designing a better kit frame.

After the team discussion Tuesday about how our build season went, one thing came out above all, making the season go faster. getting the robot done 30 min before 12:00 on bag day is not something that i would like to repeat.

after the season ended i toyed around with the idea of creating a custom kit, that could be assembled into a frame that would accommodate any drive, and any game, with only minor modification. after a couple weeks, i cannot think of any other way to improve the design.

the basic idea is to use the pocketing in the frame for more than simply removing weight. because pockets are usually included in a frame anyway, why not use them for more than just weight savings. the result is an exceedingly flexible frame, that weighs only 15 pounds, no matter what drive-train it contains. (includes frame, axles, bearing blocks)

This image shows the basic pocketing pattern for the side of the frame rails. the holes are 2 inches in diameter, and the other holes are clearance for 10-24 bolts, or 3/16 rivets.

these images show the bearing blocks that fit inside the pockets. they allow for the bearings force to be dissipated over a larger area of the tube, making it less likely to bend. the bearing hole is a-centric by 1/16 of an inch in bot the x and y directions, this allows for chain tensioning as well as the raising of the outer wheels in a tank drive by 1/8 of an inch. with this setup and a few half links, #25 chain can be kept tight without any sliding hardware.

(second image coming, photo-bucket upload is taking forever for the last 2 percent)

This image shows how the whole frame combines using gussets and rivets. the holes on the top of each frame rail are 3/8 in diameter, and have the same spacing as the kit frame, thus allowing the standard kit frame to be integrated into this frame. they also allow for bumper mounting ( the bumper has a pair of brackets that come out of it, these go on top and on bottom of the frame, and a 3/8 d clip goes around and through the center.

this frame should, with minor modification, support the following drive setups;

long orientation tank with up to 14 wheels.
short orientation tank with up to 8 wheels
long orientation mechanum with any mechanum size up to 10 inches
short orientation mechanum with any mechanum size up to 6 inches.
slider drive in any configuration

by rotating the elements of the frame 90 degrees (so that the 2 inch diameter hole is vertical) the frame can support

tank drive of any variety
holonomic drives (normal omni wheels at 45 degrees)
with the manufacture of a new top piece, a revolution swerve module can fit inside the 2 inch hole, and co axial swerve is possible.

any thoughts on how to improve this? also, would it be legal to fabricate the pieces before season, without assembling them, provided the CAD, all previous versions, and the design notes are open sourced.

thanks to teams 111, 1114, and 40, seeing their robots at CMP served as the major inspiration for this project.

Very cool idea.

No, you can’t fabricate before build season begins. Making the design open source allows you to utilize a design that was worked on before build, but not fabricate it.

Given you mentioned fabricating it in the preseason, can I assume that this is something you’d like to see mass-produced and made available to teams through retail or KOP channels (e.g. as a future kit frame)? (If you’re only planning to build them for your own team’s consumption, then this advice may not apply.)

It’s a neat idea, and it’s got great potential for modularity. I think that you’re looking at a lot of cost in fabricating it, relative to a kitbot.

The huge advantage of the last two kit frames was the ability to punch and bend them using sheet metal equipment. The raw material is relatively cheap, and provided you’re sourcing them from a shop with significant investments in turret punches and brake presses, they’re reasonably cheap to fabricate in mass production quantities too.

The problem with this design is the need for a lot of CNC setups to drill all four faces individually. (Well, you could do it manually, but that’s obviously a ton of work; not a production solution unless you can get the cost of labour way down.) So for mass production on the scale of a kitbot, I would encourage you to think of other ways that a roughly equivalent part could be made. Maybe you could get away with C-channel made from punched and bent sheet metal (like the old kit frames, but a larger section). It wouldn’t be difficult to rivet in a few smaller stiffening braces on the open side to carry the inner bearings, for example.

The universal mounting for the bearings is nice. But as you’re CNCing them anyway, maybe you’d want to consider some more hole locations for intermediate settings. That way you could be a bit more flexible in using them as tensioners. Or if you want to get a little crazy, make them in two pieces: one sets the ground clearance for raised or lowered wheels, and the other one has some sort of pattern that allows you to rivet the wheels in place in a couple positions, e.g. on centre, or plus or minus 0.125 in to compensate for chain centre distance.

What are the flaws of the current kit frame?

Have you ever had aluminum tubing bend because you directly mounted a bearing in it? I’ve never seen that problem before. Why not just punch bearing holes in the tube?

while this may not be massed produced, after our 2012 season, i will probably write a full paper, cad a sheet metal version, and submit it to Andy-Mark. you are correct that this frame has a TON of CnC work. however, the frame and gussets themselves only weigh 8.3 lbs, which of the amount of strength this frame would have, extremely light. the other advantage, is that even if we are unable to pre-fabricate the pieces, they can be fabricated durring the finalization of the design, and the CAD. which will still cut a 4-6 days off our build season.

while the 31.5 tube in this version dissalows more than 4, 6 at maximum different axle positions, the previous version used 42 tubing, and each wheel had 12 different positions. the 3*1.5 tube saves 4-5 lbs over the robot, but does make it a tiny bit less flexible. a sheet metal version could probably accommodate a larger number of axle positions.

I think it would be legal to fabricate them during the offseason, ONLY AND ONLY IF they are used for practice bot purposes or experiments during the off-season. You can’t (and shouldn’t) use the fabricated pieces that you make now for next years game.

Nice frame, btw.

the primary advantage that this has over the kit frame is how adaptable it is, it can support more wheels, more gearboxes, and more manipulators Nativity. the kit frame also weighs more, i don’t know exactly how much it weighs, but our cart, which is a kit frame, weights i would guess upwards of 20 lbs, this weighs less than half as much. the kit frame also suffers from rigidity issues, because this uses much larger members, it should hold its shape better, this should prevent the drive train from throwing chains. my last complaint about all of the rapid build options available (kit frame, 80-20) is that they don’t tend to stay exactly square. the aggressively sized gussets on this frame should keep it perfectly square, no matter what abuse it takes. i don’t dislike the current kit, it just is not quite adequate in my oppinion in terms of flexibility and strength/weight (for the weight of the kit-frame, something much much stronger can be built). That is part of the reason that the holes on this match the ones on the kit, they can be used in tandem.

the reason to use the pockets and bearing blocks is 2 fold

1, the pockets make it much, much, much lighter

2, we like to use dead axles for the outer most wheels, and live axles for the middle two. a bearing hole would not deform the frame, but a .5 inch axle hole would be pretty sketchy.

Cool idea. Very similar to some ideas that I had been meditating on. For bonus points, you could make sure the axles are spaced properly for an integer number of #25 or #35 chain links. Also be sure to offer a bearing block with a dropped center for anyone who wants to do 6/8WD.

Given the number of teams that used AndyMark Nanotubes this year, I would bet that a solid implementation of this would see plenty of interest from teams.

the bearing blocks have the bearing / axle hole off center by 1/16 of an inch in bot the x and y direction, so dropping wheels is just a matter of the orientation that you put the bearing block in. the axles are set up for an integral number of #25 chain, no matter what wheel placement you use. you can use #35 chain, however you are restricted in which holes you can use. i believe #35 chain requires a wheel every 3rd hole, while #25 chain can be run to every hole. i will probably work on making a plate that will adapt any andymark gearbox to this frame.

This concept probably isn’t patentable since it’s been thought of already (probably), and probably isn’t totally unique to FRC – but there isn’t a solution like this in a COTS form at all. (Check patentability anyways; you have 1 year…). I really like the idea of raising the wheels while also putting tension on the chain without extra hardware. If it fit within a 2x1 rail (for live axles) or 1x1 rail (for dead axles), I’d definitely consider purchasing it because simplifies the fabrication of the frame.

However, there are tradeoffs in your frame design just because of this single bearing block alone. For a ‘kit of part’ type of frame, this block appears to appears to force cantilevered wheels. It also has no use on a dead-axle setup in its current form. Cantilevered wheels force certain design considerations of extra structure for support and bumper mounting due to the bumper rules. Given that bumpers are one of the least-complied with rules at the start of inspection day at a Regional competition, I’d say we need to make any KOP frame so simple the rookies only have to slap some wood/noodles/fabric on it to comply.

Now – you could modify this frame to address everything above, and then compete with the current KOP design. Or you could work with AndyMark so you could supply the bearing blocks and they could supply a frame that is slightly modified to accommodate your bearing blocks. (this is just conjecture, and doesn’t represent AM in any way, fyi). Put 3/8" holes in the bearing blocks, simplify your bolt pattern, and see if you can make it fit into a slight variation (i.e. add holes) of the current KOP setup. Try not to change the dimensions of the U (1"x1") if you choose this method. A live-axle version that fit on the 2" face of 2x1 rails would be awesome too. Bonus points for also making a special live-axle block that directly mounts a SuperShifter, Toughbox, or Toughbox Nano to it… (though I have no idea why you’d need that ;)).

We need a little ‘light-bulb’ emoticon…

[edit] Also note that the “KOP”-type setups use #35 for a good reason. #35 chain allows more margin of error in slop / misaligned wheels, which is important for rookies in many cases. 1/16" lateral tension may not be enough for #35 – so you may also play around with dimensions that move the bearing/axle hole more laterally than vertically when it’s rotated.

If you are serious about this, make these bearing blocks asymmetrical or add an obvious feature on which way the center hole is 1/16" offset, otherwise that’s just begging to be installed/assembled incorrectly.

Take Vex and go 1/2" on center for 1/4-20 mounting holes and 1/8" thick Aluminum.


to re-state what people are saying, these features should be in the next version:

tube version - more indication/flexibility on the bearing blocks, perhaps change the hole pattern to alternating 1.125 / 2 inch holes, with the 1.125 hole placed for bearings, this could potentially help with chain placements as well.

integration of gearboxes, perhaps design an adapter to mount several standard andymark boxes (tough-box, tough-box nano, touchbox mini?

perhaps change the setup so that the bearing blocks are only ofset in 1 direction by 1/16, this way chains can be tensioned via moving the outer set of wheels, and middle wheels can be dropped without affecting the chain.

once i get the tube version to a point where i like it, i will do a sheet metal version (our team cannot weld, or use sheet metal)

Another idea: 6 bolts on each bearing block is overkill for most applications. If you slot four of those holes horizontally, you would give people the option to use either fixed or sliding bearing blocks (for tensioning). It would also help to make it easier to visually distinguish the orientation of the offset.

Team 221’s products features asymmetrical bearing blocks which you can install with varying ranges of drop.

the primary reason for using a chain tensioning and wheel lowering system that does not slide is reliability and ease of use. sliding bearing blocks are an ideal solution, however they do just that, slide. i am un-inclined to trust anything that slides in that manner. many bicycles made use an a-centric bottom bracket, one of my team-mates bicycles was actually the inspiration for these. a sliding bearing block makes it possible to over-tension chains very easily. however with these, it will be more difficult to over tension the chain, with the addition or subtraction of a half link, and the rotation of the block, any length of chain can be tensioned.

While the tubular pieces are interesting, I would be concerned with the amount of machining that they entail. In order to consider them as a KOP option, I would think they would need to be molded plastic (which could reduce the weight even further without impacting the strength). I think the real merit of the design is the modular bearing block with a-centric holes. Even a tubular frame made of 1" square tubing could utilize these bearing blocks effectively for all the same reasons that you quoted.

However, as you stated in your earlier posts, the preference is to have dead axles for all but the inner driven set. These would not need bearing blocks as they are dead axles. Using a shoulder bolt for supporting a dead axle is really all that is needed so maybe these a-centric mounts would come in various options:

  1. True bearing blocks for driven axles
  2. 1/4-20 clearance hole for the threads of a shoulder bolt
  3. 1/2" clearance hole for shaft of the shoulder bolt

Just some ideas if you really want to take this to the next level.