Chain is also used in engines, though. And teams have run center-to-center chain, most notably us this year. 
Seriously though, 118 ran (and might still, I canât recall what they did this year) chain-in-tube C-C for years.
Chain *technically *doesnât stretch, but it does wear, albeit slowly. For us, chain wear was far less of a problem than bearing blocks getting closer together due to hole clearances. Last year we ran on one tension of chain for the whole season without any problems. I would say chain is the easier of the two due to the sheer amount of forgiveness it has (and itâs ability to replace 20 pitch gears, but thatâs for another time).
There also seems to be a rare recurring problem of belt picking up field tape, although I have no idea why.
That is weird lol, though you when chain get slack it has sideways play meaning if you run chain too close to a bolt or wheel, youâll run into some problems. Belt has not play, and it can take a hit and be fine, though it really just comes down to what works best for the individual
Sometimes the chain is its own worst enemy; we had the bent-clip-to-hold-the-plate half-links in our #25 chain this year and it blew itself apart hitting the other half-link in the travel. Whoops. Switching to a Dark Soul tool and better half-links that donât have the pin made things much better. (The nice thing about a tube frame: itâs pretty hard for your chain to take a hit, except from itself apparently!)
I would be open to a belt-in-tube frame, because I definitely see the upsides from running belts in the AM14U series, but the packaging just wasnât going to happen in this yearâs robot.
You can make any design work with about any construction method⌠thought it may not be the best. So again, back to the resources.
Consider turnaround time. Tube construction? You can easily make new or modified tubes in house ifneedbe.
Sheetmetal construction? If itâs complicated geometry, especially trussed, youâve either gotta have the equipment in house or wait for your jet/laser sponsor.
Now, to those nitty gritty things⌠though sheet metal can be made to good standards, itâs much, much, much easier to make decently accurate parts in stock than sheetmetal when you lack a jet. Luckily FRC isnât demanding of high tolerances⌠usually.
We use a brake and bend our chassis every year. You just need to set up a fence and make all your flanges the same length. I doubt we have a +/- 0.010" tolerance on our bends. We bolt up the axles and use those to line up the outer walls of the chassis. Water jet rivet holes in one piece and then drill to match in the other once its all clamped up and square. Water jet has been accurate enough in my experience for bearing fits. You donât need fancy bending equipment just a pretty standard sheet metal brake.
First of all, I recommend using 1x2 x .125 aluminum tube for your chassis. We use this every single year (With the exception of last year). It allows for simple connection of gearboxes. One issue may be getting the chassis square. Our team machines 90degree gusset plates and I dont know if you guys have access to a mill.
As far as Chain vs Belt: Belt Is the way to go
My team is only starting to work with belt. The problem with chain is it stretches. Make sure you have the ability to tension your chain. One year, we couldnt tension a .25 chain and it kept snapping whenever we had quick changes in direction. Again, I dont know if anything I said applies to belts. If your team has gotten belts down, Dont go back to chain. They are heavy
-Best of luck
In 2014, for Aerial Assist, we had no mill, and a run-down (ok, vandalized) cheap drill press. Despite this, we had no problem keeping our c-channel and tube chassis square. We used steel COTS flat corner braces and tees, in the same width as the channel and tubing for gusset plates. It was heavy, but not so heavy that we had to remove any mechanisms or drill any âspeed holesâ to pass inspection. It would have been even easier if we had bought four 90 degree corner clamps (which we discovered the next summer).
Our chassis was not only square when we built it, but survived a regional of Aerial Assault with no noticeable issues, excluding only our manipulator which reached outside the frame perimeter to pick up balls (and which was NOT so well supported on the corners).
So many firm statements in here without much evidence or reasoning behind itâŚ
When it comes to winning matches and making reasonably reliable robots, for 99.9% of teams both methods are fine.
Weâve run both in drive, and have been happy with both.
More recently weâve run chain because if you already have the sprockets and chain, there is no lead time for any arbitrary length run (whereas belts are length specific obviously).