I just wrote up a length post on the Hubbub Adapter and on the Rohbox with brifter option, and the software that this website uses lost it. So, I will just say that there are lots of other drop bar options for Rohloffs at:
http://www.cyclingabout.com/rohloff-...op-handlebars/

They screwed up the photo of the HubBub on that site, it is not wood. They instead show some DIY one and say that is the HubBub.

Photo of my HubBub is attached, I used a couple V brake noodles to re-orient the cable routing.

Another article on efficiencies.
What's The Difference In Speed Between Gearbox Systems? Rohloff, Pinion, Shimano - CyclingAbout

Alff's Siskiyou uses sliding dropouts for belt tensioning, and the disc brake mount must move with the wheel to maintain proper position, so it cannot be mounted to the seatstay (SS). The only way it could be chainstay (CS) mounted was if CoMo made their non-drive-side (NDS) dropout similar to Surly Troll/Ogre dropout.

Instead it appears CoMo uses a NDS dropout insert which extends upwards and attaches the brake caliper with a standard 51mm ISO mount. You can see some of this in the first Siskiyou video from OP at 4:11", 4:37" and 4:43" (pause video and manually advance for best image). The NDS dropout insert/brake mount is likely a single piece of machined Al attaching to the frame only with the two fixing bolts of the sliding dropout. Here's an image from Gates' website with the black dropout insert/brake mount hiding between the rotor and NDS frame. It appears to extend from the ISO mount down to the slider dropout, a single piece that is otherwise unconnected to SS or CS. I am speculating a bit since there's no good photo available to offer unambiguous proof, but that's the only way the brake caliper could move in concert with the rear wheel whenever belt tensioning is required. I believe CoMo's NDS dropout insert/brake mount is very much like this Paragon Machine Works sliding dropout insert (except the linked insert is post mount).

This particular Siskiyou has wide sliding dropouts for 12x142 through axle hub, a Pinion drive with a 54mm chainline, and Gates belt which evidently requires a larger (I counted 34T) rear Gates cog, probably to improve efficiency and ensure belt longevity. I imagine this forces the frame to be wider than usual at the dropouts. The hub/wheel appears to be a stock Industry Nine. I think the brake is in the usual position relative to the center of the rear axle, as on a QR bike, it's just that everything else (SS, CS, dropouts and rack leg), is located further outboard than usual. Also, CoMo's sliding dropout insert/brake mount locates the brake caliper closer to 12 o'clock instead of the more common 1:30-2:00 position. So this is how the brake caliper is mounted behind the rack leg, almost certainly a design decision to mitigate rack/brake caliper interference. I wonder if your heels come close to striking the chainstays on a Siskiyou Pinion/Gates, especially if you don't use clipless pedals.

I believe the eccentric bottom bracket + vertical dropouts + chainstay brake mount combination used by CoMo, Rodriguez, Thorn (like your Nomad), and others is a simpler, more elegant design - but it only works for an IGH. Pinion forces Como to use the sliding dropout. My guess is they make their own sliding dropout.

I didn't know Tubus was now made in Taiwan/China. I bet the change occurred when they picked up the Racktime product line 3-4 years ago.

Interesting they left derailleur drive (DD) out of their tests. To me the results are similar to the IHPVA evaluation, in that Rohloff is likely the most efficient gear hub around, but still a couple percent short of a DD, which is yet another 2-3 percent short of a singlespeed drive.

BTW, in the chain vs. belt evaluation linked in post #69, the belt tested in 2013 was likely not the current Gates Centertrack (the first Gates belt lacked the center-tracking feature which prevents belt wandering). Current Centertrack belts require less tension (28 lbf vs 85 lbf used for 2013 test) and are thus likely to be closer to chain drive WRT efficiency/power loss than presented in the evaluation.

Seeker333, thanks for elaborating.

You are right, I was thinking of an eccentric bottom bracket to adjust belt, but you are absolutely right that you can't do it that way with the Pinion. You can only do that with dropouts and the disc brake unit then has to move.

If the new belt design requires that much less pre-load, I can easily see that being much more energy efficient than the older belt.

I am going to stick with chains for my expedition bike. I run a different larger chainring for riding around home where the bike is not loaded down. That means I need to add a few links when I come home from a trip or take out a few links when I put on the smaller chainring and go somewhere. I do not have a mountain bike so I use my expedition bike for mountain biking too, use the smaller chainring for that.

I heard from a Rohloff belt user (I do not recall if older or newer belt) said that the chainline has to be right on with a belt. I wanted a Q factor (width of pedals) to be similar on my Rohloff bike as to my derailleur bikes. Thus my bottom bracket spindle is about 10 mm shorter than would be ideal for chainline. That means that I am running about a 5mm out-of-line chainline on my Nomad because I wanted to run a bashguard in the outer position on my double crank. I do not know what the tolerance is for chainline on a belt, but a chain that is 5 mm off is not a problem at all.