Alfine 11 design
 

So the other day it occurred to me how Shimano arranged their planetary gearsets in the Alfine 11. (I might be the only person who finds this interesting.) Hopefully I am not repeating something that has already been posted. I've been goofing around with hub ratios in Excel and found http://john-s-allen.com/gears/hubratios.htm to be really helpful. Anyway, looks to me like they basically started with the design of a Nexus 7 and started improving things.

So, on Nexus 7 there are four sets of planetary gears, each one of which is only run "one way", that is, to increase or decrease the ratio, and not both. If we call these gears A, B, C and D we have:

1: 1/A
2: 1/B
3: C/A
4: C/B
5: D/B
6: C
7: D

Note that A ~= D and B ~= C.

Now for Alfine 11 we also have four sets of planetary gears, and A ~= D and B ~= C as before. They just allow 3 of the 4 to run both ways.

1: 1/A * 1/C
2: 1/A
3: 1/B
4: C/A
5: C/B
6: D/B
7: C
8: D
9: B*C
10: B*D
11: A*D

My spreadsheet can fairly closely replicate the 7 "derived" gears starting with the ratios of 2, 3, 7 and 8.

The Nexus 8 / Alfine 8 are totally different. They also have 4 sets of planetary gears, but obtain their ratios as folllows:

1: 1/A
2: B/A
3: C/A
4: D/A
5: 1.0
6: B
7: C
8: D

So there you have it, for what's it worth. Now, owners of the Alfine 11 can obsess about just how many planeteries are spinning around while they charge a hill.

So in your analysis, 7 of the 11 gears are dual stage compounded, and there's no direct drive.

I'd love to see an instrumented efficiency test on the Alfine 11.

Cool, I now know how my Alfine 8 is shifting. I had trouble visualizing the gears in the Alfine because the Shifting is much more complex than the Sturmey-Archer 3/5 speeds. I wish the direct drive was higher though for maximum cruzing efficiency, but that's not why I bought it in the first place.

But at least there seems to be no need for triple-compounded gears (such as 3rd and 5th on Rohloff).

I'm also very excited to be able to understand why the hub favors high ratios instead of low, and why there is that huge gap at the bottom. (They didn't reinforce "1/B" enough to take the torque from "1/C", and they didn't reinforce "1/A" enough to take the torque from "1/D"... only "1/C".) If they had reinforced A and B sufficiently, they could have made a symmetric 13 speed hub (with a 0.47 first gear, which is about 5th gear on a Rohloff).

Also you can sort of conclude that you can stomp on gear 2 all you like, but gear 1 is more of a risk because its actually the same final planetary stage at work, just with your torque increased by 29% on the input. (I assume that "A" and "B" are the second stage gear sets, because they typically increase torque.) Gear 3 might even be less strong than gear 2, since gear 2 is running a gear set that was significantly reinforced.

It seems that gear 2 it's not that strong:
http://forums.mtbr.com/internal-gear...cs-696480.html

Wow, that's a forum I don't think I'll try posting on.

Anyway I imagine that he just broke gear 2 ("1/A") because that is what he was using to muscle up some particular slope.

1st gear should really be placing more torque on "1/A" than 2nd gear does. I don't think Shimano would run the torque from "1/A" through "1/C" because if they where willing to do that, then it would have been no trouble to also run "1/B" through "1/C" and thereby eliminate the large gap between 1st and 2nd gears (and make a 12 speed hub). So I stand by my estimation that 1st gear should be weaker than 2nd gear.