I’ve heard it said, “don’t build it if you don’t want people to ask for it.” Which means – if you don’t want to make bikes with bent seat, top, down-tubes then don’t post pictures of a bike that has that stuff on it (duh). Oh well, too late.
I gravitate towards these types of frames but the problem is that they take a longer time to build because of the many more steps involved. Builders like Retrotec that have been doing it forever…sure, they can whip frames out no problem with all the bendy tubes you could ever ask for, but they’ve dialed in their process over the years and has nice tube benders to do it right. For the ones just starting to bend tubes it’s an exercise in patience and persistence. It’s not that the actual bending is all that hard (when you have the right tools) it’s getting the tubes to match up right with the other tubes on the bike and with each other (in the case of seat and chain stays).
So anyways, here’s some notes on this build and what worked and didn’t.
I only have a Harbor Freight tube roller with a set of SWAG off-road dies for large radius bends. I have no way of actually measuring the radius I’m bending so I just roll it until it looks about right on the drawing, or when I eye-it between the other tubes in the fixture.
Because this frame will have an internal cable 31.6 LEV dropper post, I bent a double-butted True Temper Supertherm (1/0.7/1 x 34.9mm) in the roller. To roll a butted tube I fused some extra 4130 on either side to make the tube longer so I can get a bigger bend and put the start/stop points off of the thinner or butted portions of the tube (this was not my idea, i got it from Waltworks). Once rolled, I attach a tube block to make sure I can easily phase the tube for the upcoming BB miter. Rolling slightly ovalizes the tube and leaves dimples where you stop/start on the roller when going back and forth. You can see the dimple in a couple of the photos…it’s where it transitions from going straight to curved. You can sleeve it right at the transition or leave more straight section but then you get the dimple. I needed more seatpost length for the dropper – for the longest LEV dropper you need at least 125mm of straight tube (that’s what they tell me at least). The 60mm seat collar I blogged about earlier (already slotted for the seat collar) was fusion welded around 65mm above the dimple in the tube leaving a 125mm of straight section. Once I cut the tube to size based on the BikeCAD drawing, I put another tube block by the lower end (BB end) using the flat plate to phase it with the other tube block that was on since rolling. The second tube block is what I use to load it in to the mill vise for the 90* miter for the BB. I loaded the BB and seat tube into the fixture and tacked, then took it out of the fixture to weld it all the way around with the BB heat sink in there to help prevent distortion and backpurge.
Here’s where I went manual because I wanted to see it in pencil before cutting any more tubes. I used a protractor to draw a horizontal line representing the effective chainstay length and then a line representing the *effective* seat tube angle — BB center to center of seat at the rider’s saddle height. I then traced the actual already bent seat tube from BB to the top of the seat collar so that the ‘construction line’ intersected at the correct saddle height point. I then measured the angle from BB center to the top of the bent seat tube to get the number for the frame fixture’s seat tube angle. The angle of the seat tube in the fixture is steeper for the bent seat tube than it’d be for a straight tube (stating the obvious hopefully). I used a protractor and long stainless ruler to help these measurements along. For an effective seat tube angle of 72* I set the fixture at (for this particular bend) at just over 75*. The swoop in the seat tube angles the seat post back to place the saddle in the right spot to reach the 72* spot in space.
Then, I drew the other points that rule the frame’s geometry – bottom center of the head tube, rear axle using BB drop and actual CS length, etc. Using the hand-drawn drawing of the frame I took miter angles using a protractor once I bent the top tube and marked the rough cut locations while holding the tube up to the head tube and seat tube while they were loaded in the fixture (high tech). It works pretty well but yes, you either file to fit or keep loading in the mitering fixture to take a bit more off until it is close enough to file to fit. Or you do better math than I can and just cut the bugger once…but I covered that already in the last post.
The chainstays are True Temper OX Plat S-bends. They have the lowest bend of all the stays i had. I actually mitered a pair of new Deda stays but the bend was TOO much and ended up hitting the seat tube (which is also taking up more real estate on the BB shell since it’s a 35mm diam tube). So those won’t work on all shorter stay bikes…good note to self. The True Temper CS’s worked just fine though with a good crimp.
The wishbone seatstays? Not going to go over that one…it’s a secret. No, not really. I just am not doing that again so why bother writing about it right? I want to do s-bend seatstays from here on out (unless I can come up with a design that doesn’t weigh more, is less laterally rigid, and doesn’t take 5 times as long as S-bends).
Notes on bending butted tubes:
I rolled a 31.8 x 9/6/9 top tube in the 34.9 dies to ovalize it a little and i really dig that look. I also found it to be 11% stiffer laterally and 8% more vertically compliant (source: personal observation). I found some lumps at the butt transitions, nothing huge, but even when i rolled OVER them and didn’t stop at the butted parts of the tube they still had bumps. This appears to be an artifact of this roller or how I was rolling in a bigger die. I’m not sure you can even see them if you’re not staring at the frame through my eyes (which I could probably say of most of the ‘issues’ i rant about on this blog) but I’m pretty sure they’re actually there.
The non-heat treated top tube (Verus True Temper AVRDT i think) was super easy to bend versus the Supertherm seat tube which took some serious muscle to crank the driver bolt down.
The fusion weld needs to be pretty big since the force of bending can pop it when it’s not a good weld. I used some 0.045 to make a bigger puddle on the weld…so i guess it wasn’t really a fusion weld anymore and instead a butt weld. I found I had to use longer sections of 4130 (that you can re-use) to butt weld to the butted tubes so that I didn’t start/stop over a transition or butt weld. Lots of butts in this blog. Ok, that’s plenty of words for now. Sorry so long. I should learn to write less and blog more often.