How long should a chain last.

[Yellow Saddle]

Well I must have Chris Hoy's legs because my last chain reached 1% in under 500 miles and that's on a hub-geared city commuter!

I do look after my chains, using the Mickle method, and normally get around 4000 miles before replacing so I'm wondering if I just got a duff chain. (It was a KMC Z510, my usual chain). Anyone know why I might have worn it out so quickly?

There are several reasons why that happens.

1) Measurement. If you used one of those commercial chain measurement tools, the reading is very inaccurate. It shows the chain as being shorter when it is dirty and longer when it is clean.
2) The Mickle method is inconcistend and doesn't really remove dirt from inside the chain. It's application isn't standard either, so you may have bathed, but did you wash behind your ears?
3) That particular cycle (that chain's life) may have been slightly different from previous chain's. In other words, you may have ridden through more mud or grit, even on the approximately same path.
4) Batch inconsistencies.

Chain technology is mature. No one manufacturer has technology or steel that the others don't have. The right steel is super cheap, so no-one bothers by trying to cut corners with rubbish steel. What may differ is the heat treatment of the steel. This is done after the various plate components are stamped out of a ribbon of yet untreated steel. If something goes wrong with the heating or quenching, the steel may not be as hard as it can get.

Chain life is a thing of averages.

 

[Yellow Saddle]

[QUOTE 5023602, member: 9609"]

I have that one at 2.14%
I'm curious as to how you measured it from that image? you're very close[/QUOTE]

If I tell you, will you promise to keep it a secret? Here goes, it is a long story but you did ask.

1) I assume that the photo is taken perfectly head-on. That eliminates the parallax error if the subject is reasonably small and the camera was held at a reasonable distance. If the camera was held really close but zoomed out, you would still get a parallax error. I assume you didn't do this.

2) A chain only wears on every second 1/2 link. A 1/2 link in your photo would be a black link or a silver link. If you measure the distance from the left side of a rivet on your silver link to the left side of the other rivet you would get a measurement that doesn't change with wear. That distance remains constant because the rivets don't move. I make the photo big on my screen and measure that distance. In my case it was 59mm. I note down the distance as the constant.
Now I measure the distance between the right rivet on the previous silver link and the left side rivet on the following link. I get a longer distance - 62mm in my case. That's the amount of wear between a black and silver link measured in elongation. I subtract 59 from 62 and get 3. Now I calculate the percentage wear by using the unworn link as a base i.e. 3 divided by 59 x 100 = 5.08%.

3) That figure of 5.08% is the wear over just one link. However, since we know that only ever second link wears and we care about total wear, we divide the single link wear by 2. The figure is approximately 2.5% for the example I've used.

I have often calculated chain wear via pictures on this forum. You're the first one to express curiosity.

Of course the method is not completely accurate. There's always the parallax issue as well as standard measurement flaws. But by enlarging the picture I eliminate most measurement errors. The only thing to be careful of is the rivet shape. If the rivet has a bit of a pimple on the head, as yours does have, then you should measure around it.

 

[hoopdriver]

The assumption that the photo - a photo, any photo - was taken perfectly head-on is a big one.

 

[Yellow Saddle]

The assumption that the photo - a photo, any photo - was taken perfectly head-on is a big one.

Of course it is a big one but a few common-sense checks eliminates the problem. I won't go and calculate wear from a photo that's obviously not suitable, so what's the issue?

 

[hoopdriver]

No issue at all. It's not me spending the time doing it.

 

[Ajax Bay]

My recent experience (9 speed chain, average (rural) ride distance 100 miles):
Last year one chain, run on till the cassette needed replacing as well, gave me 4694km.
This year one chain, run on till the cassette needed replacing as well, gave me 5269km.
Latest chain, also in process of being run on till the cassette needs replacing has given me 4385km.

 

[Yellow Saddle]

[QUOTE 5023711, member: 9609"]experiment with a steel rule, and compare 5mm increments near the centre of the image, and you may be surprised how accurate it is even if it is not perfectly centred or head on, I have used cameras to help measure and it does provide good results[/QUOTE]

Exactly. I doubt that the armchair experimenters and SJWs on this forum will bother. I wonder how they think aerial photography is used to make accurate calculations of objects on the ground.


It took me all of 1 minute to take this photo. It was taken from a height of 600mm (as kinda shown by the vertical steel ruler). I wonder if anyone can calculate the angular error on the extreme sides of the wooden ruler.

 

[hoopdriver]

You obviously have much more time on your hands...

 

[glasgowcyclist]

You obviously have much more time on your hands...

Not that much if he doesn't have time to do up all his shirt buttons!

 

[hoopdriver]

I hope he wasn't so rushed with his calculations...

 

[glasgowcyclist]

1) Measurement. If you used one of those commercial chain measurement tools, the reading is very inaccurate. It shows the chain as being shorter when it is dirty and longer when it is clean.

I do use such a tool but a stretched chain makes a particular sound on my bike and that's when I know to adjust the tension again. This is via an eccentric bottom bracket which, even at its maximum rotation, couldn't take up the slack that had developed. I've never got to that stage with any previous chain. That's when I put the tool on it and it dropped straight in.

2) The Mickle method is inconcistend and doesn't really remove dirt from inside the chain. It's application isn't standard either, so you may have bathed, but did you wash behind your ears?

Okay, but it's not a rushed job I do, and I do take care to clean it well and lube it just enough.

3) That particular cycle (that chain's life) may have been slightly different from previous chain's. In other words, you may have ridden through more mud or grit, even on the approximately same path.

Hmmm, my to and from routes are exactly the same and the weather has been consistent (grey) for the 6 weeks that the chain was on the bike. No salt on the roads yet either, certainly no mud.

4) Batch inconsistencies.

That was my initial thought but a variation that huge seemed unlikely (~480 miles instead of ~4,000).

 

[Yellow Saddle]

I do use such a tool but a stretched chain makes a particular sound on my bike and that's when I know to adjust the tension again. This is via an eccentric bottom bracket which, even at its maximum rotation, couldn't take up the slack that had developed. I've never got to that stage with any previous chain. That's when I put the tool on it and it dropped straight in.

Not many people are observant enough to pick up changes in chain sound, but it is a good indicator of a worn chain. The mechanism seems obscure but with a bit of scrutiny, it should come as a surprise.
The top run of a chain is under tension when the bike is pedaled. Conversely, the bottom run is under slack. When the chain is worn, it's pitch (distance between rollers) is too big for the pitch of an unworn chainring. Keep the rear sprocket out of the equation for now. When a link arrives at the chainring and the next available tooth has to enter the chain, it is obstructed by the incoming roller because the pitch is wrong. The tooth hits the roller but its rounded tip forces it to nevertheless slide past. This collision and subsequent grind sends a vibration down the chain which is within the audible range.

The same doesn't happen on the slack run. When the chain reaches the next incoming tooth there, it simply compresses a bit and conforms to the sprocket. It stays in that position until about the 10-to-12 position where it starts to "feel" some tension. Now it behaves differently from what happens up front. It slides upwards, away from the root between two teeth and settles on a position where the sprocket pitch matches that of the chain. The more the chain is worn, the higher up the chain settles. You can see this in Reiver's photo at the beginning of the thread.

If the bike is a singlespeed and the chain cannot change length courtesy of the derailer spring, then it acts a little bit like the front chainring scenario. However, even that is very variable and depends on a permutation of sprocket and chain wear.