Chain Wear

[tykeboy2003]

You have to bin the chain......this is how it works......
The chain wears, and the worn chain wears the cassette.

Quite.

[tykeboy2003]

Update, I've now done 800 miles on the new chain and measured it today by hanging it from a nail (as I did when it was new) and there is no measurable lengthening. So it looks like the application of nothing but Halfords "Drywax" chain lube has done the trick - although it's not done a winter yet......

[meic]

It probably doesnt work quite as simply as that.
The bits of the chain that rub each other are probably surface hardened, so the wear will be quite slow while that is being worn but when it gets through to the softer metal underneath the rate of wear will speed up considerably.

I dont know why they have chosen certain amounts of wear to be critical, is it because of the chain's interaction with the teeth or because that is the depth of surface hardening?

[CJ]

I dont know why they have chosen certain amounts of wear to be critical, is it because of the chain's interaction with the teeth or because that is the depth of surface hardening?

I don't really know either, but I think the critcality is mostly the chain's interaction with teeth. I think what happens as the pitch of the chain becomes effectively longer, is the load becomes increasingly concentrated on the teeth nearest to the point of (dis)engagement, in a way that doesn't change much up to a certain amount of wear and then changes much quicker. And I would speculate that when the load is shared between a critically small number of teeth, the load on each tooth may be too much for the case-hardening on the teeth, which may perhaps spall off where the chain rollers are pressing hardest. This will make the hollows in the pressure faces of the teeth that a new chain snugs into, so that it can't feed in past the now protruding tooth tips, hence it jumps.

In my case the choice of how much wear to call critical has been a matter of trial and error and has changed as I've adopted more accurate methods of measuring it (than the usual gauge). I used to think 1% elongation was okay, but after a couple of cases of a new chain jumping badly on a cassette recently vacated by a chain with exactly that much wear, I've dropped it to 0.5%. I don't always catch a chain at 0.5%, some get to 0.6 or 0.7 and sometimes there's a bit of new chain jumping on sprockets vacated by the latter - but usually only one or two sprockets and only for a short while. So I think 0.5% must be about right, as the amount of wear that lets you keep the old cassette.

And although I don't keep records of how many miles I've got from chains, I don't think I'm replacing them twice as often now I do it at half the wear. A bit more often, but not twice. So probably the chain wear also accelerates.

[531colin]

...............
In my case the choice of how much wear to call critical has been a matter of trial and error and has changed as I've adopted more accurate methods of measuring it (than the usual gauge). I used to think 1% elongation was okay, but after a couple of cases of a new chain jumping badly on a cassette recently vacated by a chain with exactly that much wear, I've dropped it to 0.5%. I don't always catch a chain at 0.5%, some get to 0.6 or 0.7 and sometimes there's a bit of new chain jumping on sprockets vacated by the latter - but usually only one or two sprockets and only for a short while. So I think 0.5% must be about right, as the amount of wear that lets you keep the old cassette. ...................

Elephant in the room?
Having lampooned wear gauges for including "the wrong sort of wear" in their estimation, the engineer's solution is to measure something which he can measure to the Nth degree of accuracy. If that accurate measurement is so laborious that you don't do it often enough, and is anyway a poor predictor of chain jumping, as in "a couple of cases out of so many chains", then in real world use its a worse measure than my chain gauge.

Last winter (or maybe the winter before) my failure to spot chain corrosion cost me a near-irreplaceable 7 speed screw-on block, so now the chain gauge hangs in the workshop next to the track pump, and I waste a couple of seconds checking the chain on every bike when I pump a tyre**. In 15 or 20 years using the gauge, I have never had a chain skip when I have replaced the chain when the gauge tells me to. (but I have on other peoples bikes, where the wear has gone over the limit)

My chain "wear" proceeds nice and slowly until a certain point when it "jumps off a cliff", and that point is corrosion, ie winter, or a few foul days when I have to ride, ie my accommodation is booked. Corrosion provides the only occasion when I have to make a decision regarding chain wear. If I get the odd link where the gauge drops right in but the rest of the chain is OK, I carry on using it. More than several links dropping through, and I replace the chain.

I expect one day I might get a chain with uniform wear according to my gauge, and then I will measure the wear for comparison, but it seems to me pointless to try to measure wear that may be uneven. (or perhaps the uneven wear is confined to the rollers, and the pins etc wear evenly?)

In the past I have replaced cassettes when I re-rim a hub, the old rim having worn through. The cassette teeth show wear, but new chains run without skipping. Now I have ceramic rims on a couple of bikes, it might be interesting.....

**edit for clarity....I mean I gauge the chain on the bike I'm pumping, not all the bikes!

[Mick F]

39" steel rule hangs on a hook in the workshop.
I take it down and lay my chains against it.
Any wear/stretch is immediately seen by the elongation at the 39"mark.

0.5% (3/16ths @ 39") is too far IMHO.
I go for 0.4% or less.

By that time, the chain is all wobbly and waggly.

[iow]

....Having lampooned wear gauges for including "the wrong sort of wear" in their estimation, the engineer's solution is to measure something which he can measure to the Nth degree of accuracy......

having read THIS i'm no longer convinced that roller wear can be ignored. the diagram appears to show sprocket wear is possible with worn rollers alone even on a chain where the pitch (as measured by a ruler) remains at 1/2 inch.
comments and thoughts appreciated.

[meic]

Interesting picture and rather compelling at first glance.

Thinking about it a bit, the wear on the rollers will allow the radius of the chains curvature to shrink a little and by being further in it should have a smaller pitch, so for that new position it has too large a pitch which is what over-elongation is all about.

However that picture is exaggerated and I wonder if the effect of 1/8 mm of roller wear on the radius will have any significant effect on the seating of the chain?

It does rather make you think that it isnt worth the effort of trying to eliminate roller wear from your measuring of chain length. I find that I have to get rid of my chains before they reach the magical elongation figure and by that time they do have a lot of roller wear.

I have a curiosity about what is really happening but it dosnt seem to have much effect on when I choose to replace my chains that is decided more empirically by what I like or can get away with.

[Mick F]

I have a curiosity about what is really happening but it dosnt seem to have much effect on when I choose to replace my chains that is decided more empirically by what I like or can get away with.

I have a curiosity about it too, but it doesn't make any difference what is happening or why.

If the chain is worn - or getting worn - chuck it out before it's too late.

[531colin]

The debate centres around how you measure that the wear exceeds the magic amount.

My Rohloff chain wear gauge is uncannily accurate, up to the mark and a new chain will run on the cassette, over the mark and it will skip, even if only on the favourite sprockets.

This suggests to me that Rohloff know exactly how a chain wears......the alternatives that I can think of are....
1) they were very lucky when they made their first gauge
2) they did it empirically, by measuring loads of worn chains and knowing if a new chain skipped on their associated cassette.

[andrewjoseph]

I find it very strange that park tools and their website are commonly recommended by the experienced riders, to everyone who wants learn how to maintain their bikes.

But not the chain gauge.

You should never, on any account, use any sort of chain gauge, instead you should use a steel rule and try to match the ruler to the very dead centre of a small rivet, while the chain is not under any tension.

I too will continue to use my chain gauge.

[CJ]

Elephant in the room?
Having lampooned wear gauges for including "the wrong sort of wear" in their estimation, the engineer's solution is to measure something which he can measure to the Nth degree of accuracy. If that accurate measurement is so laborious that you don't do it often enough, and is anyway a poor predictor of chain jumping, as in "a couple of cases out of so many chains", then in real world use its a worse measure than my chain gauge.

The proposed elephant is in fact a red herring.

I maintain more than a dozen bicycles used by a family of four, two or whom were latterly students, to whose bicycles my access was therefore intermittent, so it's been easy for them to put a load more wear onto a chain when I'm not around to catch it. They've 'boomeranged' just now, so I am in the process of addressing four or more years of neglect!

None of the several bikes used by my wife and I have escaped my more accurate wear gauge, which is no more laborious to use than an inaccurate one. I nevertheless acknowledge that the latter, which is far easier to buy at least, may be perfectly adequate if always used on the same high-quality brands of chain, that the user has found by experience to be made with the same amount of roller float as the designer of that gauge assumed to be present.

As I am maintaining so many disparate bikes however, and advising people who may be using any brand and quality of chain, I need for myself and can recommend without reservation to others, only those gauges that eliminate other factors than the all-important pin wear.

[townbikemark]

Thanks for all the replies, I'll re-check the chain with a steel ruler.

I fairly regularly clean my chain by removing it from the bike and soaking it in some WD40 type stuff I got off eBay. I then give it a good shake and then another soak/shake in some fresh stuff. I then oil it with Green Oil.

This is an interesting point of view:-

Cleaning chains is pointless.
You wash out the factory fitted lube, and there's NO WAY of stopping fine grit particles getting in. The best you can achieve is to spray it with a wax furniture polish when new and wipe with a cloth when it looks dirty.

I seem to remember reading something similar on Sheldon Brown, but it just seems to feel so wrong!

Edit - I've got a new chain on order.

I can recall seeing a Bicycle Tutor video stating the last thing you should use on your chain is WD40...this was when it was free - the vids are subscription only now. (Not sure if what you have IS WD40 notwithstanding.)

[531colin]

......... I used to think 1% elongation was okay, but after a couple of cases of a new chain jumping badly on a cassette recently vacated by a chain with exactly that much wear,.................

Sorry, I interpreted that to mean that some chains worn to 1% wore the cassettes so a new chain skips, and some didn't. Did you mean all chains worn to 1% cause skipping?

Clearly, if you can't get access to the bike, then you can't service it.

My gauge works on Sram (pc 48 up to 9 speed) Shimano, Connex, KMC, and probably others from 5 to 9 speed, I lost interest after 9.

[Big T]

39" steel rule hangs on a hook in the workshop.
I take it down and lay my chains against it.
Any wear/stretch is immediately seen by the elongation at the 39"mark.

0.5% (3/16ths @ 39") is too far IMHO.
I go for 0.4% or less.

By that time, the chain is all wobbly and waggly.

Do you have to take your chain off to measure it? (I know you do this anyway as part of your cleaning/lubrication regime).

I bought a 12" steel rule to use with the chain still on the bike, but found it too imprecise, so bought a Cyclo chain measuring device. Chains that I thought were OK, were in fact well past 1% worn, so both chain and cassette were replaced. I even had to replace the chainring on my audax bike. Chain on the touring bike partly worn (over 0.5%, but not 1%) was replaced and runs OK on a worn cassette.