Fitting belts within your frame
 

I'm currently designing a prototype frame for possible adaptation in the coming FRC season. To save space, I'm looking to run one of the 15mm belts on each side of the drive train inside of the frame element itself. I know this is commonly done with 9mm belts, but to stay on the safer side I figured we should design for 15mm. The problem I'm having is I'd like to stay with 2 inch x 1.5 inch frame without having to move up to 3 x 1.5, and this mean the only HTD pulley I can fit (accounting for a center wheel drop of 1/8 inch) in the frame is a 18T one. The 24T is just a bit too big from my calculations.

Also if it matters I'd like to do this all exact C-C. I've heard it works well and we do have access to machining that is accurate down to the thousandth, so I think its the best solution.

My question is whether or not 18T pulleys are big enough to handle the loads that an average drive train subjects them to. Has anyone used 18T pulleys on their drive train? Is 9 teeth of engagement enough?


As a side note, I'm also wondering if 6063 aluminum is good to use for frame elements over 6061. Online its about half the price, so I'm hoping we could use that since we would probably be able to buy more stock.

Re: Fitting belts within your frame
 

You're right, the biggest HTD pulley you can fit in a 2" frame is 18t. You are also right to not even consider 9mm wide an option if you try this.

If you can hold tolerance in the <10 thou range, exact CC works just fine in our experience.


(note that it's not just number of teeth engaged but the size of the moment arm that makes bigger pullies stronger)

We used 18t, 9mm belts in 2011 and had huge issues with ratcheting. If we didn't fail a belt, we were definitely about to. Going to 15mm would help, but we have never done that with 18T, so I can't speak from experience. Honestly I'd say it's a bit too close to the margin for my comfort.

I would bite the bullet and go with bigger tubing and 24T pulleys. It is a weight penalty which is annoying but you have plenty of room to rivet to both sides of the tube and the combination of 24T pulleys and 15mm belt has served us very well in the past few years.

Note that if using 1.5" wide tube, you'll either have to do some kind of clearance counterbore on the pulley or use some kind of spacer to space your bearings out of the tube. We take oversize washers and bore them to 1.125, then place them between the bearing flange and the tube. I'd like to go with the counterbore method for next year.


I can tell you that we've used 6063 in our 3x1 and 3x1.5 frames in 1/8" wall without anything resembling a failure. It's not best practice but it works.

Re: Fitting belts within your frame
 

We try to avoid 6063 aluminum as it is weaker than 6061 and much more finicky to machine. If you are just drilling holes it is probably ok, but if you are doing any end milling avoid it.

Re: Fitting belts within your frame
 

+1. 6063 has a much lower tensile strength and stiffness than 6061. If you take a hard hit to the side, 6063 bends much easier. It's cheaper, but it's not worth it; just move to 2x1.

9 teeth engagement is fine, however, I can't comment on the worth of 18t belts. It may or may not be ok; I would look up the specs of the belts online.

EDIT: Inside the tube? Not just the frame? If you are dead set on something like this, have you considered thick c-channel? That allows for upkeep and should cut down on space usage.

DOUBLEEDIT: Disregard my post about aluminum strength and see people comments below.

Re: Fitting belts within your frame
 

If you're passionate about getting your drive elements inside the tube then why not do this...



#25 chain is proven...plenty light and works great with fixed center distance.

Re: Fitting belts within your frame
 

While your first sentence is accurate, you're not speaking from experience here. Your words are just conjecture. We've taken many HARD hits and not bent 6063 in our frames.

I'm not saying it's a good choice. Definitely go 6061 if you can... I would really prefer if my team did - but going 6063 isn't going to ruin the robot in this application. If you're dead set on 6063, it should work based on my team's experience.

Re: Fitting belts within your frame
 

You would know more than me. However, I have had smaller non-frame applications (like gussets) where 6063 bends more readily than 6061, so that's where I'm coming from.

Re: Fitting belts within your frame
 

This is somewhat misleading.

6063-T6 has a 15% lower yield (35 ksi vs 40 ksi). Two identical cross sections in 6061 and 6063 have the same stiffness because stiffness is a product of Young's Modulus (which is basically 10,000 ksi for all aluminum alloys). Welded joints will be 40% weaker than in 6061, but any annealed aluminum is already incredibly weak regardless of alloy.

Our experience using square/rectangular tubing is that there is no difference. Quality 6063 should machine more easily than 6061, but I know that 973 has gotten some before that was far gummier than even 6061 (which is an impressive feat).

Re: Fitting belts within your frame
 

I read 30ksi for 6063.
Oh whoops, totally forgot about the modulus.
Both have milled well for me... it's aluminum anyway. It's when you get to the lathe that there's a hassle involved.

EDIT: Could be lower than 30ksi depending on where you buy. McMaster has two different tempers available (but I don't usually buy from there).

Re: Fitting belts within your frame
 

Something I'm missing here? 192 runs 22t gt2 pulleys and 15mm wide belt in 2x2in 1/8 wall tubing. Maybe it's the lack of flanges...

Re: Fitting belts within your frame
 

I was referring to not being able to move up to a 24T pulley, which is the next biggest size vex sells. I was hoping to use their pulleys as we've ordered from them before and they already are set up for hex shafts. The flanges dont help though, removing the flanges would make it really close if not hitting still with a 24T. If you have another supplier for 5mm HTD Hex pulleys I would you be willing to share? I know you said GT2, but were invested in HTD now, so it would be nice to stay standardized. If they aren't hex broached yet I think we could figure that out too.

Also the reason I'm trying to put one run of the belt in the frame is because I'm trying to set up a WCD w/ 15mm belts. I like the WCD design for a number of reasons, but I really want to err on the side of caution and go with 15mm belts. This means I'd need probably two 15mm pulleys inside a gearbox set up for WCD, which means making longer spacers and cantilevering the boxes even farther off the frame. To me it just seemed like a better idea to put one belt run in the frame. I'm just worried about the small engagement on the 18T.

I designed the whole thing up with 3x1.5 and it weighed ~44 lbs all said and done with the 3 CIMS and shifters (minus control system), which is pretty beefy. Moving to 2x1.5 I could save a few pounds, and since that's still structurally sound it seemed like a good way to lighten the base.

Re: Fitting belts within your frame
 

Sorry, I was assuming that rivet clearance on both sides of the frame was desired. It's also been a few months since I've done the math on pulley + belt OD. Thanks for the correction and information. Do you guys weld everything, or offset your rivets so they are on the side the belt is not on, or something else?

This is the same ballpark we were in last year. Honestly, a 44 pound drive base, complete, is not that bad. It's not crazy weight efficient like the world's best west coast drives and sheet metal frames, but if your manipulator weighs more than 70 pounds you've already messed up.

I'm curious about moving to 2x1.5 and using 22T pulleys, if it works for 192 it is something we might try next year. 3x1.5 tubing is definitely a compromise for us. We briefly explored using 2.5" tall tubing last year to try and save weight / make 4" wheels more viable, but once the game came out and the floor was perfectly flat we decided there wasn't a good reason to try something new.

Re: Fitting belts within your frame
 

If you're looking for a mostly COTs option our set up might be a bit difficult. We buy pulley rod from SDP-SI and cut it into 1.5in pulleys before doing the bores and broaching ourselves.

22t is the absolute largest you can fit in 2in 1/8 wall tubing. We offset the baseplate rivets from the belts and we do the center drop by pushing the outer wheels up 1/16 and the middle wheel down a 1/16th. I think if you did it all from just the center or just the outer you would get some rubbing. Going 3/32 either side of center to get 3/16 total should be possible though. (please check in CAD though I haven't looked at it for a while)

We chose to go with 2x2 instead of 2x1.5 because of how tight the belts are in the tube. 1.5 doesn't leave much room for things poking into the side of the tube (like gearbox bolts or the caps for access holes) and we were worried about rubbing the sides of the tubes. How well does your set up work with 1.5in wide tubing?

Re: Fitting belts within your frame
 

As I said above I was doing only one run in the tubes. I had the other runs on the outside, so fitting the pulley width wise was no issue.

Re: Fitting belts within your frame
 

This is similar to what I had to do in order to comfortably fit 24T pulleys in a 2.5" tall frame design I worked on last year. Ultimately I think we're going to go with drop tuning in 2015 (sanding down Colsons until the drop feels "just right") so this isn't an issue for us.

1.5 is a bit of a pain, but not impossible to work with. As mentioned before the main thing to keep in mind is the bearings. You either need to put a spacer between the bearing flange and the tube or counterbore the pulleys to clear the OD of a bearing. I guess VersaBlocks are also a way around this. As for bolt heads, we design around where the belts should be, aiming to put bolt holes in the middle of the tube when possible. Most things we do rivet to the top of the frame anyway. As for access hole caps, we don't even cut them anymore, but when we did we left a 1/16" thick tab to rivet a plastic cover to.