Alfine 11; how it works (long)
Posted: 11 Jan 2016, 3:23pm
The A-11 hub has been with us a few years now, and there is a fair amount of information out there. However unless I'm mistaken no-one has yet put in simple terms how the hub generates the 11 gears, which parts are used in which gears, and collected much of the available information in one place. This is (for good or ill) my stab at it. I've put it separate from the long-running A11 reliability thread mainly so that people can find it. Please let me know of any errors or omissions and I'll correct accordingly.
In a nutshell;
The hub is best thought of as a two speed (reduction or direct) gear which drives a six-speed (direct or increase, 5 options) gear.
The six-speed gear is itself a 3x2 gear, i.e. an intermediate three speed (direct or increase, 2 options) gear which in turn drives a further two speed (direct or increase) gear.
The gear ratios are selected via four selectable clutches (three lockable sun pinions, one sliding clutch) and four (automatically selected, spring preloaded) roller clutches. The eleven available gear ratios all use at least one gear train, and several gears use three gear trains in series.
The gearsets could give twelve ratios (2x6) of which only eleven are used, because one ratio is a near duplicate of another. However, there is no direct drive gear, instead a gear that gives almost the same gear ratio is used, despite the fact that this gear uses three gear trains, and would therefore presumably be less efficient.
__________________________________________________________________________________________________
The detail:
The first gear train is similar to that found in an A8 or N8 hub, in that it uses a sliding clutch (C1) to provide direct drive to the planet cage (PC1). C1 is actuated by an axial cam. When C1 isn't locked, a roller clutch (RC1) takes drive from the driver to the ring gear R1 instead. The output is always from the planet cage of this gear stage PC1. The gear ratio is the same as that found in an N8/A8 hub, the tooth counts (including stepped planet gears) are the same also, but the A11 first stage gear teeth are cut helical instead of straight-cut. The rest of the gears in the A11 hub are straight-cut.
There are four sun pinions, of which three are selectably locked. The sun pinion in the first gear train (S1) is permanently locked; the remaining suns (S2,S3,S4) are locked by pawls that pop out of the axle as per N8,A8 ( and the latest N7) hubs.
The intermediate three-speed gear has two suns (S2,S3), stepped planet pinions, and a single ring gear (It works a little like the three high gears in a SA 5s hub). The drive to this stage is input via the planet cage (PC2) from PC1. When neither sun pinion is locked, a roller clutch (RC2) transfers drive from PC2 to the ring gear R2.
R2 in turn drives the planet cage (PC4) of the final two speed gear stage. When sun S4 is locked, the ring gear R4 drives the hubshell via a roller clutch RC4. When S4 is not locked, the R2/PC4 assembly drives the hubshell directly via roller clutch RC3.
Clutch C1 is locked in gears 7-11.
The S2-S4 clutching pattern is the same in gears 2-6 as 7-11
Gear 1 uses a unique S2-S4 clutching pattern that would give direct drive gear when C1 is locked.
In the A11 hub all the planet pinions run on roller bearings (as per Alfine 8, premium Nexus 8 and Rohloff hubs), so as to decrease losses.
Once you know what it is that you are looking at, you can see nearly all the relevant parts in this sectional view
-------------------------------------------------------------------------------------------------------------------------------------
Gear configuration table:
Clutches___C1___S2___S3___S4___Hubshell Drive__Active Gearsets___Active R-Clutches__Overrun R-Clutches
Gear
11_________L_________L____L___RC4______________-,-,3,4_________RC4_____________RC1, RC2, RC3
10_________L____L_________L___RC4______________-,2,-,4_________RC4_____________RC1, RC2, RC3
9__________L_________L________RC3______________-,-,3,-_________RC3_____________RC1, RC2
8__________L____L_____________RC3______________-,2,-,-_________RC3_____________RC1, RC2
7__________L______________L___RC4______________-,-,-,4_________RC2,RC4_________RC1, RC3
6____________________L____L___RC4______________1,-,3,4_________RC1,RC4________RC2, RC3
5______________L__________L___RC4______________1,2,-,4_________RC1,RC4________RC2, RC3
4____________________L________RC3______________1,-,3,-_________RC1,RC3________RC2
3______________L______________RC3______________1,2,-,-_________RC1,RC3________RC2
2_________________________L___RC4______________1,-,-,4_________RC1,RC2,RC4____RC3
1_____________________________RC3______________1,-,-,-_________RC1,RC2,RC3__________
Notes;
a) Sun pinions are labelled from right to left in the hub
b) if the hub behaves oddly or makes noises in some gears and not others, the table above can help to identify which parts may be responsible.
c) when RC1 and RC2 are not being used to drive, they are overrun, which introduces a small parasitic drag.
d) when S2 or S3 are not locked, they turn (usually backwards) on the axle, which does not introduce an appreciable parasitic drag
e) when S4 is locked RC3 is overrun (giving a small parasitic drag)
f) when S4 is not locked (and RC3 takes drive to the hubshell) RC4 is not overrun; instead S4 turns on the axle with low parasitic drag.
---------------------------------------------------------------------------------------------------------------------------------------
Tooth counts (thanks to jb for checking these by counting.... and counting again...!)
train 1 S1 =48 Ps1 = 14 Pr1 =22 R1 = 84 (14/22 stepped planets)
train 2 S2 = 36 Ps2 = 20 R2 = 78
train 3 S3 = 39 Ps3 = 15, Pr2 = 20 (15/20 stepped planets)
train 4 S4 =39 Ps4 =24, Pr3 =14, R3 = 78. (24/14 stepped planets)
___________________________________________________________________________________
Gear Ratios in each stage (from tooth counts)
Stage 1:
1/(1 + (48x22)/(14x84)) = 0.527 or,
1.000 (direct)
Stage 2:
using S2 (1 + (36/78)) = 1.462 or,
using S3 (1 + (39x20/15x78)) = 1.667 or,
1.000 (direct)
Stage 3:
(1 + (39x14/24x78)) = 1.292 or,
1.000 (direct)
__________________________________________________________________________________
Gear Ratios (as products of gear stages)
Gear: Stage 1 x Stage 2 x Stage3 = total ratio
11: 1.000 x 1.667 x 1.292 = 2.153
10: 1.000 x 1.462 x 1.292 = 1.888
9 : 1.000 x 1.667 x 1.000 = 1.667
8 : 1.000 x 1.462 x 1.000 = 1.462
7 : 1.000 x 1.000 x 1.292 = 1.292
6 : 0.527 x 1.667 x 1.292 = 1.134
5 : 0.527 x 1.462 x 1.292 = 0.995
4 : 0.527 x 1.667 x 1.000 = 0.878
3 : 0.527 x 1.462 x 1.000 = 0.770
2 : 0.527 x 1.000 x 1.292 = 0.681
1 : 0.527 x 1.000 x 1.000 = 0.527
(give or take some rounding errors)
_______________________________________________________________________________________
Percentage intervals
Gear Ratio upshift% downshift % (Total Difference 409%)
Gear 1 0.527 ______ -22.6%
Gear 2 0.681 +29.2% -11.6%
Gear 3 0.770 +13.1% -12.3%
Gear 4 0.878 +14.0% -11.8%
Gear 5 0.995 +13.3% -12.3%
Gear 6 1.134 +14.0% -12.2%
Gear 7 1.292 +13.9% -11.6%
Gear 8 1.462 +13.2% -12.3%
Gear 9 1.667 +14.0% -11.7%
Gear 10 1.888 +13.3% -12.3%
Gear 11 2.153 +14.0% _____
average +15.2% -13.1%
__________________________________________________________________________________
Cable pulls;
There are some measurements in the A11 reliability thread but in essence all shifts are a nominal 5.0mm except for 6-7 which is longer (around 6.4mm) and 10-11 which is also longer, presumably so that the cable will run full slack in gear 11 even if it is a little draggy. Other than A11-specific shifters ( Shimano, J-tek, Versa), there are no suitable alternatives.
________________________________________________________________________________________
Comments:
I note with interest that once again (as in Nexus 7) shimano have avoided having a nice efficient direct drive gear 5 (or 6) (which would have been 'L---' clutching pattern). As it is the hub uses a single gear train in gears 1,7,8,9, thus these are likely the most efficient gears. Two gear trains are used in gears 2,3,4,10,11, (medium efficiency) and three are used in gears 5 and 6 (probable lowest efficiency). [NB Under very low loads the parasitic losses from the overrun roller clutches may be more apparent than the gear train (meshing + bearing) losses.]
If the gear ratios are set in a vaguely sensible fashion (i.e. so that the 'flat road, no wind' gear is about 70-80% of top gear) then gears 7,8,9 (which are relatively efficient) are likely to be the most used gears. On a heavily loaded touring bike gear 1 is likely to be heavily used too. This being the case the weighted average efficiency of the gear may not be as bad as the numerical average efficiency might suggest. However there is a question mark against gears 1-6, efficiency-wise, in that when the stage 1 gear is in use, there are axial thrust loads that must be supported by non-rolling element bearings.
Some folk have suggested that there ought to be more gears yet with different clutching (because there are sixteen possible permutations of four clutches) but IMHO this is not the case; S2 and S3 cannot be usefully locked at the same time (so the four '*LL*' clutching permutations are forbidden/ineffective), so the 'L---' clutch pattern is the only allowable and unused permutation.
In theory it should be nigh-on impossible to lose all drive in this hub, provided the gears are not jammed and you can still select C1, and RC2/ RC3 are still working, you should be left with a viable direct drive gear even if you can't lock S2, S3, S4 etc.
____________________________________________________________________________________
There are some very helpful instructions and photographs (prepared by jb) in the midst of the long running Alfine 11 thread which can be found here
http://forum.ctc.org.uk/viewtopic.php?f=5&t=64432&start=374
[NB; the sun locking pawls are labelled from left to right in jb's notes.]
Below these there are some diagrams which show the drive path through the hub in the various gears.
On subsequent pages there are several other useful photos and posts.
______________________________________________________________________________________
Known faults;
- the shift control is very sensitive to variations in build tolerance and adjustment. The reason for this is that the cassette joint may (because the RH hub cone is supported on two prongs, leaving gaps that the selector mechanism pokes though) only rotate about 135 degrees total, i.e. about 15 degrees per shift, tops. This makes building the hub accurately very difficult for Shimano and it also makes it sensitive to the exact external setting when in use. Unfortunately with new hubs the cable usually settles and this throws the external adjustment; worse still the build tolerances are such that the 'correct adjustment' (dots lined up in gear 6 http://www.sheldonbrown.com/nexus-mech.html) is not always the 'optimum adjustment'. 'Rofan' (see A11 reliability thread) has reported that the shift control sleeve is lighter built and longer than the equivalent A8/N8 version and this presumably makes it more likely to give trouble through wear or high shift load.
- The helical stage 1 gears were presumably added at a late stage ( the EV techdoc shows straight cut gears), I suppose to make the hub nice and quiet. However this generates an axial thrust load which can cause problems with retention of S1 on the axle (as jb has experienced in the A11 reliability thread). Arguably whoever designed the gear train screwed up here, in that when you pedal hard, the sun pinion S1 is forced rightwards i.e. against the retaining clip, rather than leftwards against a fixed shoulder. Maybe there is a good reason for this, but if they had cut the helical gear the other way (or left it straight cut) this fault may not occur in the same way or perhaps at all. If you routinely give the hub a lot of stick in gears 1-6 this is what is likely to generate this kind of problem.
-numerous users have reported skipping in some gears. This can happen even if the sun locking pawls are fully extended (in new hubs or hubs where there are burrs on the pawls) but it certainly will happen if the sun locking pawls are not fully extended. If the build tolerances of the hub are off a bit, the hub can therefore slip even when 'correctly adjusted'. The same thing is seen in numerous N8/A8 hubs. Many users have found that a little trial and error (with adjustments that are up to about 1.5mm away -in either direction- from the 'correct adjustment') can yield a more reliable setting. However if the axle assembly is examined when out of the hub the pawl lift (and therefore correct adjustment) can be checked directly. [I have seen several N8/A8 hubs where the C1 shift was not synchronised with the pawl lifts in the correct way. I do not know if the same fault occurs with A11 hubs. Note also that the sun locking pawls in a A11 hub are less likely to be overrun than in an N8/A8 hub, which means that any burrs on the pawl tips are unlikely to be knocked off when that pawl is not transferring drive. Hard (i.e. highly loaded) slips can raise burrs on, or even shatter pawls.]
- the clearances in all shimano IGHs are pretty tight; if there is any shrapnel in the hub (particles bigger than ~0.25mm) it will probably cause one or more gear trains to jam and/or be damaged. Stage 2 and stage 3 could jam entirely and you would still have some working gears, but if stage 1 starts to jam you are in trouble because the sun S1 is locked to the axle and those gears must turn at all times when you are pedalling.
- these hubs often leak oil. It isn't clear where the hub is meant to be vented; it seems to me that some leakage (most often on the RHS) is more or less inevitable.
--------------------------------------------------------------------------------
overall;
The Alfine 11 hub is (internally) only a tiny bit more complicated than a N8/A8 hub, and apart from the helical stage 1 gears there is nothing much new and different inside it per se vs the 8s hubs. So, with one or two caveats the internals ought to be reasonably reliable; an intrinsically 'bad hub' ought to become evident within the warranty period.
However the efficiency is unknown/suspect, the helical gears are probably not the best choice for high torque use in low gears, and a significant proportion of users report skipping or other problems, which can arise for several reasons.
Overall the A8/N8 hubs are both stronger and cheaper, so if they have wide enough ratios for your terrain, they probably make a better choice for any hard use. This is despite the fact that the grease lubrication in these hubs is far from perfect; running them in oil (or highly fluid grease, not shimano's) seems like a good idea to me.
cheers
In a nutshell;
The hub is best thought of as a two speed (reduction or direct) gear which drives a six-speed (direct or increase, 5 options) gear.
The six-speed gear is itself a 3x2 gear, i.e. an intermediate three speed (direct or increase, 2 options) gear which in turn drives a further two speed (direct or increase) gear.
The gear ratios are selected via four selectable clutches (three lockable sun pinions, one sliding clutch) and four (automatically selected, spring preloaded) roller clutches. The eleven available gear ratios all use at least one gear train, and several gears use three gear trains in series.
The gearsets could give twelve ratios (2x6) of which only eleven are used, because one ratio is a near duplicate of another. However, there is no direct drive gear, instead a gear that gives almost the same gear ratio is used, despite the fact that this gear uses three gear trains, and would therefore presumably be less efficient.
__________________________________________________________________________________________________
The detail:
The first gear train is similar to that found in an A8 or N8 hub, in that it uses a sliding clutch (C1) to provide direct drive to the planet cage (PC1). C1 is actuated by an axial cam. When C1 isn't locked, a roller clutch (RC1) takes drive from the driver to the ring gear R1 instead. The output is always from the planet cage of this gear stage PC1. The gear ratio is the same as that found in an N8/A8 hub, the tooth counts (including stepped planet gears) are the same also, but the A11 first stage gear teeth are cut helical instead of straight-cut. The rest of the gears in the A11 hub are straight-cut.
There are four sun pinions, of which three are selectably locked. The sun pinion in the first gear train (S1) is permanently locked; the remaining suns (S2,S3,S4) are locked by pawls that pop out of the axle as per N8,A8 ( and the latest N7) hubs.
The intermediate three-speed gear has two suns (S2,S3), stepped planet pinions, and a single ring gear (It works a little like the three high gears in a SA 5s hub). The drive to this stage is input via the planet cage (PC2) from PC1. When neither sun pinion is locked, a roller clutch (RC2) transfers drive from PC2 to the ring gear R2.
R2 in turn drives the planet cage (PC4) of the final two speed gear stage. When sun S4 is locked, the ring gear R4 drives the hubshell via a roller clutch RC4. When S4 is not locked, the R2/PC4 assembly drives the hubshell directly via roller clutch RC3.
Clutch C1 is locked in gears 7-11.
The S2-S4 clutching pattern is the same in gears 2-6 as 7-11
Gear 1 uses a unique S2-S4 clutching pattern that would give direct drive gear when C1 is locked.
In the A11 hub all the planet pinions run on roller bearings (as per Alfine 8, premium Nexus 8 and Rohloff hubs), so as to decrease losses.
Once you know what it is that you are looking at, you can see nearly all the relevant parts in this sectional view
-------------------------------------------------------------------------------------------------------------------------------------
Gear configuration table:
Clutches___C1___S2___S3___S4___Hubshell Drive__Active Gearsets___Active R-Clutches__Overrun R-Clutches
Gear
11_________L_________L____L___RC4______________-,-,3,4_________RC4_____________RC1, RC2, RC3
10_________L____L_________L___RC4______________-,2,-,4_________RC4_____________RC1, RC2, RC3
9__________L_________L________RC3______________-,-,3,-_________RC3_____________RC1, RC2
8__________L____L_____________RC3______________-,2,-,-_________RC3_____________RC1, RC2
7__________L______________L___RC4______________-,-,-,4_________RC2,RC4_________RC1, RC3
6____________________L____L___RC4______________1,-,3,4_________RC1,RC4________RC2, RC3
5______________L__________L___RC4______________1,2,-,4_________RC1,RC4________RC2, RC3
4____________________L________RC3______________1,-,3,-_________RC1,RC3________RC2
3______________L______________RC3______________1,2,-,-_________RC1,RC3________RC2
2_________________________L___RC4______________1,-,-,4_________RC1,RC2,RC4____RC3
1_____________________________RC3______________1,-,-,-_________RC1,RC2,RC3__________
Notes;
a) Sun pinions are labelled from right to left in the hub
b) if the hub behaves oddly or makes noises in some gears and not others, the table above can help to identify which parts may be responsible.
c) when RC1 and RC2 are not being used to drive, they are overrun, which introduces a small parasitic drag.
d) when S2 or S3 are not locked, they turn (usually backwards) on the axle, which does not introduce an appreciable parasitic drag
e) when S4 is locked RC3 is overrun (giving a small parasitic drag)
f) when S4 is not locked (and RC3 takes drive to the hubshell) RC4 is not overrun; instead S4 turns on the axle with low parasitic drag.
---------------------------------------------------------------------------------------------------------------------------------------
Tooth counts (thanks to jb for checking these by counting.... and counting again...!)
train 1 S1 =48 Ps1 = 14 Pr1 =22 R1 = 84 (14/22 stepped planets)
train 2 S2 = 36 Ps2 = 20 R2 = 78
train 3 S3 = 39 Ps3 = 15, Pr2 = 20 (15/20 stepped planets)
train 4 S4 =39 Ps4 =24, Pr3 =14, R3 = 78. (24/14 stepped planets)
___________________________________________________________________________________
Gear Ratios in each stage (from tooth counts)
Stage 1:
1/(1 + (48x22)/(14x84)) = 0.527 or,
1.000 (direct)
Stage 2:
using S2 (1 + (36/78)) = 1.462 or,
using S3 (1 + (39x20/15x78)) = 1.667 or,
1.000 (direct)
Stage 3:
(1 + (39x14/24x78)) = 1.292 or,
1.000 (direct)
__________________________________________________________________________________
Gear Ratios (as products of gear stages)
Gear: Stage 1 x Stage 2 x Stage3 = total ratio
11: 1.000 x 1.667 x 1.292 = 2.153
10: 1.000 x 1.462 x 1.292 = 1.888
9 : 1.000 x 1.667 x 1.000 = 1.667
8 : 1.000 x 1.462 x 1.000 = 1.462
7 : 1.000 x 1.000 x 1.292 = 1.292
6 : 0.527 x 1.667 x 1.292 = 1.134
5 : 0.527 x 1.462 x 1.292 = 0.995
4 : 0.527 x 1.667 x 1.000 = 0.878
3 : 0.527 x 1.462 x 1.000 = 0.770
2 : 0.527 x 1.000 x 1.292 = 0.681
1 : 0.527 x 1.000 x 1.000 = 0.527
(give or take some rounding errors)
_______________________________________________________________________________________
Percentage intervals
Gear Ratio upshift% downshift % (Total Difference 409%)
Gear 1 0.527 ______ -22.6%
Gear 2 0.681 +29.2% -11.6%
Gear 3 0.770 +13.1% -12.3%
Gear 4 0.878 +14.0% -11.8%
Gear 5 0.995 +13.3% -12.3%
Gear 6 1.134 +14.0% -12.2%
Gear 7 1.292 +13.9% -11.6%
Gear 8 1.462 +13.2% -12.3%
Gear 9 1.667 +14.0% -11.7%
Gear 10 1.888 +13.3% -12.3%
Gear 11 2.153 +14.0% _____
average +15.2% -13.1%
__________________________________________________________________________________
Cable pulls;
There are some measurements in the A11 reliability thread but in essence all shifts are a nominal 5.0mm except for 6-7 which is longer (around 6.4mm) and 10-11 which is also longer, presumably so that the cable will run full slack in gear 11 even if it is a little draggy. Other than A11-specific shifters ( Shimano, J-tek, Versa), there are no suitable alternatives.
________________________________________________________________________________________
Comments:
I note with interest that once again (as in Nexus 7) shimano have avoided having a nice efficient direct drive gear 5 (or 6) (which would have been 'L---' clutching pattern). As it is the hub uses a single gear train in gears 1,7,8,9, thus these are likely the most efficient gears. Two gear trains are used in gears 2,3,4,10,11, (medium efficiency) and three are used in gears 5 and 6 (probable lowest efficiency). [NB Under very low loads the parasitic losses from the overrun roller clutches may be more apparent than the gear train (meshing + bearing) losses.]
If the gear ratios are set in a vaguely sensible fashion (i.e. so that the 'flat road, no wind' gear is about 70-80% of top gear) then gears 7,8,9 (which are relatively efficient) are likely to be the most used gears. On a heavily loaded touring bike gear 1 is likely to be heavily used too. This being the case the weighted average efficiency of the gear may not be as bad as the numerical average efficiency might suggest. However there is a question mark against gears 1-6, efficiency-wise, in that when the stage 1 gear is in use, there are axial thrust loads that must be supported by non-rolling element bearings.
Some folk have suggested that there ought to be more gears yet with different clutching (because there are sixteen possible permutations of four clutches) but IMHO this is not the case; S2 and S3 cannot be usefully locked at the same time (so the four '*LL*' clutching permutations are forbidden/ineffective), so the 'L---' clutch pattern is the only allowable and unused permutation.
In theory it should be nigh-on impossible to lose all drive in this hub, provided the gears are not jammed and you can still select C1, and RC2/ RC3 are still working, you should be left with a viable direct drive gear even if you can't lock S2, S3, S4 etc.
____________________________________________________________________________________
There are some very helpful instructions and photographs (prepared by jb) in the midst of the long running Alfine 11 thread which can be found here
http://forum.ctc.org.uk/viewtopic.php?f=5&t=64432&start=374
[NB; the sun locking pawls are labelled from left to right in jb's notes.]
Below these there are some diagrams which show the drive path through the hub in the various gears.
On subsequent pages there are several other useful photos and posts.
______________________________________________________________________________________
Known faults;
- the shift control is very sensitive to variations in build tolerance and adjustment. The reason for this is that the cassette joint may (because the RH hub cone is supported on two prongs, leaving gaps that the selector mechanism pokes though) only rotate about 135 degrees total, i.e. about 15 degrees per shift, tops. This makes building the hub accurately very difficult for Shimano and it also makes it sensitive to the exact external setting when in use. Unfortunately with new hubs the cable usually settles and this throws the external adjustment; worse still the build tolerances are such that the 'correct adjustment' (dots lined up in gear 6 http://www.sheldonbrown.com/nexus-mech.html) is not always the 'optimum adjustment'. 'Rofan' (see A11 reliability thread) has reported that the shift control sleeve is lighter built and longer than the equivalent A8/N8 version and this presumably makes it more likely to give trouble through wear or high shift load.
- The helical stage 1 gears were presumably added at a late stage ( the EV techdoc shows straight cut gears), I suppose to make the hub nice and quiet. However this generates an axial thrust load which can cause problems with retention of S1 on the axle (as jb has experienced in the A11 reliability thread). Arguably whoever designed the gear train screwed up here, in that when you pedal hard, the sun pinion S1 is forced rightwards i.e. against the retaining clip, rather than leftwards against a fixed shoulder. Maybe there is a good reason for this, but if they had cut the helical gear the other way (or left it straight cut) this fault may not occur in the same way or perhaps at all. If you routinely give the hub a lot of stick in gears 1-6 this is what is likely to generate this kind of problem.
-numerous users have reported skipping in some gears. This can happen even if the sun locking pawls are fully extended (in new hubs or hubs where there are burrs on the pawls) but it certainly will happen if the sun locking pawls are not fully extended. If the build tolerances of the hub are off a bit, the hub can therefore slip even when 'correctly adjusted'. The same thing is seen in numerous N8/A8 hubs. Many users have found that a little trial and error (with adjustments that are up to about 1.5mm away -in either direction- from the 'correct adjustment') can yield a more reliable setting. However if the axle assembly is examined when out of the hub the pawl lift (and therefore correct adjustment) can be checked directly. [I have seen several N8/A8 hubs where the C1 shift was not synchronised with the pawl lifts in the correct way. I do not know if the same fault occurs with A11 hubs. Note also that the sun locking pawls in a A11 hub are less likely to be overrun than in an N8/A8 hub, which means that any burrs on the pawl tips are unlikely to be knocked off when that pawl is not transferring drive. Hard (i.e. highly loaded) slips can raise burrs on, or even shatter pawls.]
- the clearances in all shimano IGHs are pretty tight; if there is any shrapnel in the hub (particles bigger than ~0.25mm) it will probably cause one or more gear trains to jam and/or be damaged. Stage 2 and stage 3 could jam entirely and you would still have some working gears, but if stage 1 starts to jam you are in trouble because the sun S1 is locked to the axle and those gears must turn at all times when you are pedalling.
- these hubs often leak oil. It isn't clear where the hub is meant to be vented; it seems to me that some leakage (most often on the RHS) is more or less inevitable.
--------------------------------------------------------------------------------
overall;
The Alfine 11 hub is (internally) only a tiny bit more complicated than a N8/A8 hub, and apart from the helical stage 1 gears there is nothing much new and different inside it per se vs the 8s hubs. So, with one or two caveats the internals ought to be reasonably reliable; an intrinsically 'bad hub' ought to become evident within the warranty period.
However the efficiency is unknown/suspect, the helical gears are probably not the best choice for high torque use in low gears, and a significant proportion of users report skipping or other problems, which can arise for several reasons.
Overall the A8/N8 hubs are both stronger and cheaper, so if they have wide enough ratios for your terrain, they probably make a better choice for any hard use. This is despite the fact that the grease lubrication in these hubs is far from perfect; running them in oil (or highly fluid grease, not shimano's) seems like a good idea to me.
cheers