2_i 

I am now 4 days into riding on Brompton in my hilly area. I can tackle hills that I had to walk up last year when riding a local MTB. Photo belows one particularly challenging one - I could actually start from rest the parked bike and not employ any sort of zigzagging - it is actually a busy street.

In the noise of the day you cannot actually hear any cage rubbing. The only hiccups I encountered were in downshifting, particularly from middle ring to small. You cannot do it when the chain is strongly tensioned in an uphill ride - you have to slightly relax it, which you get used to pretty quickly.

In terms of outlook, seeing how things worked at different stages of the project, I know how to evolve the derailleur cage to reduce rubbing without losing current benefits. However, I will presumably start on a new cage and keep the current derailleur as a reference. I get them used so they are not overly expensive.

I looked at the yaw technology more and am somewhat puzzled by the arguments used to promote it. In that technology the cage changes alignment with the rings as you shift gears, to better match chainline direction. After all, though, chain passes through different portions of the cage for different rings - it is not a problem to deform the cage in those different regions to change alignment, making any special pivoting provisions unnecessary. I think such (and other) deformations are employed in MTB triple derailleurs. I.e. the yaw either brings benefits for road doubles, with rather primitive cages, or simply changing pivoting is not a big deal, so just as well might be done. Anyway, I will look into that later.

fietsbob

My Option , 2 speed mountain drive gearbox crank works with minimum complication.
'21.6t' and a 54.. 3 speeds used twice , but spinning that 17" low gear up a hill ,
heart rate up there, walking is just easier.. I feel no shame walking..




....

2_i 

Observations during 4 weeks in the mountains were as follows. 1. Without the triple I would not have managed. During the first week I still kept the option of a local MTB alive, but the Brompton w/triple was holding on and proving itself superior. 2. Double derailleur shifting over 3 rings was a stretch that left so narrow tolerances that every few days I needed to readjust my shifting. 3. Litepro P-frame adapter was getting deformed even with my stiffening - it lasted though longer before doing so, one week rather than minutes. 4. Pushing the inner ring close to the frame made it interfere with rear triangle when folding the bike.

It was time to go back to the drawing board. I bought the C.M. Wasson mount from @fietsbob. It basically amounts to two halves of a pipe with standard derailleur mount and cable stop braze on to one of these. The steel mount is obviously much more sturdy than poorly designed alu Litepro. However it turned out not to be all roses. The angle for the mount was all too steep and I had to bend it a little and employ the asymmetric Sram washer to get a right angle for the derailleur. I used Suntour XC Pro derailleur. It might have worked fine of its own but I could not stop myself from making the same enhancements in the cage as for the double, increasing the lower lip of the back plate of the cage and effectively pushing the middle of front plate in, improving downshift from the middle ring and upshift from the small to the middle ring. The rings are now 50-32-20 mounted on a Shimano 400CX right crank. The left crank remains the original Brompton. The BB spindle is now 122mm long and the BB is pushed to the right by 3.5mm worth of BB spacers. This pushes the fold a bit but I did not need to cut the derailleur cage.

Everything is very smooth and stable now. If it remains for a year or so, I plan to replace the Wasson (one-time owner of Foldabikes in Palo Alto) mount with a braze-on. Video shows the current operation


One more mount that I plan to explore is a Litepro Birdy mount.

tomtomtom123

The problem with upshifting is the syncing between the amount of FD movement and the gap between chainrings. On my Dahon, I initially had a problem getting my double FD to upshift the chain on a double crank, but after a lot of fiddling around with the angle of the FD, amount of cable pull, and the spacing between chainrings, I got it working fine. Sometimes the cable will slacken or the FD will shift after 1 month, but then I just turn the cable adjuster half a turn and it's back to normal.

Regarding the stability of the braze-on adapter, the one I 3D printed in bronze/steel is heavily overbuilt so it doesn't budge at all (it's starting to corrode though). But I do notice that the FD bends a little when upshifting, but maybe it's normal. It might also be that the änti-twisting screw in the FD isn't centered on the seat tube so it doesn't get good leverage. I'm not sure if my usage of the long arm FD 4700 makes any difference. The cable pull is much longer so it requires less tension in the cable, but the spring force might be the same to the short arm Shimano FDs.

Also having the 2 chainrings both on the inner side of the flange of my double crank allows me to add shims to either adjust them closer together or further apart. I spaced my chainrings 8mm center to center. 0.2mm more and the chain would not upshift. 0.2mm less and the chain will rub on the outer chainring when in 7th gear of the inner ring. The FD 4700 had a little bit too much cable pull so it would overshoot, so I slackened the cable to reduce the amount of FD movement, in combination of widening the rear part of the cage to prevent chain rub due to the short chainstay and sharp angles.

2_i 

In upshifting, it is the back plate of the cage that pushes the chain. The plate is more effective when it is wider, but the width is limited by the need for the cage to clear the smaller rings. In my case, there are big differences between the rings, so I can afford a far wider back plate than in more typical setups. In the other direction, the chain gets pushed off a ring with the front plate of the derailleur cage. My middle ring is way smaller than anticipated by the manufacturer, requiring a lower part of the front plate to work with the middle ring than anticipated by the manufacturer. For pushing the chain down, the plate protrudes into the interior and I extended that protrusion to a lower part of the cage. After these tweaks to the cage, the shifting became completely uneventful even with my humongous differences between the rings. Cable tension of course matters, but I find of lot of leeway with that tension, with shifting remaining satisfactory over a wide variation in that tension. Rubbing against the cage changes sooner than the shifting with the tension.

I still suspect that your 3D printed mount tweak would work for the triple nearly as well as the Wasson mount. The Birdy mount, that I wait for, likely puts the derailleur farther away from the seatpost than necessary.

Now I begin to ponder whether I should apply lessons from Brompton to other bikes.

tomtomtom123

My double is a 46/26 or 44/24, so a 20T jump. 16 and 18 jump shifts with much greater tolerances for misalignment. With 20 jump it has to be within 1mm of a certain spot. I suspect a 22 jump would be difficult. Also the chain catcher that I'm using is only long enough for 20 jump.

fietsbob

Fabricating another steel FD mount Made to still be in a direct 73 degree alignment with a ghost seat tube running into the BB axis

Adapting .. Because the makers of FD, are all aiming at the massive OEM market in non folding bikes.. ..

by placing it ahead of the actual seat tube on the frame parallel to the existing frame seat tube...

Moving the FD Forward, keeping it on the same quadrant of the chain ring, as if on a regular bike (11:00)

Rather than an angle much lower ( like 10:00) , as so many of your pictures show , can be done...


Brazing jig would be a start .(a Metalworking background, rather than IT). to hold the parts in place before brazing is done..







....