DISCUSSION DOCUMENT: Andre's Most Economical DIY Rohloff Drop Bar Gearshift

[Andre Jute]

Okay, sports fans, for making me a Hero of the Soviet Union, er, of the Thorn Forum, and since you insist on drop bars with your Rohloff hub gear box, here's a gearshift mechanism you can make in an hour for a couple of bucks in parts. With apologies for the quality of the sketch -- I was on my way to making a portrait when I had the idea, and had only charcoal with me to draw with, and a small corner of a well used sketchbook -- this is what it looks like:



This is a discussion sketch and document, not an engineering blueprint. The feasibility and dimensions of the device need to be discovered. Rohloff's standard rotary switch, for instance, feels like it handles a fistful of power; how much leverage will be required to match it? Eventually we'll discover whether my idea is a simple solution or merely a simplistic delusion of simplicity... Was it Einstein who said, A thing needs to be as simple as possible, but no simpler.

First off, you have to understand that the Rohloff internal gear hub doesn't use Bowden cables for the gearshift. There is no compression of the cable. The action is pull-pull, as in pull to shift down, and pull to shift up.

The components are made from a few inches of small U-section in ali for easy drilling, or steel if you have good tools, or you can bend it up from sheet if you prefer different dimensions. In addition you need, per side, a small spiral (helical) spring, a bolt, a nut, a couple of washers, and one or two O-rings or a couple of tiewraps per side. You can use the Rohloff adjusters you already have on the bike, or, since these are brake cables, use common brake adjusters. You will make two copies of Andre's Most Economical DIY Rohloff Drop Bar Gearshift.

The lever, A, you can shape any way you want. The important things about it is the position of the three holes. There are holes across it for a shaft to pivot on, and two holes in the top near one end. One hole is slotted to insert the cable, and one hole is scaled to hook your spring into.

The pivot B is hollowed out on the bottom to the shape of the handlebar and has two holes, one for the crossbar holding the lever (A), and one for the other end of the spring (C) in the crossbar left when you hollowed it out.. If want to use tiewraps rather than O-rings (E) to hold it on, you can make slots for the tiewraps.

The bolt, washers and nut are just a bolt, washers and nut suited to the width of the assembly and the diameter of the hole for the shaft.

Assemble the whole lot on your handlebars as per the sketch to the righthand side of my illustration, and attach the cable from the Rohloff to the hole in the forward end of the lever (A). Put one asssembly onto each side of your handlebars, so that you shift up on one side and down on the other side, just like on a derailleur bike. You have to press one of the levers once each time for each gear you want to shift in either direction.

There you go, a saving of at least £149.01 over some of the touted boutique versions, plus all the pride of having done the job right yourself. Or not. If not, an interesting experiment that will leave us all knowing more about our bicycles than before, and meanwhile keep us out of the bars.

Andre Jute
A few of my books in the UK and the US

"The brain of an engineer is a delicate instrument which must be protected against the unevenness of the ground." -- Wifredo-Pelayo Ricart Medina

[Danneaux]

Yes, I see your concept, Andre, and very clever it is. <nods> Yes, two levers would be required with the design, and -- because of the push-pull requirements -- one lever will retreat as the other advances.

Rather than a barrel at each lever, you're using them like teeter-totters to pull the required cable. Not mentioned but critical here is the placement of the pivot which determines the amount of cable pulled. Multiple pivot-points would allow a user-selectable change in leverage, so long as the total range of cable travel could be taken up...and that's where I'm a bit concerned.

The only fly in the ointment I can see is the considerable amount of cable travel required to shift through the Rohloff's entire range from Gears 1-14. Remember, the Rohloff shifter pulls huge amounts of cable. According to the Rohloff manual, cable travel is 7.4mm per shift or about 96.2mm total cable travel for the full shift range (13 shifts for 14 gears). This is the common source of bedevilment when creating a proper lever-type shifter whether it be a bar-end or a brifter. The barrel has to be so big compared to a derailleur's overall cable travel. To make something like a lever-type bar-end shifter for the Rohloff would require use of a "demultiplicator" as Huret once used near their bottom-bracket clamps (to change travel through use of an intermediary change in moment-arm length) or like the "Travel Agents" to make road-bike brake levers work with v-brakes.

Mr. Cresswell handled this in the Ovo Tandem shifter with a switchable advancing/retreating rachet that can "bite" as many as two gears at a time before needing to reset so it can grab as many as two more gears, and so on. The ratchet is what gives back the travel of a rotary shifter within the range of a lever -- it handles the task in little chunks (7 for the entire range at 2/shift (14.8mm per double throw)...14 if one goes a gear at a time (7.4mm per single shift). The Rohloff rotary shifter is as large as it is because of the considerable amount of cable it must pull each way.

I think your idea might work alright for several shifts, Andre, but I'm a little concerned it might not pull the large amount of cable needed to shift through the entire Rohloff range. To do so, one would need to place the cable and spring anchors at the other end of each shift lever...but that would result in less leverage available at the fingers and would make shifting effort pretty high.

I sure do salute the simplicity of the design -- it is mathematically elegant and incredibly straightforward; some real solid thinking there, Andre, and I congratulate you. I must ponder this further. Like many terrific ideas, it may be yours takes awhile to soak in before I can fully appreciate the entire scope of what you envision. I'm just tickled you have brought your considerable intellect to bear on a problem long facing us roadies. This is a real treat to ponder; many thanks!

All the best,

Dan. (...who always appreciates the elegance of approaching complex ideas simply -- how the best solutions are often found and executed)

[Andre Jute]

Thanks for the analysis, Dan. I have pedalpals waiting to go cycling, so I'll have to consider your individual points later. Meanwhile, I've added a par that was left off my original post that should make my intention  clearer:

This is a discussion sketch and document, not an engineering blueprint. The feasibility and dimensions of the device need to be discovered. Rohloff's standard rotary switch, for instance, feels like it handles a fistful of power; how much leverage will be required to match it? Eventually we'll discover whether my idea is a simple solution or merely a simplistic delusion of simplicity... Was it Einstein, or Edison, who said, A thing needs to be as simple as possible, but no simpler.

Andre Jute

[Danneaux]

Wonderful, Andre! Looking for'd to the discussion. Already greatly enjoying your topic and the idea you've put forth as a springboard for discussion. Well done!

All the best,

Dan.

[Andre Jute]

Okay, we have to provide for nearly 100mm of lever movement, a fraction short of 4 inches. Try this, which combines the two levers onto a single bracket, though it is still possible to split them to the left and right bars by making two brackets.



The levers are now flat sheet with a small bent tag for attaching the gearchange cable, and the spring if required. But this design could do without the spring I think, though the hole is there if necessary. The tag can also be soldered, riveted or bolted on. The thumbwheel is serrated to add grip. (A girlfriend drove a Ferrari on which Pininfarina had fitted "aerodynamic" hidden doorhandles made of flat sheet with just such serrations to push with a thumb to make the handles appear. If you pushed too far, the door would pop open just from the thumb-leverage on the serrations.) The whole assembly can be very narrow, which is important on drop bars with their general lack of useful space.

Andre Jute

[John Saxby]

Andre, that's heroic, as we've come to expect.  Thanks as always for your public service, and your applied thoughtfulness.

Another epigram seems à propos, which I heard/read somewhere long ago, also from engineering practice: "The first example of superior principle is usually defeated by established practice." (It might have been about Lotus' monocoque construction, or possibly Vincent's trick motorcycle-without-a-frame.)

While you've been sketching creatively, I've been trawling the internet to see what other options there may be for mounting the standard Rohloff shifter on, or close to, drop bars.  My search was prompted by the Ovo/Cresswell ratcheting shifter, and Dan's suggestions on mounts that could locate that close to drop bars.  Haven't quite found the Holy Grail yet--still rummaging around in profane rather than sacred territory--but, a little to my surprise, there may be some new-ish options in the offing. Will assemble a report when I have enough info for a report.  Price will likely be a little higher than £1.00 or so, but a long way short of £150.

I've worked on the assumption that the Rohloff shifter seems to work quite well, and that The Problem--for Drop Bar habitué(e)s at least--seems to be its 22.2 mm anchor.

J.