Touch it own it is a phrase that’s tossed around a lot, but a lot (most) teams struggle with designing something that accomplishes that, especially in V1 or V2 of their intakes. What makes an intake worthy of that that title, what are some good examples from past years, and how can teams work to design better intakes earlier?
What makes it worthy? I don’t know if there’s an objective measurement. But acquisition under crappy driving and retention under things like hard reverses and impacts sure sound like ownership to me.
Teams like Spectrum, Simbotics, and Cheesy Poofs have blogs and videos documenting the evolution of some of their intakes, which I hope someone will link (I don’t have the time to dig them up right now).
Better intakes earlier? One part stockpile, one part fast generation of mounts for things. Do you have hex shaft? Wheels (both compliant and regular)? Bearings? Motors to drive them, whether FRC-legal or just a drill that can chuck up the shaft? A few pneumatic cylinder options? Plywood and 2x4s? Surgical tubing? Something for standoffs, whether it’s PVC or threaded things or churro or peanut? Maybe even some FTC-or-VEX-scale stuff, for some games? Things like that let you determine what works and what geometry, so you can transfer that to the V1 intake. (Or if you consider the plywood to be V1, then V2.)
When I hear “Touch it own it,” it is usually hand in hand with describing a “roll-ey” or wheeled intake. I see it as basically any intake that pulls in the game piece without having to perfectly line up or driving into the piece. It also needs to hold onto the game piece after it has been collected even if defended against.
One thing that is important for this type of intake is that there has to be compliance somewhere in the system. If the game piece is hard, like a gear or cube, then the intake has to have compliance which is why you see many compliant wheels in games with hard game pieces. If the game piece is more compliant, like boulders or inner tubes, the intake is usually more rigid and relies on the game piece to comply to come in.
Pretty much all of the top teams are going to have this type of intake, so it is usually a good idea to look at games with similar game pieces in the past and see what those teams did.
IMO the question here is flawed. Don’t worry about getting intake v1 or v2 to work perfectly. Instead, iterate quickly from v1 to v2 to v3 to v15. Many small, quick iterative improvements will result in better “convergence to the best design” than a few large, slow improvements.
That being said, having a good stockpile of supplies, a thorough understanding of popular FRC intakes, and an experienced build/design/CAD team goes a long way. We also improved our design process last year by using our school’s hobby laser cutter to cut thin plywood sheets and hobby 3D printers to make parts for prototypes. We will hopefully be using them more this upcoming season, as well as using our new CNC router for prototype development.
I’ve found that a good intake has 3 things: compliance, power, and grip.
Compliance can come from either the game piece itself (such as in 2014, 2016, and arguably for 2017’s fuel balls) or from your intake assembly. Ideally, something in the intake stackup should be deflecting by about 1/8" to 1/2". I’ve found that increasing the “degrees of compliance” will improve performance only up to 2 degrees. For intakes in 2018, those two degrees would be compliant wheels and pivoting intake arms. You get diminishing returns after 2 degrees of compliance. Additionally, compliance is more important when you’re running the intake with low speed or power.
Speaking of power, power is relatively simple. As a rule of thumb, I live by “just throw 2 775pros at it.” What I mean by powered is that you are constantly able to keep your intake going, preventing the driver from manually timing any sort of grabbing action. Traditionally I call this a continuous intake, as opposed to a discrete intake (such as a single-action claw) or a passive intake (such as a box). The primary benefit of having more power is that you can collect more game pieces faster (such as in 2017) and it takes less precision to shove the game piece into your hopper/elevator/other delivery mechanism. Some teams shoot for 2x the drivetrain’s speed, but I think this only matters in the many-game-pieces case. In 2018 we ran our wheels at a 35:1 reduction and had no issues with that, and some of the teams we saw had too much speed in that their wheels would kick the cubes away.
Grip can come for free with compliance, but it’s something that’s typically forgotten when you should keep an eye on it. This is especially true for 2018, when the intake also served as the holding and release mechanism for most robots. You should have positive control over the game piece at all times during the intake process. All of your powered contact surfaces should be covered in something with high friction (Roughtop and neoprene seem to work pretty well for this) while the unpowered (i.e. not moving with the game piece) parts should be relatively slippery (bare aluminum does work here). Additionally, you probably shouldn’t just rely on the clamping force coming from just your compliance. I saw many teams this year get cube-sniped or just straight up drop the cubes because they weren’t able to grip tight enough. Pneumatics make this much easier.
Lastly: iterate, iterate, iterate. We went through 13 iterations on our intake this year, and I think that was actually a pretty small number. Our turnaround cycle was one iteration every 2 days during the first half of build season (though this slowed down as iterations got more detailed/complex in their designs).
The thing is, everyone struggles with the V1 or V2 version of their intake. You have to make several revisions, at full scale, with parts and operating conditions as close to real world as possible. The secret is failing faster and prototyping, fundamentally.
" “Try again. Fail again. Fail better.” -Samuel Beckett " -JVN
Basically, this depends on how easy it is to “push” or “kick” a game piece away from your robot. With balls, it tends to be pretty darn easy to do so. Balls will roll freely away from your robot if the intake doesn’t have a net velocity that will pull them into your robot (ie the intake isn’t spinning faster than you are driving). Other game pieces may have much more resistance on carpet, such as frisbees (2013), totes (2015), gears (2017), and powercubes (2018). With objects that don’t roll freely, you have to overcome static friction between the game piece and the carpet* in order to “kick” it away. This may allow you to run an intake at a (relatively) lower speed and still acquire game pieces. That’s not to say you won’t kick away these game pieces, but it gives you a different target velocity for your intake.
*This is also evident by watching teams attempting to intake cubes on the HDPE platforms, and how much more “slippery” those cubes became
You also have to consider the situations in which you are acquiring game pieces. Some games you are “sweeping” up game pieces as you drive across the field for various situations (think of 2017 and teams vacuuming up fuel as they drove back and forth). Other games you are racing to a game piece somewhere on the floor (think 2016 and driving to boulders on the center line). Other games you are picking up game pieces from loading areas or along walls (think of cubes along fences or from the portals in 2018). In the first two examples, being able to pick up objects at full speed (or at least typical driving speed) is highly important, and likely means you need to run your intake faster. In the latter example, your robot will likely be moving at much lower speeds when interacting near field boundaries (plus theres a field boundary to limit the motion of the game piece), so having a high speed intake is less important.
Frankly im still lost on how they came up with theirs. The build blog just kinda goes “oh yeah we built this over the weekend” with nothing about how they came up with that intake.
Have you ever built a roller intake that you really liked? If not, I can imagine the knowledge gap. If so, then hopefully much of their documentation might make you go “hmm”.
There really aren’t THAT many different styles of intakes in FRC. A wheeled intake (of some form) works pretty well a lot of times. Building one and tuning the various parameters is a good way to get to the magical “touch it, own it”.
Generally speaking, you care about a few things:
- Intake speed (really, the speed of the surface of what is touching the game piece). This can generally be less if you are reacting against the object on two opposing sides.
- Wheel type/material - How grippy, how compliant (see next note though), how large of a diameter (affects surface speed, but also acquisition area)
- How compliant - You need some compliance in the SYSTEM, as the combination of the game piece and intake. For relatively non-compliant game pieces like totes, game cubes, as well as things where the opening size has to change to account for a rotating, non-spherical object, you usually want more compliance in the intake. For somewhat compliant balls, the intake mechanism can be much stiffer.
- Acquisition area - How “aligned” do I have to be to intake the object. Various designs are better/worse at this. Out of scope of this response
- Type (bonus thought) - A roller that I can turn on, drive at the object, and when i touch it I own it, intake it, etc with no driver actions… versus having to drive up to the object and activate a “pincher” to acquire it.
All that said, the people that get to the designs more quickly just know that these variable are all important, and make reasonable guesses so V1 is closer to being correct. They still have to tune it though to make it better.
Never saw that was confirmed, alright.
Lots of good info, thanks!
I think 1360’s intake from this year embodied this concept really well. We were one of the few, if not only, teams that intook the cube diagonally, which was a strategy we had early on, because that meant we would not have to reorient the cube ever. This simple concept, along with saying we want to intake it when its on the 13" side also set a few design constraints for us, and we just went into solidworks and played with geometry until we got a cool idea we wanted to try. Lucky for us, our first idea was great, it worked nearly perfectly, so we made a polycarb one by week 2, and kept it all the way through to worlds.
I think its not really about being the most creative or experienced designer, its about setting your design constraints and just making simple sketches of what fit those constraints. If you noticed, 254’s strategy early on is to always set their design goals for their robot, and it usualy is what sets the general shape and geometry of the robot right off the bat, even if they arent in the design stage yet.
Heres our reveal which shows off the intake: https://www.youtube.com/watch?v=UbJgmk7SWtg
We watched a lot of week 1 and week 2 events this year and noticed that a lot of robots were having trouble intaking cubes. It seemed like the cubes would skitter away from them rather than being pulled in. The cubes also seemed to fall out fairly easily. We joked that it looked like teams were playing with ice cubes rather than power cubes.
One of our students noted that it appeared that the intakes were running too fast and the high surface speed of the wheel would just skid on the surface of the power cube and push the cube sideways rather than gripping the surface to suck it in. He suggested we slow down our intake to try to use grip rather than momentum transfer to suck in the cube. we changed our gear ratio on the bag motors running the intake wheels to 45:1, I think it was.
It worked great. The cube would suck in smoothly and positively. If only one wheel contacted the cube, it would push the cube sideways toward the other wheel slowly enough that the other wheel would grip it and pull it in.
The bonus was that with the high gear ratio, the force required to backdrive the motor was quite high, so the intake would hold the cube solidly without needing to run the intake motors.
“Touch it to own it” should include what happens to the game piece after it is acquired since that affects your robots ability to score the game piece accurately and consistently. Examples from the 2014 season are from 148 and 1114. Part way through the season, 148 added a “halo” to retain the yoga ball that kept it from rocking in the robot so it was in a more consistent location relative to their catapult. The team I was with that year had a robot where we could see the ball “sloshing” around, front to back and side to side by a few inches after stopping the robot. This degraded the consistency of the shot significantly compared to when the robot had not been moving before taking the shot. In 1114’s design, once the yoga ball is pulled in, it is prevented from moving relative to the puncher mechanism by the upper and lower arms and the side arms.
In 2015, the team I was with got a waitlist ticket to St. Louis so my son and I were able to look at a lot of outstanding robot designs and ask the teams how and why they chose those design elements. The head mentor for one team used the phrase “obsessive control of the game piece” after it was acquired. They had mechanical guides that kept the totes from moving side to side. The totes were always pulled in against a hard stop. This made it much possible to build stable stacks very quickly, autonomously, without driver intervention. The new tote was always in a consistent position relative to the stack already being held in their elevator so they would line up perfectly each time. They didn’t drop the stack on the raised lip of the new tote like our robot would sometimes do. We had to spend precious time to try to fix a bad stack. I suspect this design philosophy had something to do with how well 2481 did in the 2016 season.
While it is possible to iterate and refine a design to improve how well a robot “touches and owns” a game piece, there is greater benefit to put thought into how one achieves it during the brainstorming stage. In our 2014 robot, we tried to add a halo like the one 148 used but it was ineffective because our robot did not have a good place to install such a mechanism and it would have required major surgery to make a good place for mounting it.
All the suggestions here have been good and I fully agree with them. One thing that I have often seen as a characteristic of a “touch it own it” design is speed. As many have mentioned I typically think of “touch it own it” as being some type of roller design and I push my team to spin said rollers 3-4 times as fast as they initially think they should. It is really easy to slow it down if needed through programming or minor mechanical change but speeding things up can often be much harder.
In 2016, 987 had a combination of bost vertical and horizontal intake wheels to ensure the caputure if the ball. We realized that the horizontal, high speed wheels were really good at obtaining the ball and forcefully putting it in the correct position, but if alighnment was off too much, it would kick the ball away, similar to the observastions 1533 made about the power cubes this year. To combat this, we had a second roller that reached past our first set that was rotating vertically. This allowed us to grab a ball from any position in front of us, and rather than getting kicked away when out of alignment, the combination of the horizontal and vertical spinning wheels would quickly center the ball in our shooter.