Tolerances shown on the McMaster-Carr site are generous. It is a nice surprise that 2 fit into 1, and not a surprise at all that 4 doesn’t fit into 2. Results with other samples (from different lots, and thus different regions of the tolerance band) will be different.
I tried putting a piece of Thunderhex into 2. Too tight.
We used those hex tubes in two ways on our 2021 robot. This robot was tiny (22x22") and had no space for pneumatics, so we had to get a little more creative.
First was for the climber drums. This is a 3/8" thunderhex shaft inside the hex tube, surrounded by an Onyx drum. The drum’s ID is 0.635" hex (0.010" oversized). The outside hook feature was to loop the climber rope through to hold the climber down during the match. We didn’t have any problems with this 3d printed part through 4 short offseason competitions this year.
The other use was on our intake release mechanism. The intake was a 4-bar with some half inch hex shafts to run a lot of different wheels. We added some gas shocks to extend it and an Onyx “cyclone” to each side to release the mechanism as soon as the intake motor spun the shaft. The first cyclone rounded out, so we switched to the AL hex tube insert. The printed part has an ID of 0.754" (0.004" oversized, which is a noticeably tighter fit than the climber drum).
For cheaper hex tubes check out orange aluminum. We use a lot of it. We run 1/2" OD, 1/4" ID and tap to 5/16-18 for standoffs, also easy tapping for 3/8" OD shoulder bolts in a versa roller type system. Not as strong 7075 hex shaft so it depends on your application if it will work or not.
For a point of reference, as a test back in 2017 we bought all the 1/8" hex sizes from McMaster and they all telescoped nicely. A year ago we bought stock of 2 sizes for a telescoping mechanism, and they didn’t fit. I’m guessing extrusion dies are probably wearing making the fit change.
We had an FTC team use these tubes to create a really nice telescoping arm back in 2018-2019 (Rover Ruckus). I remember that the final stage was 1/2" Hex and it was sliding quite nicely in a larger aluminum hex tube. It may have been that they got lucky with how the tolerances worked out, but the whole mechanism worked perfectly. We used lots of 3D printed parts to create sliders for the larger hex tube stage relative to the base stage.
This is the best picture I can find of that robot and you really can’t see the hex tubes because the arm is fully retracted. The larger hex tube is just barely visible behind the polycarb side plate on the arm.
That robot won the Design award at Houston Championships. It was an amazing design. The CAD was amazing! It was a bit of a swiss watch to work on, but the team was meticulous in their build and the final product was a work of art.
I can’t remember whether there was a weight limit that year in FTC or not. One of the design goals that this team had was to be able to drive under the lander to avoid the congested areas on the field, so the robot was quite short (as you can tell from the pictures). That, and the use of 3D printed and Delrin parts with plenty of pocketing kept the weight down quite a bit. It was still quite dense, but weight was not a significant constraint.
I don’t recall that retracting the collector was much of an issue. The entire arm raised up to score and then lowered back down to collect, so other than the starting position, most of the time, the arm was fully extended. They used the mechanum drivebase to maneuver the end of the arm to the locations where the game pieces were to collect them rather than retracting and extending the arm. Even the climber was on the base stage of the arm. Once the hook was latched onto the lander, the robot would fold up to lift the drivebase off the ground and since the arm was still fully extended, they could score their final pieces while hanging to get in one last cycle during the endgame.
I’ve only used extensible hex shaft once (Aerial Assist Ultimate Ascent climber), and wasn’t pleased with the results. Are you looking for something that can be actively rotated, or just something that will maintain a constant orientation? If the latter, two round COTS telescopes linked at the end should be much easier to get right than a cobbled-together hex telescope.