Let’s say the game piece is 12" beach balls, and your team wants to try the 2002 team 47-style intake (not sure if they’re the first to come up with it). See also attachment.
Questions:
What materials, components, and tools would your team use to prototype this style intake? (wood vs. aluminum tube/versaframe, drills vs. motors+gearboxes, 1/2" hex standard pieces everywhere or something else…?)
What prototyping approach would you take? (variables to test, how to design prototyping to test multiple things at once, approach to testing, etc.)
Context: I think I posted once before… We’re trying to up our game this season when it comes to prototyping kit, planning, and approach. This sees like a good example to go through.
I’d honestly prototype it first with some good ol’ pvc pipe, the appropriate poly cord and rollers, and a couple of drills. Relatively inexpensive to do and saves a lot of time.
We used this intake in 2012 and had great results with it.
If we would prototype this again we’d use versaframe 1x1 and the versaroller system/tubing. Maybe even use flex wheels instead of rollers?
Also an upgrade to this setup would be to use timing belts and skip the polycord. Definitely recommend not using bevel gears on this setup and running two independent motors (775pros).
I would also try to use mecanums first to see if I could achieve similar effects lighter/simpler. (Like we did on 125 in 2014) This would likely be a static mockup without motors to see if there was spacing. that would work. Just a shaft through some 2x4s and a mock bumper.
After that, toss together a quick sketch and go to the laser/3d printer to poop out a mix of bearing plugs and a quick structure to hold things in place (or if needed to be bigger than laser, shop bot is an option). Pair of motors in VP’s and hook up to some ESCs.
Variables: Speed, grip materials, compression.
How to test multiple things at once - don’t. Treat this like an experiment, if you start mixing variables you can’t determine impact as well.
I would also consider a simpler 16 style intake given that the team I’m currently working with runs swerve which allows for lining up on pieces in a different way. Relevant because your intake system needs to take into account HOW it will be moving in relation to the game piece, what works for a skid steer may not be the same as for a swerve.
If I had to make this intake, I would probably be in the same boat as RC. Thought process:
If I had enough old wheels lying around, I’d probably try to do a sanity check using 2x4s and whatever shaft the wheels would spin on first. Just to see if I can space things out a bit and do fewer rollers. (I ball on a budget, sorry not sorry.)
This intake is going to get beat up being outside the bumpers, so I’m doing either thicker-wall aluminum tubing or polycarbonate tubing on the production version and would want to get to that structure pretty quickly lest it impact other parts of the system. The former, I would source locally. The latter, probably from VEX.
VersaRoller is wonderful stuff, and it would be my go-to here.
I agree that it will be easier to have one motor handling the centering and one motor handling the final intake to the robot, partly to avoid getting bevel gears right and partly so that the two can be tuned independently. I can see situations where having them at different speeds will help the intake move faster or avoid jams, which wasn’t really an option the last time this kind of intake was common. (In 2012, you couldn’t throw around this kind of power without using the BaneBots motors one year removed from the infamous case-shorted ones that left a team dead in the water in Einstein finals. In 2016, the 775pro was brand new. In both cases, you were using motors that probably weren’t fully understood if you wanted to get past Fisher-Price/9015/BAG power. These days, we know what the 775pro and RedLine can do.)
You can make the centering motor spin two ways with a reasonably simple gearbox; the second spur gear will go the other way. While precision is always better, even a sloppy one made on a drill press should serve the purpose here since we’re transmitting so little torque most of the time. (You could also use a couple VersaBlock Minis with 40T gears to achieve this.)
Depending on the number of beach balls on the field and possession limits, I would try to make the compliance come from the ball and maybe the joint this is pivoting out from. $6 a foot for the McMaster tubing that VEX recommends isn’t cheap, but it’s cheaper than a bunch of compliant wheels and was able to grip fuel just fine. I’d also have a backup plan in case that tubing doesn’t do the job, which would probably be compliant wheels or some other material.
Personally, I’d run polycord on this because I suspect I’ll be tuning on this during the season, and polycord is generally cheaper than timing belts.
I’m probably missing some details, but this is my initial spitballing.
Agreed with above, we went cheap on ours and used 2x4s ripped to 1x1 to test the possibility of running this for stronghold. The final version was 1x1x.0625 aluminum tubing, 3d printed rollers, polycord, and some pvc pipe. To get by the gear box idea mentioned above we used a versaplanetary on the motor and ran the polycord belt in a figure 8 to make our rollers run opposite directions.
There are so many things I’d do differently on 4329’s 2016 robot’s intake, but thanks for using it as an example :]
As far as designing this, I did most of the CAD and it was just from looking at images of old intakes like 1323’s, 973’s and 47’s. We didn’t really prototype anything, which was a bad idea. If I could redo it now I’d definitely mock up something using some scrap aluminum or wood and figure out the ideal ball compression. Do as much testing as you can ASAP and work those lessons learned into your design.
The Versaroller system was talked about, I’d definitely look at that as well. 4329 used polycarbonate tube as a roller because it was easy to put bearings into the end of it. It worked, but it also flexed a lot due to being 1/16" wall thickness, if I remember correctly. I’d go aluminum tube if I could do that over again. Check out the 1678 video on YouTube for putting rubber tubing over a roller to make it grippy.
Also do not rely on the just the polycord itself to transfer all of the motion between the rollers. We ended up using a timing belt on the first stage from inside the robot’s front roller to the 1st roller within the intake, which powered the polycord bands & all the rollers. All off one 775 motor through a versaplanetary gearbox. We had issues with bands slipping, rollers stalling & in turn the ball wouldn’t always move.
These types of intakes also require some adjustability to tune them. Our intake was probably too low to the ground, putting too much pressure on the ball. Give yourself some freedom to adjust things to find the sweet spot of compression. That’s something we didn’t really do with this design and it showed.
Here’s a video of our testing too just in case anyone is wondering how it worked: https://youtu.be/YdgZSP8JcoA Let me know if you have any more questions.
First I would ask you - what are your requirements for this intake? Do you specifically want to build a CD-2002 intake, or do you want to build an intake that can center balls while intaking them?
If it’s the former, RC has the best advice. Versatube, versarollers, and flex wheels are the way to go. You could make one of those intakes with almost no machining these days because of WCP / VEXpro parts.
If you just want an intake that can center balls while intaking, look into mecanum wheel intakes. The first I’ve seen of them is 846 in 2014, but my favorite example is 971 2016. It’s much simpler than the CD-2002 intake and has the potential to be lighter, and arguably, more compact. I’m just hoping we see some >4" diameter mecanum wheels released sometime soon.
The former is certainly a fun and interesting way to study making something really complex, because that usually opens up new avenues of thought on things.
Neither video you linked to really shows much of the mecanum intake in action; 846 was playing an inbound-to-truss strategy for much of the match, and 971 was practically right on top of the boulders and used their intake (with an omniwheel in the middle) to tap it in. 4901 experimented with a mecanum intake in 2016 and abandoned it after the first event because we couldn’t get enough centering effect. Is there some mecanum intake best practice sheet, or is it “fiddle with it until you like what you see”?
Admittedly I didn’t get a chance to view the videos too closely while in class, and chose them off of memory of what I thought would be good examples. I’ll try finding some better ones to replace those. Having seen both up close, I can confirm that they were just as if not more effective than the CD-2002 intake.
From what I’ve seen as long as you have the right compression the mecanum wheels work wonders.
Definitely interested in the simpler design with the mecanums as well. Maybe a good exercise would be to buy a multi-pack of beach balls, split our mechs into two groups, and practice prototyping by building an example of each of these two common centering intakes for the sake of comparison & experience.
The point about the mechanism getting beat up outside bumpers is valid, and it makes me think just starting in aluminum or polycarb tube for structure may be the way to go during prototyping, though the very first prototype could be in wood for the sake of speed.
Fairly balanced support for versarollers, timing belt, polycord - less for compliant wheels given the number (& $$$) that would be needed it seems for this application
The two-motor recommendation a couple of you made makes sense as well.
Probably with PVC pipe, drills, and depending the grip of the balls, either churros (for squishy balls) or pieces of rubber on 1/2 hex shafts (for harder balls).
Re: roller intakes, this should be required viewing by everyone for crafting polycord bands or putting rubber tubing over an aluminum roller: https://www.youtube.com/watch?v=fYKsIe-yMrA
I recommend checking out 1678’s other resources, tons of good stuff out there.
Consider prototyping in stages. For example:
Stage 1) quick mockup - PVC pipes and wood scraps, c-clamps and wood screws, metal scraps where needed for strength. No motors/gears/belts/chains, just manipulate parts by hand. Try multiple concepts.
Stage 2) Add quick and dirty motors/gears/belts/chains, experiment with different speeds and materials and ball compression. Refine as you go.
Stage 3) Make a more robust proto intake and hack it onto an old robot base from a previous year. Drive the mockup around chasing balls to test the pickup in a dynamic environment, try picking up balls trapped in corners, picking up a ball from a clump of balls, etc. Testing on carpet with texture and friction similiar to a tournament playing field.
