[Closed] Interesting chain wear length experiment and a Q for the engineers?

As an engineer i'd say you're delving a little too deep into this, the part that wears on a chain is the rollers, so why not just measure the wear on them over time, the wear between rollers is what will allow play to start wearing down the teeth on cassettes and chainrings?

Also, Digital Calipers are way better, used to be stipulated on the aircraft i maintained, unfortunately same with Snap-On for some reason!

Roller wear is irrelevant.

Have you done a sense check with the digital caliper on the chain? I would return it and if the next one is equally bad change brand.

rollers do wear but the increase in length is the pins wearing in the inner plates adn the holes in the plates elongating

I suspect what you are measuring here are the tolerances

A brand new chain should be exactly 12 inches long for 24 links

The tolerance of machining on each piece of the chain has very slight variation hence the differences in new length which is more noticeable over the 1.2m as you’ve noticed.

Perhaps take a set of four readings over the length of the chain and average those to get a more accurate wear picture.

The main player is wear rather than stretching of the chain links, but its not quite as simple as it seems as the wear occurs at a number of interfaces. This shows it well:
https://www.sheldonbrown.com/chain-wear.html

I've never really thought about it too much, I replace about 70% of my chains when they get noticeably sloppy then the rest when they read 0.75.

Always use identical chains.
Have an unused chain hanging on the pin as a reference chain.
Compare used chain to the reference chain.
Allow temperature to equalise first, and you have eliminated another variable.
Welcome the the nerd-den.

@tarquin the 4' of chain has 96 links, don't you think that's a good enough random sample of them to rule out that effect?

OP I don't have an answer apart from having seen brand new chains measure "work" also (cheap ones). I would query whether you can measure pin centre to pin centre, I would to it puter link end to outer link end.

Remember when measuring things like pins and holes that centre-centre is the same as edge-same edge; much easier than guessing where the centres are.

I do this, with digital calipers as chain checkers don't seem conclusive enough to me:

New chain= 5.715 inches/145.161mm
1/2% stretched (consider looking for a replacement)= 5.745 inches/145.923mm
3/4% stretched (replace)= 5.760 inches/146.304mm
1% stretched (replace immediately along with new drivetrain)= 5.775 inches/146.685mm

It seems to work.

I also buy individual sprockets on Ali-express to avoid buying entire cassettes.

I have a hierarchy of measuring for the 4 chains I rotate on my main bike.

First measure is with an Rolhoff chain checker. These push two rollers in opposite directions so measure chain wear plus roller tolerance. However, it will always think a chain is worse than it is.

Once that measure starts to look questionable, I'll move to the shimano tool with 3 tangs that pushes the measured rollers in the same direction hence removing another variable. Like all these devices though, it condemns a chain early.

Finally, I have a 600mm steel rule that I only use for chain measuring. It has three sharply trimmed pieces of tape affixed. White at 50mm, green at 558mm and red at 560.5mm. as others have said, there's no need to measure centre to centre, any consist point will do.

Chains are clean and dry for drop in tools and degreased and dried for the steel rule.

The tolerance of machining on each piece of the chain has very slight variation hence the differences in new length which is more noticeable over the 1.2m as you’ve noticed.

Which also explains why you get tight and loose spots when putting a brand new chain onto brand new direct mount chainring (should be perfectly round?) and a brand new cog (should also be perfectly round?). Not noticeable to anyone running gears but very noticeable when running single speed.

A brand new chain should be exactly 12 inches long for 24 links

I think we should petition SRAM and Shimano to make metric drivetrains.

Who’s with me?

http://www.velo-pages.com/main.php?g2_view=core.DownloadItem&g2_itemId=7098&g2_serialNumber=2

Already been done... 10mm pitch Dura ace

Too late.

http://velobase.com/ViewGroup.aspx?GroupID=cb346e28-8437-4367-88b6-a9f5b8d8a1ca

Try to avoid c-c measurements. Use a definitive edge and no loss. I it's a concentric item the distance will be the same...

I think we should petition SRAM and Shimano to make metric drivetrains.

Oh god please no. Chain pitch is about the only standard that’s not been messed around with needlessly so far.

alexnharvey
Roller wear is irrelevant.

Have you done a sense check with the digital caliper on the chain? I would return it and if the next one is equally bad change brand.

I agree regarding roller wear. I wasn't using digital though.

tjagain
rollers do wear but the increase in length is the pins wearing in the inner plates adn the holes in the plates elongating

I suspect what you are measuring here are the tolerances

A brand new chain should be exactly 12 inches long for 24 links

Agree with your first two points although I suspect the tolerances are intentional to allow space for lube. I am not so sure about "exactly" 12 inches though, maybe with a brand new chain with the factory lube in place. I didn't mention that I had marked the post at each 12" interval and there was a steady creep forward at each mark so I think the tolerances must be pretty consistent along the length of the chain.

tarquin

Perhaps take a set of four readings over the length of the chain and average those to get a more accurate wear picture.

Unfortunately, I don't have the means to measure like that in a highly accurate way, must invest in a long engineers rule. The main point of this exercise was to use the long length to magnify the measurements so that the critical wear points could easilly be seen without the need for high precision.

TroutWrestler

Have an unused chain hanging on the pin as a reference chain.
Compare used chain to the reference chain.

Welcome the the nerd-den.

Not sure I can afford to have a new chain just hanging around 🙂 and thanks for the welcome.

chickenman
Remember when measuring things like pins and holes that centre-centre is the same as edge-same edge; much easier than guessing where the centres are.

Agreed, and that is what I do sometimes when trying to measure across shorter lengths.

reeksy

I also buy individual sprockets on Ali-express to avoid buying entire cassettes.

Is that possible for fully rivetted cassettes? Or do you get the drill out?

Onzadog

Chains are clean and dry for drop in tools and degreased and dried for the steel rule.

Yes, I think degreasing is probably the best way. Trouble is, I don't usually degrease before waxing used chains so this would add an extra complication to the procedure. I might do it to the other new chains and see what the results are.

TheBrick

Try to avoid c-c measurements. Use a definitive edge and no loss. I it’s a concentric item the distance will be the same…

Agreed.

I wasn’t using digital though.

Yes I know. I meant why not also measure the chain using the digital calipers and see how that measurement compares to the others and if it is consistent with your full length measurement.

@alexnharvey I don't have any digital callipers, only vernier ones.

Ah, sorry. Sense check the full length measurement using the vernier calipers.

@alexnharvey sorry, not sure what you mean by "sense check" just assumed it was an electronics thing relating to digital calipers.

Sense as in sensible but I should or could also just say a 'double check' with a known measure. Does the X01 also measure longer than expected on the vernier caliper?

You have introduced a new method of measuring and found a result that surprises you. Therefore it is always worth going back to your old measurement method to compare new to old for verification, i.e. verifying that it's really the chain and not the device/method of measuring that is underlying the surprising result.

I've recently introduced myself headfirst to this rabbit-hole. I'm measuring chain elongation using a 12 inch digital caliper.

DML 300mm 12 Inch Digital Vernier Caliper https://www.amazon.co.uk/dp/B06X3V8VY7/ref=cm_sw_em_r_mt_dp_2SYQHJQPHEGSM4MD8AA8?_encoding=UTF8&psc=1

Methodology:

Methodology assumes roller wear is even across the chain. I take internal measurements, roller to roller on non-inner plate half links (because there's room to get the caliper anvils inside the links (even 12s)).

1. Switch units to decimal inches
2. Use internal measurement across the smallest measurable unit (three half links; outer-inner-outer). This reading is the 1.5 * (1 + elongation) - worn_size_of_two_rollers
3. Zero caliper
4. Extend out to measure the reading closest to 10 inches (23 half links of chain).

The distance between anvils: 11.5 * (1 + elongation) - worn_size_of_two_rollers
Minus the zeroed offset = 10 * (1 + elongation)

The decimal part of the reading is the percentage stretch. If the caliper reads 10.05 it is 0.5% stretched.

Caveats: a) like most calipers, the internal anvils come to a sharp edge - the sharp edge can slip off the roller and into the small clearance gap to the inner plate; b) rollers may be unevenly worn.

Chain checkers

Bicycling lore of chain replacement specifies change of chain at some sort of arbitrary number established using some sort of arbitrary checker. Is it that...

...the lore is based on the the standard type of drop-in checker (e.g Park CC-3.2, "worn at 0.5%, replace at 0.75%") and the threshold values take into account typical roller wear, in which case the elongation at replacement is actually much lower than the indicated value?

or... the lore is based on the three prong type checker, deterministically measuring elongation and not measuring roller wear, meaning that anyone using a drop in checker has been replacing their chain prematurely?

My own conclusion is that this is all typical cycling industry vagueness. There is no science. Products have appeared to meet the brochure requirement of "measure the chain". I have two drop-in chain checkers. I've spent my money and they indicate something about wear. But yet the cycling industry told me I was missing out on the science of a three pronger - I baulked and figured out the digital caliper solution instead of throwing more money into single purpose tools.

EMTB

eMTB drivetrains don't last.

With a very small mileage on my eMTB I measured (drop-in checker) the earliest indication of some wear (regardless of roller vs inner plate wear). A replacement X01 Eagle chain skipped immediately when fitted. At this point either a) the "checker" had failed to be useful... or b) the XO1 out of the box was a lemon (unlikely). I began my descent down the rabbit-hole alongside carrying on using the original fitment (worn) chain and cassette.

In hindsight the best advice seems to be a combination of rotating multiple chains from new and switching to a cost effective drivetrain technology when everything needs replacing (e.g. 10 speed).

FWIW, I like the "hang a chain on a nail" approach.

Should I go and buy a chain checker???!!!
Probably depends on what you want to measure.

https://cyclingtips.com/2019/12/the-best-bicycle-chain-durability-and-efficiency-tested/

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

Or get a free one from Renold

https://www.renold.com.au/products/sprockets-and-accessories/chain-wear-guide/standard/

@alexnharvey I understand what you mean now. Although I think the fact that all of the chains were hung and zeroed to the same fixed point and checked against another fixed point does confirm that the discrepancies exist.

@peaslaker Yes, the idea of the hang a chain method is to make it a very simple process without the need for high precision. I couldn't quite get my head around the method you are using, obviously too math heavy for my simple brain 🙂

Of course, this method will only be of use to people who remove chains on a fairly regular basis or to use occasionally as a double check when their chain checker tells them it's time.

I know I said I couldn't be arsed but I think for the sake of completeness I am going to degrease all the new chains and see what result that gives. Just to see if I can find out if there is a consistent tolerance to factor in.

It's worth noting though that this is all done with Sram chains. Shimano etc. tolerances may be different.

I couldn’t quite get my head around the method you are using, obviously too math heavy for my simple brain 🙂

Our experience of chains, I believe the ‘stretch we talk about is wear’ - the rollers and pins wear, (also when you measured the remaining grease/oil in some will make a difference), the pins wear the plate holes adding to this.. and whilst I think chains don’t (are very unlikely) to stretch on a bicycle on a race motorbike or race car, one blast off the line will make the chain measurably longer in one place over another. Yes this could be the pins elongating the plate holes but yet to find this, though I’m sure that’s there as well. Now that’s a lot of power but they are big chains, so does make me wonder if the same is a little bit possible when we stomp on a pedal uphill?
Seems to me that the relationship between the chain and sprockets and chainrings is important. We’ve had sets that won’t take a new chain. An extreme of this is my touring bike with a five speed block, three front, thick chain, I haven’t changed any of this for 10 years and all works a treat. How ever on the 11 and now 12 speed set ups we are having more regular problems with small bits of dirt (older chains must have had but didn’t notice) making the chains stiffer, catch more, being more temperamental etc..

so does make me wonder if the same is a little bit possible when we stomp on a pedal uphill?

No.

So, can I take anything useful from this, engineers of the forum? Or has it been a waste of time?

In all your rambling up there on the OP you don't actually say what it is you're trying to achieve? (other than measure a chain) Presumably it's to make your drivetrain last as long as you can? then it's lube you want and the best for that is probably that massive Plutoline thread.

Also; listen to what @tjagain tells you to do as AFAIK he's been down this rabbit hole for as longs as there's been a STW forum and what he doesn't know about longevity of drivetrains probably isn't worth knowing..

Measuring chain lengths to the degree you're doing is (weirdly) interesting, but isn't the be all and end all, as the alternative (letting it all "wear-out" together) can make it last just as long (if not longer) than rotating or changing chains constantly IME.

Also; listen to what @tjagain tells you to do as AFAIK he’s been down this rabbit hole for as longs as there’s been a STW forum and what he doesn’t know about longevity of drivetrains probably isn’t worth knowing..

Errmmm - I don't think thats right really. Its something I have looked at and thought about but I am no engineer as I am sure the engineers on here would tell you 🙂

@nickc Sorry, I thought I had made plenty allusions to what I am up to. Maybe not in the OP but since then.
So, to spell it out. I am trying to achieve maximum longevity from the drive train by rotating chains at say 0.5% or maybe sooner. I am trying to find a simple way to check that measurement that I can quickly achieve while the chain is off for waxing without resorting to using precision measuring methods.

I am already a member of the Putoline cult and it was probably the postings of @tjagain that led me in that direction and have a lot of respect for his opinions.

Letting it all wear out together is another option that I have used in the past on cheaper drive trains and whether it lasts longer that way I really couldn't say but for now I am trying this method.

Surely drivetrain longevity is for bargain basement components? (as they're often made of harder/stronger materials at the expense of weight) as the advantage of top end drivetrains is mostly weight rather than longevity, so they're often not the best for these sorts of experiments.

(letting it all “wear-out” together) can make it last just as long (if not longer)

Just as data point on this eternal argument: I have 5800 105 on my winter road bike which is used exclusively in rubbish weather. I get a little over 1,000 miles before a chain measures (with calipers/ruler) 0.7%.

Unfortunately, I let the last chain go to well over 1%, by which time it sounded horrible and the chain was jumping off the chainring if you put it in big/big.

A new chain was skipping on the new chainring, although I seem to have just got away without needing a new cassette. So that's a new £65 chainring plus a new chain to get back to working gears. Mileage on the chain? Less than 2,000.

So that's an extra 800 miles or so, the last few hundred of which involved an annoyingly noisy and increasingly unreliable drivetrain.

The run-it-into-the-ground theory might have worked better in the days of inexpensive and more robust steel chainrings, but now that chainrings are both more expensive and softer, I think the numbers just don't add up, especially if you don't like riding with noisy gears.

I run 3 chains on my Specialized Levo. The original KMC chain that came on it, and two SRAM PC1110 chains. They are labelled (1, 2 & 3) with little alloy tags on a keyring when not on the bike. They get rotated roughly every 100km, and when all 3 chains have been used they get a group bath in Putoline. All the dates and distances get written on whiteboard in the garage for each rotation.

I am 2500km into the bike, and the original chain is still on the go. I will replace the cassette and chainring along with the chains when they wear out. To put it into context, I am on the 3rd motor.

I keep the drivechain clean and well adjusted. I know how to shift when not under load. I see reports of ebikers wearing out a drivechain in less than 100km and I'm amazed. There HAS to be an element of hamfistedness and lack of mechanical sympathy for this to be the case.

Any idea what the chains are measuring after those 2500km? Is there any evidence that this works better than simply swapping before 0.75%?

I know that if I actually remember to swap the chains before 0.75%, the chainrings and cassettes will see out a lot of chains. My first chainring did more than 10,000 miles before I killed it by failing to change a chain in time.

There is another Cycling Tips piece that looks at measuring chain wear: https://cyclingtips.com/2019/08/bicycle-chain-wear-and-checking-for-it/ . It's quite a long read!

They said that when measuring a hanging chain it should be clean and weighted at the end to pull it down. So it sounds like the factory lube could affect the readings.

They recommended the three pronged chain checkers, like the Park CC-4, Pedros and Shimano tools. These also work properly with chains that have oversize rollers like SRAM Eagle.

So I got the CC-4 which is definitely better than my old chain checker and easier than the ruler method, which I could never be bothered to do regularly enough.

The measurements seem much more accurate: a part worn 9 speed chain I have is now measuring between .5 and .75 compared to over 1.0 on the old tool. The difference is almost certainly due to roller wear. Other chains I have are too new to compare.

nickc

Surely drivetrain longevity is for bargain basement components? (as they’re often made of harder/stronger materials at the expense of weight) as the advantage of top end drivetrains is mostly weight rather than longevity, so they’re often not the best for these sorts of experiments.

That's kind of the opposite to how I see things. If I have an expensive drivetrain then I want to take as much care with it as possible to put off the evil day when I have to pay for replacement parts.

@lucky13 good article, seems they tried the same hanging method that I am using and also found discrepancies between new chains.