Best internal gear hub?

[qwerkus]

[true and just as irrelevant as quibbling about the feel of a roller brake on a hub designed for a roller brake. The 3-speed hub does what it does, correctly, and so does the roller brake. If it can skid the rear wheel-- and it can-- then there is verifiably no difference in stopping power between it and any other brake that can skid the rear wheel.

Lever feel can be changed by the use of a different lever pull ratio, which usually is not an option with hydraulic brakes.

Sure - maybe I should have been more specific: if you don't care about the extra weight, the c6000 and similar rear roller brake combined with an IGH is the best system there is for a city bike. Ultra low maintenance, and all weather proof, with plenty of stopping power for a regular cyclist. I've seen 10y old bikes with this setup still running strong although never grease / cleaned once!

Yet this thread is more about IGH in high power use, like a mid drive motor. In that case, based on my experience, a rear roller brake is no match of a properly setup disc brake, and I would always recommend 180mm+ hydraulic disc brakes for anyone pouring more than 1000W into the rear hub.

 

[amberwolf]

Lever feel can be changed by the use of a different lever pull ratio,

LIke with the Avid Speed Dial levers?

 

[Chalo]

Lever feel can be changed by the use of a different lever pull ratio,

LIke with the Avid Speed Dial levers?

Yes, those do that. But even their lowest travel setting pulls more cable than Shimano's designated roller brake levers do, so Speed Dial levers will vary from firmer than normal to a lot firmer than normal.

Another variable-ratio lever I have used successfully with drum and roller brakes is the Dia Compe PC-7 EXA. These have a bell crank with a rising mechanical advantage that stops rising at a point you can set with an adjusting screw.

The most straightforward way to change leverage ratio is to switch to a lever with a different amount of cable travel.

 

[qwerkus]

Lever feel can be changed by the use of a different lever pull ratio,

LIke with the Avid Speed Dial levers?

Yes, those do that. But even their lowest travel setting pulls more cable than Shimano's designated roller brake levers do, so Speed Dial levers will vary from firmer than normal to a lot firmer than normal.

Another variable-ratio lever I have used successfully with drum and roller brakes is the Dia Compe PC-7 EXA. These have a bell crank with a rising mechanical advantage that stops rising at a point you can set with an adjusting screw.

The most straightforward way to change leverage ratio is to switch to a lever with a different amount of cable travel.

Any lever designed to work with caliper type road brakes works with shimano roller brakes. All you need is a so called "short pull" lever. Not sure if they also work with SA drum brakes; those seem to be "long pull"

 

[Chalo]

Any lever designed to work with caliper type road brakes works with shimano roller brakes. All you need is a so called "short pull" lever. Not sure if they also work with SA drum brakes; those seem to be "long pull"

I'm aware that Shimano recommends and furnishes short pull levers for use with their roller brakes. But in my experience this is incorrect.

My experience with roller brakes and most drum brakes is that they feel mushy because they're being used with short pull levers. Often, these go all the way to the bars before the brakes have given their best. On the other hand, many linear pull brake levers don't offer enough mechanical advantage for good response. I generally get the best results from long blade levers with relatively long, but not full V-brake, pull.

My e-bike destroyed two successive NuVinci hubs with braking torque from its IM70 roller brake (it overcame the side cover torque retention), using an intermediate pull locking lever and plain brake housing. My rain bike with Sachs drums works well with 1980s Deore XT 4-finger levers and compressionless housing. My chopper feels great with a Magura motocross lever pulling a Sturmey Archer front drum with normal housing. But I've also gotten great results from normal brake housing with adjustable ratio levers like Shimano Servo Wave, Dia Compe PC-7 EXA, and Avid Speed Dial.

If you have problems with the feel of any given roller or drum brake, a good first measure is to try a different lever ratio.

 

[qwerkus]

Any lever designed to work with caliper type road brakes works with shimano roller brakes. All you need is a so called "short pull" lever. Not sure if they also work with SA drum brakes; those seem to be "long pull"

I'm aware that Shimano recommends and furnishes short pull levers for use with their roller brakes. But in my experience this is incorrect.

My experience with roller brakes and most drum brakes is that they feel mushy because they're being used with short pull levers. Often, these go all the way to the bars before the brakes have given their best. On the other hand, many linear pull brake levers don't offer enough mechanical advantage for good response. I generally get the best results from long blade levers with relatively long, but not full V-brake, pull.

My e-bike destroyed two successive NuVinci hubs with braking torque from its IM70 roller brake (it overcame the side cover torque retention), using an intermediate pull locking lever and plain brake housing. My rain bike with Sachs drums works well with 1980s Deore XT 4-finger levers and compressionless housing. My chopper feels great with a Magura motocross lever pulling a Sturmey Archer front drum with normal housing. But I've also gotten great results from normal brake housing with adjustable ratio levers like Shimano Servo Wave, Dia Compe PC-7 EXA, and Avid Speed Dial.

If you have problems with the feel of any given roller or drum brake, a good first measure is to try a different lever ratio.

Now that's interesting, and would explain a lot. How about writing a short tutorial about properly setting up drum and roller brakes ? Would be great to sticky it somewhere.

 

[amberwolf]

Another solution, if you can't change the lever you use for some bike specific reason, you could probably get the Travel Agent from Problem Solvers, or make something similar yourself.

As far as setup/maintenance goes, do either of these help?

https://www.sheldonbrown.com/drum-brakes.html

https://www.sheldonbrown.com/rollerbrakes.html

There's also these, but I haven't looked at any of the result pages:
https://www.google.com/search?newwindow=1&client=firefox-b-1&ei=DCrXXKi3NOqe0gLm7omoCg&q=setup+drum+and+roller+brakes&oq=setup+drum+and+roller+brakes&gs_l=psy-ab.3...102595.102595..102829...0.0..0.84.84.1......0....1..gws-wiz.BlWz4kvi3Jc

 

[elias]

FYI - forget Sturmey Archer 5 speed hubs. I have had crap luck with my 2008 XRD5 and am going to replace it... a costly option on a $1400 Pashley Roadster.

SA hubs cannot be shifted under load without damage, so I would rule them out for motor-power use, unless you plan to throttle for it ... all the time..

I agree! I built myself a wheel around the Sturmey Archer RX-RK5 IGH which is described as "heavy duty" and ruined it in 8 days with a Bafang Ultra/52V bike:
https://endless-sphere.com/forums/viewtopic.php?f=6&t=98811
Now I'm happy with the SRAM X4 derailleur, but thats another thread I suppose...

 

[Kaemus]

Hi,

I'd like to ask for your help with my setup: for last 3 years and 10 000 kilometers (~6K miles)
it has been a BBS02 with a Nexus 8 internal gear hub.

It's been great so far but recently it crashed - as you can see on the photos





the axis in Nexus 8 hub got demaged by the rear cog. Now under a load the cog
gets pulled outside and looses the connection with the hub. I'm able to ratate
the pedals but the movement is not transmited into the hub and wheel.

I usually put around max 400W power from the motor (occasionaly for a short time up to 700W)
+ rather strong pedalling. The controller is set up for a "soft start" to avoid a huge initial
hit at the drive.

My only explanation so far is that I let the rear cog deteriorate to the level that there is
a loose space between the cog and the axis and the cog is hammerging the axis each time
I run the motor.

I'v never read about such problems with this setup - could you please comment what could
have gone wrong?

Thank you,

Kris, Poland

 

[amberwolf]

Well, these aren't really made for more than human-output power levels (and some humans can break this kind of stuff), so depending on hte gearing you have and the torque applied to it, it might be higher than it can handle over time.

For this specific problem, there is a snap ring that holds that gear against the hub. There is usually some play in it, not much, but a little, so the gear can "wobble". This can cause wear over time. If the chainline on your bike puts the front drive chainring outboard of the rear sprocket, then it could be nudging the sprocket angularly against the snapring, and wearing a little bit at a time at the three worn points. (it would be nice if there were more than just three, but this is how all of the IGHs I've had in my hands do it, from ancient Sturmey-Archer thru Sachs up to newer-but-not-recent SRAM stuff.).

If the original chainring position for your bike was moved outward (rightward) then this could be what's happening.

It could also just be too much torque for it. If your chainring size decreased in front from what the bike originally had, then it would increase the torque placed on the IGH input.

Using a larger front chainring, back to the original size, would help with that part, but the motor still outputs more torque than a human....


Now, for ways of fixing the damage...if you're really into DIY and fixing stuff, you may be able to get the part for the IGH that mounts that sprocket and snap ring--but it may be more than just a plate, and may include parts of the core of the IGH. EIther way, you would need to at least partially disassemble the IGH, and possibly completely disassemble it. There's a fair bit of loose ball bearings involved in the ones I've had open, and looked at doing so (but decided not to), and a number of small parts that need to go back just the way they came out....

Alternatively, if your next option is to replace the whole IGH anyway, then you could mill (or have milled) new indents for the sprocket tabs to fit into, offset from the original ones enough to clear any damage and to leave you room to do it again if necessary.

Then you can see about getting thicker sprockets by enough to completely fill the indents side to side, and not allow any wiggle with the snap ring. (well, if the chainline is diagonal instead of straight it's still going to do it some). They don't have to be thicker at the teeth, just the core area where the tabs are. I don't know which ones might be thicker (if any), so you would have to measure yours and then check around for measurements on other brands that would fit yours.

If you cant find thicker ones and decided to have someone make some, you could have them make ones with more sets of tabs, and then also have the IGH milled to accept them. (this would have to be a precise fit for it to make any difference in the torque transfer, or else it'll still just load the faces of the ones that touch first, and the rest won't get any loading until the other ones are worn/damaged enough to allow the sprocket to rotate enough to do that.

 

[Kaemus]

Thank you for writing a whole article on it

Well, these aren't really made for more than human-output power levels (and some humans can break this kind of stuff), so depending on hte gearing you have and the torque applied to it, it might be higher than it can handle over time.

I'm aware of that but on the other hand there are many users with very positive feedback over the years.
I was hoping to be one them

If the chain line on your bike puts the front drive chainring outboard of the rear sprocket, then it could be nudging the sprocket angularly against the snapring, and wearing a little bit at a time at the three worn points

The front sprocket that comes with the motor does move the chain rightwards. People using external gears usually buy special
(and expensive) sprocket to improve the chain line. Otherwise the chain keeps on falling down.

Your remark about the angular force caused by it led me to another thought - for the first 2, maybe 2,5 years (out of 3 years
of using it) I had constant problems with the chain until I installed something like it:



Before that the chain would get loose very often - that means that the right side of the hub was moving forward (I suppose that
more than the left part where the motor is not putting force). In consequence the whole wheel was not aligned properly (turned
left) until I corrected it. It was working in that position for a substantial part of the usage time - the angle you are referring to was even worse. Does it make sense?

Alternatively, if your next option is to replace the whole IGH anyway, then you could mill (or have milled) new indents for the sprocket tabs to fit into, offset from the original ones enough to clear any damage and to leave you room to do it again if necessary.

I think I'll follow this direction. Recently the hub was serviced (disassembled, cleaned, lubricated etc.) by the guy knowing how
to do that (I would never dare to open it). So I already invested around 1/4 of the value of new hub. Getting the replacement
part (I suppose the whole axis) would sum up to buying a new one. I'll have it milled - I hope it's not so difficult to do it precisely.

Thank you so much for your indepth analysis and tips.

I'll return with the results.

Kris.

 

[amberwolf]

Thank you for writing a whole article on it

That's what I do best. :lol:

I'm aware of that but on the other hand there are many users with very positive feedback over the years.
I was hoping to be one them

Some situations are easier on IGHs than others, and part of it depends on usage, gearing, etc. Can also depend on how the motor system operates (or is used), as the slower it ramps up the throttle the easier it is on the parts--the more suddenly power is applied, the harder it is (really hard repeated application can shear teeth off the gears inside, on many IGHs).

Before that the chain would get loose very often - that means that the right side of the hub was moving forward (I suppose that more than the left part where the motor is not putting force). In consequence the whole wheel was not aligned properly (turned left) until I corrected it. It was working in that position for a substantial part of the usage time - the angle you are referring to was even worse. Does it make sense?

Yes, and might have also resulted in more "chain lash" where the chain is yanked into tension as the motor (or pedals) start up, which is harder on the drivetrain. Taking up the tension more slowly is easier on things (which is why at least some motorcycles use something like a "cush drive" setup on the input sprocket to the rear wheel, so when the clutch engages it doesn't apply as much of a shear force to the rest of the drivetrain).

I'll have it milled - I hope it's not so difficult to do it precisely.

Hopefully anyone with milling / machining equipment and experience will also be able to do it precisely, even just matching the existing stuff. If you provide them the new sprocket to ensure a perfect fit (could even make it an "interference fit" while you're at it--they'll know what that means), then you're less likely to have issues. That way they can make it fit both the round part of the milled out area, and the width between the face of the IGH and the snapring, so there cant' be (as much) wiggling of the sprocket.


The only thing I don't know is how the milling process will affect the hardness of the metal (it's ability to resist damage from the sprocket tabs); the machinist should know something about that though.

 

[Kaemus]

Hi,

I got the part milled:



Now the bike is operating as expected.

The strange thing is - I just realised it while uploading this photo - that
not all grooves are damaged - the one on the photo is not. Before the milling I posted
a photo of damaged one:


I didn't check the third one. The wheel is installed now but I'll look at it.

Why would only one of three groves be damaged? Does that mean that all the load was
taken by one "slot"? How could that happen?

Kris.

 

[amberwolf]

Assuming the sprocket itself is *not* damaged at the two undamaged hub points, then that shows, probably, that the snap ring was moving laterally out of it's groove, pulled by the sprocket being from the chainline being at an angle, allowing the sprocket at the non-damaged areas to move off of the hub so that only the one damaged point on the hub took the whole load.

If the sprocket is damaged at it's matching points to the undamaged hub ones, then probably the hub was harder than the sprocket at those two, but the sprocket was harder at the other one.


Could be something else, but that's what the provided evidence appears to show.

 

[Kaemus]

I'll inspect carefully it after disassembly.

However I'm not very willing to do it now - the bike is finally OK again : )

Many thanks for your support and attention.

Kris.