We didn’t account for the state of the balls at our first competition this week. Not realizing that the balls would essentially have extra airholes and take a lot less energy to compress down. This ruined our ability to reliably full field shoot… only hit a few shots.
Can anyone speak from expierence n if they were able to sustain the ability to do it, or if a rework helped? We had floated the idea of trying to not rely on ball compression at all to shoot (although I am not sure how) or compressing the ball even more.
Did anyone in week 2 continue to full field shoot? I feel like a goofball for not seeing the holes in the balls membrane becoming an issue. I am curious if there were teams out there still able to do it (with balls that have seen some action). I saw a team shooting full field in my region week 1, but I would suspect they might have some difficulty playing week 2. Although I guess the balls were pretty chopped up by the end of week 1.
We were able to get about 40% accurate shooting full field. We decided it would be worth the extra 3 seconds of driving to just shoot on the closer side of the trench instead. I would love to see another team manage a high accuracy behind the control panel though.
ah. Looks like they had some aiming trouble but no issue with power! did you happen to see how they were shooting?
Were the balls as bad week 1? Our shooter never cuts up the balls… so we never saw issues at our facility. I wouldn’t be shocked if its really only a few teams that really damage the balls… and by week 2 most of the balls had seen action from those teams. Or the damage came from driving over them.
we unfortunately designed our robot higher than the trench (because we thought we’d be able to shoot full field). How are you guys shooting currently? I would take 40% accuracy at this point…
At WNE week 4 (if there is a week 4) we plan on trying full court shooting again. To compensate for lightly damaged balls were going to try increasing the flywheel speed while also increasing the hood angle so the balls have less of an arc.
Less arc, more speed makes total sense, and is also what I would do if we were still pursuing the long shot. The other thing I would try is to lengthen the acceleration path of the ball (IE: more rollers before the “main” flywheel) to decrease the final variation in exit velocity.
We have a 45-inch robot so we can’t go under the trench. In order to get accurate shots at that distance we had to add a lot of time between shots to ensure the wheel was exactly at the right speed. By going closer, we increase our accuracy and decrease our time between shots which results in no change in cycle times for us.
Our shooter has a 6lb flywheel cut out of steel plate with almost all of that weight 7 inches from the shaft to maximize inertia. That helps with our accuracy a good bit. We’re using a pair of 8-in SmoothGrip wheels from Andymark for the shooter wheels in a hooded configuration. There’s about 2 inches of compression and the back of the hood is polycarb. Another thing we do to increase accuracy is to have a guide below the shooter which ensures every ball is perfectly centered right before entering the shooter. We also feed each ball into the shooter at a consistent speed. The shooter itself is kept at speed using a falcon 500 (no reduction) with the integrated encoder and a closed loop PID. We used an encoder block on a VP gearbox at first but the velocity from that was too noisy. We also currently have 2 redlines with a 3.75:1 reduction as a kick stage just to get the wheel up to speed faster. Once it’s at speed the redlines turn off.
We have our own solution using a raspberry pi 4 instead of a limelight and the issue turned out to be a bad screw terminal on our 5v buck converted that powered the raspberry pi. It resulted in us loosing vision after we went over the bumps in the middle of the field each match.
We’re still going for it. But we need some serious improvements. our shot was getting ugly by finals.
Our guess based on some really scruffy physics (Scruffy whitepaper: Hooded Shooter Analysis) says we need more grip and/or travel length. We think we can get the former. But empirical testing has the final say.
I wish I had any helpful details to give. The weird truth is that there is no apparent “magic bullet” in the hardware. Their shooter is:
Sturdy and well-constructed out of thick 1/4" aluminum. (If anything is an important takeaway here, it’s this. Rigidity matters.)
Turreted
Adjustable hood on a Rev servo
2x 4" Fairlane shooter wheels
3.25" Colson feeder wheels
What appears to be a single 775 pro through a versaplanetary for the shooter (although it looks like there might have been a second one at some point?)
A NEO 550 on the feeder
~1/8" rubber or foam backing on the hood
Limelight for tracking
Unknown compression, but it isn’t anything incredibly out of the ordinary. I would eyeball it as 1-2" (same as pretty much everybody)
There’s no fancy dual-wheel shooter like 4414/1678/33/118, no crazy 100k RPM inertia wheel, no impossible to design Markforged parts.
Whatever they’re doing, it’s happening in software. They did say that they did a ton of tuning, but didn’t go into detail.
I would imagine that it has a lot to do with the adjustable hood. If you have an adjustable hood on a flywheel shooter, then you have two software controlled variables to play with (angle and velocity). If you check out my trajectory calculator (that I’m shamelessly plugging again) you can see that there are infinitely many combinations of velocity and power that will work for any given distance from the target, but that some of them will yield more reliable shots than others. At a guess, 6560 spent a ton of time figuring out the optimal speed and angle for each given distance to make as many shots as possible given a range of ball conditions. The result is that from close up they had a huge percentage of inner goals compared to outer goals, and from further away they were one of very few teams still reliably hitting the target.
230 was taking behind-the-control-panel shots at Waterbury this last weekend. As the event progressed, they seemed to be doing so less and less. I suspect ball damage was at least partially to blame.
I believe full field shooting is very possible. We have been able to achieve 2/5-4/5 inner goal from behind the WOF with somewhat similar balls (at home facility). We did some full court shooting at LAN with some success. But decided to stop as the ball quality left our comfortable operating window. The double double shooter mods we did at LAN helped. After Ventura we plan to test and expand on our shooter technology. We’ll see how it goes!