I’ve noticed something on a couple of frames I’ve built and I’m trying to work out whether it’s coming from my chainstay setup or the dropout geometry.
I’m using PMW DR0103 UDH dropouts with a 12 × 142 mm rear end. I’ve been using the same chainstay profile on both sides and setting the BB ends an equal distance either side of the frame centreline.
With the wheel correctly centred in the frame, however, I’m consistently ending up with the drive-side chainstay noticeably closer to the tyre than the non-drive side.
The stays are equally spaced at the BB, so I’m wondering whether I’m missing something at the dropout end.
I know the UDH interface changes the geometry on the drive side compared with a conventional dropout, but what I don’t know is how the DR0103 handles that at the actual chainstay attachment point.
Does the DR0103 place the drive-side and non-drive-side chainstay attachment points symmetrically about the frame centreline when correctly fixtured, or is there an offset between them that needs to be accounted for when laying out the stays?
Alternatively, if the dropout chainstay positions are symmetrical, then presumably I’m introducing the difference through the bend, phasing or orientation of the stays themselves.
Has anyone built with the DR0103 and seen this? I’m basically trying to understand whether using identical stays, equally spaced at the BB, should give roughly equal tyre clearance with these dropouts, or whether the two sides need to be treated independently. (couple of images of the issue and a fancy short of one of the bikes :))
Thru axles are really challenging with metal. Your miters have to be tight, weld sequence has to be dialed, chakra aligned etc etc.
If your wheel is off center by 1mm (measured as 2mm), that means you have a problem of about .3mm at the dropout - which isn’t very much. In old days, we could file the dropout a few strokes to correct this but with thru-axles, you can’t really do that. The Syntace system had an eccentric that helped fix it.
Without putting hands on your fixture, alignment tools, seeing your jointing etc, it’s hard to say from here. I can say you’re looking for a very small problem.
If it’s consistently the same error, that gives some clues to mitering process or fixture alignment.
It’s only happening on frames where I’ve used the PMW UDH dropouts. I’ve built with Polydrops and other dropouts using the same fixtures and don’t see the issue — if anything my mitering/process has improved since those earlier frames.
The wheel itself is centred in the frame and the bikes ride perfectly normally. The issue is really aesthetic: with a centred wheel I have noticeably more tyre-to-chainstay clearance on the NDS than the DS, which makes the rear triangle look as though it’s out of alignment even though the wheel isn’t.
I can make the gaps visually equal by moving the NDS chainstay inward at the BB rather than setting the two chainstays equally either side of BB centreline. That works, but it feels slightly counterintuitive, which is why I’m wondering whether there’s something inherent in the PMW UDH dropout/stay geometry that means a symmetrical BB-end layout doesn’t produce symmetrical tyre clearance or I’m totally missing something
you can actually see the issue the bikecad design with the dropout:
Gotcha, I’m not familiar enough with those dropouts to comment. When I add them to a BikeCAD model, without any tubing alignment adjustments, I don’t see the offset you show. I also exported to FreeCAD and don’t see the offset. I wonder if you’re doing something at the tube/dropout join to cause it. Interesting problem!
Thanks for taking the time for looking into that, appreciated! Must be something I’m doing while phasing the CS? My model is an approximation of the bends in the stays, i’ve been print a full size template of the rear end and sorting fit that way…maybe i need a more accurate why of modeling the Reynolds double bend stays?
At a bit of a loss, just in the middle of another build with these dropouts, would nice to figure it out.
Are you using pre-bent stays? Are you dimpling (just) one of the stays for chainring/crank clearance? If so, that may be your problem. Dimpling the stay anywhere near the bend will change the profile of that stay (straightening it out).
Yes, using pre-bent stays. I dimpled stays myself on one bike that has the issue and then on the next i used pre-bent and pre-dimped to see it that was the issue, but its arguably worse on the 2nd bike.
I’ll have a proper look at the jig calibration tomorrow, just to double check…but Cobra’s jigs are pretty accurate??
I would measure the offset from the centreline of each chainstay at the dropout end to the actual face of the dropout that the wheel sits against. A difference there is the only explanation I can think of, assuming the wheel is dished such as to centre it between those two points.
It seems odd that the biggest offset is so close to the BB (which we can easily assume is on centre and the stays are spaced equally on it). I feel like the tyre gap difference would be much smaller if the problem was caused by the dropout. Seems to me like the issue should be either in the stays themselves, or maybe the BB alignment.
And at the risk of teaching you to suck eggs, have you done a full sanity check? Is the BB actually centred and symmetrical? Are the stays actually the exact same shape? Is your wheel actually properly dished? I don’t know how many times I’ve had a head-scratcher like this and triple-checked everything, only to find I needed to actually check everything that I “knew” was already right.
Sucking eggs is my forte! I’ve been chasing my tail all day, even drove around and bought a laser
But I think the solution is so stupidly simple it’s embarrassing…if I’m correct? Bonus, I’ve verified that my jig is aligned!
I’ve added some photos to illustrate the findings when added the chainstays against a known reference of the bb edge. The big clue is the significant different lengths of the dropouts and its effect on the diagonal. See if you guys draw the same conclusion as me (the measurements are +/- 1mm or so, my conclusion below):
Spoiler: Because the NDS dropout is so much longer you have to move the CS out of a mirrored position to give sufficient space to slot the dropout.
I can’t place that stay in the true mirrored position of the opposite side. I have to move it roughly 6–8 mm farther outboard. That then carries through the whole chainstay path and gives noticeably more tyre clearance on that side, even though the wheel itself is centred and the frame aligns and rides correctly.
So the apparent “misalignment” is really being introduced by the asymmetric dropout geometry rather than by the fixture or the rear wheel…now time to fit it cause it looks crap
Fixes?:
1, Bend the chainstay, instead of moving out of mirror
2, Move the chainstay inboard
3, Smaller/no dipple on the NDS chainstay
I’m more inclined to do 2, sound like a good plan?
Oh hey we had that issue last week, the easiest fix is to rotate the nds chainstay inwards at the bb, so its closer to the centerline at the shell and towards the outside at the dropout, but at the right distance at the tire. If you’re using the cobra jig you’ll have to cut them one at a time if you’re using a double bend tube, its a bit finicky. You can also do asymetric dimples afterwards if you mess up by a few mmn
