Chain line and chain wear...

Where I live I am very happy with a 11-21 cassette... I like having
very small gaps between the sprockets... I do spend quite a bit of time
in the 19 and 21 when I climb seated... We don't have any hills where I
live that I can't climb seated with the 21 or standing with the 17 or
18... None of our hills is more than 200-300 feet tall so we aren't
talking more than a minute or two... I have noticed that the 19 and 21
sprockets show accelerated wear though... I don't mean they start
skipping or anything... They just get real ugly marks on them from the
chain hitting them at an angle... I also noticed I wear out chains a
lot faster... I noticed this because I am temporarily running a 12-25
right now and my chain has lasted easily twice as long as my old chains
did with the 11-21... I'm still using a max of 21 so the 23 and 25
never get touched... With the 11-21 I get around 2000 miles out of a
chain but with the 12-25 the chain is still measuring like new at 2000
miles... I usually replace the cassette at 4000 miles just to be
safe...

I could change the cassette to a 13-26 and have a straight block from
13 to 19, and use the 23 and 26 as spacers to clean up the chain line,
but this seems a bit wasteful, since I would never use those gears and
I really would like to have the 12... I have no use for an 11 but it
does clean up the chain line a bit when I am in the small ring... I
plan to switch to a Chorus cassette next time I buy one and I have
figured out how to make an all steel 11-21 by taking an existing 11-23
Chorus cassette and replacing the last two pinned sprocket pairs with a
10S-78C and a 10S-91C... The 12-25 I have right now works fine but I do
miss the 18...

My questions:

1) Will the steel 17, 18, 19 and 21 wear a lot better than the titanium
versions, or will they still wear out fast due to the chain line not
being optimal? Well, I know they will wear better of course, but will
they wear a lot better? Or am I asking too much from the system with
the crazy chain angles?

2) In the big picture, what "ranking" does chain cleanliness, chain
line, sprocket size and torque applied to the chain affect wear? I have
read that the smaller sprockets wear faster due to having less teeth to
distribute the wear and because the chain has to flex more when it
passes over them... I am guessing that how hard you pedal affects
sprocket wear too... I just wonder if torque is an order of magnitude
more important to chain/sprocket wear than chain line, or vice versa...
Like, would you say the wear breakdown is torque 70%, sprocket size
15%, chain line $10% and cleanliness 5%?

3) Will the spacer effect from a 13-26 be worth the trouble given I
never will use those gears?

There are several reasons why I like a corn cob cassette... It allows
me to run a very short chain (104 versus 108) which makes shifting seem
a bit cleaner, and it keeps the chain from whacking around in some gear
combos when there is some slack... I can also dial in the B screw to
get the upper pulley real close to the cassette which seems to make
shifting even cleaner... I will also admit to liking the corn cob look
better than a pie pan look, and like I mentioned before, a 21 is all I
really need...

PS - I really don't need the 11... I am considering the 12-23 as well,
except it is impossible to make an all steel 12-23...

PPS - I'm not cross chaining from the 53 very much... On my bike the
chainstays are so short I get pretty big angles going from the 39...
And yes, it is set up properly...

Thanks!

--
Mike (Xyzzy)

On 24 Jun 2005 08:29:11 -0700, "Xyzzy"
wrote:

[snip]

PPS - I'm not cross chaining from the 53 very much... On my bike the
chainstays are so short I get pretty big angles going from the 39...
And yes, it is set up properly...

Thanks!

Dear Xyzzy,

It's unlikely that the "pretty big angles" caused by a
shorter-than-usual chainstay between the 39 tooth and the
larger rear cogs are a problem.

In fact, the shortness of the chainstays is unlikely to make
a visible difference in the angle of the chain.

To expand on an example from a recent thread about shifting
problems being attributed to short chainstays . . .

If chainstay distance (bottom-bracket-to-axle) shrinks only
13 mm (412 mm to 399 mm), and the largest-to-smallest rear
cog width is about 40 mm, then the maximum chain arc widens
from 5.545 degrees for the long 412 mm chainstay to 5.725
degrees for the short 399 mm chainstay--only 0.18 degrees:

5.725 degrees 5.545 degrees
| |
.__. highest gear | 40 mm
front_______________________.__. lowest gear |
0 399 412 mm

Here's a list that gives some idea of the degrees of chain
arc for various chainstay lengths with a rear cluster 40 mm
wide:

chain chain
stay arc
length degrees
412 5.545
402 5.682
392 5.826
382 5.978
372 6.137
362 6.305
352 6.483
342 6.671

Starting with a fairly normal 412 mm chainstay, we have to
shorten the chainstay about 70 mm (almost 3 inches) to
increase the maximum chain angle 1 degree.

If you have a midget chainstay only 302 mm long, the arc
increases 2 degrees to 7.545 degrees

But these are maximum cross-chaining arcs--if you don't do
much cross-chaining, then these tiny differences are muc
smaller.

The whole derailleur concept took off only after people
realized that chain alignment was nowhere near as crucial as
originally thought:

"The rest of the 1930's derailleurs suffered from the false
assumptions of the period: Chains have to be perfectly
aligned. (Although the wide, stiff chains of the period gave
some support to this theory, it is not true in general.)"

--Frank Berto, "The Dancing Chain," p. 142, 2nd edition

Whatever's wearing out your gears, it's not likely that the
short chainstay is the culprit.

Good luck,

Carl Fogel

2) In the big picture, what "ranking" does chain cleanliness, chain
line, sprocket size and torque applied to the chain affect wear? I have
read that the smaller sprockets wear faster due to having less teeth to
distribute the wear and because the chain has to flex more when it
passes over them... I am guessing that how hard you pedal affects
sprocket wear too... I just wonder if torque is an order of magnitude
more important to chain/sprocket wear than chain line, or vice versa...
Like, would you say the wear breakdown is torque 70%, sprocket size
15%, chain line $10% and cleanliness 5%?


Two fairly recent tests were done, one by Spicer:
http://www.sdearthtimes.com/et1199/et1199s13.html,

and another by Kyle and Berto:
http://www.ihpva.org/pubs/HP52.pdf

Both showed that chainline made a negligible difference in the
efficiency... ie it was too small to quantify. Kyle and Berto tested a
Shimano LX system (MTB triple) in the big/big and small/small combos,
so the chainline should have been more extreme than you are seeing.

Seems strange to me... if it's not inefficient, then why does it make
so much *noise*? I avoid the two largest cogs with the big ring, and
the two smallest with the small ring, because it just sounds like
something is wrong. I suppose it is possible that wear could be
accelerated when cross-chaining even though there aren't increased
mechanical losses... though I'm having a hard time imagining how that
could be.

Spicer also found that efficiency *increases* with higher chain tension
(with power held constant), instead of decreasing or staying constant
as I would expect. In fact, this effect was so great that it more than
offset the increased losses from using smaller cogs/rings (ie a 39/14
may be more efficient than the equivalent 53/19, or a compact 50/34
crank may be more efficient than a 53/39). Though Kyle and Berto's data
also shows a slight trend in the same direction, it is less convincing
than Spicer's.

Spicer ran a test with a clean unlubricated chain as well, and guess
what... the efficiency didn't go down! On that one I suspect that he
did not get the chain *completely* clean... just a *little* residual
lubrication would have lasted the duration of the test.

I hope that helps... even though the results are a bit
counter-intuitive.

-Ron Ruff

I just discovered that the Spicer test report can also be downloaded:

http://www.ihpva.org/HParchive/PDF/hp50-2000.pdf

Power inefficiency (wear ) requires force (tension) and relative motion.


wrote in message
oups.com...


Spicer also found that efficiency *increases* with higher chain tension
(with power held constant), instead of decreasing or staying constant
as I would expect.

I've taken a look at the Spicer report, and it appears that friction
and wear are *not* the primary cause of inefficiency in a bicycle chain
drive. In the conclusions he states:

"From the results of this study, it
appears that the efficiency of the bicycle
chain drive depends intimately on
the chain operation as it engages and
departs from the sprockets on the hightension
part of the drive. Owing to the
high efficiencies measured under high
chain tensions, friction can account for
only a few percent of the overall losses.
Most probably, mechanical losses that
are not converted to thermal energy in
the drive account for the remainder of
the measured loss."

He found that drive losses actually went *down* with higher tensions,
even though the frictional losses increased (based on theory and
infrared imaging).

Unfortunately, he doesn't address what these "mechanical losses that
are not converted to thermal energy" might be. It looks like he
originally assumed that the losses would be frictional, then found that
this was not true and left it at that.

Could it be vibration due to chordal variations? This is much greater
with small cogs, and would match the trend in that respect (smaller
cogs = lower efficiency). Possibly, this effect would be reduced by
higher tensions, matching the other major trend (high tension = higher
efficiency).

I've also thought about "losses" attributed to changing the momentum
(direction) of the chain as it passes over the cogs and pulleys... but
not enough to know if this makes sense.

I've had no luck finding information on the internet. I would have
thought that the efficiency and wear of chain drives would have been
thoroughly studied and tested in industry since they are so common, but
I've not found any published test data or modeling. Does anyone here
have knowledge of this?


Getting back to the OP's questions:

Based on the test data, cross chaining, using small cogs, and failing
to lube your chain... will *not* slow you down... at least not enough
to notice. However, all of these things *should* wear out your cogs and
chain more quickly... but I don't know exactly how much...

I won't let that stop me from making some semi-educated guesses,
though.

Regarding the effect of cog size:

In Spicer's pre-test theoretical analysis he gave 63% higher frictional
loss to a 52/11 compared to a 52/21, and 52/16 was 28% higher than a
52/21. This was with the chains in line. This should be a good estimate
of relative wear.

Concerning cross-chaining, he stated:

"the frictional effects of offset
should be small compared with
pin/bushing losses. Any effect would
necessarily appear in the largest offset
conditions"

So, he is saying it "should" be small... but I doubt that is true if
the cross-chaining is extreme. He doesn't report having tested this
case with thermal imaging; only power/efficiency measurements. As far
as I know, the quantitative effects of cross-chaining on wear have not
been modeled or tested. You should be ok in any case using all but the
small ring with 2 smallest cogs, and the big ring with 2 largest cogs.
You can even violate this "rule" occasionally... if you can stand the
sound.

Regarding lubrication:

He found no appreciable rise in friction (based on thermal imaging)
even with a dry chain. Again, I doubt that his chain was truly "dry",
so I wouldn't put much faith in that result. He also didn't subject the
chain to dirt, mud, water, etc like a real chain would see.

-Ron Ruff