[DampRobot]

Tensioning is the primary reason for using bearing blocks (or wheel trucks, or whatever you want to call them). There are a number of ways to tension your chain/belt/whatever in your drivetrain, but in my mind bearing blocks are far and away the best solution.

Solution 1: exact center to center design. I think this is what the OP is referring to in terms of "just drilling a hole," but if it was, he left out a lot of the necessary detail. Basically, you design holes into your DT frame that are exactly the diameter of your bearings, at exactly the right distance apart to keep your drive belts/chains perfectly tensioned.

Advantages: low part count, lighter, simpler. Low maintenance (potentially).

Disadvantages: very tight tolerances. You need to get bearing holes to withing -.002/+.000 IIRC (it's been a while) to get a good fit, and center to center distances probably need to be +/- .005 for belt and +/- .01 for 25 chain. (It's been a while, and I'm mostly pulling these numbers out of my behind, but these should give you an idea of the tolerances required.) If you get it wrong, you have to remake everything. Generally harder to assemble and to maintain if it breaks. It often requires a heavier drivetrain, as you must use .125" tubing to properly support the bearings instead of much lighter .0625" tubing. Getting an efficient system is pretty hit and miss.

Solution 2: Tension the belt/chain without sliding a bearing. You can put an idler in to change the chain path and adjust the tension by changing the position of the idler. You can also physically change the length of the chain belt by putting a tensioner in instead of chain links (see the 221 product, or for example the chain that moved 971's 2012 intake arm). Some teams like to shove a floating sprocket into the middle of their chain runs to spread the chain apart and tension the chain run.

Advantages: A lot lower tolerances than solution 1. You can choose exactly where you want the endpoints of your drive system. Easy to do "sloppily", so it often works well for prototypes.

Disadvantages: higher part count than solution 1, and almost always the lowest efficiency of the three solutions (you have an extra idler just adding drag). Lacks a lot of elegance. Depending on the idler design, can be more complex, and the idler can slip over time.


Solution 3: slide one of the endpoints of your system. Almost always, this means a sliding bearing block. See VersaTrucks for a COTS way to implement this system, or 254's DT for the design that continues to inspire teams. Often synonymous with WCD in DTs.

Advantages: You can dial in tension (which means efficiency) after everything is machined. Lower tolerance requirements than solution 1, more localized tolerances (for example, +/- .002 over 2", instead of over 14"). More elegant than solution 2. Easy to fix/modify. Used by a lot of top teams.

Disadvantages: higher parts counts, you can't choose exactly where both endpoints are. Sometimes requires maintenance if you don't use cams/screws to keep the bearing blocks from slipping.

Maybe I'm biased, but solution 3 always appealed the most to me. You get an efficient system that's easy to maintain and easier to machine than exact c-c designs, at a minimum cost of parts count and complexity. COTS solutions like the VersaTruck have made this so easy and accessible that many of the tolerance/machining time constraints have been eliminated.