What mileage to replace a chain?

[PCR]

Is there a rule of thumb of when to replace a chain? Specificly Shimano 105 group? I don't have a chain gauge :( Thanks all.

[Louis]

I don't have a chain gauge

If I were you I'd buy one, they aren't that expensive. The only guidelines I know of are based on % of chain stretch.

Alternatively, you can measure with a ruler, but the go-no go tools are so simple.

Park Info (http://www.parktool.com/products/detail.asp?cat=5&item=CC%2D3)

ITEM # CC-3
Chain Wear Indicator

A worn chain shifts poorly and wears sprockets at an accelerated rate. The CC-3 is a "go, no go" gauge designed to accurately indicate when a chain reaches .75% and 1% wear, the points at which most chain manufacturers suggest replacement. Made from precision, laser cut steel.

http://www.parktool.com/images/products/productimages/det_CC-3-60_2006628_79503.jpg

[xjoex]

I second the cc-3. Its easy to use and cheaper than a new chain:
http://4.bp.blogspot.com/_Jz4f840sc_c/SJpT5craT2I/AAAAAAAAA1E/juvz_6CYQ_Y/s400/100_2138.JPG

-Joe

[TimD]

Four thousand miles if you take good care of it (e.g., keep it out of the rain, really clean, and adequately lubricated). Two thousand miles if you don't.

Or once per season, whichever comes first!

YMMV.
Tim

[KeithS]

A worn chain shifts poorly and wears sprockets at an accelerated rate. The CC-3 is a "go, no go" gauge designed to accurately indicate when a chain reaches .75% and 1% wear, the points at which most chain manufacturers suggest replacement. Made from precision, laser cut steel.



To answer the mileage question. There isn't one. It's probably over a thousand but less than 3 thousand. There are a lot of factors, what kind of riding you do, how clean you keep your gear, lubrication, weather, cassette wear. Last year I got 1000 miles on one chain and have over 2000 on the next one. I know I need a new one now, shifting on the trainer is showing the familiar signs. I have DA and run Ultegra chains.

So my answer is the same as Louis' buy the guage or do what I do - ask the man who owns one.

[Pete Serotta]

To answer the mileage question. There isn't one. It's probably over a thousand but less than 3 thousand. There are a lot of factors, what kind of riding you do, how clean you keep your gear, lubrication, weather, cassette wear. Last year I got 1000 miles on one chain and have over 2000 on the next one. I know I need a new one now, shifting on the trainer is showing the familiar signs. I have DA and run Ultegra chains.

So my answer is the same as Louis' buy the guage or do what I do - ask the man who owns one.


Yep, that is what I do.

The chain used for a week of rain everyday last Oct in Vermont on dirt (MUD) roads lasted a very short time.....In Raleigh on very little wet roads and cleaning it, I use 2500 miles as a change point. That is me and no science involved. I just like the crispness of a new chain in shifting and it is cheaper than cassettes (which will wear faster on worn (not worn out) chains.
Mega milers and folks who religiously clean and lube chain will not agree.....
but it works for me. :D

[oldpotatoe]

Is there a rule of thumb of when to replace a chain? Specificly Shimano 105 group? I don't have a chain gauge :( Thanks all.

Do ya have a ruler?

1/16 inch stretch over 12 links, pin to pin, over 1 foot->replace. use the least expensive compatible chain.

[RPS]

As others have said, I wouldn’t go by miles because wear rate is dependent of many factors; including chain tension. I ride mostly on very flat terrain most of the year and my chains last a lot of miles. And because I wear the cassette out in the middle much quicker I’ve found that replacing chains early to avoid cassette replacement doesn’t same me much if any money. I therefore ride the bike until shifting is not up to my expectation and then replace the Ultegra chain and cassette at the same time. If I were the type that changed wheels (and therefore cassettes) often between bikes I’d probably take a different approach but I usually don't.

Also, as 9-speed Ultegra cassettes become more expensive and harder to find compared to chains I might start replacing chains a little more often, but for now I prefer to go by what feels and works right (or good enough) rather than chain elongation.

[znfdl]

The only guidelines I know of are based on % of chain stretch.


A worn chain shifts poorly and wears sprockets at an accelerated rate. The CC-3 is a "go, no go" gauge designed to accurately indicate when a chain reaches .75% and 1% wear, the points at which most chain manufacturers suggest replacement. Made from precision, laser cut steel.


Chains do not stretch, the rollers get worn. Get a chain checker tool, one will last a lifetime.

[Dave]

Chains don't stretch, as in the plates becoming longer, but the length of the chain does increase. That is not due to roller wear, but due to wear between the pins and the bushings formed into the inner sideplates. Roller wear has no effect on chain length. A chain with 1% elongation will be a full link or about 1/2 inch longer over it's full length.

Those chain checker tools are not accurate, since they add roller wear at each end of the measurement and all chains do not measure the same between the rollers, when new, due to difference in the roller diameter and clearance with the bushing. Some chain tools will show a new Shimano chain to be worn by .25%. Even if you discount that initial false wear, those tools can still report close to twice the actual elongation.

All you need to measure the elongation of a chain is a ruler, but a precision 12" machinists rule is even better. Measuring elongation works for most chain brands, but not all. I've used Campy chains for 6,000 miles and measured about .2% elongation. By that measurement, the chain should have been good for a lot more use, but in reality it was shot. The rollers were severely worn and the side clearance was nearly twice that of a new chain. With a Campy chain I go now set my use limit at a roller spacing of .240 inch, compared to a new reading in the .200-.205 range. That can only be done by alternating the use of several chains on one cassette. If a single chain is used to that wear point, it's quite likely that a second new chain would skip on one or more cogs.

I'm just now getting some reports on 11 speed chain life. I usually get more miles from my chain than the average person, probably due to more frequently lubing, but I've got a couple of reports from users who took caliper readings of the roller wear and came up with less wear than I'm seeing (.215 inch spacing at 2500 miles). They also report the same sort of elongation that I get - not even measureable after 2500 miles. I'm thinking that the gritty conditions I ride in are contributing to the faster roller wear that I get. I also do a lot more climbing than the average person with at least 4,000' of climbing on every 50 mile ride. I spend a lot of time in my 34T little ring that creates more chain tension.

[jscottyk]

As per usual, Sheldon summed it up nicely (http://www.sheldonbrown.com/chains.html#stretch).

And Pardo's commentary (http://pardo.net/bike/pic/fail-004/000.html) is very helpful too.

Short answer, go with the ruler method. It's cheaper and more accurate.

[Louis]

We may be splitting hairs about whether or not chains "stretch", but if when new the distance between pins for n links is x inches, and two months later the distance between those same n links is x+dx inches, then in my opinion the chain has stretched. The change does occur in the rollers, not the side plates, but the net effect is that the chain is longer, which for me is the definition of stretching.

Louis

[Dave]

Louis...

The rollers have absolutely nothing to do with the chain becoming longer, as measured between the pins. The additional length is due to wear between the pins and the "holes" (bushings) that are formed in the inner side plates. I could take a chain apart, remove all of the rollers and put it back together. It would still measure the same length.

A 50 inch long chain (when new) will be 50.5 inches long when the wear on each of the 100 pins and bushings totals .005 inch. That is 1% elongation or stretch. That .005 inch of wear might be an equal amount of wear on the pin and the bushing or a very unequal amount. It does not matter if the pin wear is .001 inch and the bushing .004 - it's the total that adds up to .5 inch.

Calling this stretch or elongation does not matter; it's just semantics. Some people just think of stretch as meaning plastic deformation. While it is possible to pull hard enough on a piece of steel to make it permanently longer (plastic deformation) that is not the cause of chain stretch.

[Louis]

Calling this stretch or elongation does not matter; it's just semantics.

Agreed.

The statement was made that chains to not stretch. All I was trying to say is that they do.

Honestly, I couldn't care less about the details of how it happens. I only care about how the changes in the chain might affect the operation of the drivetrain and the other hardware on the bike. I'm not really into tribology, but agree that it's an important field of research that can help improve our daily lives.

[godfrey1112000]

I have averaged over the last 3-4 years about 2750 miles per chain

Or when ever it starts to slip up the last steep hill before home

keep it lubed and clean all the time, always shimano on a shimano 7900 crank set

bottom bracket about 25,000-27,500 just replaced my fsa ceramic bb, it lasted two and a half years



Park Info

ITEM # CC-3
Chain Wear Indicator

A worn chain shifts poorly and wears sprockets at an accelerated rate. The CC-3 is a "go, no go" gauge designed to accurately indicate when a chain reaches .75% and 1% wear, the points at which most chain manufacturers suggest replacement. Made from precision, laser cut steel.

[AndrewS]

Whether chains actually stretch or not is immaterial to the wear that's important. And that wear is measured at the cog contact points - the chain rollers. Even if the links haven't stretched, the rollers may have worn enough to develop slack. That's why I measure with a chain wear tool, not a ruler.

I read a discussion once where it was suggested that roller wear doesn't matter since the rollers are still going to have the same spacing as the links. Which is true, but only in a straight line. When the chain is wrapped around a cog, the links settle to the minimum circumference while the worn rollers can move out to a greater radius and therefore circumference. That means the rollers and links will be spaced differently while actually touching the cogs. And since increased roller spacing is the devil that prematurely wears cassette cogs, that's the spacing that must be checked.

So either replace your chain more often than necessary, or use a chain wear indicator. A ruler is likely to give you an indication too late.

I don't understand why Sheldon Brown correctly points out that the elongation is primarily in the rollers and then tells you to measure the links. Maybe he was thinking about the rollers in a straight line configuration and didn't realize how they would behave in a curve.

[markie]

I think I have 10,000 miles on my current chain and it has been through a couple of hard Winters with nasty salty commutes.

I wonder if I ride fixed because I cannot afford gears? :p

[Dave]

AndrewS....

Nice theory about the rollers, but using a chain checker is still not accurate. If you want to monitor roller wear it should be done by placing caliper tips between the rollers at each end of one pair of inner plates. Otherwise you are combining some amount of elongation and some amount of roller wear and have no idea what you are measuring.

Too add some facts to your conjecture, most chain brands will elongate fast enough that the roller wear will not be excessive before .5% elongation is reached, so nothing is gained by measuring roller wear alone. If you use a common chain checker, the roller wear can be nearly as much as the elongation, so the tool reports twice the actual elongation. If the initial false wear that most of these tools report is not subtracted, then the reading is even more inaccurate. The tool might report .75% elongation, when it's really only .25% actual elongation by adding .25% roller wear and .25% of false wear due to the roller diameter and clearances of a new Shimano chain.

You are correct however, that with certain chains - Campy is the only one I've found - measuring elongation with a scale is not a good choice. The pins are much harder and the chain may elongate very little, while the rollers wear at a rate the is at least similar to other brands. When I left a chain in use for 6,000 miles, it showed very little elongation, but the roller spacing increased from .200 inch to about .240 and it caused chain skip when a new chain was installed. I took the chain apart and found the roller ID to be at least .010 inch larger and the OD about .005 inch smaller. When you measure the change in roller spacing, you're measuring the wear at three places, mulitplied times two - (bushing OD, roller OD and roller ID).

[palincss]

AndrewS....

Too add some facts to your conjecture, most chain brands will elongate fast enough that the roller wear will not be excessive before .5% elongation is reached, so nothing is gained by measuring roller wear alone. If you use a common chain checker, the roller wear can be nearly as much as the elongation, so the tool reports twice the actual elongation. If the initial false wear that most of these tools report is not subtracted, then the reading is even more inaccurate. The tool might report .75% elongation, when it's really only .25% actual elongation by adding .25% roller wear and .25% of false wear due to the roller diameter and clearances of a new Shimano chain.

If I understand you correctly, the worst thing that can happen if you rely on the common Park-type chain checker is that you'll replace the chain prematurely, i.e., the tool will indicate more wear than is really there. That's fairly harmless compared to either not replacing the chain soon enough, or relying solely on mileage without measuring wear at all, correct?

[Ti Designs]

Nice theory about the rollers, but using a chain checker is still not accurate. If you want to monitor roller wear is should be done by placing caliper tips between the rollers at each end of one pair of inner plates. Otherwise you are combining some amount of elongation and some amount of roller wear and have no idea what you are measuring.


You can also wake up every morning before sunrise and go to some observation point just to make sure the sun really does rise. Most people just assume it'll rise 'cause it's always done it before. It's a lot like setting a mainatinance schedule for replacing the chain - all my other chains have worn, I know the expense of letting it go too long...

The bicycle is a simple machine, please don't change that.

[Dave]

If I understand you correctly, the worst thing that can happen if you rely on the common Park-type chain checker is that you'll replace the chain prematurely, i.e., the tool will indicate more wear than is really there. That's fairly harmless compared to either not replacing the chain soon enough, or relying solely on mileage without measuring wear at all, correct?

Yes. The worst that can happen is you throw away chains that may be half worn, or perhaps even 1/4 worn out.

The whole idea of changing chains at only .5% elongation is to keep the sprocket teeth shaped such that they will still engage with a brand new chain. Roller wear without excessive elongation can cause the same wear problem. When I wore out only one sprocket by using a single chain for 6,000 miles, I did not throw it away. I just continued to use the cassette with slightly used chains. It may only take 3-500 miles of use on a new or newer cassette to make a chain compatible with a sprocket that does not engage with a new chain.

Rather than toss chains when partially worn, I use several chains in a rotation so the sprockets are alwasy mated to chains with similar amounts of wear. This may allow the chains to be used far longer. That decision is up to the user. You still have to get each chain in the rotation onto the cassette before so much wear occurs that the last chain skips. If you do that and continue rotating the chains rather than tossing them, you should get far more miles from a given number of chains and one cassette.

Side wear can affect the shifting and some people will toss a chain based soley on that, even if they never measure it. They just feel that the shifting is not up to par and toss the chain, even if the real problem is the need for a shift cable tension adjustment. I measured the side clearance of my chain with 6,000 miles on it and found it to be around .013 inch, which is nearly twice that of a new chain. FWIW, KMC chains have more side clerance when new than most other brands.

[Dave]

You can also wake up every morning before sunrise and go to some observation point just to make sure the sun really does rise. Most people just assume it'll rise 'cause it's always done it before. It's a lot like setting a mainatinance schedule for replacing the chain - all my other chains have worn, I know the expense of letting it go too long...

The bicycle is a simple machine, please don't change that.

The same type of maintenance schedule can be used with a chain rotation, except the chains are cleaned, lubed and kept for reuse instead of being tossed in the trash with lots of life left in them. You're making up complications that don't really exist.

The expense of tossing half worn chains can easily exceed the cost of replacing a cassette. It all depends on the cost of the cassette compared to the chains.

I use all steel cogs and my cassettes cost about the same as three chains. If I use 3 chains and one cassette to ride 15,000 miles and some other person uses 6 chains to do the same thing, the other guy has already spent enough to buy a whole new cassette. If the cassette is a real expensive SR 11 model, then he's just spent an extra $120-150 and got nothing for it.

I don't spend any more time measuring chains than others do with a chain checker. For the ultimate in simplicity, a simple plug gage can indicate the limit of roller wear. I made my own gage by grinding down a 6mm hex wrench to about .070 in thick. If that gage ever drops between the rollers, the chain gets tossed. If I was using Shimano or KMC chains, I might use a 1/4 inch gage instead of 6mm, since the roller spacing is larger to begin with.

[Ti Designs]

interesting concept - you keep multiple chains in rotation which means the chain wear is going to match the cog wear assuming the cogs wear much slower. You're also smart enough to make a vert simple go/no-go gauge to check the chain (people who use digital calipers on chains scare me, they tend to be the same ones who show up at the bike shop with digital scales and ask to see out entire collection of 5mm hardware).

I'm far too lazy. I run 9-speed, and I have a large collection of cassettes. I've found that when 10-speed came out people were more than happy to give away their 9-speed stuff, most of which was pretty much unused. I have a half dozen 11-21 cassettes 'cause as one guy put it "you can't ride that around here". I also have a 100:1 rule - ride the bike 100 hours for every hour I work on it. Cleaning chains doesn't fit into that plan.

My track bike is a different story. I have multiple chains of different lengths, kept clean and in bags with the number of links written on the side. The 100:1 rule doesn't work for track 'cause there's so much time waiting around...

My fixed gear is the worst case. New England roads and the amount of salt they put down makes chain maintainance pointless. I put a new chain on in the fall (an $11 expense) and run it 'til it don't work no more. I had one chain with so many stiff and rusted links that it heated up while I was riding and started to smoke. I did what any New Englader would do if they noticed their engine smoking, I added oil.

[AndrewS]

Dave,

I don't understand your objection. A chain wear tool will measure the total amount of increase in roller distance from a combination of two factors: Link elongation and roller wear. And that's a good thing.

Your cogs get worn prematurely because the chain rollers (the only real contact point between chain and cogs) are no longer settling into the correct 1/2" per tooth distance. When that happens the teeth are eroded. The teeth can't tell why the rollers are no longer 1/2" apart, they just receive the results.

So if your rollers can now settle at a distance of 1/2" + X, you are going to have damage done to your cassette.

I think maybe the confusion on this comes from a misunderstanding of why you are supposed to replace a worn chain in the first place. It is purely for the freewheel/cassette longevity. A well worn chain, elongated or not, is not going to break or damage anything else but the cassette cogs.

[Dave]

AndrewS....

Combining roller wear and actual elongation into a jumbled-up and meaningless reading is not a good thing. If you think that roller wear changes the pitch of the chain, you're wrong. You could cut out all the rollers and the chain would still have the same pin to pin measurement and overall length.

When a chain reaches 1% elongation, the wear on a each pin and bushing pair is only .005 inch. A chain checker usually measures no more than 12 pins, or a total of .060 inch. The increase in roller spacing may be 10 times the wear on a pin and bushing, so just one roller pair can have about the same increase in spacing as 10 pins. Mixing the wear of many pins that actually increase the chain pitch with one pair of rollers that does NOT change the chain pitch makes no sense at all. If roller wear had the same effect as change in pitch, a chain wouldn't last even 500 miles.

The amount of chain wear that each person's cassette can handle is unique. That's because we all ride on differing terrains at different speeds and with different power outputs. When I rode moderately rolling terrain around Kansas City, my cassettes lasted far longer because I did not use any one cog as often. Riding in the mountains, I use only a few cogs for a long time and wear just a couple of them out far more quickly than the others. It's more critical that my chains not be left in use for too long before switching to a new one. There is no one amount of wear that will guarantee that you'll be able to use 3, 4 ,5 or 6 chains before chain skip occurs with a new chain. The suggestion to toss chains at .5% elongation or any specific amount or roller spacing increase will not be valid for every rider. It may be perfect for some and far too conservative for others. That's why I prefer the rotation method because it totally eliminates the chance of having chain skip.

Here's some more info on chains and chan wear. Just keep in mind that all modern chains are now the "bushingless" type. The other type hasn't been in wide use for at least 30 years.

http://pardo.net/bike/pic/fail-004/000.html

[AndrewS]

You could cut out all the rollers and the chain would still have the same pin to pin measurement and overall length.


You're missing my point. There has to be a reason to declare a chain "worn out". Overall length of a chain, by itself, is a completely arbitrary measure that doesn't relate to anything else on the bicycle.

3 questions to clarify the problem:
Would you agree that we replace worn chains to increase cassette life?


Would you agree that the only important contact between the cassette cogs and the chain are the rollers?


Would you agree that the geometry of the the interaction between the chain and cassette cogs are the root cause of worn chain related cassette wear?

If you can't answer yes to all, please explain.

[Louis]

Would you agree that we replace worn chains to increase cassette life?

As an aside, this is really two questions in one:

a) Do we do this to protect the more expensive cassette?

(I think most folks would say, yes, that is what I've been told and why I don't wait until the thing can't shift and the chain skips.)

b) Perhaps the more important question: Does this really extend the useful life of the cassette, or is is just an "old mechanic's tale" ? I have no idea if this is really true and have seen no rigorous scientific data to say one way or the other.

Louis

[palincss]

All this focus on the cassette -- does no one care about the chain rings as well?

[Louis]

All this focus on the cassette -- does no one care about the chain rings as well?

I haven't seen the numbers on this, but I think they tend to last longer (more teeth to share the load) and the shifting in front is not affected by wear as much as in back.

[AndrewS]

Care? It's all relative. I have an old Raleigh I bought for commuting. The old chain and freewheel were incredibly worn and out of spec. The chain rings were fine.

There's a lot more surface area on a chain ring and that seems to make them more resistant to chain wear, despite being aluminum. That's why I say that chain to cog wear is the only important factor in chain replacement.

[Dave]

3 questions to clarify the problem:
Would you agree that we replace worn chains to increase cassette life?


Would you agree that the only important contact between the cassette cogs and the chain are the rollers?


Would you agree that the geometry of the the interaction between the chain and cassette cogs are the root cause of worn chain related cassette wear?

If you can't answer yes to all, please explain.

1) Chains are changed to increase cassette life but not always with a new chain. I rotate several chains rather than tossing half worn ones.

2) Of course, but you are confused when it comes to distinguishing roller spacing from chain pitch. A new Shimano chain may have .010 inch more roller spacing than a new Campy chain, but both have exactly the same pitch. Despite the larger roller spacing, the Shimano chain is not partially worn when new.

3) Your last statement is far too vague to respond to. Please read the articles that I've linked. Cogs will eventually wear out, not matter how often a chanin is changed. Someone recently posted that they install a new chain every 1000 miles. They will use 15 chains for 15,000 miles but probly get no more cog life than I do with 3 chains. Eventually a new chain will skip on the most worn cogs.

In a nutshell, any chain wear pattern that differs too much from that of the next chain installed may cause chain skip, but only on the the most worn cog(s).

If you've got some better idea on how to manage chains and improve cassette life, please explain it. I've got no idea what you're driving at.

[Louis]

1) Chains are changed to increase cassette life but not always with a new chain. I rotate several chains rather than tossing half worn ones.

Assuming that chains wear faster than cassettes or chain rings it makes sense to rotate through several chains as they and the other parts the chains work with age. That way, much like a married couple, then grow old together and are most similar.

I think it comes down to time and hassle vs money. How much extra life does one buy and how much money do you save if you're willing to spend the extra time rotating chains more often? In addition, given the up-front cost of buying multiple chains, it's not a simple problem to figure out.

I'd bet that the two most important factors are 1) Do you enjoy working on the bike? (I don't mind, after all, I'm an engineer, but I'd rather be riding) and 2) Are you the methodical type who likes to have a system? For some things I like to do this, for others I don't bother and take life as it comes.

Louis

[AndrewS]

Dave, the article you posted lays it all out, yet still comes to this extraordinary conclusion "However, bushing/roller wear does not affect chain performance... Thus, bushing/roller wear should not be included in overall wear measurements."

This conclusion is wrong because rollers have their own pitch. When the chain is relatively new the rollers have the same pitch as pins and bushings because they are all concentric.

BUT, as the rollers wear down on the inside, they are free to move fore and aft or up and down. While a chain tool measures this as fore and aft, the cassette experiences the roller movement as up or toward the outside of their circumference.

In other words, when a cog, pin, bushing and roller are all new, they all have exactly the same pitch because all the chain parts are concentric. But if the rollers have wear on the inside, the pressure on the chain will force the pins down toward the center of the cog and the rollers will move out toward a greater circumference. That means that the center points of the rollers are now further away from each other than the pins or teeth they're supposed to be sitting between. This effectively increases their pitch. And the links are now sitting lower between the cogs, effectively decreasing their pitch. And since the cogs only contact the rollers, the increased spacing on the rollers is what is going to wear the teeth, as illustrated in that article.


And just to put it another very different way, as the worn rollers come off the jockey pulley they are less likely to cleanly mate with the cassette cog because their location is variable. So they have to run over the teeth and change position to settle between them. They aren't dropping into position - the teeth are forcing them into their spots. Unworn rollers drop cleaning between the teeth because they are centered over the pin and the pin has the same pitch (or spacing) as the teeth.


As to Shimano Vs. Campy Vs. whoever, I've never measured a new or moderately used chain by any brand that failed a Chain Wear Tool test prematurely. But I'm guessing you have?


Whatever the mechanics, the rollers are the things that come in contact with and wear the cogs. I don't see how any sort of explanation can posit that the rollers themselves are immaterial to this process and not worthy of measure.

[AndrewS]

As to whether I have a "better idea to improve cassette life": Nope. I replace the chain when the wear tool says to.

What I really find convoluted is your mixing worn chains idea. How does giving a chain a breather help? Chains and cogs don't wear together - they wear in different ways that are not complimentary. So how is replacing a worn chain with another worn chain going to increase the life of a cassette vs. putting a new chain on?

[Louis]

So how is replacing a worn chain with another worn chain going to increase the life of a cassette vs. putting a new chain on?

It comes down to this:

Can a worn chain and a worn cassette together perform better than that same worn cassette and a new chain? I've been told they can. I've never done a scientific experiment to find out.

[AndrewS]

I can't imagine how. As the article demonstrates, worn chains have increased pitch. A worn cog has the same pitch it always had, but rounded off teeth. Putting an increased pitch chain in that circumstance is only going to cause the cog to wear faster because the chain is going to spending more time close to the top of the teeth.

Sometimes a worn out cassette will only work with its worn out chain, but that's after they're both beyond spec.

[GBD]

Hi Andrew,

I am with you on the chain wear logic.
The measuring methods discussed may be precise, but too complicated for me.

I ride Campi record chain and Chorus 10 speed cassettes, FSA chain rings.

Chains lasts over 8K miles, cassettes last longer. I use PG2000 lubricant and probably lube every 250 miles.
Use Park CC2 tool.

The chain is no cleaned often, maybe twice a year.

I have no shifting or skipping problems what so ever.
So if it works, why change?

My cadence is usually over 90, so perhaps this extends the life of chain.

[Dave]

As to whether I have a "better idea to improve cassette life": Nope. I replace the chain when the wear tool says to.

What I really find convoluted is your mixing worn chains idea. How does giving a chain a breather help? Chains and cogs don't wear together - they wear in different ways that are not complimentary. So how is replacing a worn chain with another worn chain going to increase the life of a cassette vs. putting a new chain on?


You're just not comprehending the information provided. Chains and cogs do wear-in together. When you state that they don't you're ignoring the facts. Rollers don't have "their own pitch". As they wear, the roller OD gets smaller and the ID gets larger. Those two dimensions also vary quite a bit between different brands of new chains, but the pitch is not different because of it. I've measured new chains with roller spacing varying from .200 inch (Campy) to .215 inch (KMC). They still have the same pitch.

Using chains in a rotation does not necessarily increase cog life, but it could reduce the number of chains that you purchase by one half or even more. In my examples, I don't talk about getting more life from a casssette, just a similar life with far fewer chains. In a previous example, I compared using 3 chains in a rotation to using 6 brand new chains in succession. With my method, the cassette and chains are retired at my discretion, after several rotations. With the all-new chain method, you might toss each chain when a chain checker reads .5% or .75% elongation, but when you install the 7th new chain (or maybe only the 4th) it may skip, ending the cassette's life. You will have spent twice as much for chains and not ridden any further. If you were smart and not tossed all of those old chains, you could start reusing them and get a lot more miles from the cassette. That most-worn cog is not worn beyond use, it's just too worn to mate with a new chain.

If you change chains when the chain tool "says to", is that at .5%, .75% or 1%? I've already explained why a Shimano chain will indicate that it's .25% "worn" when new, even though it has no wear. The tool is misreading what is mostly a larger roller ID as elongation, but the chain is not worn, and the rollers do not have a larger pitch. You are completely wrong if you think that a new Campy chain with .200 inch roller spacing has rollers with a different pitch than a new Shimano chain with .210 inch roller spacing. The length across any given number of pins will be identical and the pitch will be the same, when new. Put a chain checker on those two chains it will show the Campy chain to have zero wear and the Shimano .25%.

Regardless of how often you change a chain, it will still wear out the most-used cogs at some point. You might go through 6, 8, or 10 chains, but at some point, a new chain will skip on one or more of the cogs and you will then consider that cassette to be trashed. The cost of using too many chains can easily exceed the price of a new cassette.

There is no difference between removing a partially worn chain and replacing it with a new one and rotating a group of chains, all closely matched for wear. All you're doing is maintaining a small difference in wear between the chain in-use and the next chain to be used on the cassette. If you allow any chain to remain in use for too long, you can still encounter chain skip because the cog teeth have worn-in to fit that chain and are mismatched to the next chain that is installed.

The chains in a rotation are not being given "a breather" any more than the new chains sitting in a box at the store.

It is true that chains with a large increase in pitch will accelerate cog wear, so if all the chains in a rotation approach a true 1% elongation, cog wear will increase. At some point, the most-used teeth can become hooked and not properly engage even a worn chain, but at that point, the cassette should have seen a great many miles of use.

What I suggest to you is to stop throwing away the chains that you are removing when the chain checker tells you to. Keep them until the day comes when a new chain skips on your worn cassette. Remove that new chain and keep it for use with a new cassette, some time in the future. Then start reusing those old chains. They won't skip and you'll get thousands of free miles instead of a pile of trash.

[Louis]

I run Shimano 9-spd and replace my chains with a new one when my "go, no-go" tool says that most sections I measure are above 0.75% and one or two are above 1%.

I never have a problem with skipping, but have noticed that toward the end of a given chain's life the shifting has deteriorated and can not be improved through adjustment. Once the new chain has been installed the shifting magically improves and stays that way for a long, long time. Whenever that happens I tell myself "Gee, I should have changed that chain a while ago."

I conclude that in my case even with a cassette that has seen a lot of miles a new chain seems to work better than an older one. YMMV

Louis

PS Nearly all of my riding is on hilly terrain, so I use cogs up and down the range, not just one or two in the middle of the cassette. I assume this extends the useful life the cassette.

[Dave]

Louis...

There's nothing wrong with your approach to changing chains, but eventually you will find out which cogs get the most use when they skip right after installing a new chain. I'd been riding for 20 years before I had it happen. I change bikes and equipment so often that I never wore out a cassette. I've put a third chain on a cassette at 8,000 miles and not had a chain skip problem, but that was back when I rode more moderate terrain. It was also the last new chain I put on that bike, right before I sold it. I always do a major overhaul before selling a used bike (new tires, chain, cables, tape, hub service).

My first chain skip occurred when I deliberately kept using one chain far longer than normal - 6,000 miles. The elongation was minimal at that point, but the rollers were shot and the side clearance was large. A new chain skipped on the 19T cog. That's how I found out that you can create chain skip even with elongation as low as .2%.

[AndrewS]

Dave, I think we're more alike than different, so do understand that I appreciate your thinking on this. I will summerize the problem, as I see it, once more and leave it at that:

Geometry: An 18 tooth cog has a nominal circumference of 9 inches. 18 chain links is 9 inches. The rollers support the links from the bottom so when they are wrapped around the cog they also have a circumference of 9 inches.

On the bike, the chain touches 180 degrees of the cog, or 4.5 inches of a 18 tooth cog. If the rollers are not supporting the chain under tension, the pins will seek the lowest points in the cogs where the circumference of the cog is LESS THAN 4.5 inches. But since the links do total 4.5 inches, the links coming on and off the cog are not going to be centered over the teeth, causing tooth wear.

I know you understand this concept since I have read your posts about using Shimano shifting on Campy cogs. It's no different. Unworn rollers are the interface that make sure the links and cogs are interfacing at the same radius.



As for chain rotation, I just don't understand how you manage to declare one chain worn enough for temporary retirement (to be used later), and another chain TOO WORN for continued use on the well worn cassette. It would seem to me that you would have to either remove the first chain before it actually appears worn, or keep the second one so long that it might be the culprit in wearing out the cassette.

Of course, if one were to have a half worn chain sitting around for some reason, might as well try it before tossing the cassette. But I can't see how anyone with one bike is able to put this idea into action with any precision.

[SPOKE]

Yep, that is what I do.

The chain used for a week of rain everyday last Oct in Vermont on dirt (MUD) roads lasted a very short time.....In Raleigh on very little wet roads and cleaning it, I use 2500 miles as a change point. That is me and no science involved. I just like the crispness of a new chain in shifting and it is cheaper than cassettes (which will wear faster on worn (not worn out) chains.
Mega milers and folks who religiously clean and lube chain will not agree.....
but it works for me. :D

i'm his mechanic......about 2000-2500 for replacement of 10spd or 11spd chains. helps keep cassettes and chainrings in good shape and shifting stays crisp.

[Dave]

AndrewS...

You're making this much too difficult. Rather than try to analyze the cog to chain relationship, forget about it and pay attention to how chain and cogs really work together. The most important thing to remember is that a new chain skips on overly-worn cogs. You put new chains on as often as you like, but eventually, one of them will skip on the most-worn cogs. A worn chain should not just start skipping while in use, with cogs that it has worn-in with.

I thought this chain rotation thing was kind of a no-brainer. Buy a new cassette and 3 chains. Let's say the chains are Shimano, so they elongate rapidly. I'd use a scale to check true elongation and remove that first chain before it reaches .5% elongation. The exact amount is not that critical. Do the same for each of the three chains. Being conservative, you should not have a problem with chain skip when that third new chain goes into use. If you do, then the chains should have been rotated sooner. The key is to pick a conservative amount of wear that you're sure will not cause a skipping problem when that third new chain is put into use. With Campy 11 having thinner cogs (same 1.6mm as Shimano 10), I decided to rotate mine at 1500-2000 miles.

If you want to use a chain checker, that's fine too, just keep in mind that even .75% by the chain checker is not likely to be more than .3-.4% actual elongation. If you know that a particular amount of wear on the chain checker allows you to use 3 or more chains on a cassette, change them at that point.

After all three chains have seen some use, start using them again, just don't leave any one chain in use any longer than it was the first time - pay attention to how many miles it took to produce that elongation. You could also rotate the chains at more frequent intervals - it's not that critical.

Toss the chains whenever you think that shifting is deteriorating or they reach 1% true elongation (by scale measurement). You should get about twice the useage from each chain and just as many miles from your cassette.

With Campy 10 chains, I rotated my chains every 2000-2500 miles, on the first useage. I judged chain wear by roller spacing since there will be very little elongation at that mileage. A spacing of .220 inch is about half worn, IMO and a good time to rotate. After that, I use a master link and might change chains every couple of weeks or whenever I decided that the chain needed cleaning. If you keep the other chains cleaned and lubed, you just slap a new one on and clean the other one later. I'll trash the chains when the roller spacing gets to about .240 inch, unless I feel that the side wear is excessive and causing shifting problems.

The only difference between my procedure and what most people do is I don't put on a 4th, 5th or 6th new chain until I finally get enough wear on the cassette that I getting skipping when one of those new chains is installed. My 4th, 5th and 6th chains are half-used, not new. The practice of trashing chains long before they are really worn, in an effort to lengthen cassette life just costs you a lot more money in chains and may yield no more cassette life.

[Jack Brunk]

My head hurts big time. Over a chain really?

Can we get a engineer section only please. :beer:

[sspielman]

My head hurts big time. Over a chain really?

Can we get a engineer section only please. :beer:


Sort of sheds new light on Toyota's problems, eh?....

[Dave]

Some folks need a reading comprehension class or less imagination. They get some silly idea in their head about how things work and won't give up on it. Sometimes if you explain the same thing six different ways, a light finally comes on. Sometimes not.

I believe in the results of real world experience, even if it contradicts something an "expert" has written. In this case at least one expert says that elongation (real elongation, measured with a scale) is the only chain wear factor to be concerned with. I proved that to be wrong twice in one year. After wearing out one cog on a steel cassette in 6,000 miles, I wore out two Ti cogs on another cassette in only 4,000 miles. Even the chain with 6,000 miles on it had no more than .2% elongation. Both chains were lubed every 100 miles or less. IMO, it's simply a case of cogs becoming worn to fit smaller diameter rollers that causes a new chain (with larger rollers) to skip. The bottom line is not to leave any one chain in use for too long. In my case, the fast wear could also be attributed to riding in the mountains with a 28T chainring that creates a high chain tension.

[bobscott]

I always read Dave's posts knowing his ideas arise not from myth but from testing theories.

It is clear to me that this approach has put him in a place that he understands chains at another level from the rest of us.

We're very lucky to have his input.

bobscott

[Dave]

Apparently I left AndrewS wanting for more information.

About that chain and cog relationship - I can only speculate on exactly what occurs when a chain with lots of roller wear, but little elongation creates a wear pattern that will make a new chain skip, but as I stated earlier, it has to be the shape worn into the teeth from the smaller rollers that won't mate with a new chain having larger rollers. A chain with only a few hundred miles of use on it, with little roller wear and little elongation will not skip on that same "worn-out" cog, while a new one will. I know this because I tried it. I just accept the facts of how the parts really work together. I don't need a scientific explanation for everything. You don't have to accept any of this as fact.

If you read Jobst Brandt's article about chains, he even claims that in theory a new chain and a new cogs won't mate, but provides no explanation:

"Skipping Chain

That a new chain does not want to engage used sprockets may be
obvious, but in theory a new chain cannot freely engage a new rear
sprocket under load even though it has the same pitch as the chain.
That is because the tooth being engaged would be under load and this
is a contradiction in itself. Therefore, a slightly worn sprocket,
that has pockets in its load bearing face is even more reticent to
engage a new chain with perfect 1/2 " pitch.

Sprockets, by the way, do not change pitch when they wear, only the
tooth form changes because the number of teeth remains the same and
the base circle remains essentially unchanged for normal sprocket
wear. On fixed gear bicycles that are ridden until the chain is 5%
out of pitch, the base circle may be slightly reduced but this only
makes the fit of a new chain worse.

Without a strong chain tensioner or a non derailleur gear, the chain
has insufficient tension on its slack run to engage the sprocket when
under tension. In contrast, engagement on the tension side, as on the
crank sprocket, generally succeeds even with substantial tooth wear
because the tension encourages engagement. This condition, however,
enhances "chainsuck", the failure of the chain to disengage the
chainwheel. This can generally not occur without a long arm
derailleur, common to most MTB's and therefore, road bicycles normally
experience a grunchy disengagement instead of a chain jam.

A new chain has a pitch of exactly one half inch. A sprocket, worn by
a longer pitch (worn) chain, has hooked teeth but with the correct
pitch. The hooked profile is formed by the rollers of a worn chain as
they exit the sprocket under load, whereas rollers of a new chain with
correct pitch exit under no load because the load is transferred to
the next roller before disengagement. However, with hooked sprockets
the new chain cannot engage under load because its pitch doesn't allow
it to get over the hook and into the next pocket. These differences
are only a few thousandths of an inch but enough to prevent engagement
when the previous roller is fully engaged.

As a chain wears it concentrates more of its load on the last tooth of
a sprocket before disengagement because the chain pitch no longer
matches the pitch of the sprocket. This effect sometimes breaks off
sprocket teeth. The load concentration on the sprocket also
accelerates wear and is another reason to replace a chain at 1/16th
inch wear.

Jobst Brandt <jbrandt@hpl.hp.com> "