pic: Looking for feedback on belt driven WCD design

Our team is looking to move to a WCD drivebase next year (after 5 years of using the KoP drivebase) and plan to prototype one in the offseason.

Some details:

[LIST][]5mm HTD belt drive using 30T pulleys. Back run is on inside of chassis, front run is inside tube.
]2.5" x 1.5" x .125" side rails. Fits 30T pulleys with a 1/16" raise for front/back wheels and 1/16" drop for center wheels.
[]Vex 3CIM Ballshifter attached with through bolts and in-tube nylon spacers
]4" Vex Traction Wheels
[]Welded chassis design, though may use 221’s Chassis Tube Connectors as well (don’t know the pitfalls of a welded chassis yet)
]Direct C-C belt run with .002" center add
[*]Riveted bellypan, possibly Delrin

We have liked the KoP drivebase but want to move to something a little more flexible. We are used to bullet proof drivetrains though (thanks AM!), so I wanted to make this as maintenance free and robust as possible.
I wanted to stick with belts for lack of maintenance, but was worried about the large cantilever that running two 15mm belts inside the chassis would require. So I ended up running one inside the chassis, and one inside the tube and out of the way.
This thread, and others, led me to go for 15mm if I wanted to be bulletproof.
I considered VersaBlocks, but I didn’t like the need for the cam when I never intend to adjust tension. VBs without the cam means relying solely on friction to hold them in place, which doesn’t fit the bulletproof requirement; so it was press-fit bearings into the tube.

More pictures found here

Now for some questions:

[LIST][]How safe is it to run bearings in 1/8" tube wall in heavy defense games? I have seen problems with 1/16th" wall, but I’m not sure about 1/8"?
]Will my bumper mounts be sturdy enough if welded? I have seen a lot of designs that have a top bar connecting each mount, though I don’t know how much resilience that adds.
[]Are nylon spacers inside the tube appropriate for attaching the gearbox to the rails? Would just having the bolts run through the inner tubing wall work better?
]Is 2 thou center add reasonable for a drivetrain? I have heard of people going without a center add, but I’m used to preloading.
[*]How big should washers be on the end of my shaft to hold the wheels on? I am thinking mine are a little undersized.

Any other comments or suggestions are definitely welcome! I would like to get some solid feedback before we start machining.
CAD available here

Also, my second link is messed up :confused: Here are the other pictures.

Looks great!! I really like the use of churro shaft between the gearboxes. Why are your drive rails 1.5" wide? It seems that it would be fairly easy to run the more commonly used 2 by 1. I’ve only seen 1.5x2 tube used for a drive rail here: https://www.chiefdelphi.com/forums/showthread.php?t=148743 and in that design (even though Ty used chain) he was able to fit both chains inside the drive rail.

Not sure if this fits on the gearbox side, but on the other side it looks like you can definitely improve the strength on your L brackets by adding a “rib”, aka connecting the two open corners so that rather than just an L, it looks more like a:


Hopefully that makes some sense, or someone else has a picture/cad readily available to explain this.

Delrin on the bellypan is a great call. Definitely worth looking into using Delrin, or anything else over time-consuming heavy-machining lightened 1/8 aluminum. It’ll save you lots of machine time during season.

Overall, great job on the design! Good luck prototyping it this offseason,


First - move the transmissions to the center wheels. In the instance (WHEN) something goes wrong with a belt or pulley, having the trans driving the rear wheels means you won’t be moving.

Next - delrin is not the right material for a belly pan. It is a great material with regards to wear, but it is also heavy and quite flexible. This is not the right application for it. Polycarbonate, aluminum, or expanded / foamed PVC are better options.

If you’re going to put the belts from the middle wheels to the fronts inside the tubes, do the same for the rear and lose the external stage on the ball shifters to save width - unless you simply have to have the gear ratio change.

If you really have to have a belt run not in the tube, then move your rails inward and put the belt run on the outside. That makes belt maintenance much much easier, and you can use the pulleys that are made to fasten directly to the wheels. It also keeps the moving component (belt) away from your wiring. This is a preference, not a rule - teams do it both ways.

I’m not a fan of “West Coast-style” drives generally, but beyond that I don’t have any complaints about this design. Looks solid, well reinforced, and space efficient.

I suppose I could suggest moving the gearboxes to the center to help with weight distribution and center wheel redundancy, but that sort of thing can also be fairly game-specific (since you might have to put gearboxes on the end if the game requires a mechanism in the middle).

Vex 15mm pulleys are .866" wide, making one too wide to fit inside a 1" tube. It would be possible for me to make this work with a custom pulley or one straight from SDP-SI, but I would rather keep things as COTS as possible, and Vex lets me do that. Thanks for the gusset idea too, I think I’ll use that!

Part of doing this was to see how I would put a gearbox at the back if it was useful for a design. On the upside, one wheel will always be powered, and as long as I design the rest of the robot to always rest on the back, I would only be mildly screwed.

Next - delrin is not the right material for a belly pan.

Cool, I was thinking about aluminum, but was hoping to save some weight.

If you’re going to put the belts from the middle wheels to the fronts inside the tubes, do the same for the rear and lose the external stage on the ball shifters to save width - unless you simply have to have the gear ratio change.
I would end up with a pretty massive tube fitting both belts in there on COTS pulleys AND lose the extra reduction provided by the 3rd stage. I think getting the best speeds with the wheel size I want is more important.

I actually don’t mind this at all. Gearboxes in the back gives you more design flexibility even when you don’t have to have stuff right in the middle of your bot. Plus with the four or six CIMs back there, it’s not even hard to “lean into” the off-center weight distribution (just find a way to stick your battery back there, and you’re done). Sure, in the resting state, the back wheel probably still isn’t carrying more than half your robot weight*. Also when you’re on the middle and front wheels, one belt will be carrying the torque for your whole weight. Nevertheless, I like the compactness of having the drive gearboxes near the edge of the robot.

*it’s a (literal) balancing act because if the CoM goes too far back, you’ll be liable to tipping in that direction

If you’re looking for alternate bellypan materials, I’d highly recommend 1/4" baltic birch plywood.

1836 has used 6mm baltic birch wood for bellypans the past few years with no problems. Laser cuts in a few mins tops with electronics mounting holes and everything. Definitely recommend using it if Delrin is not a good choice as Tom Line mentioned.

We like laser cut 5052 .1" aluminum sheet. Design placement in CAD, remove excess material (we use a diamond lightening pattern but I’ve seen teams have good luck with square patterns and thinner crossmembers), and you’ve got one good looking, strong, and functional bellypan. When I get back home I can send the CAD for ours if you want.

I find that cutting out a rectangle with a few holes is so much easier than a lightening pattern (both to effectively design and manufacture) on top of that that its worth the arguable strength loss or weight gain to use wood/other alternatives that don’t need lightening just so its done much faster.

Great start! Belt in tube drives are nice and they have a few key advantages here.

If you stick with this setup, you can reduce the tube width to the standard 2x1 by going with a custom pulley. If you can machine a 16mm wide pulley without flanges, it will fit in a 1" wide tube. What won’t fit are the bearings - you either have to counterbore the pulley by 1/8" on each side, or you can add oversize washers / shims between the bearing and the tube. Either way, it’s worth the effort to get a thinner tube if you are doing this exact setup.

You should be able to safely run 24 tooth pulleys with 15mm wide belts and 4" wheels. The only reason I say should and don’t guarantee it is because you are essentially driving two wheels through one belt (the center to middle belt), and I don’t have direct experience verifying that setup, but otherwise I have run many, many drivetrains with 24T pulleys and 15mm belts and never had anything close to a failure. This would let you go down to 2x1 as long as you don’t rivet to the tube above the belt / pulley line.

Answering your specific questions:

  1. It’s very safe to run bearings directly in a 1/8" tube wall. Have done it for years without any issues.

  2. Bumper mounts will definitely be fine - this is essentially what 2791 did in 2011 and 2012 and they worked well.

  3. The only issue with the nylon spacers would be that it would be difficult to assemble that sort of system. I usually would just drill the outer hole for tool clearance and use only the inner hole to mount the gearbox. Try and go for >2 screws holding the gearbox on, ideally not all colinear, for best results - but with the cross bracing you have you should be fine.

  4. 2 thou is reasonable. I usually run zero but 2 thou won’t be a problem. Don’t be surprised if the tension “seems loose” - as long as it never ratchets it is supposed to be that way.

  5. Honestly, as long as they clear your shaft they should be fine… I’ve done as small as 3/4" for sure. I would definitely be sure those bolts are Loctited in place, and check them periodically during the season.

All that said, do consider doing a 3x1.5 tube setup and running both belts inside the tube with 24T pulleys. The 3" tube and smaller pulleys let you rivet or bolt to the space above and below the belts, making it easy to mount stuff to it. Two belts in tube isn’t much harder than 1 belt in tube, and it ends up saving space since you don’t need any flanges on pulleys that are in a tube. If you want, you can do 2x1.5 tubing with 24T pulleys, and two rows of rivet holes - just use the row opposite that side’s belt for mounting stuff.

If you have any more questions about this style of drive, let me know - I’ve built drives just like this for many seasons.

I will only add that we have entertained belts, but end up switching to chain. On the other hand we tend to be a bit “pushy” and don’t like being slippy.

A properly designed belt system never slips, and has no issues with pushing.

Entire DT torque through belts can be tough on the belt in a pushing match. The traction limited case for a 154 lb FRC robot running 30t belts on 4" wheels exceeds Gates HTD belts tensile rating. We ran a 24t off our gearbox this year and failed the belts at champs after 2 districts and district champs. It is common practice in FRC to run belts over max tension since they are designed for a much longer life than FRC entails.

Traction limited wheel torque:
(154 lbs) / 2 (half your robot weight) * (2 in wheel radius) * (CoF of 1) = 154 in lbs

Traction limited belt tension:
With 30 tooth pulleys this puts your pitch radius at 0.94 inches
154 in lbs of torque /0.94 = 163.8 lbf

The rated working tension for 5mm pitch x 15mm wide is 60.3 lbf according to SDP-SI (pg 50)

Direct driving the center wheel gets you a lot closer to the rated tension by halving the load on the belt between the middle and rear wheels. Alternatively you can run a larger pulley or wider belt from the drive gearbox.

Belts don’t slip if under proper tension.

This looks great, on the whole. I’m not a huge fan of the belt-in-tube, just because I like having easy access to things in case they break, though belts (especially 15mm) are quite reliable, so it’s a perfectly justifiable design choice.

You should consider using Vex bearing blocks instead of mounting everything directly to the tubing. They deal with loads a great deal better. https://www.vexrobotics.com/bearingblocks.html

Additionally, try using chain instead of belt (#25 works) since it’s a bit better for higher load applications and some 4" colson wheels instead of the treaded ones. They word a bit better on carpet and are less finicky (they’re one piece).

I’ve never had or heard of any problems with bearings in 1/8’’ tube.

Belt is perfectly fine in FRC drives so long as you design correctly, and neither treaded wheels nor Colsons are clearly “better” than the other (there are design trade-offs, and it depends on your constraints).

When giving advice, please don’t state something as if it were an absolute rule unless you are very confident that it is correct in all cases.

Yeah, I like the idea of reducing size even more, but I like keeping as much of this as COTS as possible. COTS lets us start assembling as soon as we know what we want to make without machining or sending things out for manufacturing. Though the benefit of narrower tubes (and no spacers, heh) is a real one.
The main reason I went with 30 over 24 was so that I could step up to 6" wheels without any design changes. 8" might be feasible, but I couldn’t find much info on it. I have also read of some people snapping belts with 24T pulleys, though I never quite found a concrete reason for that. The 30T gives me a little safety factor to cover the unknowns.

All that said, do consider doing a 3x1.5 tube setup and running both belts inside the tube with 24T pulleys. The 3" tube and smaller pulleys let you rivet or bolt to the space above and below the belts, making it easy to mount stuff to it.
I may step up to 3" regardless just for the mounting opportunities. Loosing out on being able to rivet or bolt anything to the front top/bottom kind of sucks.

Reading comments about the full force of the robot on the back two belts makes me thing that chain may be a better choice specifically for rear-mounted gearboxes. While I like the simplicity of belts, the additional strain in this application may not be worth the risk. And thanks for the math analysis, I’ll be passing that on to my students.

Thanks! I didn’t really want to go belt-in-tube for the exact reason you mention, but I wanted to avoid the almost inch of cantilevered shaft that would be taking up internal space.

I mention in the OP why I went away from Versablocks. If I was doing chain I might use them if I wanted to run the chain in the tube.

It’s worth noting that running HTD belts “out of spec” is pretty common practice in FRC - the specs you’ll get from Gates are formulated for industry, which requires orders-of-magnitude longer service life than what we need in FRC.

You will almost certainly be fine with 30T pulleys and 15mm belt on 4’’ wheels, in my experience, even with the gearbox in the back of the robot.