Simple guide to worn rear cassette

For discussions about bikes and equipment.
User avatar
CJ
Posts: 2981
Joined: 15 Jan 2007, 9:55pm

Re: Simple guide to worn rear cassette

Postby CJ » 26 May 2019, 1:24pm

thelawnet wrote:If you get 3 chains @ 0.5% to 1 cassette, rather than 1 chain @ 1% to 1 cassette, then for the same (3%) cumulative chain wear the equation is

2 (3Cn + S) < 3 (Cn + Cs)

i.e.

Cs > 3Cn

So in that case the cassette needs to cost more than 3 times as much as the chain to justify changing the chain twice as often. This will rarely be the case, though there will be sometimes that it is true.

Of course if the chainrings wear out when wearing a chain to 1%, that will also affect the equation, and wearing to 0.75% rather than 0.5% is clearly not as dramatic as 0.5% vs 1%.

This calculation assumes that chains wear at a constant rate versus mileage, whereas my engineering intuition, matched by empirical observation, leads me to expect that wear accellerates as the chain wears. The particular engineering reasons for that I'll leave for another post, but the result is that the 0.5% to 1% phase will happen quicker than 0 to 0.5% and you don't get twice the mileage by leaving the chain to wear twice as much.

Let's assume it wears twice as fast during the 0.5 to 1% phase, so you get 1.5 times the mileage. In that case you'll buy two chains and two cassettes in the same mileage as running three chains to 0.5% on one cassette. The break-even equation then becomes:

2 chains + 2 cassettes >= 3 chains + 1 cassette
So: 2 cassettes - 1 cassette >= 3 chains - 2 chains
So: 1 cassette >= 1 chain

In this case a cassette only has to cost slightly more than a chain, for it to be more economical to replace chains at 0.5% wear. In practice they cost about twice as much and we break even at a less rapid rate of chain wear accelleration; which I'll not bother to calcualte because this is mostly guesswork anyway! And because in pactice, by replacing chains at 0.5%, I've had cassettes outlast more than three chains.
Chris Juden
One lady owner, never raced or jumped.

User avatar
Gattonero
Posts: 3540
Joined: 31 Jan 2016, 1:35pm
Location: London
Contact:

Re: Simple guide to worn rear cassette

Postby Gattonero » 27 May 2019, 9:34am

IMO, the most wise approach is to consider the use made of the parts and their price.

I.e. on my Brompton the sprockets are so cheap that I won't bother doing any calculation nor measurements: when they reach the 0.75% wear mark on the Park Tool CC2.2 checker (whatever are the measurements error this implies), both chain and sprockets go in the bin.

Other bikes I have will use single freewheels, 7 speed freewheels, 8 speed Campagnolo, 9 speed Shimano and 10 and 11 speed Campagnolo. The last two, do get more attention due to the cost of the cassettes. I replace the chain just before it can get to the 0.5% wear mark and can get 2-3 chains out of the same cassette.
The above brings to a very important thing to consider: how many sprockets are actually used, and the pedalling style of the rider. I'm not too heavy, I use almost all the sprockets in a cassette (clearly, the middle ones get some more use) and especially I do not "stomp" on the pedals so the wear on my cassettes is fairly even.
It is by riding a bicycle that you learn the contours of a country best,
since you have to sweat up the hills and coast down them.
Thus you remember them as they actually are...

User avatar
Sweep
Posts: 5012
Joined: 20 Oct 2011, 4:57pm
Location: London

Re: Simple guide to worn rear cassette

Postby Sweep » 27 May 2019, 11:50am

Have come across a few folk gatto who change at 0.5 per cent but am puzzled. If that is good practice why do chain checkers commonly have just 0.75 and 1 per cent markers? My KMC checker just has a single marker 0.8 per cent.
I tend to change chains at 0.75 per cent, hope for the best on the chainrings, change cassette after 3 or so chains (in truth i don't count) though as my initial post implies i have no great idea of when a cassette is worn.
Sweep

Brucey
Posts: 33816
Joined: 4 Jan 2012, 6:25pm

Re: Simple guide to worn rear cassette

Postby Brucey » 27 May 2019, 12:21pm

Gattonero wrote:…..The above brings to a very important thing to consider: how many sprockets are actually used, and the pedalling style of the rider....


this is a good point. Between this and the vagaries of actual measurement, it is no surprise that folk successfully use different strategies when maintaining their transmissions.

Some folk manage to knacker a (favoured, small) sprocket such that it won't take a new chain before the chain itself is worn to 0.5%.

Chains vary too; it seems to me that chains with slack-fitting rollers are often a bit more tolerant of worn sprockets.

I'd throw another one in there which is that because folk pedal differently, they may or may not suffer skipping even with the same combination of parts.

I've seen cassettes that didn't look especially worn even though the chain measured about 1% wear; I'd associate this with riders that pedal swiftly and smoothly rather than those that pedal forcefully.

Bottom line is that eventually you will figure out what works for you if you try enough different things...?

cheers
-------------------------------------------------------------------------------
~~~~~~~~~~~~~~~~~~~~~~Brucey~~~~~~~~~~~~~~~~~~~~~~~~

mercalia
Posts: 10305
Joined: 22 Sep 2013, 10:03pm
Location: london South

Re: Simple guide to worn rear cassette

Postby mercalia » 27 May 2019, 3:42pm

I seem to have a magic chain. I have used it alot not sure of the mileage, but when I do the 12" ruler test hasnt worn by 1/32" though my Parks tool suggests maybe some where between .5 and .75%. or maybe its the Carlube spray grease I am using that is doing a good job. Any way I have no complaints

User avatar
CJ
Posts: 2981
Joined: 15 Jan 2007, 9:55pm

Re: Simple guide to worn rear cassette

Postby CJ » 27 May 2019, 4:25pm

Sweep wrote:Have come across a few folk gatto who change at 0.5 per cent but am puzzled. If that is good practice why do chain checkers commonly have just 0.75 and 1 per cent markers? My KMC checker just has a single marker 0.8 per cent.
I tend to change chains at 0.75 per cent, hope for the best on the chainrings, change cassette after 3 or so chains (in truth i don't count) though as my initial post implies i have no great idea of when a cassette is worn.

The numbers on chain measurement tools provide a false impression of precision. These tools are quite short, four inches in the case of Rohloff's Caliber-2 and they simply insert between the rollers. The fault with that is it adds two random doses of roller slop (of which all chains have varying amounts even when new) onto only eight doses of the all-important pin/bush wear it's pretending to measure!

Perhaps these tools are made to allow for the typical roller slop of a good-quality slightly-worn chain, but one can only guess. I can tell you for sure that a less-good-quality chain can have such sloppy rollers that these tools will declare a band new chain as over "1%" elongated, even when the actual length of the whole chain, measured under tension, comes out spot on! And when I fitted that chain (at my friend's request to his bike), notwithstanding it's palpably sloppy rollers, it worked just like a new chain should! (But maybe not for long.)

I put "1%" in quotes not only because it's inaccurate, but also because when you look closer at the Rohloff tool, it actually says 0.1mm, with 0.075mm for the other side. I guess that must be mm per link, meaning 0.8% and 0.6% respectively. Whatever: it's nonsense since I invariably find chains to be less worn than these short-arsed, roller-pushing tools say they are. How much less you will discover only when you hang the chain from a nail and measure a much greater length. I have such a nail (same thickness as a chain rivet and with the head cut off) near the top of my shed doorframe, and a couple of marks exactly 50 and 50.5 inches below it. When the end rivet of an outer link gets halfway to the second mark, I reckon it's time to replace the chain.

I think that with everyone using these woefully inaccurate tools, differences in roller slop probably account for much of the variation in the amount of chain wear people find they can get away with, since the chain they think has "1%" wear, may well have only 0.5% actually. I replace chains at 0.5% of ACTUAL elongation.
Chris Juden
One lady owner, never raced or jumped.

Brucey
Posts: 33816
Joined: 4 Jan 2012, 6:25pm

Re: Simple guide to worn rear cassette

Postby Brucey » 27 May 2019, 6:11pm

cheap verniers allow on-the-bike chain measurement without roller slop interference;

https://forum.cyclinguk.org/viewtopic.php?f=5&t=115336

cheers
-------------------------------------------------------------------------------
~~~~~~~~~~~~~~~~~~~~~~Brucey~~~~~~~~~~~~~~~~~~~~~~~~

User avatar
Gattonero
Posts: 3540
Joined: 31 Jan 2016, 1:35pm
Location: London
Contact:

Re: Simple guide to worn rear cassette

Postby Gattonero » 29 May 2019, 6:23am

CJ wrote:
Sweep wrote:Have come across a few folk gatto who change at 0.5 per cent but am puzzled. If that is good practice why do chain checkers commonly have just 0.75 and 1 per cent markers? My KMC checker just has a single marker 0.8 per cent.
I tend to change chains at 0.75 per cent, hope for the best on the chainrings, change cassette after 3 or so chains (in truth i don't count) though as my initial post implies i have no great idea of when a cassette is worn.

The numbers on chain measurement tools provide a false impression of precision. These tools are quite short, four inches in the case of Rohloff's Caliber-2 and they simply insert between the rollers. The fault with that is it adds two random doses of roller slop (of which all chains have varying amounts even when new) onto only eight doses of the all-important pin/bush wear it's pretending to measure!

Perhaps these tools are made to allow for the typical roller slop of a good-quality slightly-worn chain, but one can only guess. I can tell you for sure that a less-good-quality chain can have such sloppy rollers that these tools will declare a band new chain as over "1%" elongated, even when the actual length of the whole chain, measured under tension, comes out spot on! And when I fitted that chain (at my friend's request to his bike), notwithstanding it's palpably sloppy rollers, it worked just like a new chain should! (But maybe not for long.)

I put "1%" in quotes not only because it's inaccurate, but also because when you look closer at the Rohloff tool, it actually says 0.1mm, with 0.075mm for the other side. I guess that must be mm per link, meaning 0.8% and 0.6% respectively. Whatever: it's nonsense since I invariably find chains to be less worn than these short-arsed, roller-pushing tools say they are. How much less you will discover only when you hang the chain from a nail and measure a much greater length. I have such a nail (same thickness as a chain rivet and with the head cut off) near the top of my shed doorframe, and a couple of marks exactly 50 and 50.5 inches below it. When the end rivet of an outer link gets halfway to the second mark, I reckon it's time to replace the chain.

I think that with everyone using these woefully inaccurate tools, differences in roller slop probably account for much of the variation in the amount of chain wear people find they can get away with, since the chain they think has "1%" wear, may well have only 0.5% actually. I replace chains at 0.5% of ACTUAL elongation.


There's too much thinking about how to get the most perfect measurement of chain wear.
But the main problem I see is: what are you going to do with the data?

The Average Joe will have zero benefit from a measurement that includes "roller wear" and one that doesn't, because this number has no meaning for him: not having replaced enough chains or cassettes, there's no comparison to make.
Indeed, as I said above, wear in the sprockets can easily be uneven and not easy to spot; plus the rider may have a high or low power output that is applied to the pedals on a short part of the sprocket revolution, which implies different wear in different parts of the same sprocket.
It is by riding a bicycle that you learn the contours of a country best,
since you have to sweat up the hills and coast down them.
Thus you remember them as they actually are...

User avatar
Gattonero
Posts: 3540
Joined: 31 Jan 2016, 1:35pm
Location: London
Contact:

Re: Simple guide to worn rear cassette

Postby Gattonero » 29 May 2019, 6:29am

Sweep wrote:Have come across a few folk gatto who change at 0.5 per cent but am puzzled. If that is good practice why do chain checkers commonly have just 0.75 and 1 per cent markers? My KMC checker just has a single marker 0.8 per cent.
I tend to change chains at 0.75 per cent, hope for the best on the chainrings, change cassette after 3 or so chains (in truth i don't count) though as my initial post implies i have no great idea of when a cassette is worn.


The CC2.2 has the wear marks with 170.5mm for the 0.75% and 170.0mm for the 0.5%
Even the last one is too optimistic, most riders and especially the heavy ones, will get a new chain to skip even on a cassette that looks evenly worn (the amount of oil/grease on each sprocket can give clues) and had not had other chains.
IMO, the lowest mark should be set to 169.5mm
It is by riding a bicycle that you learn the contours of a country best,
since you have to sweat up the hills and coast down them.
Thus you remember them as they actually are...

Brucey
Posts: 33816
Joined: 4 Jan 2012, 6:25pm

Re: Simple guide to worn rear cassette

Postby Brucey » 29 May 2019, 7:45am

what are you going to do with the data?

If you use a chain checking tool (any chain checking tool more or less) after a few goes you will learn for your riding style/chain choice/etc what elongation allows you to safely change the chain before the cassette won't take a new chain without skipping. It might be a reading of 0.5% , it might be a reading of 0.75%, and those readings may or may not be 'real' elongations of the chain. For this use it actually doesn't matter if the calibration is absolute, not until you compare readings with other folks; discrepancies can arise through errors in chain measurement and/or variations in the way the cassette wears with any given chain (length).

The scientist in me fundamentally abhors any inaccuracy in measurement but the pragmatist in me sees that you can use almost any method of checking chain wear as a guide to when to change the chain.

In a bike shop it seems to me that a chain checking tool can be a godsend; a customer may query a mechanic's opinion but if a tool gives a clear reading that says 'worn out' then there is little argument. Several LBSs near to me have used Park Tool CC1 for many years

Image

and when it hits red they recommend a new chain and cassette, the (IMHO sound) argument being that the chainrings are preserved this way. They will only change the chain alone if the wear is less than the 'red mark' and the customer both specifically requests it and understands that it may not work. This is in good part a reflection of the needs of their customers; they are usually more interested in the bike leaving the shop 100% definitely working rather than saving a few quid and taking a small chance that there might be skipping. Maybe one in a hundred customers has a new chain only and this only works some of the time.

The Park Tool CC1 has the major advantage that it appears to give a very clear indication of 'good vs 'no good'; a customer can (unless colour-blind I suppose) see that very easily and will usually accept the verdict. Of course it isn't as simple as that really, but it does make for an easy basis to make a decision on. In point of fact it seems to me that most modern chains are lucky if they go more than ~1500 miles before going near or into the 'red' on that tool. AFAICT the number of times a chain alone has been changed (when the gauge reads at or very near the red) and there has been no skipping is really tiny, and then it has been folk who pedal smoothly rather than aggressively.

Out of interest I have taken numerous discarded chains and cassettes and examined them; individual riders may wear their cassettes and chains similarly each time but there is almost no absolute relationship between the elongation of the chain and the state of the worst sprocket.

cheers
-------------------------------------------------------------------------------
~~~~~~~~~~~~~~~~~~~~~~Brucey~~~~~~~~~~~~~~~~~~~~~~~~

User avatar
Gattonero
Posts: 3540
Joined: 31 Jan 2016, 1:35pm
Location: London
Contact:

Re: Simple guide to worn rear cassette

Postby Gattonero » 1 Jun 2019, 10:46am

There is the rub: before the method, is the type of wear the user will get, the most important thing to consider. For this fact, one needs to try several times with different grades of wear of chain and cassette, and assuming this is done with a similar bike setup and on similar journeys (i.e. commuting is very different from a long sunday ride).

And if anything, the Park Tool CC1 has to be the least tool to use: apart from taking a measurement of only 7 pins (IIRC) vs the 11 pins used by the CC2.2, the reading dial moves on a bolt that is bound to have a noticeable play, which in turn can give a more pronounced error of the effective wear.
I have compared hundreds of times the CC1 with the CC2.2 and many times a chain that would be in the "safe zone" with the former, turned out to be already in the 0.5% wear mark of the latter, hence not safe to be replaced alone in most cases.

On the other hand, the home mechanic has often a good amount of time to double-check a given chain on a given cassette, and make a comparison (and take note, it helps) if/when a sprocket skips.
Of course, it all goes down the drain if a new setup is used, i.e. if moving from 10 to 11 speed. Ought to be one of the reasons why the bikes I use most are either single-speed or have 10sp Campagnolo, amd my Mtb's are still 2/3x9 :lol:
It is by riding a bicycle that you learn the contours of a country best,
since you have to sweat up the hills and coast down them.
Thus you remember them as they actually are...

User avatar
CJ
Posts: 2981
Joined: 15 Jan 2007, 9:55pm

Re: Simple guide to worn rear cassette

Postby CJ » 11 Jun 2019, 7:11pm

Gattonero wrote:if anything, the Park Tool CC1 has to be the least tool to use: apart from taking a measurement of only 7 pins (IIRC) vs the 11 pins used by the CC2.2, the reading dial moves on a bolt that is bound to have a noticeable play, which in turn can give a more pronounced error of the effective wear.
I have compared hundreds of times the CC1 with the CC2.2 and many times a chain that would be in the "safe zone" with the former, turned out to be already in the 0.5% wear mark of the latter, hence not safe to be replaced alone in most cases.

I agree, a short measuring tool is bad and I'd also suspect moving parts of introducing further error, but given that it's measuring inside the links, those errors should all trend towards over-reading, failing good chains rather than passing bad ones, so I'm puzzled that your better tool fails chains that the bad one passes. I guess the CC1 is just completely up the creek! But whatever, Park don't make that tool anymore. They seem to have replaced it with the CC2, which is longer (which is better) but still has a moving dial, although differently arranged so perhaps less wobble-prone.

As for CC2.2, I'm afraid I can't find it anywhere. Do you mean CC3.2? I've got the old version of that gauge, the CC3, and that measures 10 links/pins, so maybe that's what you've got too. Its replacement (CC3.2) measures 14 links/pins: even better!

None of these exactly correspond with your odd numbers of pins however. Are you sure you've got that right? Remember that the end of the tool engaging an inner link merely presses the roller onto its bush and does not have any effect on the corresponding pin/bush joint, so that one doesn't count. All of the tools I've checked in the flesh and on the interweb measure an EVEN number of links and instances of pin wear. As pointed out by Jobst Brandt in his interesting page on this subject, since pin wear affects only the outer links (across an inner link the rollers may get sloppy, but stay the same distance apart when pushed in the same direction), for consistent results in chain measurement this must be over an even number of links.

And here's a thing: I've been making some measurements of bits of chain (in response to someone else's suggestion that roller slop is somehow equivalent to pin wear - it isn't) and found that whilst dismantled inner chain links, pin-hole to pin-hole, are as exactly as I can measure, half an inch (12.7mm) long, outer links are slightly shorter than that: by an amount equal to the difference in diameter between the pins and those holes. The difference was 0.05mm in this case, which means each pin can shift 0.025mm off-centre, so two inner links can pull apart by 0.05mm across the outer link that joins them, which measures 12.65mm across pin centres. Thus the new chain is just right when you pull on it. Neat eh? I also found that the rollers, even on this brand new chain of good quality (Shimano HG73) were almost ten times more sloppy than the pins: being able to shift to and fro on their bushes by an average of 0.4mm total movement (i.e. can shift off-centre by 0.2mm)!

Two questions I was tying to answer with my measurements is how much difference roller slop actually makes to the use of a chain gauge, and do Park Tool make any allowance for it when they design such a gauge. As already noted, the Park CC3 measures 10 links which should measure exactly 127mm, pin centre to pin centre. But the gauge fits between rollers so subtract one roller diameter (7.67mm) = 119.33mm. The 1% side of the Park Tool measures 120.6mm, which is 1.27mm longer than that, which very neatly is 1% of 127mm. QED! However: when you fit this chain into the tool it will push those rollers apart some 0.4mm, which is 0.3% of 127, so the chain will fail this side of the tool when only 0.7% elongated. And that's assuming the rollers don't get any more sloppy than when the chain was new.

So the answers are that roller slop makes a substantial difference to the measurement, that Park Tool doesn't allow for at all.
Chris Juden
One lady owner, never raced or jumped.

Brucey
Posts: 33816
Joined: 4 Jan 2012, 6:25pm

Re: Simple guide to worn rear cassette

Postby Brucey » 11 Jun 2019, 8:34pm

IIRC the rollers on any one inner half-link (IHL) don't quite stay the same distance apart when pushed in the same direction; the reason for this is that the half bushings which support the rollers also wear unevenly. The extent to which this occurs seems to vary with the chain and the usage conditions. I think it is most likely that IHL rollers stay about the same distance apart if the chainring and (single) sprocket both have even tooth counts, because this will allow every other tooth to wear as the pin joints wear. However in a derailleur system or with odd tooth counts, the chain can sit any way it feels like on the sprocket/chainring teeth and there is no way for the teeth to wear that makes it favourable for IHL rollers to remain the same distance apart; should this happen the rollers are immediately unevenly loaded so as to promote wear in the half-bushings.

Another factor that is seldom discussed is that once the chain rollers are worn (and to some extent on new chains with sloppy rollers) it is clear that the chain links articulate around the pin centres, but the pin centres are no longer coincident with the roller centres. This is (to my mind) guaranteed to make a mess of the relationship between the actual tooth geometry and the involute curve described by the pin centre of each half link as it engages and disengages with the sprocket. That this is associated with bad (hooking/ramping) wear of the sprocket teeth is clear enough, but for the above reason are sloppy rollers (even in chains of the correct pitch) also associated with faster tooth wear too?

cheers
-------------------------------------------------------------------------------
~~~~~~~~~~~~~~~~~~~~~~Brucey~~~~~~~~~~~~~~~~~~~~~~~~

User avatar
tykeboy2003
Posts: 1026
Joined: 19 Jul 2010, 2:51pm
Location: Swadlincote, South Derbyshire

Re: Simple guide to worn rear cassette

Postby tykeboy2003 » 12 Jun 2019, 7:07am

Sweep wrote:Am sure that this must have been covered before, at least with folks' individual pics of cassettes they are concerned about.

But for simplicity can some kind soul/s point me at a simple online guide with good pics that tells you how to identify a cassette that is worn and needs changing?

Am pretty up to speed on chains and front rings I think but have never been to sure about what to look for on cassettes.

Being a meanie, would hate to think that I was binning them too soon.


You don't need any pics, just put a new chain on and if it skips the cassette is too badly worn, if not it's ok.

User avatar
Sweep
Posts: 5012
Joined: 20 Oct 2011, 4:57pm
Location: London

Re: Simple guide to worn rear cassette

Postby Sweep » 12 Jun 2019, 9:53am

tykeboy2003 wrote:
Sweep wrote:Am sure that this must have been covered before, at least with folks' individual pics of cassettes they are concerned about.

But for simplicity can some kind soul/s point me at a simple online guide with good pics that tells you how to identify a cassette that is worn and needs changing?

Am pretty up to speed on chains and front rings I think but have never been to sure about what to look for on cassettes.

Being a meanie, would hate to think that I was binning them too soon.


You don't need any pics, just put a new chain on and if it skips the cassette is too badly worn, if not it's ok.


Interesting and simple method.
Anyone any views on this?
Sweep