One of the best ways to quickly ruin expensive drivetrain parts is to keep pedaling with a worn chain. Most manufacturers agree that once a chain reaches 1% elongation, it’s considered worn out. Continuing to use the chain past that point will typically destroy cassettes and chainrings and lead to skipping if you install a new chain. So the longer it takes for a chain to reach 1% elongation, the better, right?

Wippermann 11 speed wear test finds Connex 11SX chain lasts longest

That’s what Wippermann is hoping to highlight with their latest chain wear test. Obviously, any test run by a manufacturer that claims their own product is best has to be viewed with a bit of skepticism, but Wippermann’s test does seem pretty thorough. The test protocol involves a run in period, then various degrees of offset chain angle, oil, water, and sand additions for wear accelerators, and a 600 N load.

Wippermann 11 speed wear test finds Connex 11SX chain lasts longest Wippermann 11 speed wear test finds Connex 11SX chain lasts longest

After running the test and then running each chain to 1% stretch, the results are as shown above. Not surprisingly, the Wippermann Connex 11SX chain tested the best, but the Campagnolo Record, KMC X11SL Gold, and Shimano HG 901 were up there too. Obviously, the test says nothing about quality of shifting – only outright durability. If anything though, the test does show that higher end chains often do offer better durability.

The winning chain, the Connex 11SX is a stainless steel chain with a nickel coating and Connex Link that is compatible with all 11 speed road and mountain bike drivetrains and is available from outfits like Cantitoe Road for $89.96.



  1. Campy gets a lot of flack, but in my opinion, their metallurgy and general level of quality is very high. For me, their stuff has always lasted or maintain a high level of function longer than comparable OEM components.

        • Not in my experience. Only the Record chains last longer due to the stainless rollers. The cheaper options sold under Chorus and Veloce don’t last nearly as long. Campy owners tend to be more particular wrt keeping their bike jewelry clean

          • I’ve never run the cheaper Campy stuff so you may be correct. I do try to keep my Campy bike cleaner, but I’m not obsessive about it, I’m happy to keep riding Campy chains. I do ride other chains on other bikes, and they seem to not last nearly as long.

      • No metallurgy? So they do not specify the type of steel and its heat treat?
        Do you go to a steel supplier and say “I want steel sheets”?

        No, because if you do, you are going to be surprised at what you get. They know what grade of steel works for their design and specify that like any other end user of steel products. It’s likely Campy has a metallurgist on staff for this, especially since they are a valued components maker of non-bike items (they have a lot of machining knowledge in Italy).

        What you are basically stating is like saying there is no metallurgy going on in the building of bridges, skyscrapers, pressure vessel, etc just because those people aren’t actually making and treating the steel themselves. This is simply untrue.

      • This is total crap. I don’t care to defend Campagnolo for countless reasons, but the idea that they don’t know or care anything about metallurgy is just absolutely ignorant and makes no sense whatsoever.

    • you may need to do some work but that would be interesting because those chains are expensive! Just enter that chart above into excel, add a cost column, and add a hours/dollar column using the maths. You could figure out the most economical chain.

    • Keep in mind the the tests were at 600N @ 90 RPM. That’s ~960 – 990 watts for most crank arm lengths. That’s a lot of “juice” for 100 +/- hours.

      • This. Plus “then various degrees of offset chain angle, oil, water, and sand additions for wear accelerators…”

        And that is kind of what makes the test poor. The high load itself it very abnormal for chain use and that will accelerate wear in a manner that I am not sure is representative of actual chain use/wear/life.

        • That’s the point of accelerated wear tests. The results are transferable to real life ie if you ran the chains on a bike at quarter of the power you would still find the chains would finish in the same positions just with proportionally higher hours on each chain.

          • Tom, you’ve got a lot of assumptions baked into that statement. Do all of the various parts of these chains have linear deterioration rates relative to the enhanced load? Without data, you don’t actually know… Besides, accelerated wear tests aren’t even necessary in this space. Best case a road chain lasts a few thousand miles and you can easily get that on a machine cranking on it at simulated speeds/loads in under a week.

              • snip snip snip “Do all of the various parts of these chains have linear deterioration rates relative to the enhanced load? Without data, you don’t actually know…”
                As with all scientific and engineering testing an assumption has to be made. In the case of the above discussed testing protocol it is reasonable to assume that each component sees the same load, speed and environmental conditions since the test protocol is exactly the same for each chain. Stripping each apart would deliver nothing special since you would then need to do this to each chain before it was tested, measure individual components, record their dimensions and position in the reassembled chain and repeat after the test.

                This gives you a thousand extra activities for zero gain in test objective which is to identify which system (complete chain) lasts longer under constant load in an accelerated wear test.

                So for intents and purposes there isn’t anything wrong with the test methodology

                • There is nothing wrong with the test, but it mainly tells you how the chains respond under the loads provided (which are high for a bike chain).

                  Wear rates are not always linear with respect to load. I wouldn’t expect a huge shuffle of order on the results, however, the percentage difference in performance may change significantly at lower (more realistic) load scenarios.

                  Furthermore, the angularity they used is very high. Which is fine as its stated, but it may highlight wear modes on certain chains that never actually occur in real life (rough calc shows its nearly 100% greater than what most bike run, even crossed up on a double)

                  • snip snip snip “Wear rates are not always linear with respect to load.”
                    This will vary in terms of material condition and tolerance but essentially steel’s coefficient of friction varies very little across alloys unless we are talking duplex stainless steels which I would think are complete overkill for bicycle chains.
                    Based on this a linear wear rate can be assumed but typical martensite stainless steels are going to be tougher and more wear resistant than plain carbon steels and austenite stainless steels.
                    Material selection and condition are part of the design and determine quality of the final product that is being compared to other similar products.
                    An Accelerated wear test will always favour the better designed and manufactured for longevity product an that is what Wipperman is trying to highlight.

                    SO even if a chain fitted to a drivetrain may never see the drive offsets used in the test or the loads this is irrelevant because the test is trying to discriminate between products quality wrt longevity and all test specimens see the same conditions.

                    So even if the worst chain lasts 10,000miles the best one will still outperform by at least the same factor. In fact the light the real life loads the bigger the performance gap will be

                    • You speak like a metallurgist so I may defer to your expertise. Not sure I’ll fully agree with that last paragraph though 😉

                    • First, I must say that this is the most nested response I’ve ever seen on BR and apparently we’ve reach the limit. That it itself is impressive.

                      Secondly, it’s entirely false to assume that the accelerated wear conditions translate to normal use. One can introduce accelerated degradation mechanisms that don’t exist in the real world (or don’t exist to the extent of the test) so as to make the accelerated conditions non-representative of the real world.

                      For one example: suppose you were to put these chains on a track bike. Okay, you wouldn’t put an 11 speed chain on track bike but stay with me for a moment: you’d never have any angle to the chain, and it wouldn’t exposed to water or sand in any significant way. So you’d never see two of the accelerated factors that were put into the test.

                      Backtracking a bit, just ’cause the chain in the real world is exposed to a little water/sand and some angles doesn’t mean the accelerated test can be applied linearly to a real world scenario.

                    • I’m basically on board with Tim on this testing here. Like the limits we’ve reached with the Comment Board, good testing usually involves pushing a component towards the extreme to derive some meaningful results, provided the underlying assumptions are logical.

                      It may not be perfect, but in life-testing, it makes little sense to rely solely upon a 5 or 25-year real time test before putting a product out. Chain wear is pretty straight forward compared to complex systems or those with various interactions (like food).

                      The alternative would be to conduct real life chain tests over 2-10k miles maintaining the EXACT same conditions on the relatively few samples, or provide thousands of testers.

                      Even then, I believe JBikes and Feldy would much more easily find the variation in “riding styles,’ shifting, time of day, weather changes, maintenance, etc. etc. would confound the results much more so than this test here.

    • They should have included the KMC chains with their DLC coating in test as well, since they’re comparably priced. I’ve never paid anywhere near $90 for chain. When I can get three KMC X10SL / X11SL silver or gold chains for just a little more than that why would I spend the extra for the Wipperman?

        • note that the SRAM XX1 wich is identical to the RED does not last any longer than the 14$ chain (only 2%).

          89$ ? ive never payed more than 36€ for a Wipperman Connex SX or a Record chain. Approx same as DureAce/XTR but it last twice as long, approx same price as a RED/XX1 but it last 3-4 times as long.

    • I don’t see any issue here. These companies are in the best position to know the wear profiles and proper maintenance schedules for their products. They wouldn’t try to mislead consumers for their own financial benefit.

      ~ Guy who changes oil in his 2018 car every 3,000 miles.

    • less conflict of intrest, more chains worn past 1% on a modern drive trains cut into the cassette and chainrings.

      The diistances between the links and spacing of the chainrings are prescise. 1% doesn’t sound like much but that 1% of the total length of the chain mashes right into the chainrings and cassette causing premature wear.

  2. How did they measure? Had all the chains the same lenght at the same number of links? It is well known, that different brands have different real lenght of the chains at the same amount of links. Thus the chains that are shorter when new, seem to last longer when measured by a chain caliper. Which does not have to be true.

  3. Based on personal chain use I think that chart is pretty accurate. I still opt for the kmc gold because it’s the best bang for my buck

  4. Wipperman connex chains are the bees knees. Back in the 10speed days I had a customer that would stretch and sometimes even snap a 7800 chain in a weekend. That was when wipperman was 3x the cost of dura ace and the customer ate it up because of the safety factor.

  5. I’m going to be that guy, who cares? Chains these days are so damn good even Race oriented chains like the KMC x11sl that is on 2 of my frames last about 3,000 miles between replacement. I ride about 10,000 miles a year on super light chains from kmc(which are like 35 bucks ordered from Taiwan through ebay) and have about the same wear as my bikes with $17 Sram chains. Make a quieter, smoother, less prone to dirt build up (which is why I like hollow plate/pin chains) chain and you’ve got my money otherwise it’s just a commodity good.

  6. …….or, you could not replace your chain every couple thousand miles. Instead, use the entire drivetrain together until its completely worn and then replace everything at once. Two different schools of thought.

    • works fine if you are on older equipment and can get the spares cheap. Not so good if you are on new, high end equipment, where new rings, cassette and chain could run you close to $500.

      • Not necessarily. All drivetrains parts, even new high end equipment, will still wear out and need to be replaced. Regularily and obsessively replacing chains does not mean your $500 cassette and chain rings will magically last forever. You’re actually spending more money doing it this way.

        • I used your method once, and ended up with STI shifting that wouldn’t work going up or down the cassette. Never mind the risk of so badly wearing out a chain that it breaks when you are leaning on it. On the other hand, if I change out my chain at 1%, I don’t have to change the other stuff. I’m riding cassettes and rings that are 10 years old, and I don’t baby my stuff.

  7. How does it stack up against sram’s other chains (1130-50-70 etc) or are the chains on the chart each company’s best performer???
    Also, are there any non-biased test teams out there or do parts change so often that tests are moot.

    • From personal experiences Sram chains are the absolute worst. In this order Campagnolo, KMC, and then Shimano are the order I’d buy/use.

      • I would agree that SRAM’s 11s speed chains are horrible in terms of wear. They are quiet and run smooth though but I get about 1500km on mtb and 5000km on my road bike with PC1X. That’s about 1000km short of what I get with KMC X11EL Gold on MTB and 3000km short of what I get on road bike.
        The SRAM Eagle chains are superb though. Over 5000km on an MTB is impressive

  8. Has anyone else broken theseConnex chains regularly. Bought the rustprooof version for my fat bike, snapped it twice in less than 200 miles. Threw it in the trash cause it sticks to fix a chain in sub-zero weather. Just me?

  9. h € €/h
    53 10,0€ 0,19€ sram 1110
    67 13,0€ 0,19€ shimano HG601
    187 43,0€ 0,23€ connex 11sX
    58 16,3€ 0,28€ kmc X11-93
    132 43,0€ 0,33€ campagnolo record
    70 24,0€ 0,34€ shimano hg701
    125 43,0€ 0,34€ connex 11sB
    87 32,5€ 0,37€ Shimano HG901
    112 47,0€ 0,42€ KMC X11sl gold
    57 35,0€ 0,61€ sram xx1
    74 60,0€ 0,81€ FSA N1
    per google prices

    conclusions :
    For the lower cost per hour, get the cheapest chain (sram 1110/shimano hg601) unless you value the few grams shed.
    The connex 11sX has an attractive hourly cost though.
    The sram XX1 is surprisingly bad (the worse FSA N1 price is the K-Force Light one, so not sure)

  10. I would like to know how many test protocols they tried before they found the one to put them on top. Paying for results is quite common in some industries.

      • No, there are 12 levels of a single factor, with what looks to be zero replication. So n = 1. Given there is no reported variance, and no statistical testing, this article is basically worthless. They may well have repeated the test with different chains until they got the results they wanted, and only reported these.

        • actually 12 subjects in a test thus n=12. For a more reliable result 5 chains of each brand could have been tested that would have then been n=60 and a statistical significance could have been demonstrated between the different brands. using one of each brand gives rise to a possible bias due to sampling error. ie the wipperman chain just happens to be a good one from the batch taken from the extremity of the bell curve.

  11. Am I the only one who doesn’t like the name engraved on the side? Seems like a magnet for dirt and would get dirty just looking at it.

  12. I put three HG701 chains on my bike a year, one with a new cassette in the spring, one before the one race I do a year and one in the fall to last me to spring. That costs less than the cost of one Whipperman chain, and I have excellent shifting all year and do roughly 10k km a year.

  13. wouldve been better if the load was on a chainring and cassette.. advanced chains are optimized for their working conditions not side loads.

    my xx1 eagle seems indestructible for example. my red 22? not so. but much ligther and still lasts 2500kms for my usage

    • That’s an interesting point. I find KMC chains to resist linear elongation better than any other I’ve used, but they wear out in another way – laterally. This manifests in inconsistent shift engagement or “floating” adjustment. There seems to be a point where if the chain is too laterally flexible, it doesn’t engage as positively with adjacent cogs, which I find somewhat counter intuitive.

      • no chain elongates linearly. The rollers wear out and therefore the pitch between mating surfaces increases. The outer plates stay the same length

        • This^ Chains don’t stretch, they wear. Rollers and pins decrease in diameter as they wear. This effectively makes the rollers farther apart inside the chain plates which will cause them to chisel away at the teeth of cogs and rings.

    • The test simulated chain use on a cassette via the induced angularity. Any multi-gear bike induces side loads on chains as you shift off the nominal chainline centerline.

      They used a high one though at +/- 5 deg. Typical chainlines will keep max chain angularity at almost half of this so I take the overall percent differences between chains with a grain of salt.

  14. Hmmm… Shimano HG901 chain lasts 47% of the life of the Connex 11sx, but can be easily found for about $36 USD in a lot of locations – even under $40 for RIGHT NOW at your LBS. With the excellent shifting performance of Shimano on a Shimano drivetrain, and a real-life lower cost-per-mile, why go with the Connex? While Wipperman does make a great chain, and the general shifting performance is good, I have found that the shifting degrades fairly quickly with them after just a few months, whereas a well-maintained Shimano chain seems to shift better after they have “worn in” a little… A really impressive test, for whoever may want to take it on, is to compare shifting performance at various points in the test, and consider that as well as wear, and then rank by performance/price to come up with a list of the best-wearing/best-performing/best value per-mile (or hour) chains.

    • interesting that the perception of improved shifting performance of Shimano chain on Shimano Cassette being better than generic brand chain on Shimano cassette. I have found this to not be true. I have found differences in cassette clatter from the chain running over the teeth. Shimano chains do seem to have a slightly smoother running on their drive trains. I have not found the shifting to be improved over a Connex, KMC 11XL or even SRAM PC-X1, XX1, or RED 22 or 1170. There is a marked difference over the cheaper chains across the board, i.e. you get what you pay for in a higher priced chain in that shifting, quietness and longevity seem better balanced and therefore more consistent in performance no matter what groupset they’re running on than the cheapest chains.

      So my philosophy is to buy the best 11speed chains that are on close out at my LBS. I’ll wait till they have close out deals on the KMC’s 11XL EC Gold, Shimano Dura Ace or Ultegra or Sram RED or 1170.

      Interesting thought that the test does not include TAYA chains

  15. Not surprised to see a SRAM wear item at the bottom (or Campagnolo at the top) – one can only wounder how they managed to destroy the reputation of the SEDI (later Sachs) chain.

  16. I think the $/hr is a good metric, However, the data presented is in a sterile environment. It is highly likely that the real world (i.e. dirt) will greatly impact the life of the chain. Maybe it won’t change the rankings, but it will almost certainly change the $/hr results.

  17. Given the recommendation is to change an 11 speed chain at 50% wear, there are only 2 clear winners on the time to wear front – the KMX SL and the Connex SX – and some of the better to 100% wear do much worse than the others to 50% wear.

    The cost to 50% wear is more relevant than the cost to 100% figures.

COMMENT HERE: (For best results, log in through Wordpress or your social media account. Anonymous/fake email comments may be unapproved or deleted. ALL first-time commenter's posts are held for moderation. Check our Comment Policy for full details.)

This site uses Akismet to reduce spam. Learn how your comment data is processed.