Are those custom Falcon500 Shafts in the drive gearbox? if so what material did you make them out of??
Hello, how do you shoot 11 power cells in the autonomous period , can you provide this part of code?Thanks!
Strategy and Architecture
Our climber had the unique challenge of not only needing to reach the bar, but also need to retract so far that we could fit as tall of a robot below us as possible for the fork system, and then retracting a bit further to lift that robot on our forks off the ground.
It is important to note this since Reach and Stroke are two different metrics. For example, you can easily get more Reach by having your hook flip up, but once flipped up it won’t retract, so you get Reach without Stroke. In our case we needed both Reach and Stroke so that pushed us into having a 3-moving-stage climber.
Trying to package the climber into such a small area, to maximize the realestate on the robot for the Serializer meant than instead of normal tube arrangement with 1/2" difference between each size, we tried for only 1/4". This meant we could not extend with constant force springs, fit cabling between each stage, or use bearings to serve as sliders.
The Climber is similar to the Forks described above. It consists of the same 2", 1.75", 1.5", and 1.25" 1/16" wall square tubes with 1/16" adhesive-backed-teflon serving as the sliders between each stage.
Early architecture layout sketch shows the extended climber reaching the hanging bar with the forks dropped to accommodate another 28" robot underneath us.
Note the climber extends so high so that it can reach the far tip of a bar that is tilted away from us (strategic advantage to being able to buzzer-beater climb at the end, maybe less relevant in hindsight). Disregard the feeder and intake roller/4-bar placements pictured there.
To extend, we use a combination of a pneumatic cylinder and compression springs. The 0/75" Bore, 18" stroke cylinder lifts the first (1.75") stage along with the 1.5" and 1.25" stages. Then the compression springs inside push up the 1.5" and 1.25" stages off of the 1.75" stage.
The compression springs are 1" OD, 0.816" ID and come from McMaster at 36" long but I think we cut them slightly shorter to like 30". We have 2 springs that we twist/interleave into each other inside of each climber. This got us enough stroke (need to extend 2 stages worth which is 2*18.375 = 36.75" while also still being compressed and having enough force to hold up the climber at the end of that extension.
One problem we face was with the springs buckling and getting all twisted up. When the climber is extended, the bottom of the spring is not in the 1.125" square hole of the inner stage but rather the 1.875" hole of the outermost base stage. To prevent buckling, we put 3/4" and 5/8" OD round tubes that start nested inside of each other and then as the climber extended the smaller tube would move up with it and help prevent the spring from buckling up top. The anti-buckling tubes, were critical to constraining the compression springs so that this design would actually work.
Overall climber cross-section in collapse state.
Close up cross section of bottom of climber showing the Spring anti-buckling tubes.
The climber retracts via a 1/8" diameter dyneema cord that is wrapped around a 0.6875" diameter spool that is driven by the 1/2" hex shaft output from the drivegearbox. It connects to the climber by being wrapped and knotted around a standoff at the top of the innermost tube.
Close up of spool on and pneumatic cylinder mount on bottom of climber
Close up of top of climber showing end-plug in top of 1.25" tube for string to tie off and attachment for pneumatic cylinder
Control Panel Mechanism / Wheel of Fortune (WOF)
Named after the Wheel of Fortune game show, our Control Panel Mechanism consists of a Falcon 500 that drives a 16:18T belt reduction that via a 3/8" hex shaft inside the structure drives a 18:54T belt reduction that has a 3" gray flex wheel that spins the side of the Control Panel.
The structure consists of a 1x1x1/16" tube with Markforged endcaps to hold the pivot/motor and the wheel/color sensors.
The arm needs to be low and flat so that the shooter can shoot over it.
The arm is lifted by a 9/16" stroke, 4" stroke cylinder. An additional steel tensile cable (not modeled but visible in the robot photo) connects the cylinder-foot-block and tube to act as a safety tensile member and hopefully prevent the cylinder from breaking if we drive too hard into the control panel.
Cross section shows the 3/8" hex shaft inside the tube and the pulley powertrain. The wires for the color sensors ran inside this tube (taped against the side of the wall far from the shaft). The shaft is held by 3/8" thunderhex bearings that are pressed into the printed endcaps.
Knowing that we still weren’t confident in our ability to build a reliable and well-driven swerve drive, we were left with making our usual West Coast drive. Without a swerve, that pushed us away from the front-court “death cycler” and towards driving across the field, cycling from the feeder station.
This strategic decision made us want to have an extremely powerful drivetrain that would be unbeatable and just hauling butt across the field and crossing the bumps in the middle with no problem, motivating us to opt for 6" pneumatic tires.
We also never wanted to be current limited, so having 6 Falcons gave us plenty of confidence that we could drive as fast and push as hard as we wanted while pulling minimal current per motor and thus not being limited by the per-motor breakers. The overall breaker and battery would obviously still be limits, but with datasheets and testing showing it can spike above that current for brief periods during short pushes/accelerations, we weren’t that worried.
So the stage was set for a 3 Falcon gearbox, which by itself is already large/complex. On top of that. we wanted a 2-speed DOG shifter to give us Low Gear so that we could have even more torque (you’d be surprised how what you think is a traction-limited robot can suddenly draw more current in pushing scenarios when another robot gets slightly on top of yours and gives you much more downforce/friction/traction).
On top of the 2-speed shifter, we did a quick motor count of our subsystems and determined that if we wanted 2 motors for the Climber (we were planning to be lifting both ourselves and a partner) that those motors would have to be PTOd off the drivegearbox. Thus we now have a second cylinder for PTO on the drivegearbox, shifting power off one of the motors to instead go into the climber.
The packaging of the PTO is really cleave in my opinion. Instead of being like our 2013 more-traditional PTO which just took the already spinning gearbox and allowed it to also output to other gears to drive the climber, this PTO take the 3rd motor’s power and directs it away from the drivetrain and into the climber. This keeps the other 2 drive motors isolated from the climber, allowing us to both control drive and climber at the same time. In the 2013 climber, for example, if you wanted to extend the climber you also have to be driving the wheels.
This required machining a custom, extra-long, extension rod (the piece that goes from the cylinder to the push the roll pin that push the dog itself.
Lastly, we needed some sort of engageable brake to stay up at the end of the match. Our options were either an engageable ratchet/pawl like we did on 2019 Elevator/Arm, a Friction Brake like in our 2018 Elevator, or in our case we opted for an engageable dog brake because of packaging we couldn’t get the friction brake early enough in the geartrain to have adequate torque to resist the weight of 2-robots, and the Dog packaged more cleanly than the ratchet.
All of these dog shifting shafts and components we purchased from West Coast Products (PNs 217-3635 and 217-3628).
Gearbox Plate Design / Layout
I’m surprised I didn’t post this earlier when posting the Technical Binder, but I also wanted to post the Presentation our students created for the 2021 Infinite Recharge at Home judges presentation.
I think they did a really good job, there’s a lot of details about the parts comprising each subsystem.
Unfortunately, we do not have a recording of the students presenting it, so this is now mostly just a visual aid.
Thanks for sharing all the info, Torrance! My favorite part is seeing all the geometry layout sketches y’all do and how you design a robot with sketches. It’s something I’ve been placing a lot of importance on with my students and I like seeing how other teams do it.
Caveat: I’m not associated with the team. But I do like to stare at their code a lot.
2020/2021 code isn’t realeased on their github page yet.
From the 2019 codebase, I suspect the answer to your question is going to be a very long and involved one, involving much more than the source code alone.
Some of the core components are here, if you are interested to familiarize yourself with some of the concepts:
- What compressor did you use? The chassis image does not look like the Viair 90C.
- I see the intake prototype page of the presentation shows a variable compression prototype. how do you make it variable.
Where do you buy 1.75” and 1.25” X 1/16” wall square tubing? What alloy is it?
Our compressor we have been using for the past years has been the Thomas 215 Series WOB-L
The “variable” compression there is just the fact that we used flex wheels on a smaller shaft, instead of having thin rubber sleeves over hard tubes. This allowed both the ball and the flex wheels to compress, meaning the compression could vary along the path of the ball helping it get up and over the bumper more smoothly/quickly.
I don’t see 1.75” or 1.25” X 1/16” on their website. I have not been able to find a source for these sizes in 6061 alloy.
If you’re in California, call them. If you’re outside their delivery radius, look for a similar outfit near you and call them. If they tell you no, ask how much business you need to bring in for them to look into it. Many of the high profile west coast teams have two decades of yearly orders with Coast, like anyone else they’ll look into things if they know you.
Online catalogs are where options start, and in this case represents what Coast can normally get next day.
I just reopened the presentation and realized I misremembered and incorrectly described the variable compression intake prototype. The other one called the Flex Wheel prototype is the one that had flex wheels.
For the Variable Compression Intake pictured there, the compression is not varied while intaking but rather by adjusting the rollers’ position through assembly. We have the rollers on bearings that are held on an additional plate that be slid around and clamped in a new position (look on left side of lowest roller). The Versaplanetary driving the roller via a belt also has adjustability so we can keep the belt tensioned.
We have been making intake prototypes of this style for years. The ability to quickly adjust the roller positions without having to remake the entire assembly is really good for fine tuning a prototype. You’d be surprised how much a 1/2" change in compression can affect how far a ball flies high and over your robot, versus just coming in perfectly and landing right into the Serializer/feeder/hopper.
Yes, this. They can get a lot of things they don’t list in their catalog. 6063 will usually be available in 1/16 wall as well (but be aware its significantly weaker than 6061).
I’m not familiar with vendors in your neck of the woods, but I know EMJ Is nationwide and likely would be able to get it for you.
We recently bought some 1/16" wall 6061 tubing in those sizes from mcmaster. Tolerances were much better than the ±0.025" spec.
Patrick brings up a good point. We also bought the 1.75" and 1.25’ in 1/16" wall from McMaster. Coast Aluminum didn’t have those sizes., what I stated above is incorrect.
We do try to use Coast for our other, more common, stock since it is cheaper and can be delivered in 20ft lengths to our shop.
We use Coast too! They’ve been really great suppliers with excellent prices.