Bicycle Mechanics - Chain wear gauge -- worth it?

I've always just use a ruler to measure my chain wear. Does anyone have a reason to switch to one of these chain wear indicators, like this one?: http://www.parktool.com/products/detail.asp?cat=5&item=CC%2D3 . Is there a benefit to using them versus rulers? (other than maybe being quicker). Thanks!

This was brought up in another very, very recent thread

Oops! I've been reading regularly for the last couple of months, so didn't bother with a search. Must've missed that one. Should've searched to begin with. It was a spur of the moment question. Sorry for the duplicate thread. Thanks for the help.

The campus bike co-op has two Park chain check tools, supposed to be the best ones available. One saidthat my chain is worn out, the other said it is almost new. They have been used for years by college students who don't know what they're doing, so I'll cut them slack on the quality of the tool. But there is no visual indication that they aren't working right.

I use a steel ruler with 1/64th inch markings at home, and that gets the job done.

^^^^ it was actually discussed just oposite. Rullers, when held parallel to a chain show the distance from the outter pin to outter pin, but do not show the wear on the rollers themselvers where the cogs contact the chain.

Use an undamaged chain wear tool. Simple.

^^^^ it was actually discussed just oposite. Rullers, when held parallel to a chain show the distance from the outter pin to outter pin, but do not show the wear on the rollers themselvers where the cogs contact the chain.


You're joking right?

I don't see a need for one. This was brought up in another very, very recent thread (a search always helps), and the conclusion was that rulers are just as, and probably more effective in determining chainwear without the need for the expense of a chainwear tool that also may be susceptible to damage in the toolbox so it gives incorrect readings.

I have that Park chain checker, and thought it was the silver bullet.

I developed a habit of installing a new chain when stretch reaches the .75% criteria (3/32"). Most say to replace at 1% criteria (1/8" stretch). When I installed the third chain on the same cassette, the cassette had excessively worn. Some possibilities exists that could explain the prematurely worn cassette: the chain gauge is out of whack, the abrasive off-road conditions accelerate wear regardless of chain stretch, or I'm not using the tool properly. I do lubricate my chain after every 4 hours of ride time with Prolink (this is probably excessive, but I like a quiet chain).

FWIW - the chain and cassette were both SRAM. Now, I'm running a Shimano XT cassette with SRAM PC991 chain.

Some say to replace at a 0.50% (1/16") criteria, some say to replace the chain and cassette together every time.

I'm back to using a steel ruler.

I have that Park chain checker, and thought it was the silver bullet.

I developed a habit of installing a new chain when stretch reaches the .75% criteria (3/32"). Most say to replace at 1% criteria (1/8" stretch). When I installed the third chain on the same cassette, the cassette had excessively worn. Some possibilities exists that could explain the prematurely worn cassette: the chain gauge is out of whack, the abrasive off-road conditions accelerate wear regardless of chain stretch, or I'm not using the tool properly. I do lubricate my chain after every 4 hours of ride time with Prolink (this is probably excessive, but I like a quiet chain).

FWIW - the chain and cassette were both SRAM. Now, I'm running a Shimano XT cassette with SRAM PC991 chain.

Some say to replace at a 0.50% (1/16") criteria, some say to replace the chain and cassette together every time.

I'm back to using a steel ruler.


i don't understand what you are saying. are you saying the park tool is at fault?

for years i used a rohloff caliber and now i use the park version, which costs about 12 bucks. i get about two chains per cassette. before when i was riding hard a chain would last about 2k miles. these days maybe twice that.

the park tool is inexpensive and it's a no-brainer.

ed rader

I don't see a need for one. This was brought up in another very, very recent thread (a search always helps), and the conclusion was that rulers are just as, and probably more effective in determining chainwear without the need for the expense of a chainwear tool that also may be susceptible to damage in the toolbox so it gives incorrect readings.


the park tool, which copies the rohloff caliber which has been used in europe for years, is about 12 bucks and is a solid piece of metal just like a steel ruler. what tool are you talking about?

ed rader

there's the "Go-no go" drop in solid one, and the swing bar one with the scale read thru a little window in the handle.

i can't count well so a ruler is tougher on me than the swing arm gauge.

Here's what I do: after a couple of thousand miles, I buy a new chain. But I don't install it, I just compare it to my old chain by lining them up with a nail or stick. If the old chain is stretched, I install the new one. Otherwise, the new one goes in the parts box for another 500 miles. Wash. Rinse. Repeat.

$12 isn't much, I know, but dollar for dollar, the chainchecker just doesn't stack up against a good engineer's steel ruler. A steel rule is good for a whole range of other measurements. Plus, the rule can tell you reasonably precisely the *amount* of wear that a no-go chainchecker cannot. If you want Park, get the one that measures four items instead of just one -- spokes, bearing balls, crank cotters, chain length. Plus maybe drop out width and a few other items along the way. But the Park rule I have used has the measurements printed on aluminium and they wear away with use... unlike an etched rule.

The chainchecker that I referred to was the swingbar one (Park CC2).

My favorite wrench did not trust them. He encouraged use of a simple ruler and measuring the links.

there's the "Go-no go" drop in solid one, and the swing bar one with the scale read thru a little window in the handle.

i can't count well so a ruler is tougher on me than the swing arm gauge.

+1 on the swing arm one. The Park CC-2 is a little more expensive at about $20 but might be better. The pins seem to be hardened steel drill rod that, with reasonable care, should not get knocked out of whack. I would recommend it over a precision steel rule, because of the wear being on the inner rollers (as someone else mentioned). I believe that you can see the wear on the inner rollers, its visible to the naked eye, but to also quantify the % of chain stretch on the Park CC-2 just helps clarify when its time (to replace the chain) that much more. My current method of cleaning my chain, where I put the dirty chain in a pan and scrub the individual links with a toothbrush soaked in clean mineral spirits and then the bottom of the pan wiped clean from time to time with clean paper towels that are thrown away -- has helped me to observe what the "dirt" is partly made up of. There are lots of little silver bits of metal in there. After cleaning my chain this way, its helped me to stay on top of keeping the drivetrain (individual cassette cogs, individual chainrings and chain) cleaned and lubed (I use prolink or white lightning epic currently), and also to get a new chain when stretch is at 0.75% or greater.

I use a cheap metal vernier caliper, available at Home Despot for less than $10. It measures to .1mm (which is .004 in.). First, I get a base measurement from a new chain. I use the inside measurement, set at 106mm, insert it against two rollers and measure. Thats the measurement for a new chain. Wear limit is .5 mm more.

You're joking right?

Explain

i don't understand what you are saying. are you saying the park tool is at fault?

for years i used a rohloff caliber and now i use the park version, which costs about 12 bucks. i get about two chains per cassette. before when i was riding hard a chain would last about 2k miles. these days maybe twice that.

the park tool is inexpensive and it's a no-brainer.

ed rader


I can't precisely fault the Park gauge (I have the one piece, go- no go checker), since there are many variable involved. For example, I ride off-road and perhaps the dusty/sandy condition in Southern CA will wear cassettes (like sandpaper) more than a stretched chain. Another variable is how I sample the measurements. I usually take three or four measurement around the chain. However, there could be a few spots that I don't measure where the stretch is beyond acceptable criteria, and those culprits could be causing premature wear. A chain has hundreds of moving parts, so variation is expected

When installing new chains at the 0.75% criteria, I was able to get two chains per cassette. I was hoping to get more chains per cassette. But you mention you get two chains per cassette, so perhaps my experience is normal.

Explain
Here's my take on it:

Wear on the outside of the chain rollers isn't significant. If you watch a chain roller engaging a cog tooth, you'll see that the chain roller doesn't really slide much on the tooth, the force is basically perpendicular to the contact surface. The place where there is a LOT of sliding/rubbing friction is between the chain pin and the side plates or bushings, and that's where most of the chain wear occurs.

While it's true that the roller-to-roller distance is what actually defines the pitch of the chain, the wear on the rollers progresses much more slowly than the wear on the pins. If you find a badly worn chain (e.g. one that's stretched by 2%), you'll see that the pins are visibly deformed, while the outside surface of the rollers is basically unscathed. Look at this photo from Sheldon Brown's site:
http://sheldonbrown.com/images/chain_wornpin.gif

There's no difference between measuring roller-to-roller distance or pin-to-pin distance: wear on the pins will cause the chain to elongate LONG before wear on the rollers becomes an issue.

Why does that big knotch look like its right in the center of the pin?

Speaking of Sheldon (http://sheldonbrown.com/chains.html)(a la the picture in moxfiyre's post), he has a good explanation of such things, as well as another method for determining chain/sprocket wear on the bike.

Why does that big knotch look like its right in the center of the pin?
Good eye! I have no non-bull**** response to that ... dangit :p

If you go to Sheldon's page, www.sheldonbrown.com/chains, his explanation asserts that most of the wear occurs where the pins rub against the side plates or bushings. Yet the picture strongly suggests that most of the wear has occurred where the pins rub against the inside of the rollers, as you have observed.

Maybe Sheldon is wrong on this one? I am now thoroughly confused...

Mox, I used to think exactly what you did, that the wear was mostly on the plates and pins, and not the rollers. BUT my own experience, at least with SRAM chains, has been contrary to that. And it's not the *outside* of the rollers that wears, I think it's the *inside* of the rollers, rubbing against the pseudo-bushings that are created by the plates. That makes sense too, because it's probably the first place where grit combines with lube to form grinding compound. If the rollers are made of softer metal than the plates, as seems to be the case with the SRAM chains, they are what will wear, on their inside surfaces. If the plates are softer, *they* will wear. If the pins are softer than the plates, the plates will wear them away like in the picture. But my chains still had pristine, perfect pins and measured at exactly 12 inches, when the rollers had worn enough to have significantly more space between them.

Of all the methods discussed above, I think that San Rensho's is best. It takes the exact measurement of the gap between adjacent rollers, which is the bottom line, regardless of what exactly creates that gap. I like gcl8a's comparison of new chain/old chain, too.

Mox, I used to think exactly what you did, that the wear was mostly on the plates and pins, and not the rollers. BUT my own experience, at least with SRAM chains, has been contrary to that. And it's not the *outside* of the rollers that wears, I think it's the *inside* of the rollers, rubbing against the pseudo-bushings that are created by the plates. That makes sense too, because it's probably the first place where grit combines with lube to form grinding compound. If the rollers are made of softer metal than the plates, as seems to be the case with the SRAM chains, they are what will wear, on their inside surfaces. If the plates are softer, *they* will wear. If the pins are softer than the plates, the plates will wear them away like in the picture. But my chains still had pristine, perfect pins and measured at exactly 12 inches, when the rollers had worn enough to have significantly more space between them.

Of all the methods discussed above, I think that San Rensho's is best. It takes the exact measurement of the gap between adjacent rollers, which is the bottom line, regardless of what exactly creates that gap. I like gcl8a's comparison of new chain/old chain, too.
Well, this has been very educational I must say. It sounds like wear on the insides of the rollers is a very significant part of chain stretch. I had no idea, but now I stand corrected :) Maybe I should point this thread out to Sheldon Brown? It seems like his own photo disagrees with the text on his page...

I have steel vernier calipers, so I think I'll start using San Rensho's method to measure chain wear. Seems accurate and pretty darn foolproof.

You have to look at an inner sideplate that doesn't have a rivet through it, as in the second picture. It is that shoulder that wears, I think, along with the roller to produce the elongation. Sheldon points out that the smooth radius of this shoulder on the inner sideplate is what makes bushinless chains last longer. I think, and I may be wrong, but the rivets are fixed in the outer sideplates (hence the reason you need a chain tool to get them past the friction fit), but the "bushing" of the inner sideplate as well as the roller must be free to move. It is the movement of the rivet on the inner links that causes the wear and elongation. One of the problems in all this description is that people assume a chain link is made up of only one rivet and the bits ahead of it, when in fact it comprises two outer plates, two inner plates, and a roller and a rivet in the middle to hold the lot together. Six parts altogether.

The wear per link is miniscule in normal circumstances, but is magnified somewhat when when spread over 12 full links or 12", so enabling a rule to be used for an accurate assessment of the average wear across the chain.

Well, this has been very educational I must say. It sounds like wear on the insides of the rollers is a very significant part of chain stretch. I had no idea, but now I stand corrected :) Maybe I should point this thread out to Sheldon Brown? It seems like his own photo disagrees with the text on his page...


But you see, that's just the point. That wear on the insides of the rollers doesn't have *anything* to do with lengthening the chain! That's why my chain still measured a perfect 12 inches, even though the rollers had worn inside enough to allow a larger gap between them! My point is that measuring the lengthening or "stretch" of a chain is really not a universal method of determining wear.

Mox, I used to think exactly what you did, that the wear was mostly on the plates and pins, and not the rollers. BUT my own experience, at least with SRAM chains, has been contrary to that. And it's not the *outside* of the rollers that wears, I think it's the *inside* of the rollers, rubbing against the pseudo-bushings that are created by the plates. That makes sense too, because it's probably the first place where grit combines with lube to form grinding compound. If the rollers are made of softer metal than the plates, as seems to be the case with the SRAM chains, they are what will wear, on their inside surfaces. If the plates are softer, *they* will wear. If the pins are softer than the plates, the plates will wear them away like in the picture. But my chains still had pristine, perfect pins and measured at exactly 12 inches, when the rollers had worn enough to have significantly more space between them.

Of all the methods discussed above, I think that San Rensho's is best. It takes the exact measurement of the gap between adjacent rollers, which is the bottom line, regardless of what exactly creates that gap. I like gcl8a's comparison of new chain/old chain, too.
The wear is in the pseudo bushing of the inner sideplates, not in the rollers. The rollers and rivets are either hardened or surface-hardened steel (ence their shiny surface). The sideplates are untreated, unhardened steel, I think. The shoulders on the inside of the inner sideplates wear both themselves and the rollers. Sheldon does not deny that the rollers wear somewhat, too. Your wear of rollers and non-elongation of the chain might be feasible, but... how was your shifting performance?

Generally, all this stuff is pretty esoteric and in practice, good chain maintenance and regular checking (irrespective of tool) will help extend the life of cogs and chainrings. Or you can be like me with one bike, and leave it all to work and wear together with fairly regular cleaning and get 12,000km out of 8sp Sora gearset :D.

But you see, that's just the point. That wear on the insides of the rollers doesn't have *anything* to do with lengthening the chain! That's why my chain still measured a perfect 12 inches, even though the rollers had worn inside enough to allow a larger gap between them! My point is that measuring the lengthening or "stretch" of a chain is really not a universal method of determining wear.
Well, it's true that the pitch of your chain has not changed with respect to the spacing of the pins. But there's now a lot of "slop" in the rollers, so that when the chain engages the cog teeth, the rollers will not be concentric with the pins. Will this cause accelerated wear of the cog teeth?? It doesn't seem like it should, actually, but I have been known to be dead wrong before :)

I have had one chain like yours which seemed to be barely worn when measured with a steel ruler, but when replaced the cassette would skip like crazy.

You have to look at an inner sideplate that doesn't have a rivet through it, as in the second picture. It is that shoulder that wears, I think, along with the roller to produce the elongation. ...One of the problems in all this description is that people assume a chain link is made up of only one rivet and the bits ahead of it, when in fact it comprises two outer plates, two inner plates, and a roller and a rivet in the middle to hold the lot together. Six parts altogether.


Right, when those are the parts that are wearing the fastest. But in my SRAM chains, it seems that the pins were super hard, because they showed no visible wear. The pseudo-bushings of the inner plates must also have been quite durable, more than the rollers at least, since the gap between rollers increased significantly, but not the length overall. Lengthening only shows up, as you said, when the shoulder or pseudo-bushing of the inner plate and/or the pin are the worn elements. The rollers have nothing whatsoever to do with the length of the chain. They're just along for the ride.

Well, this has been very educational I must say. It sounds like wear on the insides of the rollers is a very significant part of chain stretch. I had no idea, but now I stand corrected :) Maybe I should point this thread out to Sheldon Brown? It seems like his own photo disagrees with the text on his page...

No his photo doesn't necessarily disagree with what he writes. But you have to understand that the inner sideplates of a chainlink have those shoulders on them to keep the roller in place. The rollers don't contribute to the elongation of the chain. They *may* contribute to improper meshing of the chain with the sprockets and chainrings, but because the actual pitch of the chain might be unchanged, really that effect, I think, would be insignificant. Now noisy... that the chain might be.

The wear is in the pseudo bushing of the inner sideplates, not in the rollers. The rollers and rivets are either hardened or surface-hardened steel (ence their shiny surface). The sideplates are untreated, unhardened steel, I think. The shoulders on the inside of the inner sideplates wear both themselves and the rollers. Sheldon does not deny that the rollers wear somewhat, too. Your wear of rollers and non-elongation of the chain might be feasible, but... how was your shifting performance?
Yeah, this is basically what I thought happened before reading this thread. But if you look at the photo above, it's pretty clear that a LOT of wear has occurred on the inside of the roller and the mating surface of the rivet. As I mentioned, I don't quite see how this will cause accelerated cassette wear.

Perhaps roller wear CAN cause poor shifting: when the chain engages the cogs teeth, the rollers will have a lot of slop and so their position on the teeth may slip to one where they do not have as much positive contact. That is only speculation though...

No his photo doesn't necessarily disagree with what he writes. But you have to understand that the inner sideplates of a chainlink have those shoulders on them to keep the roller in place. The rollers don't contribute to the elongation of the chain. They *may* contribute to improper meshing of the chain with the sprockets and chainrings, but because the actual pitch of the chain might be unchanged, really that effect, I think, would be insigificant. Now noisy... that the chain might be.
Yes, I think you've stated it correctly. The roller wear won't contribute to an overall elongation, but may contribute to less-than-perfect engagement of the chain with the sprockets.

This also explains why the roller-to-roller measurement of the Park CC2 tool can be misleading, since it isn't actually measuring the chain pitch in terms of pin-to-pin measurement.

So, which way of measuring is correct in terms of prolonging the life of the cassette? That's the $1,000,000 question :) I think most of us wouldn't care too much about what's going on with the chain if we knew how to tell when it was damaging the cassette.

Well, it's true that the pitch of your chain has not changed with respect to the spacing of the pins. But there's now a lot of "slop" in the rollers, so that when the chain engages the cog teeth, the rollers will not be concentric with the pins. Will this cause accelerated wear of the cog teeth?? It doesn't seem like it should, actually, but I have been known to be dead wrong before :)

I have had one chain like yours which seemed to be barely worn when measured with a steel ruler, but when replaced the cassette would skip like crazy.

Hmmm. I need to think about this some more. I felt that it would accelerate wear on the teeth, because the gap was larger from roller to roller, but I see your point--that simply because each roller is no longer concentric doesn't necessarily mean that it would accelerate wear, and if the distance between the leading contact surface of each roller is still regulation, then it shouldn't matter if the roller isn't concentric. So I see what you mean. But then your experience with the one you replaced, would seem to indicate otherwise. My brain hurts.

Yeah, this is basically what I thought happened before reading this thread. But if you look at the photo above, it's pretty clear that a LOT of wear has occurred on the inside of the roller and the mating surface of the rivet. As I mentioned, I don't quite see how this will cause accelerated cassette wear.

Perhaps roller wear CAN cause poor shifting: when the chain engages the cogs teeth, the rollers will have a lot of slop and so their position on the teeth may slip to one where they do not have as much positive contact. That is only speculation though...
The shoulders on the inner plates are designed to come pretty close together. If you read the article about the history of chain design, the shoulders on old chains were bushings that met in the middle and provided a guide for the rollers, but also allowed the inner plates to move independetly of the outer plates and the rivet. The shoulders on the inner plates of modern chain replaced the bushing. The result was a reduced number of parts to manufacture together, and therefore manufacturing efficiency. The benefit appears to be improved longevity of chains. However, the role of the shoulders doesn't change, and because they are moving around the rivet, wear will occur on the rivet, although less so than on the shoulders.

lawkd, I think you can only determine your wear rates by using a micrometer to measure the diameters of all the moving components before fitting and at interval or when changing over chains.

The shoulders on the inner plates are designed to come pretty close together. If you read the article about the history of chain design, the shoulders on old chains were bushings that met in the middle and provided a guide for the rollers, but also allowed the inner plates to move independetly of the outer plates and the rivet. The shoulders on the inner plates of modern chain replaced the bushing. The result was a reduced number of parts of manufacture together, and therefore manufacturing efficiency. The benefit appears to be improved longevity of chains. However, the role of the shoulders doesn't change, and because they are moving around the rivet, wear will occur on the rivet, although less so than on the shoulders.
Yes, I understand how this type of wear works, and why it causes an overall elongation of the chain, and why that in turn wears out the cassette.

But the issue in my mind now is: does the OTHER kind of wear, that of the rollers, affect the longevity of the cassette as well?

Yes, I understand how this type of wear works, and why it causes an overall elongation of the chain, and why that in turn wears out the cassette.

But the issue in my mind now is: does the OTHER kind of wear, that of the rollers, affect the longevity of the cassette as well?
Probably not that much because the pitch of the chain hasn't changed enough for the rollers to ride up over the teeth and fail to engage. Remember, too, that wear generally is dependent on lubrication. And there is always a degree of "slop" in chains to allow them to "bend" sideways when they move away from perfect chainline.

Roller wear is not an issue that I have read raised before like this.

Probably not that much because the pitch of the chain hasn't changed enough for the rollers to ride up over the teeth and fail to engage. Remember, too, that wear generally is dependent on lubrication. And there is always a degree of "slop" in chains to allow them to "bend" sideways when they move away from perfect chainline.

Roller wear is not an issue that I have read raised before like this.
So I guess that means the Park CC2 tool is misleading since it doesn't actually measure the pin-to-pin chain pitch. I'm inclined to agree with you... other than the anecdotal evidence of my cassette which was totally worn out by a chain which appeared to be <<1/16" past the 12" mark.

Chain and sprocket wear works in mysterious ways, I guess. My brain hurts too now :p

So, which way of measuring is correct in terms of prolonging the life of the cassette? That's the $1,000,000 question :) I think most of us wouldn't care too much about what's going on with the chain if we knew how to tell when it was damaging the cassette.

That's what it all comes down to, absolutely.

I'm starting to get the feeling that some of you don't trust the merit of measuring the distance between several rollers [6,12...]. The point of measuring multiple rollers is because the slop is magnified and is therefore much more precievable then just measuring the distance between two rollers.

A go no-go chain wear tool does this better then any other method I can think of.

I'm checking chains for wear every day at my shop, and have never been let down [yet] by the 0.75% side of my park tool.

If there are still fans of the steel ruller, please help me get it thru my thick head why thats better?

Here's my take on it:

Wear on the outside of the chain rollers isn't significant. If you watch a chain roller engaging a cog tooth, you'll see that the chain roller doesn't really slide much on the tooth, the force is basically perpendicular to the contact surface. The place where there is a LOT of sliding/rubbing friction is between the chain pin and the side plates or bushings, and that's where most of the chain wear occurs.

While it's true that the roller-to-roller distance is what actually defines the pitch of the chain, the wear on the rollers progresses much more slowly than the wear on the pins. If you find a badly worn chain (e.g. one that's stretched by 2%), you'll see that the pins are visibly deformed, while the outside surface of the rollers is basically unscathed. Look at this photo from Sheldon Brown's site:
http://sheldonbrown.com/images/chain_wornpin.gif

There's no difference between measuring roller-to-roller distance or pin-to-pin distance: wear on the pins will cause the chain to elongate LONG before wear on the rollers becomes an issue.

And let me throw some more gasoline on the fire.

I remember a while ago, someone posted that you can extend the life of your chain by periodically turning it around. The concensus was the guy was crazy.

But if you look at the photo posted by Mox, the wear is on one side of the pin only, the drive side. So doesn't it make sense that if you flip the chain over, you will start with a fresh pin for each link and the chain will wear longer?

I'm starting to get the feeling that some of you don't trust the merit of measuring the distance between several rollers [6,12...]. The point of measuring multiple rollers is because the slop is magnified and is therefore much more precievable then just measuring the distance between two rollers.

A go no-go chain wear tool does this better then any other method I can think of.

I'm checking chains for wear every day at my shop, and have never been let down [yet] by the 0.75% side of my park tool.

If there are still fans of the steel ruller, please help me get it thru my thick head why thats better?
To my mind, at the moment, it is the pitch of the chain as defined by the rivets that determines how well the chain meshes with the teeth. That pitch starts at 1", and as the shoulders wear, that 1" increases so eventually, the rollers cannot engage without riding up.

As to roller slop, have you checked a new chain? I just did, and there is vertical freeplay in the rollers. The reasons, I believe, are simply to enable the chain to "bend" sideways on multi-gear systems, and to overcome allowances in the machining or stamping of cog and chainring teeth.

I also think that if there was no freeplay in the rollers, you would have an unbendable chain, in any direction. Take the case of when a rivet is inserted and binds the plates together; usually, the roller is caught in there as well. The link cannot bend because of the friction between the outer and inner sideplates and the roller.

I'm starting to get the feeling that some of you don't trust the merit of measuring the distance between several rollers [6,12...]. The point of measuring multiple rollers is because the slop is magnified and is therefore much more precievable then just measuring the distance between two rollers.

A go no-go chain wear tool does this better then any other method I can think of.

I'm checking chains for wear every day at my shop, and have never been let down [yet] by the 0.75% side of my park tool.

If there are still fans of the steel ruller, please help me get it thru my thick head why thats better?

i've got to agree -- K.I.S.S. i'll continue to use a drop-in chain checker.

ed rader

And let me throw some more gasoline on the fire.

I remember a while ago, someone posted that you can extend the life of your chain by periodically turning it around. The concensus was the guy was crazy.

But if you look at the photo posted by Mox, the wear is on one side of the pin only, the drive side. So doesn't it make sense that if you flip the chain over, you will start with a fresh pin for each link and the chain will wear longer?
Hahaha, no.

The reason why the pin is worn on one side only is because that's the side that the adjacent inner plates contact under pedaling pressure. If you flip the chain over, the inner plates ("bushing" area) will still be contacting the same area. Maybe a bit off it, mostly overlapping (as the chain wraps around the cog/chainring, it bends. If you reverse the bend direction, the angle of the two adjacent links changes)
On second thought, this angle difference might just be big enough to be somewhat relevant.

I agree that the insides of the rollers may wear, and that this wear will not show up in measuring chain "stretch". However, I don't think that wear on the inside of rollers will wear out the cogs, because so long as the rollers are still the proper distance apart (let's say they're all shifted .25mm to one side, to take San Rensho's measurement). There is now more play in the rollers, but so long as they all shift to the same side in unison (which they will under pressure from chainring or sprocket teeth) then the pitch of the chain where it contacts the teeth (which is the thing that matters for wear) is not changed by worn rollers.

In the picture from Sheldon's site, the wear against the pin is caused by the inner chain plate (which rotates against the pin), not by the roller (which contacts the inner sideplate, not the pin directly). The inner sideplate is curved inward where it contacts the pin, in order to function as a semi-bushing. So that's why the wear in the picture is toward the middle of the pin.

I'm not sure how much the inside of the rollers wears. The initial impetus in this thread to think that it's wearing a lot is the improper interpretation of the picture, thinking that the wear on the pin was caused by the roller and not by the "bushing" part of the inner chain plate. I'm not sure that the rollers do experience significant wear. But if they do, moxfyre is correct that this adds an error term into the typical chain checker's measurement of chain stretch, an error term that doesn't happen when using a ruler.

As to San Rensho's bringing up the method of running the chain backwards, wouldn't change anything about chain where, and here's why: the chain is always under tension. So if you reverse the direction of the chain, the links are still pulling on each other, and so the same surface of the pin is wearing. You'd either need to change bicycle design so that the chain is under compression (which is obviously ludicrous) or go through with a chain tool and an extreme amount of patience and rotate each pin 180 degrees with respect to the outer chain plate in which it is embedded.

I'm not sure how much the inside of the rollers wears. The initial impetus in this thread to think that it's wearing a lot is the improper interpretation of the picture, thinking that the wear on the pin was caused by the roller and not by the "bushing" part of the inner chain plate. I'm not sure that the rollers do experience significant wear. But if they do, moxfyre is correct that this adds an error term into the typical chain checker's measurement of chain stretch, an error term that doesn't happen when using a ruler.


Actually, I believe that was me. But thanks to Rowan, I have re-thought that, because he's right that the rollers and pins are hardened steel, whereas the plates are not. So what I think is actually happening, is that the inside of each roller is wearing away the semi-pseudo-bushing that it turns against (and, there is much more movement of the roller against that side of the pseudo-bushing, than there would be movement of the pseudo-bushing against the pin). The roller's wear against the pseudo-bushing has the same effect of producing slop in the rollers and increasing the gap between rollers, but NOT lengthening the chain. And I agree with all of you who have pointed out that as long as the pins remain the proper distance apart, and the rollers & bushings have worn uniformly, the chain should not accelerate wear, because the pitch is still correct. Therefore, since the ruler is the only measuring tool that *only* measures the pitch of the pins, I think that's going to be my instrument of choice in the future.

I'm not sure how much the inside of the rollers wears. The initial impetus in this thread to think that it's wearing a lot is the improper interpretation of the picture, thinking that the wear on the pin was caused by the roller and not by the "bushing" part of the inner chain plate. I'm not sure that the rollers do experience significant wear. But if they do, moxfyre is correct that this adds an error term into the typical chain checker's measurement of chain stretch, an error term that doesn't happen when using a ruler.
Good summary!

One quibble: I don't think it's wrong to interpret the picture the way Re-Cycle did. If the pseudo-bushings (good word :)) were what was causing the most wear on the pin, I would expect to see it worn down in two sections extending about 1/3 of the way in from either end. The picture, however, shows that something is wearing down the middle section of the pin. Nothing contacts that part of the pin except the roller.

I think you're totally right that it is the bushing-vs.-pin wear which actually causes "chain stretch". But the wear on the rollers appears to be surprisingly large. I think we need more data points. I have a few old chains in my basement... I'll crack a couple this evening and take some photos.

The picture...shows that something is wearing down the middle section of the pin. Nothing contacts that part of the pin except the roller.
...
I think we need more data points. I have a few old chains in my basement... I'll crack a couple this evening and take some photos.

I'll look forward to hearing about what you find! On my SRAM chains, the pseudo-bushings practically meet in the center, so there's effectively no way for the roller to ever contact the pin. I think most chains are made that way. Maybe Sheldon's chain wasn't, but every other one I've seen has been.

One quibble: I don't think it's wrong to interpret the picture the way Re-Cycle did. If the pseudo-bushings (good word :)) were what was causing the most wear on the pin, I would expect to see it worn down in two sections extending about 1/3 of the way in from either end. The picture, however, shows that something is wearing down the middle section of the pin. Nothing contacts that part of the pin except the roller.

On my SRAM chains, the pseudo-bushings practically meet in the center, so there's effectively no way for the roller to ever contact the pin. I think most chains are made that way. Maybe Sheldon's chain wasn't, but every other one I've seen has been.
I thought that the wear on Sheldon's pictured chain was from the pseudo-bushings meeting in the center, not from the roller.
Even if they didn't meet in the center, the roller sits against the pseudo-bushings, and has no way to contact the center of the pin. So I conclude that the wear on the center of the pin in Sheldon's pictured chain is from the pseudo-bushings, which practically meet in the center and would thus leave a clean and continuous section of wear.
I actually doubt the rollers wear very much, because there's not necessarily a lot of friction that they experience. The pseudo-bushing vs. pin contact area has lots of friction because that's where the chain bends. But there's little-to-no reason (as far as applied/experienced forces are concerned) for the roller to have to rotate against the pseudo-bushing.
So I'd expect moxfyre's dissected chain to show that the roller is in pretty good shape on old chains as well as new.

Look again at the picture and that pin that is pushed out--the deepest wear is *not* at the center of the pin. If you consider the entire length of the pin, including what part of it would be inside the outer plate, and the end that still *is* inside the outer plate, that deep area of wear is not at the center, but beside the center point. Right where the pseudo-bushing would contact it. IMHO. :)

Look again at the picture and that pin that is pushed out--the deepest wear is *not* at the center of the pin. If you consider the entire length of the pin, including what part of it would be inside the outer plate, and the end that still *is* inside the outer plate, that deep area of wear is not at the center, but beside the center point. Right where the pseudo-bushing would contact it. IMHO. :)
Yeah, okay. Dang it, I'm wrong again :D

Roller clearance or wear is independent of what you want to measure, which is chain pitch increase as a result of wear. The park guages measure the combinatation of pitch increase together with roller clearanance. Not a valid measure of chain wear. Use a ruler.