We spent the day gathering all the parts for the competition robot. We still don't have the sheet metal back yet but that should be Monday. We want to have everything else ready to go when we get it back, and it should go together like a big Lego set.
We also replaced the belt on our shooter with sprockets and chain. We now have 2 CIMs and 2 BB775s driving a 5.33:1 VEXpro gearbox, that has a 12 tooth sprocket driving at 26 tooth sprocket that share a shaft with 2 15 tooth sprockets that drive 26 tooth sprockets on each arm of the shooter. Our final gear ratio is right at about 20:1. With the chain we have to be much more careful not to stall the motors into a hard stop.
We also started testing different release angles by adding a bar to the launcher arm at different points. This dramatically alters our shot trajectory and can provide us a much further shot. We're going to play with this more to find the optimal position.
A few minor things as well, we got more air tanks and the real mounting brackets from Pneuaire, we mounted them on the practice bot and had to modify the mounts slightly to make them work the way we wanted. We also put on the 3D printed pot adapter for the launcher and that didn't work as well as the makeshift one we had before. We'll have redesign a bit to make sure that it's very robust. We changed the outward angle of the wings by adding a small spacer to the cylinders, which prevents them from pulling any further out.
Here is a quick video of us testing the ability to drive with two balls. I wonder why we would want to do that?
- Spectrum
“When in doubt, always err on the side of generosity!” - Ping Fu
We're sort of just waiting on the parts to come back from powder coat at this point. Lots of little things going on but not much major work. We're reworking the launcher and have added springs to make it easier on the motor to accelerate to top speed. We still have a lot of programming to do before we're sure we can make all the shots we need.
We're also making sure that we will have all the parts for the competition robot ready to go once it comes back.
We also discussed our "artist formally know as the withholding allowance". The GDC has changed the rule to 45lbs due to inclement weather around the country. We actually weren't planning to withhold that much this year, but with the new rules we probably won't be bagging our collector and launcher bar. We'll have a full drive train and tower in the bag, but we should be able to make lots of modifications to the other parts of the robot.
These are some of the 3D printed parts that we have been using this year. There will be a few more as well. We're finding lots of interesting ways to make parts with the 3D printer this year.
- Spectrum
"So often people are working hard at the wrong thing. Working on the right thing is probably more important than working hard." - Caterina
Sorry for no post yesterday, we had a really short meeting since there were parent-teacher conferences. Mostly worked on chairman's award stuff.
We'll start with a video today of us doing some catch testing. We're not really planning to catch that often but we think we're pretty decent at it. If there is defense it might be a different story, which is why we might not be catching very often.
We also received our new proper 4" mecanum rollers today. We're still not sure if we will be switching back to octocanum or not but the new rollers look a lot nicer than the ones for 6" wheels. We'll probably do some side by side tests once we have the competition robot up and driving.
Our shooter has been slowly getting worse since late last week and we weren't really sure why. We had been making some changes on the upper assembly, but no one really looked down at the first couple stages because we figured well, how could those be breaking? We were very wrong.
That is a picture of the belt that drives the first stage of the launcher. Apparently you really shouldn't ever allow your belt to ratchet since you can just rip the teeth clean off the belt backing, we now know that from experience. We moved the belt around to where the bad section wasn't engaged during launch and we got all our distance back. The ratcheting was happening when we didn't have limit switches or other sensors installed on the launcher and we were running it into a hard stop. Now that we have more sensor feedback we shouldn't have this issue anymore. We still might move to chain just to be safe.
We received our bigger bore cylinders,1.5" bore x 1" stroke, today for our drive train. Somewhere in the design of our drive train we forgot to do some basic math on the leverage needed to lift our robot and now were paying the price by having to redesign to fit the larger cylinders. Had we just moved the point on the modules that the cylinders push down, we wouldn't be in this mess. We made a pretty poor system today which will get us through some testing, but we'll have a more complete fix ready for the competition robot.
We spoke with our powder coat sponsor and they say it will be done by Friday or Monday. We're hoping for Friday but we know we can assemble the robot in the last week if we have to.
We've fixed most of the problems that cropped up over the weekend and we're happy with where the competition robot should be once we get her built.
- Spectrum
"The willingness to share does not make one charitable; it makes one free." ~Robert Brault
This morning we got a call from VEX explaining that they had intentionally shipped the incorrect rollers with the 4" mecanum wheels so that they could get the product out faster to teams. The rollers they shipped are for the 6" wheels and thus don't provide a completely smooth motion. This is why we were able to break one since 100% of the load ends up on the tips of the current rollers during part of their revolution.
We worked more on collection today. We strongly believe that collection will be more important than shooting in this year's game. Having a fast and reliable collection system alone will allow you to play in eliminations.
Here is some quick video of our robot catching bouncing balls. We're hoping to be able to chase down balls that are bouncing anywhere on the field and catch them instead of having to collect them, which should be faster.
We're still using the mecanum wheel collector. We tried to build one out of 1/2" schedule 80 PVC but the mecanum wheels out performed it by a lot. We're going to keep trying different wheels and rollers but right now the mecanum wheels are working really well.
- Spectrum
“The best creative solutions don't come from finding good answers to the questions that are presented. They come from inventing new questions.” - Seth Godin
Well today was interesting. We started the day by taking off the mecanum wheels from our drive train and replacing them with the omniwheels.
We have the robot set up so that we can remove the end caps of the chassis easily to work on certain parts of it.
We also made our collector not pivot outwards. Originally we thought the collector would need to expand to allow the ball through, but after testing we started reducing the swing more and more. Now we think we can build in enough flex and compress the ball enough that the system won't need to expand. This will eliminate a degree of freedom and make the system much simpler.
While doing this we decided to test different wheels for collection. We tried the set of the mecanum wheels that we had taken off the drive train, because we're trying to make the most of the items we have on hand. Even though they were intended for a much different purpose, they worked surprisingly well. The mecanum wheels are very grippy on the game ball and they are actually making the ball go into our robot better than some of the other wheel choices. We still have a lot of testing to do before we decide on a final roller, but mecanums are currently in the lead.
That last shot shows our backwards tipping point on this robot, so we don't think we'll have a tipping problem. The wings are still very rough, but they are doing their job.
We put the potentiometer on the launcher today and should be able to get the encoder on in the near future. The plan is to use the encoder as a velocity sensor and the potentiometer to set the release point.
We added an encoder to one side of the drive train and may add one to the other if we find a need for it.
- Spectrum
“You may encounter many defeats, but you must not be defeated. In fact, it may be necessary to encounter the defeats, so you can know who you are, what you can rise from, how you can still come out of it.” - Maya Angelou
This is a hard post to write. It's just been one of those days where nothing goes well. We had some problems yesterday but we were pretty optimistic we would be able to fix them without too much trouble. Things didn't get better today.
From bad crimps to breaking a handle on the robot to diagnosing a drive train problem for 40 minutes only to realize it was a dead battery, nothing went our way today. We would try to improve things only to make them worse. The current state of just about every subsystem is markedly worse then it was 36 hours ago.
The biggest problem of the night came at around midnight when we were testing our latest change to the collection system and getting ready to test driver centric control. We were driving along fine, we strafed right and then when we went to strafe back left we saw the robot sort of dip and then just completely stop moving. We strafed back and it sort of worked but in a big arc. We thought we might have unplugged a speed controller or gotten something jammed in a gear. When we turned the robot on it's back to see the drive train we discovered that we had broken one of the rollers clean off our VEXpro 4" mecanum wheel.
We haven't done much to these wheels at all, we have only had them for 6 days and have driven less then an hour and half on them on FRC carpet only. They haven't been put up against other robots or had any pushing tests done with them, we have just driven around and tested collection for the most part. After looking at the broken wheel we found that the roller didn't spin at all, we checked all the other wheels and have found at least five other rollers that either don't spin or are very tight.
The six of us that were still at the shop at midnight discussed it and unless there is some very strong reasoning by the rest of the team to stay with mecanums, we will be switching to butterfly for this year's drive train. If this were to happen during a match we wouldn't be nearly as effective and the ability to strafe isn't worth the risk or breaking wheels and having to make repairs.
To top off the day of everything going wrong, on our way down from the third floor where our practice area is to our lab we ended up on the first floor. Our lab is on the 2nd floor, literally nothing went right today.
Tomorrow we'll change the wheels back to omniwheels, and we'll halt development on holonomic control and sensors. This may turn out to be a good thing, without worrying about mecanum wheels and holonomic control we'll have more time to focus on other systems. That's at least how we're going to spin it.
15 days left to get it right, tomorrow should be better since we get to watch the Superbowl as a team and have some fun.
- Spectrum
"Sometimes adversity is what you need to face in order to become successful." - Zig Ziglar
We started the day off with a visit from one of our local FLL teams, from The St. Thomas More School. These students were all 5th through 7th grades and they were all very interested in the robot. Surprisingly some of the questions took a very technical turn and we started talking about the difference between an FPGA and a general purpose CPU. One of the students even told us about his favorite Linux distribution and was pleased to hear that we would be switching to developing in Eclipse next year.
Here is some video of the robot collecting and shooting for the students. Our practice area only has 10ft ceilings so we couldn't shoot very hard but you can get an idea of what the competition robot will hopefully be able to do.
After the demo we started investigating the problem we found yesterday with the pneumatics on our octocanum modules. We believe we didn't account for the lever action of the module and the amount of force we put on each cylinder. Our plan is to move to 1.5" bore cylinders but that will take a bit of rework on the mounting, though not too much. Luckily we have a working robot this early so we can catch major and minor problems before we get to competition.
We worked on adding our wings to the robot, which will help us secure the ball while driving, prevent it from slipping out while collecting, and allow us to catch passes. The current wings are made of hardboard but we are planning to move to corrugated polycarbonate. The corrugated polycarbonate was suggested to us by our friends James Tonthat and Scott Rippetoe from Texas Torque FRC#1477. It's apparently used very often in greenhouses but you can get it at a variety of plastic distributors. It's rigid like 1/4" Lexan but much lighter.
We are working on an extremely cool way to hook up an IMU to the cRIO to get an accurate heading. When it's all said and done it should only take a single PWM cable hooked up to a digital input pin on the cRIO. We'll have more on this later.
The last thing we did for the day was work on improving our collector more. We found that it had a lot of wobble because the CAD model had the wrong size holes for the bushing we were going to use. We ended up drilling out the holes for bearings and then packing them with a little bit of electrical tape to stop the wiggle. This fix will not be on the real robot but sometimes you have to do things like that to get your prototypes testable. There is still a very long list of variables that we don't have solved for our collector, here is just a sample.
Wheel type & diameter (we might move to a roller system instead of wheels)
Number of wheels and spacing
Speed
Angle to deploy down
Angle to hard stop the rotation in and out
There are plenty more too, all of this will get tested and iterated until we find the setup that collects best. We will probably keep changing some of these throughout the entire season.
Week 4 has come and gone and we're left with 17 days to build the competition robot and hopefully find answers to many of the variables in our robot.
- Spectrum
"An expert is a person who has made all the mistakes that can be made in a very narrow field." - Niels Bohr
