Shimano Alfine 11 - Longer term reliability

[jb]

I was under the impression that the shift assist was an integral part of the design. The selector mechanism on its own is not strong enough in certain circumstances to operate the shifter and will result in it being bent.

[vanGoose]

I was under the impression that the shift assist was an integral part of the design. The selector mechanism on its own is not strong enough in certain circumstances to operate the shifter and will result in it being bent.

With the pedals loaded, shifting UP simply wouldn't do, all you get is a loose cable. But no excessive force would be able to enter into the mechanism.
With the pedals loaded, shifting DOWN would still be done by activation of the shift assist pawls, but at the price of unlimited (internal) force acting to rip out previous gear though pedal force tends to keep the currently active pawl latched.
I disabled the shift assist PAWLS only. As a result, the forced rip-out of the previous gear does no longer happen. All i do by shifting down with the pedals loaded now is that i tension the shift assist return spring. When i later release pedals then the shift assist spring will change to the lower gear - that's it.

The original shift assist mechanism IS an integral part of the hub, but not to protect the mechanism from excessive forces. It USES external, even excessive force (from the pedals/driver unit) to do what one would better not do.
My "mod" quits two things: Assisted down-shifting and risk of excessive force inside the mechanism. Shft assist feature becomes a shift preselect feature.

There is one downfall though in my mod: If you would shift down MORE than 2 gears without ever releasing the pedals, then you could wind up the shift assist return spring so much that finally the shifter clutch touches the other edge of the shift assist pawls carrier plate and that would be the time where the spring can no longer protect the cable from being ripped off or the mechanism being bent.

[jb]

I wouldn't say it was excessive force, you are in effect driving a wedge between the pawl and the sun wheel in order to push the pawl out of mesh with the ratchet and retract it. This will be a finite force that has been calculated.

Its not like your trying to push something 'into' mesh, which could result in excessive force.

The danger of not having the assist is like you say, going too far without a gear change taking place, which might happen with an unexpected climb appearing resulting in crashing down the gears quickly.

[Brucey]

the loading within the pawl control parts is effectively unlimited once the shift servo is engaged; the pawl goes 'overcentre' to provide reliable drive in any one gear and may need a very high force to push it out of engagement with the sun pinion. I don't recall there being any spring like element that can delay this shift action, it is just forced. IIRC I have seen the shift sleeve distorted (twisted lengthwise) in some hubs, in a way that suggests that it can be overloaded, presumably by such forced downshifts.

cheers

[jb]

Sure, it will be capable of delivering a very high force, but what I mean is the force required to shift the pawl out of engagement in the worst case scenario can be fairly accurately estimated I would have thought. Thus the mechanism should be designed to cope with a degree of safety.

Other than a foreign body getting stuck under the pawl there isn't much that would stop it from disengaging.

[Brucey]

the force required to push the pawl out of engagement will increase whenever you are pushing hard on the pedals. The force through the shift servo will also increase in a similar way. There is no limit to either force, they just scale with pedal pressure (and input gear ratio). However there is very much a finite limit to the strength of the shift sleeve etc. so I think the system is very much vulnerable to being overloaded.

cheers

[jb]

the force required to push the pawl out of engagement will increase whenever you are pushing hard on the pedals. /

cheers

Yes, and this should have been accounted for in the design

The heaviest conceivable person on the steepest conceivable hill trying to change down will be the force it should have been designed to cope with. There is nothing else to add unexpected load, unlike for instance, trying to force a car gear box into mesh with a worn synchromesh.

Also the sun gears that do the changing are geared up not down so the force is not as great as it would be on the step down gear, hence why you should always turn the axle in that direction when fitting to remove the slack as you've mentioned before.

[Brucey]

I see what you mean but I still think that there is no practical limit to the force, not least because wear on the parts may make the pawl 'dig in' further and thus increase the force required to push it out.

The N8/A8 is easier to look at because there are fewer gears with compound gear trains in action and it shares a similar shift servo layout. Practically speaking in (say) 6th gear (which has the lowest pawl loadings) the pawl has to transmit a torque that is about 25% of the input torque. I think this thrust is (very roughly) about equal to the chain tension (because the radius of a sprocket might be about x4 the radius of action of the pawl). To get the pawl to release there will be two sources of drag; friction and the effect of the ramp angle. I'd guess at a coefficient of friction of about 0.1 and a ramp angle that is also in the region of x0.1 force-wise. The force on the shift sleeve fingers is about double that on the end of the pawl because of the position of the pawl tab vs the end of the pawl.

Thus even if this estimate might be wrong by a factor of x2 then for a chain tension of ~400 lbs then you are looking at a (radial) force of 80 to 160lbs on the pawl tab to push the pawl out of engagement. This sounds like 'a lot' to me. This force will be even higher than that (for the same chain tension) in gears 2,3,4,7,8.

cheers

[vanGoose]

The discussion becomes somewhat theoretical. Reality and expierience are proof enough, i honestly think:

i) There ARE hubs out there which have been damaged in a way that excessive forces upon down-shifting degraded their inner circuit.
ii) We MAY think that the design should have considered what force a human can introduce through pedaling upon doing the shift, but apparently they took assumptions which are a little bit too optimistic.
iii) Some 99% of A-11 users seem not to damage their hubs this way, so the design can be considered fairly ok or the feature is not frequently used.
iv) If one manages to damage the inner circuit this way, however, then - to my today's knowledge - there is no repair possible. (I doubt that replacement of axle unit is guaranteed to repair since it may be uneasy to get the right axle unit hardware revision for a specific hub, or more inner parts might suffer than an axle unit replacement actually covers.)

To me, this is reason enough to disable the shift assist pawls, hopefully BEFORE i damaged anything else...

v) With shift assist pawls disabled, a new risk exists that one can continue to shift down without ever unloading the pedals until shift assist return spring is exhausted and can no longer protect the mechanism. Then serial troubles are unavoidable. But is that very likely to happen?? The benefit, however, (and i rate this benefit higher) is that the continued risk of damage upon EVERY loaded down-shift is now gone.

vi) With shift assist pawls disabled, shifting down behaviour is exactly same as shifting up behaviour:

- If pedals are (almost) unloaded then it shifts.
- If pedals are sufficiently loaded then nothing happens. Shifting will be delayed until pedals are eventually released.

[jb]

I'll warrant the assist mechanism will be used to some extent on every down change; its duration of activation will be mille seconds because as soon as the pawls are engaged the move will be initiated and the spring will take over thus retracting the assist pawls again.
If you choose to disable it then fair enough, you know what your doing - and what not to do. But for the general public I'd say it does what it says on the tin.
The ultimate fault with the A11 is that the select mechanism is too compact and whilst it probably works faultlessly for the majority it shows its weaknesses when used for the more arduous tasks.

[vanGoose]

I'll warrant the assist mechanism will be used to some extent on every down change; its duration of activation will be mille seconds because as soon as the pawls are engaged the move will be initiated and the spring will take over thus retracting the assist pawls again.

Sorry this is not true. The shift assist return spring needs to be tensioned sufficiently for the shift assist pawls to activate, but it will only be tensioned at all when the pawl of previous gear is "locked" through the driver's force on the pedals. If you like to test: you will feel and hear the difference. "Feel" because you will notice the jump of the gears at your foot. "Hear" because the snap-back of the shift assist pawls carrier plate back to retracted can be heard easily. This is of course even easier when the shift assist pawls themselves have no more chance to produce their own "sound", as is now the case with mine...

If you choose to disable it then fair enough, you know what your doing - and what not to do. But for the general public I'd say it does what it says on the tin.

Right! I am not saying that everybody shall file off pawls. I did it and my personal feel is that it was nice for me. But if someone else finds his mechanism worn like mine was then they'd know now how to put into service again what remained of their hub .

The ultimate fault with the A11 is that the select mechanism is too compact and whilst it probably works faultlessly for the majority it shows its weaknesses when used for the more arduous tasks.

Yep! I suspect (and will - maybe - analyse further next winter...), however, a part of that weakness to sit with the outer rotating parts, not with the axle unit only. Like this weird US guy would say: Let's wait and see what happens!

[radlr]

Hi, long time no see...

Calling just to say: "Second's Out!".

My book grew to 52 pages, contains more photographs and - maybe - can start further discussions about the "magic" within pawl engagement profiles...

Here is your link for everybody who is curious:

https://c.gmx.net/@329602864890517692/P ... EjXCnpZyCw

A few notes:

5. The link is valid until 2018-06-21.

Comments welcome - enjoy!

Hi vanGoose, I guess I missed the time to download your book on A11 disection. Would you please make another link?

Thank you very much.

-Radlr

[maax]

Good morning to all those in the forum, I just joined. I am Italian and here we are not so familiar with the changes to the hub as abroad. I just wanted to bring my testimony with the hub 11. Finally I mounted the 11 on a race bike with belt drive and up to now I have traveled 21,000.00 Km. I have always done the oil change after the first 1,000.00 Km every 5.000,00 Km as well as instructions using Shimano oil.
I must say that for now I had no problem and I'm very happy with the end 11 considering also that when I go out with the group of friends I mistreated enough in the changes without lightening the ride.
 
For my experience, the synchronization of the transmission wire is essential, after which it is ready for any use.

[meandros]

Good morning to all those in the forum, I just joined. I am Italian and here we are not so familiar with the changes to the hub as abroad. I just wanted to bring my testimony with the hub 11. Finally I mounted the 11 on a race bike with belt drive and up to now I have traveled 21,000.00 Km. I have always done the oil change after the first 1,000.00 Km every 5.000,00 Km as well as instructions using Shimano oil.
I must say that for now I had no problem and I'm very happy with the end 11 considering also that when I go out with the group of friends I mistreated enough in the changes without lightening the ride.
 
For my experience, the synchronization of the transmission wire is essential, after which it is ready for any use.

Dear Maax,

If you're rockin' the newer SG-S7001-11 hub, you should be good, it is indeed very reliable out of the box as per shimano's own instruction manuals.

However, the matter gets dicey once the older SG-S700 is concerned. Including two of my own, I have now serviced four different such hubs, all of which have different _true_ alignment. Of all four A11s I've worked on, only a friend's hub had the true 6th gear alignment where shimano marks it, and that is the infamous yella lines' aligning. Real quirky stuff, I know, but it is what it is. However, I have found the older hub version (SGS700) to be a solid design, even with the multitude of flaws , and, as everything, once understood, it doesn't pose any more issues.

Btw, as maintenance goes, taking care of the cassette joint is key. Although shimano do not recommend cleaning/lubing it (they gotta sell'em things, right?) I have found good ole wd40 works very well in cleaning and leaving a superficial oily film just tenacious enough to work out for the CJ. Anywho, just thought I'd drop in some of mine into the ole proverbial sp!t bucket. Yeehaw!

[Fredrik]

Hello Alfine 11 Users.

Forgive me for my bad English. (School English and google translate helps me!)
My old Alfine 8 hub broke down so I Decided to buy a new one, long story about a broken parcel and the SG-7000-8 hub went out of stock and so on, so i thought I buy 11 gears instead, a SG-S700-11 hub.
Search the internet for differences between 700 and 7001 hub and found this great forum.
But reading briefly through this page it seems like a better choice to choose the SG-S7001-11 Hub.
I have a waiting refund at an online store (for the broken parcel) but the company doesn't have the SG-S7001 hub in a wheel set.
But a brand new SG-S700-11 hub could not be bad from start by it self or could it?
Yes I can see your testemony about oil leaks and narrow adjustment windows and cable thickness differences and "they should be replacing _all_ S700s ...a flawed ~400$ product." and so on.. i know you probably advise me to get the 7001 hub ..but is the s700 realy that bad?


Best regards
Fredrik.