Saturday, May 28, 2011

15 Bicycle Chain Stretch Test & Results

Do bicycle chains get stretch marks? Will smearing cocoa butter on them be a step in preventive maintenance for future? I don't know, but hold that thought for a moment.

I, like many, am a fan of chains. For bikes, they present a technology that is  ubiquitous, economical, and proven to work almost seamlessly with external shifting systems. Belts are slowly staking their claim in the single speed road and mountain bike arena, however I have to be honest - show me a more simpler, self cleaning power transmission mechanism that doesn't load up shafts and bearings as much as a belt does, and I'll be sold on other ideas. 

".. the less stretch, the more responsive the bike becomes.."
But in cycling, as we all know, equilibrium is rare. Everything has to get scrutinized more thoroughly than a coroner would do a murder victim, from the pimple on our skin that's disturbing laminar air flow to the secret ingredients used to make those mundane Presta valves and you know, that's what keeps our world a bit interesting (or not).

Now if you just may recall, Wipperman was getting fancy in the recent past by testing a host of chains in order to rate them according to their wear rates. You can read that blog post here which described the test protocol, the results they came out with and so on.

Recently, I was told that the company commissioned a different test on a similar selection of chains to test for elongation under load. Tom Petrie of Cantitoe Road - a chain test data center - passed along some literature that said the following :

"Wippermann recently tested a number of popular 10-speed chains for stretch under load. For a reference point, each chain’s length was measured under a nominal load of 10 kg. Then each chain was measured under 75kg and 150 kgloads, and the results recorded. Not surprisingly, chains with cut-out plates and hollow pins stretched more than those with solid plates and pins. And, the chains that stretched least were the Wippermann’s Connex 10 series with solid plates and pins!

How much a chain stretches under load affects how quickly the load is transferred to the driven cog. The less stretch, the more responsive the bike becomes. And, less stretch means less energy is lost to stretching the chain! Especially in sprint, time trial, or hill climb events, reducing these losses is critical. 

Wippermann tested 31-link sections of chain. This is the average number of links under load between chain ring and cog. While the actual amount of stretch is small (from 1.10 to 2.15 mm) the differences are substantial. Among the various chains tested, the “stretchiest” stretched almost 100% more than Wippermann Connex!"

After testing, the data was cobbled up into a table to make sense of the results. They follow :

Summary of chain stretch test data

Elongation vs load plot

The document went on to make light of these  :

"In addition to raw material and proprietary heat-treating processes, the shape of Wippermann Connex outer plate is largely responsible for its resistance to stretch. Note that chains featuring elaborate side-plate cut-outs and hollow pins are the “stretchiest” while chains with solid plates and solid pins stretch less. But even solid-plate solid-pin chainswith sculpted “figure 8” outer plates stretch more than Wippermann Connex. The extra-strong rectangular outer plates on Connex chains contribute significantly to their resistance to stretch."

This stretching they're talking about should be nowhere big enough to cause a yielding in the chain material. So the material attains its original shape after unloading like a spring, and the real question then becomes - how will a 100 thou inch change in chain length in the worst case scenario affects overall power transmission efficiency? Is it any more significant than the normal vibrations introduced into the chain due to tensile load changes and sprocket tooth effect? Does the stretching get better or worsen in weaker chains when the chain is misaligned/cross chained? Finally in the big scheme of things, how will cyclic stretching/unstretching react with notorious elements like salt water? Could it possibly accelerate the failure of cut-out chains in those circumstances?

What do you think? While you sip your coffee, you may also be interested in glancing at a "shifting performance" study done on chains through Wippermann, the hardly surprising conclusion of which was that there is no observable correlation between a worn chain and shifting performance compared to a new one. You just may not want to break your bank over a chain.

*  *  *


  1. Dave Kina1:48 AM

    I ride a Wippermann 10 sx. I have been for about 4 seasons. one of them lasted about 7,000 miles. great chains and an easy connex link for removal to thoroughly clean the chain futher preventing unneccessary friction from dirt etc. BE SURE TO PUT THE CONNEX LINK ON THE CORRECT SIDE OR ELSE IT WILL SKIP IN THE SMALLER COGS. I haven't broken one of these chains.

  2. Steve Lubanski1:49 AM

    In this era of stupid, stretch does not concern me nearly as much as the lack of pin overlap to the plates. Esp. with shifting systems then have the ability to change two or three gears at a time.

  3. Huh? I’m not sure I understand what you’re trying to say when you say “lack of pin overlap to the plates”. With respect to “stupid”, I sort of agree with you. The amount of stretch is small. I don’t know if you can really feel it, but what fun would bicycle parts be if you weren’t comparing theoretical advantages of, for example, the aerodynamic efficiency of the tire bead / rim interface of one tire versus another, or the like. The point is that some chains stretch almost twice as much as Wippermann chains do. Wippermann chains are stronger, wear longer, shift as well, and stretch less than the competition. They also feature the advantage of the Connex quick link. And yes, full disclosure, I import and sell them." - Tom

  4. Steve Lubanski1:50 AM

    What I meant as pin overlap, is how much of the pin protrudes beyond the plate. Many bridges de designed with a flush style system like used on 10 and 11 speed chains failed or were retrofited because of the bad design.

  5. Steve Lubanski1:51 AM

    what I love is Campy 10 and 11 spd chains. When putting them together, there is a warning, Do not remove links from this side. After the chain is together, Where is this side!!!!!!!!!!!!!!

  6. Wheel less in Sestriere after leaving the front wheel outside my mate's appartment in Torino !

    Last Saturday the Dureace chain lost a link after a season of use . New magic pin lassted a few hours and then the link opened again . Sunday on Passo Gaiai the chain received more heavy abuse and repeated three times except the second time it carried away the front mech . Turned the chain around and put in another pin and ridden the etappe ( google Parrabuddy for the story ) each day since .

    Seems that the strain was on the front but works when you fix it at the rear .

    Note that i carry lots of pins but that the chain has survived around 20k of heavy use and looks like now that i have a new derailleur i will let "sleeping dogs" lie a while longer .

  7. Are the new chains are "stretching", or is it just play between the plates and the pins. I was under the impression "stretch" in a worn chain was due to wear on the plate/pin interface.

  8. Anonymous1:07 PM

    I wonder how much of this measured stretch is due to lube being squeezed out from between the load bearing surfaces in the chain? Assuming these were all factory new chains they would have all had the (generally rather thick) proprietary factory lube on them.

  9. Roof mounted bicycle racks. If positioned securely, roof mounted bike racks will carry a minimum 4 bicycles and can even accommodate more reliant on your car type.

  10. It would be interesting to plot elongation versus chain mass. The Weight-Weenie-favorite KCNC-SL's aren't doing so well here.

  11. I have a question for the physics-minded folks on here, regarding this issue: to my simple mind, it seems that this chain stretch, if it is truly elastic and is happening at the side plates, would not actually effect TOTAL power transmission. What I mean is, pretend the chain were more like a very strong rubber band. If you pulled on it, at first it would not move the rear cog, but once you put enough force on it and its resistance to stretching got high enough, it would pull the cog forward at the same rate you were pulling on the rubber band. then, once you start to lessen the force of your pull, the rubber band would still pull for a little while longer until it returned elastically to its original length. So wouldn't the total force transmitted to the cog be the same? In fact, could this stretching actually be beneficial, since the elasticity would smooth out peaks in power and spread the force over a greater time period, thus reducing slippage in poor conditions?

    i'm probably wrong, but it's just something that's always bugged me about complaints of frame or chain flex.

  12. Spend your dollars based on primarily how long will you be using it or you require it to stay with you; good quality bicycle car racks can last up to several years, however if you need it only for a very rare trips, just buy the lower priced variety.

  13. I find it very interesting and useful especially to those who love to ride on their bicycle. It is very important to know how strong your bike chain. It's for the assurance that you are safe during travel.

  14. Anonymous5:10 AM

    very interesting post! I got my bike chain from

  15. If the chain does indeed deform elastically (stretch) as Wipperman attests, then it could have significant affect on chain wear. The mechanism for accelerated wear at maximum stretch would be pitch mismatch. Lets assume the chain is slightly worn and just about approaching the allowable 0.5% elongation before the chain and sprocket pitch are our of synch, and it then stretches, the 0.5% limit would be exceeded at the incoming cog (tooth) would then strike the incoming roller and produce wear.
    Further, stretched links at the beginning side of the sprocket rotation would concentrate load on the first two or three links and cogs, further accelerating wear.

    I'm surprised that Wipperman focused on nonsense like "responsiveness" and other intangibles rather than accelerated wear on lightweight chains towards the end of the chain's life.

    In summary, I think chain stretch would produce an exponential chain wear curve.

    Any thoughts on this?


Thank you. I read every single comment.