pic: FrameDrive 6wd Concept

The main theme here is integrating a gear box into a frame rail, and the second is eliminating tensioners from a WCD setup.

Highlights are 2 3 cim gearboxes, single speed geared at 7.64:1, 4 wheels gear driven, 2 wheels belt driven at C-C +.005" spacing.

Some more CADing details need to be worked out, but everything was able to be packaged within FRC constraints. It’s not the lightest drive train out there, but everything can be machined with manual mills and lathes.

Any feedback is appreciated.

Very elegant. How much does it weigh? Are those 3x1.5 tubes?

I would still put in some tensioners. You may not think you need them, but an initial tensioning of the belt would be beneficial. You don’t need to retension with belts, but having cams and some small slots in there would be a good idea.
Just my opinion.

Side note: how is stuff being attached? Rivets? If so, you might have trouble with some of those connections.

I am not sure how much it weighs, I need to go through the densities and weight values of the components yet. I will put post them after I go through them. They are 3x1.5" tubes, and when I did a quick check before, going to 3x1.5" from 2x1" adds a few pounds.

My thought on the tensioning, if it is like you say that once they are tensioned the belts are all set. Why not design the proper tension in the c-c distance between the axles from the start? I don’t have as much experience with belts as others do in FRC, but that doesn’t seem like too much of a stretch (no pun intended ;)). If there is other experience out there saying differently I would appreciate the correction.

Attachment is a mix of rivets and bolts. I did not go through the detail of adding those into the CAD yet. Basically the front and rear frame rails will be bolted to the main side rails, and the bumper rails are riveted in place.

Ah, but that is a bit of a stretch. Belts and chains both stretch over time, especially for the first little bit right after they’re put on as they wear in. Having no provision for dealing with that is asking for trouble. If you design for the final distance, the initial belt tension is going to be a bit tight, running the potential risk of a catastrophic stretch. If you design for the initial distance, you’re going to be fine at first, but a little bit sloppy/slack later. Split the difference and get both sets of issues. It won’t be much, but it will probably be enough to be noticeable.

There are a number of ways to deal with this–sliding bearing blocks come to mind as the typical WCD solution–but if you don’t want to design tensioners in, probably the easiest solution to add later is the “floating idler” tensioner–drop a slightly larger-than-the-rest pulley/sprocket into the middle with no anchor and let it find a spot where the tension is right. Second easiest is a small spring-loaded pulley; install and use the spring to push the top of the chain or belt upwards a touch to increase the tension.

How much stretch should be expected in belts and chain?

I’m not entirely certain with belts–I don’t have much experience with them.

For chain, it depends somewhat on the chain, but it’s not altogether uncommon to need to adjust the chain length by up to a half-link (usually by tensioning the chain instead of removing a half-link). Adjusting by more than a half-link is a lot less common, and probably means somebody didn’t calculate the chain length right in the first place.

A 15mm HTD belt put in C-C distance should not stretch during an FRC lifetime to the point where it loses tension. IIRC 2791 runs their belts C-C in a WCD-style setup with two 15mm belts, and hasn’t had tensioning problems yet, though if Chris could weigh in on this I’m sure he can describe their experiences better than I can.

All in all, you should be fine with C-C tensioning for belt in an FRC application. While I cannot speak for C-C, our belts this year had tensioners designed in that we never used because the belts didn’t gain any slack (that is, while they were still in one piece - but that was in part our fault).

Just like in this thread I seem to get mixed answers about belt tensioning as well. Some say tensioning is needed, others say a proper c-c is sufficient for belts.

I did asked a lot of teams that ran the AM14U kit bot chassis this past year about the belts, and they all loved them. I never heard a complaint, and to my knowledge there is no tensioning on that chassis.

I know roller chains need tensioning when they are ran over long distances, but you can get away with c-c on shorter runs of chain. We ran c-c on our intake this year with no tensioners, for a short run of roller chain, and had zero issues with it.

If you’re running 9mm wide belts you should have adjustment to tension them correctly. If you’re running 15mm wide fixed c-c would be fine. I would suggest adding up to 10 thou to the calculated c-c depending on belt run length, pulley size, and the specific belts you’ll be using.

In my experience, they’re really just not necessary for a drivetrain. We’ve used 15mm HTD or GT2 belts at exact center distances, no offset, for 3 seasons without any failures. The calculated center distance works essentially forever in terms of FRC robot run times (a few hundred hours at most). While belts may eventually stretch, it takes thousands of hours under load for it to become noticeable.

I’ve never seen a properly tensioned HTD or GT2 timing belt stretch in an FRC robot. We’ve used 9mm belts at exact centers for mechanisms and 15mm belts at exact centers for drivetrains and they have stayed the same tension throughout a season and beyond.

Basically, we put in the belts week 3 or 4 of build season and we’re done. We never need to adjust them through the competition season, they just work maintenance free. When we’ve disassembled old robots, the belts show very little visible wear. As long as you properly size the belts for the application (use 15mm and it’s a lot harder to go wrong) they’re a very robust and simple solution for power transmission.

As for the gearbox in tube, I think the difficulty of assembly and maintenance for the gearbox would offset the packaging advantage. We used to run “one plate” gearboxes integrated with one side of the drive tube, but it made assembly harder and repairs a nightmare. Inside the tube would probably be worse. Belts in the tube I do recommend as there are a lot fewer parts to align and belt failures are quite rare with proper sizing.

Assembly and maintenance is a concern of mine as well. I would like to build and test before using during the season, if that is a direction the team wants to go. It is tempting because it opens up quite a bit of room and lowers the cg a bit.

I went through and checked the weight, and it is 34 pounds with the motors (without the electronics). I’m not sure if that is higher or lower compared to other drives, it seems around the same weight we normally see.

34lbs is pretty good for a 4-cim drive. Optimal would be 30lbs or even less, but 34lbs is quite repectable conidering you have the tubing that hangs over the wheels for mounting stuff. Sub-30lb drives tend to be just the base and motors.