Chain Stretch

I've been thinking about chain wear, sometimes called chain stretch,
and have done a bit of research on the subject.

One method is to lay the chain on a flat surface and measure the wear
over, perhaps 12 inches of chain length, from the head of one pin to
another. But modern multi speed chains are a bit more complex then the
old fashioned chains and the rollers on a modern chain are not
supported by the pins but by protrusions on the inner surface of the
inner links thus does the distance from roller to roller relate to
distance from pin to pin?

Another method is to ignore the pin to pin distance and simply measure
the roller to roller distance using a chain gauge. But I have also
read that when comparing roller to roller measurement to pin to pin
measurement there is not necessarily a correlation, or in other words
a pin to pin measurement might show one thing while the roller to
roller might show a totally different wear pattern. In addition I read
that in at least one case the roller to roller wear was not constant
and varied from place to place in the length of the chain

Brandt, I believe, wrote a treatise on chain measuring gauges and
argued that nearly all of them gave an incorrect figure for wear, or
perhaps, did it the wrong way.

So the question is what is the best system to use to avoid unnecessary
sprocket wear, assuming that sprockets cost more and are more trouble
to change than chains.

There seems to be three options. One, to use a ruler and measure from
pin to pin. Two to use a chain tool and measure from roller to roller.
Or three, to use some combination of the two.

Or perhaps there is a fourth - ignore the whole thing as a tempest in
a tea pot :-)
--
Cheers,

John B.

On 9/15/2017 4:48 AM, John B. wrote:

I've been thinking about chain wear, sometimes called chain stretch,
and have done a bit of research on the subject.

One method is to lay the chain on a flat surface and measure the wear
over, perhaps 12 inches of chain length, from the head of one pin to
another. But modern multi speed chains are a bit more complex then the
old fashioned chains and the rollers on a modern chain are not
supported by the pins but by protrusions on the inner surface of the
inner links thus does the distance from roller to roller relate to
distance from pin to pin?

Another method is to ignore the pin to pin distance and simply measure
the roller to roller distance using a chain gauge. But I have also
read that when comparing roller to roller measurement to pin to pin
measurement there is not necessarily a correlation, or in other words
a pin to pin measurement might show one thing while the roller to
roller might show a totally different wear pattern. In addition I read
that in at least one case the roller to roller wear was not constant
and varied from place to place in the length of the chain

Brandt, I believe, wrote a treatise on chain measuring gauges and
argued that nearly all of them gave an incorrect figure for wear, or
perhaps, did it the wrong way.

So the question is what is the best system to use to avoid unnecessary
sprocket wear, assuming that sprockets cost more and are more trouble
to change than chains.

There seems to be three options. One, to use a ruler and measure from
pin to pin. Two to use a chain tool and measure from roller to roller.
Or three, to use some combination of the two.

Or perhaps there is a fourth - ignore the whole thing as a tempest in
a tea pot :-)

Interrupted sideplate chain does indeed wear faster than
full roller chain. However both economy of manufacture and
side flex (for index shifting) are better with interrupted
sideplates.

Generally, chain wear is measured with enough tension to
take up any slack, not merely laid out on a table.

The outer plates are joined by the rivet. The innies float
and exhibit wear. By measuring 24 rivets' worth of slop we
can effectively get an expanded 'vernier scale' of the very
small per-rivet clearance change. Since our functional
aspect for chain-to-sprocket efficiency is pitch, a
rivet-t-rivet measurement seems right to me and all our
gauges here measure that.

See section #8d.2 he
http://www.faqs.org/faqs/bicycles-faq/part3/

--
Andrew Muzi
www.yellowjersey.org/
Open every day since 1 April, 1971

On Fri, 15 Sep 2017 07:44:29 -0500, AMuzi wrote:

On 9/15/2017 4:48 AM, John B. wrote:

I've been thinking about chain wear, sometimes called chain stretch,
and have done a bit of research on the subject.

One method is to lay the chain on a flat surface and measure the wear
over, perhaps 12 inches of chain length, from the head of one pin to
another. But modern multi speed chains are a bit more complex then the
old fashioned chains and the rollers on a modern chain are not
supported by the pins but by protrusions on the inner surface of the
inner links thus does the distance from roller to roller relate to
distance from pin to pin?

Another method is to ignore the pin to pin distance and simply measure
the roller to roller distance using a chain gauge. But I have also
read that when comparing roller to roller measurement to pin to pin
measurement there is not necessarily a correlation, or in other words
a pin to pin measurement might show one thing while the roller to
roller might show a totally different wear pattern. In addition I read
that in at least one case the roller to roller wear was not constant
and varied from place to place in the length of the chain

Brandt, I believe, wrote a treatise on chain measuring gauges and
argued that nearly all of them gave an incorrect figure for wear, or
perhaps, did it the wrong way.

So the question is what is the best system to use to avoid unnecessary
sprocket wear, assuming that sprockets cost more and are more trouble
to change than chains.

There seems to be three options. One, to use a ruler and measure from
pin to pin. Two to use a chain tool and measure from roller to roller.
Or three, to use some combination of the two.

Or perhaps there is a fourth - ignore the whole thing as a tempest in
a tea pot :-)

Interrupted sideplate chain does indeed wear faster than
full roller chain. However both economy of manufacture and
side flex (for index shifting) are better with interrupted
sideplates.

Generally, chain wear is measured with enough tension to
take up any slack, not merely laid out on a table.

The outer plates are joined by the rivet. The innies float
and exhibit wear. By measuring 24 rivets' worth of slop we
can effectively get an expanded 'vernier scale' of the very
small per-rivet clearance change. Since our functional
aspect for chain-to-sprocket efficiency is pitch, a
rivet-t-rivet measurement seems right to me and all our
gauges here measure that.

See section #8d.2 he
http://www.faqs.org/faqs/bicycles-faq/part3/

Yup, Brandt (in all his glory :-)

I've been using an 18" stainless scale (ruler) which assuming a 1%
wear limit is 3/16". (old eyes need big marks :-)

I recently came across another chain measuring scheme that seemed to
make good sense. Simply pull on the chain at the front of the chain
ring forwards to see how much it moves away from the sprocket teeth.
--
Cheers,

John B.

On 9/15/2017 10:20 PM, John B. wrote:
On Fri, 15 Sep 2017 07:44:29 -0500, AMuzi wrote:

On 9/15/2017 4:48 AM, John B. wrote:

I've been thinking about chain wear, sometimes called chain stretch,
and have done a bit of research on the subject.

One method is to lay the chain on a flat surface and measure the wear
over, perhaps 12 inches of chain length, from the head of one pin to
another. But modern multi speed chains are a bit more complex then the
old fashioned chains and the rollers on a modern chain are not
supported by the pins but by protrusions on the inner surface of the
inner links thus does the distance from roller to roller relate to
distance from pin to pin?

Another method is to ignore the pin to pin distance and simply measure
the roller to roller distance using a chain gauge. But I have also
read that when comparing roller to roller measurement to pin to pin
measurement there is not necessarily a correlation, or in other words
a pin to pin measurement might show one thing while the roller to
roller might show a totally different wear pattern. In addition I read
that in at least one case the roller to roller wear was not constant
and varied from place to place in the length of the chain

Brandt, I believe, wrote a treatise on chain measuring gauges and
argued that nearly all of them gave an incorrect figure for wear, or
perhaps, did it the wrong way.

So the question is what is the best system to use to avoid unnecessary
sprocket wear, assuming that sprockets cost more and are more trouble
to change than chains.

There seems to be three options. One, to use a ruler and measure from
pin to pin. Two to use a chain tool and measure from roller to roller.
Or three, to use some combination of the two.

Or perhaps there is a fourth - ignore the whole thing as a tempest in
a tea pot :-)

Interrupted sideplate chain does indeed wear faster than
full roller chain. However both economy of manufacture and
side flex (for index shifting) are better with interrupted
sideplates.

Generally, chain wear is measured with enough tension to
take up any slack, not merely laid out on a table.

The outer plates are joined by the rivet. The innies float
and exhibit wear. By measuring 24 rivets' worth of slop we
can effectively get an expanded 'vernier scale' of the very
small per-rivet clearance change. Since our functional
aspect for chain-to-sprocket efficiency is pitch, a
rivet-t-rivet measurement seems right to me and all our
gauges here measure that.

See section #8d.2 he
http://www.faqs.org/faqs/bicycles-faq/part3/

Yup, Brandt (in all his glory :-)

I've been using an 18" stainless scale (ruler) which assuming a 1%
wear limit is 3/16". (old eyes need big marks :-)

I recently came across another chain measuring scheme that seemed to
make good sense. Simply pull on the chain at the front of the chain
ring forwards to see how much it moves away from the sprocket teeth.


The ancient rule of thumb for that is replace chain when a
4mm key will slip under the chain. Index shifting will be
poor when a 5mm key fits. You cannot stand on the pedals
when a 6mm key slides under the links. That's a very rough
gradient and not always accurate, but a starting point anyway.

--
Andrew Muzi
www.yellowjersey.org/
Open every day since 1 April, 1971

On Sat, 16 Sep 2017 08:00:38 -0500, AMuzi wrote:

On 9/15/2017 10:20 PM, John B. wrote:
On Fri, 15 Sep 2017 07:44:29 -0500, AMuzi wrote:

On 9/15/2017 4:48 AM, John B. wrote:

I've been thinking about chain wear, sometimes called chain stretch,
and have done a bit of research on the subject.

One method is to lay the chain on a flat surface and measure the wear
over, perhaps 12 inches of chain length, from the head of one pin to
another. But modern multi speed chains are a bit more complex then the
old fashioned chains and the rollers on a modern chain are not
supported by the pins but by protrusions on the inner surface of the
inner links thus does the distance from roller to roller relate to
distance from pin to pin?

Another method is to ignore the pin to pin distance and simply measure
the roller to roller distance using a chain gauge. But I have also
read that when comparing roller to roller measurement to pin to pin
measurement there is not necessarily a correlation, or in other words
a pin to pin measurement might show one thing while the roller to
roller might show a totally different wear pattern. In addition I read
that in at least one case the roller to roller wear was not constant
and varied from place to place in the length of the chain

Brandt, I believe, wrote a treatise on chain measuring gauges and
argued that nearly all of them gave an incorrect figure for wear, or
perhaps, did it the wrong way.

So the question is what is the best system to use to avoid unnecessary
sprocket wear, assuming that sprockets cost more and are more trouble
to change than chains.

There seems to be three options. One, to use a ruler and measure from
pin to pin. Two to use a chain tool and measure from roller to roller.
Or three, to use some combination of the two.

Or perhaps there is a fourth - ignore the whole thing as a tempest in
a tea pot :-)

Interrupted sideplate chain does indeed wear faster than
full roller chain. However both economy of manufacture and
side flex (for index shifting) are better with interrupted
sideplates.

Generally, chain wear is measured with enough tension to
take up any slack, not merely laid out on a table.

The outer plates are joined by the rivet. The innies float
and exhibit wear. By measuring 24 rivets' worth of slop we
can effectively get an expanded 'vernier scale' of the very
small per-rivet clearance change. Since our functional
aspect for chain-to-sprocket efficiency is pitch, a
rivet-t-rivet measurement seems right to me and all our
gauges here measure that.

See section #8d.2 he
http://www.faqs.org/faqs/bicycles-faq/part3/

Yup, Brandt (in all his glory :-)

I've been using an 18" stainless scale (ruler) which assuming a 1%
wear limit is 3/16". (old eyes need big marks :-)

I recently came across another chain measuring scheme that seemed to
make good sense. Simply pull on the chain at the front of the chain
ring forwards to see how much it moves away from the sprocket teeth.


The ancient rule of thumb for that is replace chain when a
4mm key will slip under the chain. Index shifting will be
poor when a 5mm key fits. You cannot stand on the pedals
when a 6mm key slides under the links. That's a very rough
gradient and not always accurate, but a starting point anyway.

Interesting. Thanks.
--
Cheers,

John B.

On Friday, September 15, 2017 at 8:21:14 PM UTC-7, John B. wrote:
On Fri, 15 Sep 2017 07:44:29 -0500, AMuzi wrote:

On 9/15/2017 4:48 AM, John B. wrote:

I've been thinking about chain wear, sometimes called chain stretch,
and have done a bit of research on the subject.

One method is to lay the chain on a flat surface and measure the wear
over, perhaps 12 inches of chain length, from the head of one pin to
another. But modern multi speed chains are a bit more complex then the
old fashioned chains and the rollers on a modern chain are not
supported by the pins but by protrusions on the inner surface of the
inner links thus does the distance from roller to roller relate to
distance from pin to pin?

Another method is to ignore the pin to pin distance and simply measure
the roller to roller distance using a chain gauge. But I have also
read that when comparing roller to roller measurement to pin to pin
measurement there is not necessarily a correlation, or in other words
a pin to pin measurement might show one thing while the roller to
roller might show a totally different wear pattern. In addition I read
that in at least one case the roller to roller wear was not constant
and varied from place to place in the length of the chain

Brandt, I believe, wrote a treatise on chain measuring gauges and
argued that nearly all of them gave an incorrect figure for wear, or
perhaps, did it the wrong way.

So the question is what is the best system to use to avoid unnecessary
sprocket wear, assuming that sprockets cost more and are more trouble
to change than chains.

There seems to be three options. One, to use a ruler and measure from
pin to pin. Two to use a chain tool and measure from roller to roller.
Or three, to use some combination of the two.

Or perhaps there is a fourth - ignore the whole thing as a tempest in
a tea pot :-)

Interrupted sideplate chain does indeed wear faster than
full roller chain. However both economy of manufacture and
side flex (for index shifting) are better with interrupted
sideplates.

Generally, chain wear is measured with enough tension to
take up any slack, not merely laid out on a table.

The outer plates are joined by the rivet. The innies float
and exhibit wear. By measuring 24 rivets' worth of slop we
can effectively get an expanded 'vernier scale' of the very
small per-rivet clearance change. Since our functional
aspect for chain-to-sprocket efficiency is pitch, a
rivet-t-rivet measurement seems right to me and all our
gauges here measure that.

See section #8d.2 he
http://www.faqs.org/faqs/bicycles-faq/part3/

Yup, Brandt (in all his glory :-)

I've been using an 18" stainless scale (ruler) which assuming a 1%
wear limit is 3/16". (old eyes need big marks :-)

I recently came across another chain measuring scheme that seemed to
make good sense. Simply pull on the chain at the front of the chain
ring forwards to see how much it moves away from the sprocket teeth.

That test shows more the wear on the sprockets than that on the chain.

On Sat, 16 Sep 2017 13:32:45 -0700 (PDT), wrote:

On Friday, September 15, 2017 at 8:21:14 PM UTC-7, John B. wrote:
On Fri, 15 Sep 2017 07:44:29 -0500, AMuzi wrote:

On 9/15/2017 4:48 AM, John B. wrote:

I've been thinking about chain wear, sometimes called chain stretch,
and have done a bit of research on the subject.

One method is to lay the chain on a flat surface and measure the wear
over, perhaps 12 inches of chain length, from the head of one pin to
another. But modern multi speed chains are a bit more complex then the
old fashioned chains and the rollers on a modern chain are not
supported by the pins but by protrusions on the inner surface of the
inner links thus does the distance from roller to roller relate to
distance from pin to pin?

Another method is to ignore the pin to pin distance and simply measure
the roller to roller distance using a chain gauge. But I have also
read that when comparing roller to roller measurement to pin to pin
measurement there is not necessarily a correlation, or in other words
a pin to pin measurement might show one thing while the roller to
roller might show a totally different wear pattern. In addition I read
that in at least one case the roller to roller wear was not constant
and varied from place to place in the length of the chain

Brandt, I believe, wrote a treatise on chain measuring gauges and
argued that nearly all of them gave an incorrect figure for wear, or
perhaps, did it the wrong way.

So the question is what is the best system to use to avoid unnecessary
sprocket wear, assuming that sprockets cost more and are more trouble
to change than chains.

There seems to be three options. One, to use a ruler and measure from
pin to pin. Two to use a chain tool and measure from roller to roller.
Or three, to use some combination of the two.

Or perhaps there is a fourth - ignore the whole thing as a tempest in
a tea pot :-)

Interrupted sideplate chain does indeed wear faster than
full roller chain. However both economy of manufacture and
side flex (for index shifting) are better with interrupted
sideplates.

Generally, chain wear is measured with enough tension to
take up any slack, not merely laid out on a table.

The outer plates are joined by the rivet. The innies float
and exhibit wear. By measuring 24 rivets' worth of slop we
can effectively get an expanded 'vernier scale' of the very
small per-rivet clearance change. Since our functional
aspect for chain-to-sprocket efficiency is pitch, a
rivet-t-rivet measurement seems right to me and all our
gauges here measure that.

See section #8d.2 he
http://www.faqs.org/faqs/bicycles-faq/part3/

Yup, Brandt (in all his glory :-)

I've been using an 18" stainless scale (ruler) which assuming a 1%
wear limit is 3/16". (old eyes need big marks :-)

I recently came across another chain measuring scheme that seemed to
make good sense. Simply pull on the chain at the front of the chain
ring forwards to see how much it moves away from the sprocket teeth.

That test shows more the wear on the sprockets than that on the chain.

Yes, it may well do that.

--
Cheers,

John B.

On 9/17/2017 1:28 AM, John B. wrote:
On Sat, 16 Sep 2017 13:32:45 -0700 (PDT), wrote:

On Friday, September 15, 2017 at 8:21:14 PM UTC-7, John B. wrote:
On Fri, 15 Sep 2017 07:44:29 -0500, AMuzi wrote:

On 9/15/2017 4:48 AM, John B. wrote:

I've been thinking about chain wear, sometimes called chain stretch,
and have done a bit of research on the subject.

One method is to lay the chain on a flat surface and measure the wear
over, perhaps 12 inches of chain length, from the head of one pin to
another. But modern multi speed chains are a bit more complex then the
old fashioned chains and the rollers on a modern chain are not
supported by the pins but by protrusions on the inner surface of the
inner links thus does the distance from roller to roller relate to
distance from pin to pin?

Another method is to ignore the pin to pin distance and simply measure
the roller to roller distance using a chain gauge. But I have also
read that when comparing roller to roller measurement to pin to pin
measurement there is not necessarily a correlation, or in other words
a pin to pin measurement might show one thing while the roller to
roller might show a totally different wear pattern. In addition I read
that in at least one case the roller to roller wear was not constant
and varied from place to place in the length of the chain

Brandt, I believe, wrote a treatise on chain measuring gauges and
argued that nearly all of them gave an incorrect figure for wear, or
perhaps, did it the wrong way.

So the question is what is the best system to use to avoid unnecessary
sprocket wear, assuming that sprockets cost more and are more trouble
to change than chains.

There seems to be three options. One, to use a ruler and measure from
pin to pin. Two to use a chain tool and measure from roller to roller.
Or three, to use some combination of the two.

Or perhaps there is a fourth - ignore the whole thing as a tempest in
a tea pot :-)

Interrupted sideplate chain does indeed wear faster than
full roller chain. However both economy of manufacture and
side flex (for index shifting) are better with interrupted
sideplates.

Generally, chain wear is measured with enough tension to
take up any slack, not merely laid out on a table.

The outer plates are joined by the rivet. The innies float
and exhibit wear. By measuring 24 rivets' worth of slop we
can effectively get an expanded 'vernier scale' of the very
small per-rivet clearance change. Since our functional
aspect for chain-to-sprocket efficiency is pitch, a
rivet-t-rivet measurement seems right to me and all our
gauges here measure that.

See section #8d.2 he
http://www.faqs.org/faqs/bicycles-faq/part3/

Yup, Brandt (in all his glory :-)

I've been using an 18" stainless scale (ruler) which assuming a 1%
wear limit is 3/16". (old eyes need big marks :-)

I recently came across another chain measuring scheme that seemed to
make good sense. Simply pull on the chain at the front of the chain
ring forwards to see how much it moves away from the sprocket teeth.

That test shows more the wear on the sprockets than that on the chain.

Yes, it may well do that.

Or not.

I happen to have some extreme "bad example" chainrings with
less than 20% tooth height [1] so I measured the root
diameter against a new chainring just now. Jobst is correct.
The wear is against the loaded side of the tooth, almost no
change at root. New 40tt ring root=155mm, utterly worn
out=154.8mm

[1]replaced on customers' bikes, used here for demo
purposes. The one I measured is more worn than this one

http://www.yellowjersey.org/photosfromthepast/worn2.jpg


--
Andrew Muzi
www.yellowjersey.org/
Open every day since 1 April, 1971

On Friday, September 15, 2017 at 8:21:14 PM UTC-7, John B. wrote:
On Fri, 15 Sep 2017 07:44:29 -0500, AMuzi wrote:

On 9/15/2017 4:48 AM, John B. wrote:

I've been thinking about chain wear, sometimes called ch
See section #8d.2 he
http://www.faqs.org/faqs/bicycles-faq/part3/

Yup, Brandt (in all his glory :-)

Disagree. Completely. Brandt shines much, much more brightly, where the subject has more depth. These few words, although correct and useful, are most assuredly NOT Brandt at his finest.

On Tue, 19 Sep 2017 11:11:00 -0700 (PDT), Doug Landau
wrote:

On Friday, September 15, 2017 at 8:21:14 PM UTC-7, John B. wrote:
On Fri, 15 Sep 2017 07:44:29 -0500, AMuzi wrote:

On 9/15/2017 4:48 AM, John B. wrote:

I've been thinking about chain wear, sometimes called ch
See section #8d.2 he
http://www.faqs.org/faqs/bicycles-faq/part3/

Yup, Brandt (in all his glory :-)

Disagree. Completely. Brandt shines much, much more brightly, where the subject has more depth. These few words, although correct and useful, are most assuredly NOT Brandt at his finest.

The term was meant to be sardonic.
--
Cheers,

John B.