ccollins <
[email protected]> wrote in message news:<
[email protected]>...
> Originally posted by Gwhite
>
> > No problem. If you are still insistant on the setup, I
> > can give a few more tips.
>
> Need you have asked? With my hard-head and affinity for
> using old stuff. Yes, this is my intent.
>
> So, on E-bay (for $10.00) I just bought a NOS front
> derailuer. It is a FD-MT62 (half-step) "Deore II"
I don't know that model in particular. If it is a 1/2+granny
front, then it will be
1. "deep," so that the granny can be used. (A "shallow" one
would be okay for just a 1/2-step w/ no granny, but that
is not what you're doing.)
2. the inner plate "hangs" only scarcely lower than the
outer plate.
> I am making a fully loaded touring bike.
>
> The 110/74 cranks are my pallette on which I'll place
> good gears.
The right choice for nearly any bike in my opinion.
> From this thread it seems I am going to have the outer two
> be half-step distance and a bailout (For that big hill
> just south of Molera State Park in Big Sur).
By saying "bailout," in reference to the crank and rings, I
assume you are talking about the granny chainring. Anywhere
from 24t to 28t is fine, depending on your low gear
requirements.
> The 170mm cranks need, I guess, 50/44 as the outers and
> something small like a 26.
Hey now. No "guessing" is needed or desired. The gears are
defined by a balancing of mathematical relationship, as the
1/2-step is defined, and actual parts available. Don't
guess. Design or re-use that already designed. A 50-44 is
unlikely to meet the "half" criteria. A 50-46 is more likely
to be mathematically "correct."
> What thought process goes into these decisions?
I think this has been noted in this and countless other
threads to date. The mathematical definition of a half-step
is that the *step* between the front chainrings is *one-
half* that of the *step* between the rear cogs. Thus a shift
in the back is a "step," and that in the front is a "one-
half step." The relationship of steps is a so-called
"geometric" one. Another absolutely equivalent way of saying
this is that we strive to have certain "percent difference"
between gears.
Tooth counts are obviously integers (no 15.23 tooth cogs,
hah hah). The impact of this that one can only approximate
the ideal mathematical numbers representing the steps when
constrained by integers and the actual cogs those integers
represent. This approximation is practically unimportant:
the approximations are easily "close enough."
So the system can only be designed by understanding the
mathematical foundation. You don't have to do that because
it has already been done.
> For the rear freewheel, shouldn't I go with the Mega 7
> (11,13,15,18,21,24,24)?.
I presume (11,13,15,18,21,24,_28_)?.
We seek _equal_ geometric steps for the rear cog tooth
gradation. This is because we have only one geometric step
in the front (the granny is dismissed from design
consideration at this point because it is not a part of the
half-step system).
Computing the steps:
>> [meanStep, compavggeo] =
>> geomeanstep([11,13,15,18,21,24,28])
meanStep = .1672 .1432 .1826 .1543 .1336 .1543
compavggeo = .1559
So we would seek a chainring step of about half
"compavggeo," which is .1559/2 = 0.078 (7.8%).
Choosing some common big rings (50,48, & 46), we have
For 50 big ring middleRing = 46.2 round to 46
For 48 big ring middleRing = 44.4 round to 44
For 46 big ring middleRing = 42.5 round to 44 or 45
So let's arbitrarily test the result for the 48-44 combo:
44 48 11 106.000 115.636 13 89.692 97.846 15 77.733 84.800
18 64.778 70.667 21 55.524 60.571 24 48.583 53.000 28
41.643 45.429
For this we obtain (from low to high): meanStep = .0870
.0672 .0870 .0465 .0870 .0672 .0870 .0953 .0870 .0561 .0870
.0801 .0870
I don't like it. Note the step variance can get high.
There is a
3.65% and a 9.53% step.
Now let's look at the middle six cogs of my design and a
46-42 ring set. 42 46 12 92.750 101.583 14 79.500 87.071
17 65.471 71.706 20 55.650 60.950 24 46.375 50.792 29
38.379 42.034
For this we obtain (from low to high): meanStep = .0910
.0983 .0910 .0914 .0910 .0716 .0910 .1032 .0910 .0632 .0910
The step variance is a bit lower in this design. I think it
is superior.
If you are stuck with a 7sp freewheel, I would go with a 12-14-17-20-24-29-
34 rather than the 11-13-15-18-21-24-28. The "need" for a
heavily loaded touring gear above 100 inches is dubious --
it is easily sacrificed. On the other hand, a 34 tooth cog
will get used on tough climbs.
(12-14-17-20-24-28-34 & 12-14-17-20-24-28-32 are fine too.)
> sheldonbrown.com writes of it as good technology and
> solid. The gearing is unique.
Unique itself doesn't count.
> Assuming it works with my shifters, and framespacing, how
> does this current proposed combo effect my choice of rear
> deraillier?
It will be MTB and any modern 43t wrapup Shimano is fine.
> I have 7-SIS downtube shifters and the framespacing
> *is* 126mm.
>
> So, I am going to jump through hoops, and search for odd
> old equiptment, and the result will be a good drivetrain
> for loaded touring.
The only "odd" old part truly required is the front
derailleur. Most everything else can be had brand new.
I should note another "trouble" with this setup. The middle
and inner rings should be closer together than they are on
any other modern setup. Failure to "make this happen" will
result in occasional "bad" big-ring to middle-ring shifts in
the front. The chain will sometimes "freewheel" on the
middle ring if spaced with "normal" rings. How does one get
them closer? This is not easily done. What I did was buy a
_middle_ 46t ring and use it as the outer of my 46-42 combo.
This is the only simple way I found to solve this problem.
Rivendell did for a time have 46t "middles" for 110 cranks.
I know of no other easy and cheap solution. The "pickup
section/ramp" of "normal" outer rings places the tooth
centers too far apart. This is one of those subtleties that
makes me advise against the system.
You could make life easier you know. I still think you
should question your direction. I have some stuff you can
buy if you still insist. I have a Sachs 7sp 12-14-17-20-24-28-
32 freewheel (shimano 7sp SIS compatible) and a 126 mm
touring wheelset with a 40 spoke rear and 36 spoke front in
excellent condition.
> This is esoteric, but I appreciate the feedback. (and
> lessons).