Chain stretch dilema.

[Old Codger]

I have three flat bar chain checkers which I use and change my chains at 0.5%. I recently purchased one of those with the small pins and sliding gauge and printed on it it said that new chains can be between 0.25 and 0.5%.
Thinking this was strange I checked three new Shimano chains, one 12 speed and two 11 speed with all four checkers and sure enough all three chains were at 0.25%.
Does this mean that I should be going up to 0.75% before changing chains? What do people think?

 

[Pentti]

New chains are 0.00% - If your "checker" tells you something else it is not right for whatever reason.
Measuring roller centres is the only failsafe/accurate way to do it IMO.
I now use one of these stupidly priced, but deadly accurate devices....

LL Chain Wear Tool

Chain wear has become a little more complicated ever since 1x drivetrains have gained popularity. These drivetrains are designed to cross chain, something most of us were taught to avoid when we started riding bikes as a serious endeavor. Add to that the variables that flat top chains have...

www.abbeybiketools.com

 

[steve_sordy]

I always measure a new chain over 100 links, ie 50 inches. Then if it is different, (it never is) I can compare later.

Years ago, I used to measure with an internal vernier caliper gauge. The distance I measured was 5.7". A new chain was always 5.700". But when I started using 11 and 12 speed chains, the start measurement started to vary. (I suspect that the rollers were a different diameter to previous). Initially, if I recorded say 5.706" then that was the measurement I used to compare the same chain as it wore. But then I noticed that despite very careful measurement, often repeating the same measurent several times to ensure that I was as accurate as possible, the chain would vary along its length. What! Yes, using a vernier caliper, the chain would vary along its length, implying differential wear. It didn't matter how many times I measured the same section of chain I always got the same result, so at least I was consistent. Move the crank half a turn and measure again, different! This should be impossible as it cannot be manufacturing variation within the length of a single chain. I know how chains are made and it is unlikely to the Nth degree. Maybe the rollers are barrel-shaped, or maybe they do wear differentially?

So I stopped using my vernier caliper (more accurate than a go/no-go chain gauge.) Instead I started removing the chain and measuring across 100 links. This method averages out any section-to-section wear and reduces the error impact of any measuring-device-to-chain-contact variation by nearly nine fold.

What this says to me is that the short distance over which chain gauges operate cause them to be inheritantly inaccurate and can cause a chain to be retired long before it is necessary.

So if you have a chain that is due to be replaced according to your chain gauage, check the chain over 100 links and then make your decision. Measure with a metal tape measure, with the chain laid flat over a flat surface and pulled tight.
50" = new chain
50-1/8" = 0.25% = OK
50-1/4" = 0.5% = OK (As far as I'm concerned, but I'd buy a replacement ready for action. As long as the chain is running OK).
50-3/8" = 0.75% = replacement asap.

But you don't have to measure to the nearest one eighth of an inch. Most metal tape meaures have increments of 1/16". That is to 0.125%. That is more than accurate enough, but if you are careful, it is quite easy to meaure to 1/32" i(0.06%)! The choice is yours.

 

[irie]

I always measure a new chain over 100 links, ie 50 inches. Then if it is different, (it never is) I can compare later.

Years ago, I used to measure with an internal vernier caliper gauge. The distance I measured was 5.7". A new chain was always 5.700". But when I started using 11 and 12 speed chains, the start measurement started to vary. (I suspect that the rollers were a different diameter to previous). Initially, if I recorded say 5.706" then that was the measurement I used to compare the same chain as it wore. But then I noticed that despite very careful measurement, often repeating the same measurent several times to ensure that I was as accurate as possible, the chain would vary along its length. What! Yes, using a vernier caliper, the chain would vary along its length, implying differential wear. It didn't matter how many times I measured the same section of chain I always got the same result, so at least I was consistent. Move the crank half a turn and measure again, different! This should be impossible as it cannot be manufacturing variation within the length of a single chain. I know how chains are made and it is unlikely to the Nth degree. Maybe the rollers are barrel-shaped, or maybe they do wear differentially?

So I stopped using my vernier caliper (more accurate than a go/no-go chain gauge.) Instead I started removing the chain and measuring across 100 links. This method averages out any section-to-section wear and reduces the error impact of any measuring-device-to-chain-contact variation by nearly nine fold.

What this says to me is that the short distance over which chain gauges operate cause them to be inheritantly inaccurate and can cause a chain to be retired long before it is necessary.

So if you have a chain that is due to be replaced according to your chain gauage, check the chain over 100 links and then make your decision. Measure with a metal tape measure, with the chain laid flat over a flat surface and pulled tight.
50" = new chain
50-1/8" = 0.25% = OK
50-1/4" = 0.5% = OK (As far as I'm concerned, but I'd buy a replacement ready for action. As long as the chain is running OK).
50-3/8" = 0.75% = replacement asap.

But you don't have to measure to the nearest one eighth of an inch. Most metal tape meaures have increments of 1/16". That is to 0.125%. That is more than accurate enough, but if you are careful, it is quite easy to meaure to 1/32" i(0.06%)! The choice is yours.

I check a new chain (usually Shimano CN-9100) with a Park Tool CC-2 in several places then change the chain when it stretches more than 0.5% from that baseline (normally measures 0.0 when new). Simples.

 

[bmwpowere36m3]

It'swe not uncommon for new chains to register more than 0.0% wear, up to 0.25%... depends on the checker. No, you should replace at 0.5%, regardless of starting point.

Some checkers only measure a short distance (4-6 inches), so ERROR is greatly magnified.

 

[TheKaiser]

I have three flat bar chain checkers which I use and change my chains at 0.5%. I recently purchased one of those with the small pins and sliding gauge and printed on it it said that new chains can be between 0.25 and 0.5%.
Thinking this was strange I checked three new Shimano chains, one 12 speed and two 11 speed with all four checkers and sure enough all three chains were at 0.25%.
Does this mean that I should be going up to 0.75% before changing chains? What do people think?

I'm not sure if you have ever checked out all of the chain wear testing, and chain lube testing (vis a vie wear prevention) on the ZeroFrictionCycling site, but if you haven't done so, it is well worth a read and really allows one to geek out on the tech nuances of this stuff.

Since wear testing is such a big part of his chain and lube testing, he has also done a deep dive into how best to actually measure wear, and has written extensively on the topic. Interestingly, he has found that many wear gauges are inaccurate out of the box, or for ones that require the user to apply some degree of pressure "by feel" to the tool to gain the reading, introduce a substantial amount of inaccuracy based on how hard the user is pressing. I had always thought that @steve_sordy 's method of measuring the whole chain made the most sense, as the stack up of wear along the whole length of chain would make wear easier to determine, as the total number of millimeters of wear will be greater on a full chain than the 100-150mm that most checkers measure. Interestingly, according to the ZeroFrictionCycling guy, that is not the best method, because a given chain will not wear totally evenly along it's whole length, and the most worn sections will be the sections that cause wear to the teeth on your cogs. In other words, you may throw the gauge on the chain and get a reading of 0.4%, leading one to think you can run it a bit longer, but another section of that same chain may be past the limit at 0.6%. He suggests using an accurate tool, and measuring 3-5 separate sections of the chain, using the worst reading to determine if the chain should be replaced.

Here is a link to a whitepaper he published testing the accuracy of different chain gauges against a known 0.5% worn chain: https://zerofrictioncycling.com.au/wp-content/uploads/2024/03/Chain-Wear-Checkers-Table.pdf

He has also written extensively about the effect of Sram's oversize rollers on some of their chains messing with gauge accuracy, and also about the differential rate of roller wear (which doesn't actually lead to "stretch"), and pin/plate wear (which does lead to chain "stretch"). Many tools include roller wear in their measurement, so if you have a chain model with very hard wearing pins and plates, but relatively soft fast wearing rollers, then you will get a measurement on your gauge that indicates it needs to be replaced, even though the actual pin-to-pin distance is still well within the acceptable range. He also emphasizes the need to measure a chain that is relatively clean, but not freshly wax lubed, because both grit and a fresh coat of wax will take up some of the extra space in the pin/link interface, leading to the appearance of a chain that measures out as less worn than it truly is.

Anyway, hopefully that is of interest and not TMI.

 

[Pyr0]

Why on earth is the SRAM flat top chain checker so expensive?
£30 for a little sliver of metal seems a bit much.
Anybody seen them cheaper?

 

[irie]

As said in post #4

I check a new chain (usually Shimano CN-9100) with a Park Tool CC-2 in several places then change the chain when it stretches more than 0.5% from that baseline (normally measures 0.0 when new). Simples.

 

[steve_sordy]

@TheKaiser At last it was not a figment of my imagination! Somebody else agrees with me that chain can wear diferentially olong its length!

"....Interestingly, according to the ZeroFrictionCycling guy, that is not the best method, because a given chain will not wear totally evenly along it's whole length, and the most worn sections will be the sections that cause wear to the teeth on your cogs. In other words, you may throw the gauge on the chain and get a reading of 0.4%, leading one to think you can run it a bit longer, but another section of that same chain may be past the limit at 0.6%. He suggests using an accurate tool, and measuring 3-5 separate sections of the chain, using the worst reading to determine if the chain should be replaced....."​

When I first noticed it, I started to measure every half turn of the crank until I'd done the whole chain and then averaged them out. At first I could not believe it and I assumed that it was my measurement technique at fault. But I got the same result each time I measured the same section of chain. I have no explanation for this variation. It is why I prefer to remove the chain and measure over 100 links. If I am going to average the chain, why not do it properly and much more quickly?

Whilst I noticed it, I have no idea why it happens. Chains are assembled from randomly mixed parts on a machine. Unless you have seen how chains are made you cannot imagine just how randomly they are mixed. (I did 6 weeks in a chain factory as part of my engineering training). The machine could vary of course, but it would have to be drifting in and out of the standard settings on a very rapid basis. When the chains wear differentially, it can be produce different measurements within half a crank, lets say 17 teeth, or 8.5". Not feasible.

Just had a thought! Maybe chainwear is a bit like braking bumps. One person creates a slight skid, the next person digs out a little bit more. Soon every body's wheels are digging in the same hole, and the hole gets deeper rapidly. If you stamp on the pedals, or do something that causes a little bit of wear on a tooth or on rthe pins of a few links, then maybe they attract further wear later and it accelerates! Just a thought, not a well thought through theory. Can anyone else come up with a proposal, or do they actually KNOW?

 

[Montana St Alum]

I replace a chain if it starts shifting badly. I've never made it to .5%.

 

[irie]

@TheKaiser At last it was not a figment of my imagination! Somebody else agrees with me that chain can wear diferentially olong its length!

No component comprised of over 100 supposedly identical components will perform with perfect homogeneity.