Rohloff makes the case for Cyber Nexus

"Andre Jute" wrote...

[...]
I'm sure you can easily imagine several occasions in which
gearchanging based on speed wouldn't be ideal...

Let's see if your examples are persuasive:

I could be going down a long hill at high speed. I wouldn't want to
switch to a lower gear each time I slow down for a difficult bend.

Why ever not? You'd be in the most optimal gear for accelerating after
the corner.

When you set the "pedalling effort" on the Cyber Nexus, you probably
do that on the basis of your muscle strength when you are on the flats.
When you are going downhill and you need to accelerate, there is
gravity which is helping you in the task; as a consequence you have
some "spare" muscle force which normally you want to use for
extra acceleration. To do this (ie use the same force you'd use on
the flats to accelerate when you're going downhill) you need to
switch to a higher-than-normal gear, in accordance to the
steepness of the descent.
Unfortunately, Cyber Nexus knows only speed and you would
therefore constantly be using a gear lower than the one you need.
As a conclusion: in this case CN is not able to select the optimal
gear. It would if it measured force or if it knew slope and reacted
to it.

Of course a similar "mirror" reasoning could be done when you are
accelerating while going uphill.

I could prefer different cadences for different conditions. For example,
I like to go easy uphill but have tougher gears on the flats.

Then going uphill you pedal slower and on the flat pedal faster. Cyber
Nexus will give you the gear of your heart's desire.

Mind, I was talking about "cadences"!
When after a hill I reach a flat and start to pedal faster, the speed will
increase and Cyber Nexus will switch to a harder gear. At this point
(or after some further gear switching) I would need to slow down my
pedalling cadence, otherwise the Cyber Nexus would keep switching
to harder gears and I would keep going faster and faster until I couldn't
bear the effort.
But my original point is that I might prefer a higher cadence on the hill
and a lower cadence on the flat. CN wouldn't allow me to achieve that.

On the other hand, I might want low gears when facing a long ride uphill
in order to avoid my legs getting tired, but I might be ready for a
tougher
effort when I know that the hill is going to be short.

In either case you merely pedal as your legs dictate and the
electronics will put you in the correct gear. (..)

Here you are right.

I could be facing a short steep hill, having little force left in the
legs and willing to maintain constant speed. Normally I would
switch to a lower gear and increase cadence.. Cyber Nexus would
just compel me to slow down.

No. I have this one too. It is what Cyber Nexus is great at, switching
down and letting your cadence pick up. It matches your desired effort
to the gear required. Your cadence increases but the box doesn't
change up.

This situation is similar to the one I described above, for downhill
acceleration.
In order for Cyber Nexus to switch to a lower gear you need to slow
down first, ie you need to decrease your cadence. When CN
switches to lower gear you need to increase your cadence again
otherwise you would keep slowing down after each gear change and
CN would react each time with a further switch!

My point was: if you're facing a hill and your legs are a bit sore,
you might still be able to keep going at the usual speed by swhitching
to lower gear and increasing cadence. Unfortunately CN doesn't
know your legs are sore, it just knows about your "desired effort"
on the flat you set at some point. With CN the only way to ease to load
on your legs in this case is slowing down.

off saddle cut

What if I were training the muscles in my legs and I were alternating
stretches with great efforts with others at minimal effort?

Then you click a switch and take over manual control of the Cyber
Nexus gearbox. The same answer suffices for all the situations above

In this case you'd be in "electronic assisted" gearchange, which I'm sure
is great, but we were discussing about "full automatic" gear
changing, weren't we?

[...]
If I were going shopping or on a leasurely ride, I think that all of the
above would hardly apply. The contrary would happen if I were in a
more sporty mood. This is why I mentioned comfort cycling.

This is why the Cyber Nexus has a manual control mode.

see right above.

snip on rotary control

but I believe you if you say that the Cyber Nexus is better, even in
manual control I imagine.

Well, there's a big difference between turning a very stiff control
and clicking a tumber switch.

I've never used either, but I'm inclined to agree thoroughly.

For the record. the Cyber Nexus doesn't
have a manual control at all, in the sense of a cable directly
connected to the gearbox. What Cyber Nexus has is electronically
assisted gearchanges under manual control. The rider decides when the
shift happens but the electronics make the shift. That is exactly what
the Dura-Ace electronically assisted shifting does too; it is merely a
cut-down Cyber Nexus system with battery rather than dynohub power.

This seems reasonable to me. See both above and below :-)

In fact, I can't even see that the drastically cut-down Cyber Nexus
which is the Dura-Ace electronically assisted shifting for racers has
any advantage over the full Cyber Nexus -- even for racers -- except
perhaps weight, which I suspect is rather smokey argument.

Cycle racing is a sport of surges, in which stretches at medium
power alternate with other high-power high-force stretches;
this cannot be kept under control by an automatic gear, even
if its electronics were based on exerted power rather than speed.
Just to make an example, in a sprint racers produce an all-out
effort which would be unsustainable for the whole duration
of the race. Should they change the settings of the Cyber Nexus
just before sprinting? Even if they did, I doubt they would be more
efficient than with manual controls.

That sounds reasonable. But I remember that when I turned up at a
track with a Ford (...)

It doesn't make sense to compare a vehicle with a human!
A vehicle doesn't have more or less force (or power) according to
how long it travelled during the day, on how much it ate the night
before, how well it slept, how much it flamed on the Usenet,
how much lactate is getting around, etc. And a vehicle doesn't
have limited amount of fuel for very high power efforts which
should be used with care when the occasion (or the race) requests
it.
A vehicle goes on as it wants until it has fuel and then it stops.

It is pure and simple macho bull**** to claim, as has been
repeatedly claimed here by the wannabe Lances, that a human can
shift faster than electronics,

I'm sure that, when the moment to change gear comes, Cyber
Nexus is quicker than my finger.
But electronics reacts to changes in speed, ie it necessarily lags
behind the human...

I've heard a lot about intelligent human anticipation. It sounds good
until you ask what actually happens. And what actually happens is that
the human, in anticipation, switches out of the optimum gear
(presuming he had the smarts to be in it in the first instance) yards
before it is necessary, and in that space, in a less than optimal
gear, loses fractions of a second, or a second. Changing too early is
as bad from an efficiency viewpoint as hanging on to the wrong cogs
too long. Both cost time.

I would tend to agree with you, but the point I've been trying to make
in all my posts is that: (a) you know when you need/want to exert
more or less effort, CN only knows the effort level on the flat
you set when you started riding, (b) you know how you want to part
exerted power between cadence and muscles in different instants,
CN doesn't.

Still, CN switches *after* you slow down / speed up. Ideally a
human might be quicker by switching exactly when needed...

[...]
and the claim that on a road bike a human shifts more
appropriately than electronics is, if true at all, only temporarily
true --

only temporarily true? What does that mean?

It means that if it is true, which you believe and I don't (and I
have just demonstrated above why it is untrue), it remains true
only until the electronics catch up to measuring force as well, or
whatever you want measured.

Devices to measure power are already well available (powertap etc),
but they tend to be quite expensive.

I wouldn't word what you meant with "only temporarily true".
CN doesn't measure force; maybe in the future it will,
maybe not.

There is no intrinsic reason that an electronic control
on a bike can't have a menu from which you choose the parameters
to be measured, and assign them varying weights, in your particular
desired gearchange. Any good HRM/bike computer, say the
Ciclosport HAC4, already includes enough measuring and
calculation power.

Yes. I do use a HAC4 and I love it. Nonetheless I have
to say its measurements of power are (obviously) garbage.

On the other hand, it might be enough just to put the desired force
control, (...)
Did I mention that three effort maps are already under the rider's
thumb on the Cyber Nexus: (...)
There is no reason on earth you shouldn't have many more
programmes each of which puts the shift points, even for
individual gears, at particular desired points.

....and which would basically turn your fully automatic gear
changer into an electronically assisted gear changer, which
takes us back to the Dura Ace you've been mentioning.

After all, you already remember a lot of esoteric info about which
crankwheel and which sprocket you're on, and what the ratio is, for
each of the two dozen or so gears of your derailleur bike...

I did actually ponder about the appropriate ratios when I chose
chainrings and cogs (I'm lazy, I like a triple, so I have 3x9 on
my touring bike). When I cycle, of course, I know only vaguely
which gear I'm in because my brain automatically selects the
most appropriate one ;-)

You do remember all that stuff, don't you, Gennaro?
Because if you don't, if you forget it even once a month, then you
*need* full-auto Cyber Nexus.

I do insist on the fact that it is often the case that your desired
cadence and your desired effort vary in the course of a single ride.
a) CN is basically a speed-controlled constant-cadence gear
switcher. Its problem is that you set your desired
effort (on the flat!) once and for all (save the D, Ds and L setting)
and then CN keeps the cadence (roughly?) constant measuring
your speed.
When the road ahead gets steeper CN knows nothing about it and
all you can do to shift to easier gears is slow down.
b) measuring and controlling force might be useful in that sense but...
c) ...the need to continually tell CN your desired force (or power)
output makes automatic gearchanging very similar to
electronically controlled gearchanging. This is the reasoning
behind the 'electronic' Dura Ace I guess.

[...]
Humans may want to change according to many other factors,

Humans may want to do a lot of irrational things. Whether they
should be allowed to pretend they change gear better than
electronics can is a matter to be decided by team managers once
the policy decision is taken that efficiency is what counts, not what
"humans may want".

As refined as your electronics can get it will never be able to know,
for example, when an athlete should deplete his limited resources
needed to produce a short effort at very high power...


Humans may
want to change according to many other factors, see above.

See above about electronics being able to measure and react to
whatever you want your gearbox to react to.

it will be welcome

[...]

Let's see that again: "In a couple of years Shimano will charge
racers a *huge* premium for putting the full auto back on the
Dura-Ace gruppo."

I'm convinced "full auto", if ever, will be used only when it allows
full control over power output (therefore not really being 'full
auto'....)

Andre Jute

bye
Gennaro

On May 8, 6:38*am, Andre Jute wrote:

I haven't heard the slightest argument even to begin to persuade me
that a Shimano Cyber Nexus type automatic system wouldn't be
superior on a Rohloff to the obstreperous Rohloff rotary control,
certainly for road use.

Depends on your criteria for what's superior, but--
Imagine it deciding to do the 7-8 shift in the
middle of a power stroke.

...and dropping you into 11th? That is a problem of the Rohloff box
design, not of the Cyber Nexus type electronic shift.

You're the one who couldn't see how auto shifting wouldn't
be superior "on a Rohloff".

For those who haven't tried a Rohloff: describing the 7-8 shift
as dropping you into 11th gear (or 14th for older specimens) if
done under load doesn't fully describe the experience. It's not
a smooth shift if done under load.

I like the Rohloff; I'd buy one again. But I'd never want it
autoshifted.

Tom Ace

I think we can leave it there, Gennaro. If you were to ride the Cyber
Nexus, you'd find it much more useful and pleasant than you presently
expect, partly because of slop in the system, what you describe as
"then CN keeps the cadence (roughly?) constant measuring your speed".
This has been a good discussion but, as several contributors have
already pointed out, we're past the point of information or even
opinion and well into personal preference. Ciao. -- Andre Jute


On May 8, 10:14*pm, "Gennaro" wrote:
"Andre Jute" wrote...

[...]

I'm sure you can easily imagine several occasions in which
gearchanging based on speed wouldn't be ideal...

Let's see if your examples are persuasive:

I could be going down a long hill at high speed. I wouldn't want to
switch to a lower gear each time I slow down for a difficult bend.

Why ever not? You'd be in the most optimal gear for accelerating after
the corner.

When you set the "pedalling effort" on the Cyber Nexus, you probably
do that on the basis of your muscle strength when you are on the flats.
When you are going downhill and you need to accelerate, there is
gravity which is helping you in the task; as a consequence you have
some "spare" muscle force which normally you want to use for
extra acceleration. To do this (ie use the same force you'd use on
the flats to accelerate when you're going downhill) you need to
switch to a higher-than-normal gear, in accordance to the
steepness of the descent.
Unfortunately, Cyber Nexus knows only speed and you would
therefore constantly be using a gear lower than the one you need.
As a conclusion: in this case CN is not able to select the optimal
gear. It would if it measured force or if it knew slope and reacted
to it.

Of course a similar "mirror" reasoning could be done when you are
accelerating while going uphill.

I could prefer different cadences for different conditions. For example,
I like to go easy uphill but have tougher gears on the flats.

Then going uphill you pedal slower and on the flat pedal faster. Cyber
Nexus will give you the gear of your heart's desire.

Mind, I was talking about "cadences"!
When after a hill I reach a flat and start to pedal faster, the speed will
increase and Cyber Nexus will switch to a harder gear. At this point
(or after some further gear switching) I would need to slow down my
pedalling cadence, otherwise the Cyber Nexus would keep switching
to harder gears and I would keep going faster and faster until I couldn't
bear the effort.
But my original point is that I might prefer a higher cadence on the hill
and a lower cadence on the flat. CN wouldn't allow me to achieve that.

On the other hand, I might want low gears when facing a long ride uphill
in order to avoid my legs getting tired, but I might be ready for a
tougher
effort when I know that the hill is going to be short.

In either case you merely pedal as your legs dictate and the
electronics will put you in the correct gear. (..)

Here you are right.

I could be facing a short steep hill, having little force left in the
legs and willing to maintain constant speed. Normally I would
switch to a lower gear and increase cadence.. Cyber Nexus would
just compel me to slow down.

No. I have this one too. It is what Cyber Nexus is great at, switching
down and letting your cadence pick up. It matches your desired effort
to the gear required. Your cadence increases but the box doesn't
change up.

This situation is similar to the one I described above, for downhill
acceleration.
In order for Cyber Nexus to switch to a lower gear you need to slow
down first, ie you need to decrease your cadence. When CN
switches to lower gear you need to increase your cadence again
otherwise you would keep slowing down after each gear change and
CN would react each time with a further switch!

My point was: if you're facing a hill and your legs are a bit sore,
you might still be able to keep going at the usual speed by swhitching
to lower gear and increasing cadence. Unfortunately CN doesn't
know your legs are sore, it just knows about your "desired effort"
on the flat you set at some point. With CN the only way to ease to load
on your legs in this case is slowing down.

off saddle cut

What if I were training the muscles in my legs and I were alternating
stretches with great efforts with others at minimal effort?

Then you click a switch and take over manual control of the Cyber
Nexus gearbox. The same answer suffices for all the situations above

In this case you'd be in "electronic assisted" gearchange, which I'm sure
is great, but we were discussing about "full automatic" gear
changing, weren't we?

[...]

If I were going shopping or on a leasurely ride, I think that all of the
above would hardly apply. The contrary would happen if I were in a
more sporty mood. This is why I mentioned comfort cycling.

This is why the Cyber Nexus has a manual control mode.

see right above.

snip on rotary control

but I believe you if you say that the Cyber Nexus is better, even in
manual control I imagine.

Well, there's a big difference between turning a very stiff control
and clicking a tumber switch.

I've never used either, but I'm inclined to agree thoroughly.

For the record. the Cyber Nexus doesn't
have a manual control at all, in the sense of a cable directly
connected to the gearbox. What Cyber Nexus has is electronically
assisted gearchanges under manual control. The rider decides when the
shift happens but the electronics make the shift. That is exactly what
the Dura-Ace electronically assisted shifting does too; it is merely a
cut-down Cyber Nexus system with battery rather than dynohub power.

This seems reasonable to me. See both above and below :-)

In fact, I can't even see that the drastically cut-down Cyber Nexus
which is the Dura-Ace electronically assisted shifting for racers has
any advantage over the full Cyber Nexus -- even for racers -- except
perhaps weight, which I suspect is rather smokey argument.

Cycle racing is a sport of surges, in which stretches at medium
power alternate with other high-power high-force stretches;
this cannot be kept under control by an automatic gear, even
if its electronics were based on exerted power rather than speed.
Just to make an example, in a sprint racers produce an all-out
effort which would be unsustainable for the whole duration
of the race. Should they change the settings of the Cyber Nexus
just before sprinting? Even if they did, I doubt they would be more
efficient than with manual controls.

That sounds reasonable. But I remember that when I turned up at a
track with a Ford (...)

It doesn't make sense to compare a vehicle with a human!
A vehicle doesn't have more or less force (or power) according to
how long it travelled during the day, on how much it ate the night
before, how well it slept, how much it flamed on the Usenet,
how much lactate is getting around, etc. And a vehicle doesn't
have limited amount of fuel for very high power efforts which
should be used with care when the occasion (or the race) requests
it.
A vehicle goes on as it wants until it has fuel and then it stops.

It is pure and simple macho bull**** to claim, as has been
repeatedly claimed here by the wannabe Lances, that a human can
shift faster than electronics,

I'm sure that, when the moment to change gear comes, Cyber
Nexus is quicker than my finger.
But electronics reacts to changes in speed, ie it necessarily lags
behind the human...

I've heard a lot about intelligent human anticipation. It sounds good
until you ask what actually happens. And what actually happens is that
the human, in anticipation, switches out of the optimum gear
(presuming he had the smarts to be in it in the first instance) yards
before it is necessary, and in that space, in a less than optimal
gear, loses fractions of a second, or a second. Changing too early is
as bad from an efficiency viewpoint as hanging on to the wrong cogs
too long. Both cost time.

I would tend to agree with you, but the point I've been trying to make
in all my posts is that: (a) you know when you need/want to exert
more or less effort, CN only knows the effort level on the flat
you set when you started riding, (b) you know how you want to part
exerted power between cadence and muscles in different instants,
CN doesn't.

Still, CN switches *after* you slow down / speed up. Ideally a
human might be quicker by switching exactly when needed...

[...]

and the claim that on a road bike a human shifts more
appropriately than electronics is, if true at all, only temporarily
true --

only temporarily true? What does that mean?

It means that if it is true, which you believe and I don't (and I
have just demonstrated above why it is untrue), it remains true
only until the electronics catch up to measuring force as well, or
whatever you want measured.

Devices to measure power are already well available (powertap etc),
but they tend to be quite *expensive.

I wouldn't word what you meant with "only temporarily true".
CN doesn't measure force; maybe in the future it will,
maybe not.

There is no intrinsic reason that an electronic control
on a bike can't have a menu from which you choose the parameters
to be measured, and assign them varying weights, in your particular
desired gearchange. Any good HRM/bike computer, say the
Ciclosport HAC4, already includes enough measuring and
calculation power.

Yes. I do use a HAC4 and I love it. Nonetheless I have
to say its measurements of power are (obviously) garbage.

On the other hand, it might be enough just to put the desired force
control, (...)
Did I mention that three effort maps are already under the rider's
thumb on the Cyber Nexus: (...)
There is no reason on earth you shouldn't have many more
programmes each of which puts the shift points, even for
individual gears, at particular desired points.

...and which would basically turn your fully automatic gear
changer into an electronically assisted gear changer, which
takes us back to the Dura Ace you've been mentioning.

After all, you already remember a lot of esoteric info about which
crankwheel and which sprocket you're on, and what the ratio is, for
each of the two dozen or so gears of your derailleur bike...

I did actually ponder about the appropriate ratios when I chose
chainrings and cogs (I'm lazy, I like a triple, so I have 3x9 on
my touring bike). When I cycle, of course, I know only vaguely
which gear I'm in because my brain automatically selects the
most appropriate one ;-)

You do remember all that stuff, don't you, Gennaro?
Because if you don't, if you forget it even once a month, then you
*need* full-auto Cyber Nexus.

I do insist on the fact that it is often the case that your desired
cadence and your desired effort vary in the course of a single ride.
a) CN is basically a speed-controlled constant-cadence gear
switcher. Its problem is that you set your desired
effort (on the flat!) once and for all (save the D, Ds and L setting)
and then CN keeps the cadence (roughly?) constant measuring
your speed.
When the road ahead gets steeper CN knows nothing about it and
all you can do to shift to easier gears is slow down.
b) measuring and controlling force might be useful in that sense but...
c) ...the need to continually tell CN your desired force (or power)
output makes automatic gearchanging *very similar to
electronically controlled gearchanging. This is the reasoning
behind the 'electronic' Dura Ace I guess.

[...]

Humans may want to change according to many other factors,

Humans may want to do a lot of irrational things. Whether they
should be allowed to pretend they change gear better than
electronics can is a matter to be decided by team managers once
the policy decision is taken that efficiency is what counts, not what
"humans may want".

As refined as your electronics can get it will never be able to know,
for example, when an athlete should deplete his limited resources
needed to produce a short effort at very high power...

Humans may
want to change according to many other factors, see above.

See above about electronics being able to measure and react to
whatever you want your gearbox to react to.

it will be welcome

[...]

Let's see that again: "In a couple of years Shimano will charge
racers a *huge* premium for putting the full auto back on the
Dura-Ace gruppo."

I'm convinced "full auto", if ever, will be used only when it allows
full control over power output (therefore not really being 'full
auto'....)

Andre Jute

bye
Gennaro

On May 8, 11:25*pm, Tom Ace wrote:
On May 8, 6:38*am, Andre Jute wrote:

I haven't heard the slightest argument even to begin to persuade me
that a Shimano Cyber Nexus type automatic system wouldn't be
superior on a Rohloff to the obstreperous Rohloff rotary control,
certainly for road use.

Depends on your criteria for what's superior, but--
Imagine it deciding to do the 7-8 shift in the
middle of a power stroke.

...and dropping you into 11th? That is a problem of the Rohloff box
design, not of the Cyber Nexus type electronic shift.

You're the one who couldn't see how auto shifting wouldn't
be superior "on a Rohloff".

For those who haven't tried a Rohloff: *describing the 7-8 shift
as dropping you into 11th gear (or 14th for older specimens) if
done under load doesn't fully describe the experience. *It's not
a smooth shift if done under load.

The "dropping into 11th or 14th" only happens exceptionally. I have no
experience, only what I've heard here on RBT. I'd be interested to
hear if it has happened to any of the resident Rohloff owners and how
often.

I like the Rohloff; I'd buy one again. *But I'd never want it
autoshifted.

Two points:

1. How will a manual change on a Rohloff that goes wrong be better
than an autochange that goes wrong?

2. Is that a visceral objection to an autobox in principle, or merely
because the 7-8 problem on the Rohloff? Tell us, Tom, if the 7-8
problem were fixed, would you consider an automatic Rohloff, or would
the whole idea be anathema, or, of course, unsuitable for the way you
ride?

I too would buy another Rohloff. It is amazing how useful that big
range is.

Andre Jute
Just wondering

Per Tom Ace:
Depends on your criteria for what's superior, but--
Imagine it deciding to do the 7-8 shift in the
middle of a power stroke.

I did a few of those in the beginning, but it soon became a
non-issue.

Maybe I don't ride hard enough, but even the few times it did
happen it was far preferable to the times I took the saddle horn
up my butt from missed der shifts.
--
PeteCresswell

On May 8, 3:58*pm, Andre Jute wrote:

1. How will a manual change on a Rohloff that goes wrong be better
than an autochange that goes wrong?

If I were obstinate enough to shift 7-8 manually under
heavy load, I could make just as crunchy a shift as
an automated system could impose on me. Even so,
I'd be expecting it. The auto could serve it up at
just the wrong moment in traffic.

2. Is that a visceral objection to an autobox in principle, or merely
because the 7-8 problem on the Rohloff? Tell us, Tom, if the 7-8
problem were fixed, would you consider an automatic Rohloff, or would
the whole idea be anathema, or, of course, unsuitable for the way you
ride?

Would I want auto if the 7-8 shift were smoother?
Probably not. There are a couple issues--

Even the smoother shifts (i.e. not 7-8) in a Rohloff
are not totally smooth if done under load. You feel
the crank jump a couple degrees as backlash in the
new gear is taken up. I doubt that this could be
eliminated altogether (but feel free to tell me
how well the Nexus performs in this regard).
I don't like that feel of discontinuity under load,
and I can avoid it by easing up on the pedals or
shifting when the cranks are vertical. I wouldn't
want auto shifting to impose it on me.

Continuous variability would do away with some
of the annoying idiosyncrasies of automatics.
In 1996, I rode a bike with a prototype
continuously variable automatic transmission.
See http://minortriad.com/cvt.html
Jobst would quibble over whether it's truly
continuous, but there were no abrupt shifts
from gear to gear. It did have other problems;
I think it's no accident that it's not in production.

Setting aside the issues of smooth shifting,
there's also the question of whether I want a
machine picking the ratio. I can see pros and
cons to auto shifting--but overall I'd rather
choose for myself, as what gear feels best
to me at any moment depends on more
variables than a machine knows about.

Tom Ace

On May 9, 6:24*am, Tom Ace wrote:
On May 8, 3:58*pm, Andre Jute wrote:

1. How will a manual change on a Rohloff that goes wrong be better
than an autochange that goes wrong?

If I were obstinate enough to shift 7-8 manually under
heavy load, I could make just as crunchy a shift as
an automated system could impose on me. *Even so,
I'd be expecting it. *The auto could serve it up at
just the wrong moment in traffic.

Ah. Okay. I thought you meant something lifethreatening, like crank
going "slack" just when you're standing up on the pedal. I don't mind
crunches; the Rohloff is an agricultural implement, meant to crunch.
The literature tells us so. Question: how many miles does your Rohloff
have on it?

2. Is that a visceral objection to an autobox in principle, or merely
because the 7-8 problem on the Rohloff? Tell us, Tom, if the 7-8
problem were fixed, would you consider an automatic Rohloff, or would
the whole idea be anathema, or, of course, unsuitable for the way you
ride?

Would I want auto if the 7-8 shift were smoother?
Probably not. *There are a couple issues--

Even the smoother shifts (i.e. not 7-8) in a Rohloff
are not totally smooth if done under load. *You feel
the crank jump a couple degrees as backlash in the
new gear is taken up. *I doubt that this could be
eliminated altogether (but feel free to tell me
how well the Nexus performs in this regard).

You just made an involuntary joke, Tom. Sure, Shimano's Nexus 8 speed
change, manual or auto, is vastly smoother than the Rohloff; anyone
who has experience of both will tell you that. And it happens
precisely because of what we're discussing, because the Rohloff is a
precision-made instrument intended for agricultural use for 100,000km
(and now that they have experience, a 100k miles is being mentioned).
The Shimano is less over-built and therefore changes more easily
because there is more slack in the system. Or you could say the
Shimano, built for a third the service life of a Rohloff, changes
easier because it isn't intended to last so long. (It's not a fair
comparison: the street price of a Shimano Nexus box with all its
fittings is about 15% of the price of a comparable Rohloff kit.) You
pay for the strength and long life of the Rohloff in the quality of
the shift.

I don't like that feel of discontinuity under load,
and I can avoid it by easing up on the pedals or
shifting when the cranks are vertical. *I wouldn't
want auto shifting to impose it on me.

I'm trying to relearn shifting when the cranks are vertical, a skill
lost in the Nexus years...

Continuous variability would do away with some
of the annoying idiosyncrasies of automatics.
In 1996, I rode a bike with a prototype
continuously variable automatic transmission.
See *http://minortriad.com/cvt.html
Jobst would quibble over whether it's truly
continuous, but there were no abrupt shifts
from gear to gear. *It did have other problems;
I think it's no accident that it's not in production.

Thanks for the interesting pic.

Setting aside the issues of smooth shifting,
there's also the question of whether I want a
machine picking the ratio. *I can see pros and
cons to auto shifting--but overall I'd rather
choose for myself, as what gear feels best
to me at any moment depends on more
variables than a machine knows about.

And now we're back to personal preference.

Andre Jute
You can ride only one bike at a time

On May 9, 8:55*am, Andre Jute wrote:

The Shimano is less over-built and therefore changes more easily
because there is more slack in the system. Or you could say the
Shimano, built for a third the service life of a Rohloff, changes
easier because it isn't intended to last so long.

I think you're guessing. There are different
types of mechanisms inside, the Nexus hubs
using roller clutches that Rohloff doesn't.

Tom Ace

On May 9, 7:47*pm, Tom Ace wrote:
On May 9, 8:55*am, Andre Jute wrote:

The Shimano is less over-built and therefore changes more easily
because there is more slack in the system. Or you could say the
Shimano, built for a third the service life of a Rohloff, changes
easier because it isn't intended to last so long.

I think you're guessing. *There are different
types of mechanisms inside, the Nexus hubs
using roller clutches that Rohloff doesn't.

Tom Ace

Actually, Rohloff in the FAQs says that, because of the way the thing
is built, microscopic pieces of the gears will be knocked off in the
first 1000 kilometres, after which it should settle down. That's why I
asked you how many miles on your Rohloff hub -- Andre Jute.

On May 9, 1:27*pm, Andre Jute wrote:

The Shimano is less over-built and therefore changes more easily
because there is more slack in the system. Or you could say the
Shimano, built for a third the service life of a Rohloff, changes
easier because it isn't intended to last so long.

I think you're guessing. *There are different
types of mechanisms inside, the Nexus hubs
using roller clutches that Rohloff doesn't.

Tom Ace

Actually, Rohloff in the FAQs says that, because of the way the *thing
is built, microscopic pieces of the gears will be knocked off in the
first 1000 kilometres, after which it should settle down. That's why I
asked you how many miles on your Rohloff hub -- Andre Jute.

And I don't know because I don't count miles.
It quieted down years ago, meaning less noise
in 1-7 but no difference in shifting.

What made you think the Nexus shifting smoother
had anything to do with play, or in being less overbuilt?

Tom Ace