.

- Allen Gregory
www.spectrum3847.org

"If you're not making mistakes, then you're not doing anything. I'm positive that a doer makes mistakes." - John Wooden


Thursday, February 2, 2012

Build Day 27: It Works


We mounted the electrical panels and wired up the robot today. It still needs some wire care to make sure all the wires are tucked away nicely and won't cause us problem later. But for now we got it working, the collector feeds the balls up to the shooter and it's able to fire them. We aren't able to do real distance shoots until the weekend because we don't have a tall enough ceiling in the classroom. All the current system on the robot check out under manual control. We have a lot of automation to put on this thing before it will work perfectly.

We also continued work on our bridge device that will be mounted to the back of the robot. Hopefully we will have it on tomorrow.

We still are working on the bumpers they always seem to take longer than they should. Does anyone have a good source for 1000 Denier Cordura in the Houston area?

We've also began the duplication process for this robot. Most of the parts we have CADed so it's easy to make our replicas. They are going very quickly because we aren't really worried about them looking nice just as long as they function.

Programming has decided that the use of an off board processor will make our vision system much more reliable. The current plan is to use a Beaglebone running openCV to process the images from the camera and then send back the aiming information to cRIO.

- Allen Gregory

"Gentlemen, we are going to relentlessly chase perfection, knowing full well we will not catch it, because nothing is perfect. But we are going to relentlessly chase it, because in the process we will catch excellence. I am not remotely interested in just being good." - V. Lombardi

Wednesday, February 1, 2012

Build Day 26: Starting to Build Spares

We aren't finished with our robot yet, but we began the process of building the practice robot parts today. We know certain systems won't change drastically so we are beginning to manufacture the spare parts that we will use on our practice robot.

The focus for the main robot is now on the controls team both programming and electrical. The electrical will get finished tomorrow or Friday, we ran out of wire and have to get more but we will have it tomorrow. The controls team is working through vision problems and we are now considering off loading the vision processing to a separate computer on the robot. The most likely candidate is a beagleboard-xm running OpenCV. It's a far more powerful computer than the cRIO that controls the robot and will have no problem doing vision processing at fast enough frame rates for our aiming. However this adds another system that we have to program and another thing that could break on the robot.

The mechanical team is also working on designing some of the body work for the robot, to cover the shooter and other chain runs. Its also nice to have smooth surfaces for our logos and sponsors.

The bridge device started it's construction process today, it's sort of a guess and check design. We know the motor setup will work but since we still don't have a bridge we can't test the actual device, this is something that may have several iterations before we find one that works as well as we require.

- Allen Gregory

"Teamwork is the ability to work together toward a common vision. The ability to direct individual accomplishments toward organizational objectives. It is the fuel that allows common people to attain uncommon results. " - Andrew Carnegie