Killer Bees - 2014

[z_beeblebrox]

Originally Posted by apalrd

About the drivetrain:
We initially used Vex encoders with Vex 2.75" omni wheels as floating odometry wheels, but quickly removed them for weight savings and ran time-based autonomous programs (we did do voltage compensation to try to keep it reasonably consistent). There are no sensors in the drivetrain.

Omni wheels are for position measurement.

[yash101]

Originally Posted by z_beeblebrox

Omni wheels are for position measurement.

I believe you can use omni wheels to create quite creative drivetrains. They are not solely for position measurements. You can make an omni drive, a kiwi drive, and possibly even a tank drive capable of drifting!

Jim posted a picture, showing how omni wheels were used for the drivetrain, not just odometry!

[tim-tim]

Originally Posted by AlecMataloni

So, what's the standard bend radius on the .050 sheet you guys use? I've read that 6061-T6 bends poorly/ cracks below a bend radius of double the material thickness. Is this the rule of thumb you follow?

I can't speak for 33, but the typical rule of thumb is twice the thickness for 6061-T6. We don't use much .050" or 6061-T6 on our robot, but at work we typically just use a 1/8" bend radius or 3/32" (.094") if we are in a pinch.

[Phyrxes]

Interesting read, thank you for posting. This was our first year using a sheetmetal frame and our sponsor suggested 5052 sheet over 6061 as thats what they use for most of their paying customers and can generally make our parts of of scrap. Any insight on what alloy properties to look for in this application?

[cadandcookies]

Originally Posted by yash101

I believe you can use omni wheels to create quite creative drivetrains. They are not solely for position measurements. You can make an omni drive, a kiwi drive, and possibly even a tank drive capable of drifting!

Jim posted a picture, showing how omni wheels were used for the drivetrain, not just odometry!

I think he was referring to the question about the purpose of the green omnis in the middle of said picture, which do not appear to be driven.

Another fantastic robot from 33. I remember seeing your 2012 robot at Championships and being immediately inspired to push my team to create more aesthetically pleasing robots. Your robots have been excellent examples of both form and function in a killer package.

[apalrd]

The green wheels were intended for odometry but were later removed. The gray omni wheels are the drive wheels.

We use 6061 T6 as it is a harder than 5052, in FRC this translates into the need for a thinner sheet for the same overall stiffness. In the thicknesses we use (0.050" and 0.063") it does not crack when bent properly, but we have issues bending anything thicker (we have successfully bent a few 0.090" pieces, but also cracked several). 5052 is usually used by sheet metal shops as it is easier to work with, many teams use 5052 successfully in FRC (some on CD will say it's impossible to bend 6061, that is not true). I have only ever seen 6061 and 5052 used for FRC sheet metal applications.


We can in fact be pushed, we intentionally do not resist and instead spin out. That was part of the strategy.

[DampRobot]

Any details on how you implemented your catapult? I'm interested in how you packaged the springs in what looked to be a really tiny space, and how you designed for an optimized shot/designed for adjustability of the shot.

[apalrd]

Originally Posted by DampRobot

Any details on how you implemented your catapult? I'm interested in how you packaged the springs in what looked to be a really tiny space, and how you designed for an optimized shot/designed for adjustability of the shot.

I attached some pictures and here is a description.

The moving portion of the catapult itself is a right triangle (or very close, it's a 95 degree corner). One leg spans the length of the robot, and has several applied pieces which hold the ball, a cross bar for the end stop strap (a seatbelt strap), and a cross bar for the pull down strap (a high strength rope). The other leg has mounts for the springs, the hypotenuse is a stiffening member, and the axle is located on the right angle. There are two welded triangles spaced 2" apart with the pull down and end stop strap axles in between, as well as structural pieces.

The fixed portion makes a 'Z' (really an 'N') with a vertical member on the front and rear of the machine connected by a diagonal member, with additional stiffening members triangulating the corners.

There are 6 springs. 2 are 2" diameter and 4 are 1" diameter. The 2" springs have a 1" spring inside of them. All of the springs mount to the ends of the moving portion. 2 of the 1" springs are mounted directly to the fixed portion, the 2 2" and 2 1" springs inside are mounted to a 'banana bar' which is adjusted with a lead screw. The lead screw is driven via a CIM and Versaplanetary 3:1.

The Choo Choo is remote mounted via the orange cable. We learned that we could play with the choo choo geometry to flatten the motor torque as the catapult is pulled in (as the spring force will increase with pull in distance, but we can also decrease the lever arm of the second hump to equalize the torque). The choo choo is driven via a Dewalt gearbox and then 7.2:1 in additional gear and chain reduction (total of 86.4:1). We did this instead of a rachet as we wanted to be able to reverse the choo choo, but this did not end up working. The dewalt gearbox was still handy for packaging, but it's a giant PITA to make and I don't recommend it at all.

As for adjustability, we have 3 primary points of adjustment: end position, start position, and spring length.
End Position: There is a black seatbelt strap which controls the end stop. At the fixed end it is wrapped through a hex shaft which is secured from rotating by a wrench-looking piece (it is visible in the 3rd picture). By removing the wrench-looking piece and using a pair of 1/2" wrenches we can adjust the end position under load.

Start position: The eye bolt which secures the choo choo cable to the catapult arm can be adjusted to set the start position. We do this to calibrate the dwell time, among other things.

Spring length: This is directly adjusted via the spring adjuster lead screw. We can also adjust the 2 fixed springs via their eye bolts, and have several springs which we can swap for major adjustments.


http://imgur.com/a/iQznZ

[BJC]

Originally Posted by DampRobot

Any details on how you implemented your catapult? I'm interested in how you packaged the springs in what looked to be a really tiny space, and how you designed for an optimized shot/designed for adjustability of the shot.

I don't have any pictures but I will do my best to explain.

We put the pivot point of the throwing arm all the way at the front of the robot. In order to have a reaction point for extension springs (we looked into others but ultimately were most comfortable with extension) the arm is actually an L shape. There is about 600lbs of springs attached to a 9" lever arm. Their were 6 total springs. Two were 2" OD and the other four were 1" OD. To save space we actually put two of the smaller springs down the center of the two larger. Those four springs were attached to a fork towards the back of the robot. One side of the fork is fixed to the robot while the other side could be driven up and down by a lead screw to finely adjust spring tension. The remaining two springs were attached statically to the robot. In order to pull back the catapult we packaged a choo choo mechanism in the back center of the robot. The second link was a loss link made out of dyneema cord that went forward around a pulley then up attaching to the throwing arm. A seat belt strap could be finely adjusted to change the exit angle of the catapult.

We had three co-dependant variables when tuning the shot:
-spring tension (lead screw)
-angle of thrower when choo choo overcentered (dyneema string length)
-exit angle of thrower (seat belt length)

There were a number of things that I think were key to our catapult's success. The first is the enormous amount of potential energy it had stored combined with an incredibly short change in angle. A very short dwell time in the robot meant that other physics - such as the robot moving, being bumped, or hit - did not meaningfully affect the shot. By the end of the season we were also actually shooting through our ball hoop. This meant that the ball really couldn't move around at all within the robot even while firing. Other small things like using springs instead of surgical tubing meant that the shot never changed once we dialed it in even over multiple competitions.

Glad you like the robot, hopefully Jim or someone will pop in with thrower pictures to better explain.

Cheers, Bryan


Edit: Beaten by Andrew -- not fast enough on the draw

[DampRobot]

Originally Posted by apalrd

I attached some pictures and here is a description.

...

http://imgur.com/a/iQznZ

Originally Posted by BJC

I don't have any pictures but I will do my best to explain.

...

Cheers, Bryan

Thank you so much guys! I really love reading about your engineering, and I'm really grateful you guys took the time to describe it in such detail. I know I'm not the first one to tell you this, but the omnis and the incredible shot on Buzz really made this one of the best robots in the world. Thanks again for sharing.

[yash101]

Originally Posted by cadandcookies

I think he was referring to the question about the purpose of the green omnis in the middle of said picture, which do not appear to be driven.

Another fantastic robot from 33. I remember seeing your 2012 robot at Championships and being immediately inspired to push my team to create more aesthetically pleasing robots. Your robots have been excellent examples of both form and function in a killer package.

I figured, but didn't want anyone else to get confused like me .

That was a great robot, but it's pickup didn't seem as powerful when I saw some match videos!

[Abhishek R]

Originally Posted by yash101

I figured, but didn't want anyone else to get confused like me .

That was a great robot, but it's pickup didn't seem as powerful when I saw some match videos!

Pickup seemed fine to me... What about it didn't seem "powerful?"

[yash101]

Well, it looked like it was struggling a little to pick up the ball. However, I shouldn't dock off for this as it was better than ours . However, I would post that robot in the "Scary Robots" thread because it's quite a beast!

[mman1506]

Did you ever experiment using a brake plate on your robot to prevent the robot from strafing while your in shooting position?

[apalrd]

Originally Posted by mman1506

Did you ever experiment using a brake plate on your robot to prevent the robot from strafing while your in shooting position?

No.

Strafing while in the shooting position is irrelevant, the goal is the whole width of the field.