The mismatched wheel is actually what solved some of the in taking issues for us. We did some other geometry changes as well but that was what finally let the cube spin in to place so nicely. The real intake will likely have 2 mismatched wheels we just haven’t tested in taking from the other side yet.
Looking good! Cool to see the side roller design work at that big of an angle. Significance of the white sheet in front of the CIMs? Just as protection?
Do you have plans for a photoelectric sensor or limit switch to identify if you have a cube?
Great looking intake. Can’t wait to see it on the final robot.
The white sheet stops the top corner of the cube from getting caught on a motor and preventing it from rotating. The real intake won’t have motors in that location.
Yes there will be a sensor to identify if we have a cube, possibly two to identify if it’s seated fully or not, false positives may be an issue with only one sensor.
I see the pneumatic cylinders on the intake, but it appears as though when you’re intaking a cube you’re approaching it with the intake “closed”. If I’m seeing things correctly the cylinders could effectively be swapped out for springs.
The intake also has to drop off cubes and a pneumatic release can possibly do that cleaner than a spring. Especially if we need to shoot the cube and have the wheels quickly move away from the side of the cube.
The cylinders allow us to adjust compression using a regulator. It’s also possible to have multiple pressures with a shuttle value as per the Q&A this year, so one pressure for intaking and one pressure for holding tight to the cube.
Does this new revision work equally well in the 11" wide orientation?
Yes, it works well in 11" orientation from the testing we have done.
Why not just utilize one sensor and look at the current being applied through CAN? Once you hit stall, you should effectively be in possession of a game piece. (Unless your robot is driving against a wall trying to pick up the field borders.)
That is an option as well, we used current sensing to know we had the gears last year. The intake doesn’t really completely stall though and I’m pretty sure if we got it wrong the Fairlane wheels could possibly rip the fabric (we haven’t done it yet) so we want to be sure that isn’t a problem.
We haven’t graphed current during the intaking yet to know if it will only have a current spike once we completely have a cube. It could spike during peak compression as the intake goes around the cube corner.
Current is an option and one I like but it may not be the best option, more prototypes and testing are needed before we know for sure.
Yet another very elegant design from Spectrum.
Do the mismatched wheels cause the Cube to spin primarily in one direction if it is not “square” to the front perimeter or your chassis?
Amazing
Yes it always spins towards the fairlane wheel, even if it were “square” it spins it around at least 90 degrees unless it’s against a wall and we hit square.
It looks like your design overcomes the issue of jamming (depending on arm length) that can occur with some of the other designs published here on CD.
What cylinder sizes are you guys using? Great work
Edit: also, what belt sizes? We’re thinking about doing something similar
3/4" bore, 1in stroke, around 20ish PSI for intaking but it works higher and lower than that as well.
Various belt sizes, I didn’t design/build this so this is just from our documentation but it should be right.
Wheel to pulley: 42:36, - 70t
Pulley to pulley 36:36, - 90t
Pulley to motor 36:18 - 80t
All 9mm, with the two 36t pulleys being 15mm width pulleys with two 9mm belts on them.
VP is geared down to 35:1 I think, but we have swapped in multiple ratios so I don’t know what was installed for that video.
Thank you for sharing! 
Version 8 photos and gifs are up on our blog
http://blog.spectrum3847.org/2018/02/day-29-30-31-intake-updates-robot.html
