Chain wear, just reassure me.

[Samuel D]

I suppose the point of a dedicated chain-wear indicator is that it gives a binary good-or-bad result that needs no interpretation. When measuring with a ruler you always have to be sure that the chain is taught (straight) and that the zero mark on the ruler hasn’t slipped by the time you read the 12-inch end. Some people – mainly those who grew up without operating any kind of mechanical toys or machinery – find this more difficult than others. You have to clamp the ruler to prevent slippage, stabilise the hand against the chainstay while taking the measurement, consider parallax error, etc. These skills do not come naturally to some cyclists.

Even though I have a couple of chain-wear indicators, I end up using a ruler to be sure. With hindsight, I wouldn’t have bought the indicators.

[De Sisti]

Yes, but I'm sure the other chain measuring tools still give a good indication when a chain needs replacing.

well, sort of... if you only knew for sure what the tool read when the chain was new....

With my cynical head on, I'd perhaps suggest that the Park tool chain gauge is ideal for bike shops because I have yet to see a chain that had been used for more than five minutes (or sometimes at all) that wasn't "in need of replacement sir"....

cheers

They still give an indication as to whether the chain is worn; so people will still rely on them.

[mercalia]

I have a 39inch steel rule that I use to measure mine.

I understand that Mr Shimano had a metric chain once, but it dies a death. All chains are imperial now.

even in the EU I bet the French love that

[mattsccm]

1/16th over. Smack on. (Well as far as my eyes can tell) My steel rule is marked in 32nds and both were clamped together. That's going in the "it'll do on a worn cassette box and a new one going on.

[Brucey]

1/16th over. Smack on. (Well as far as my eyes can tell) My steel rule is marked in 32nds and both were clamped together. That's going in the "it'll do on a worn cassette box and a new one going on.

so was that measured over 12", second time around?

If so that would put the actual chain wear at 0.5%, somewhat less than the chain checker suggested? Normally it is OK to fit a new chain on a cassette that has seen a 0.5% worn chain; the exception to this would be if you very much favour one of the smaller sprockets, which may be more worn than the others.

....They still give an indication as to whether the chain is worn; so people will still rely on them.

I guess so, but there is no need to buy a chain checker that is likely to give a false reading; the shimano chain checker tools don't and they are readily available. I mean, you wouldn't deliberately buy a chain that wears out prematurely, so why buy a tool that tells you the same thing...?

cheers

[mattsccm]

Tells me that the checker is pessimistic. New chain sits perfectly. The cassette is near new actually and the chain in question was fitted/measured to see if it worked. Seemed fine (did a couple of miles) but a new one makes sense.

I have a pile of near new cassettes that are only of use on this bike and I run the chains until knackered which gives me about 3 to a cassette which is still working. Last one did about 10,000 miles through 2 complete years as an almost pure gravel bike. Caked in mud most of the time. Neither chain or cassette were new before fitted. Shifting wasn't an issue but I had reached the point where I was worried to stand up on a steep hill.

[Gattonero]

I'm not sure how people does use the chain checkers.
For a start, is a "go/no go" not "push it in until it fits".

Also, "what" is people going to do with the numbers given by a gauge or a vernier? For all I've seen written around, including the "untouchable" opinions, no one seems to offer at least a 3-digit experience of "what" exactly a 0.5% or 0.75% or 1% means in real life.

Gathering evidence to destroy this myth that chain-checkers are not useful, and only a few people are right while everyone else -including chain manufacturers- are wrong.

Stay tuned

[Brucey]

Tells me that the checker is pessimistic....

it could be that with a new chain and still be wrong on a more worn one. The whole point is that

a) there is no guarantee that the scale on the checker is linear and
b) once one roller's worth of free play is included in the 'wear measurement' all bets are off; chains are manufactured with differing amounts of free play in the roller bushings.

A worn chain with tight roller bushings will measure the same as an unworn chain with slack roller bushings, using such chain checkers....

So it isn't as simple as 'right or wrong' or 'optimistic vs pessemistic'; the vast majority of chain checkers include something that (if it isn't completely irrelevant) isn't that closely related to chain wear per se. At best you might be able to make some kind of comparison between two chains of the same type, if you could be certain that they were made to the same tolerance.

Note that when chains are used in other industrial applications, the methods recommended to assess wear and the tools supplied to measure wear do not include any allowance for roller bushing wear; they use pin centre to pin centre measurements, as recommended by the Pardo link and as achieved by a small proportion of bicycle chain checkers and anyone with a ruler....

cheers

[Gattonero]

OTOH, apparently is the rollers that would work directly on the sprocket
As much as pin wear is important, so is the wear on the rollers.
Insisting on the pin wear only makes no sense. In fact, that guy does not include the wear that modern chains do get while working on an angle, he does assume that the chain does always work straight.

[Brucey]

there are a lot of ifs buts and maybes here. One has to question the motivation to use a chain checker anyway. I think there are really only two reasons;

a) you want to change the chain before the cassette wears too much and/or
b) you want to change the chain (and cassette) before the chainrings wear too much.

roller wear plays a small role in the former and (relative to pin wear) almost no role in the latter, provided chainrings of reasonable size are used.

If it were otherwise then rates of sprocket/chainring wear would vary hugely with chain type, even when said chains are 'unworn', because there is significant variation in roller bushing tolerance in new/part worn chains.

When all is said and done the best chain measurement becomes an approximation to the wear state of the sprockets; a 0.5% worn chain that is nearly always used on one sprocket may well have worn that sprocket (especially if it is a small one) so that a new chain won't run on it under load. The usual assumption is that the sprockets share the wear in some reasonable fashion, but this isn't always the case....

If such measurements are anyway an approximation, you might ask why do folk get antsy about the numbers that chain checkers produce? Well IMHO it is in good part because they are often pretending to give something they cannot; a reading of the accurate percentage of chain wear. If they simply gave a 'good' vs 'bad' reading that would be better, even if it wouldn't necessarily be accurate/meaningful with every type of chain for the reasons discussed.

cheers

[Brucey]

I have a little box with many, many offcuts of different new chains in it. I have kept them "because they might come in handy one day...."... they don't, obviously....

But maybe I found a use for them....

Using the KISS principle (Keep It Simple, Stupid) I have carried out some very simple measurements that can be reproduced by anyone who also has a half-decent digital Vernier caliper.

I just measured the outside measurement across two neighbouring rollers. I chose this measurement because it (or something very much like it) is included in the 'measurements' that are taken using 'bad' chain checkers. Depending on how the chain is constructed, variations in this measurement can also appear as variations in the apparent chain wear when using 'bad' checkers.

The measurement is sensitive to

1) the pin to pin spacing in that link
2) the roller diameter
3) the slack in the roller bushings

Because the jaws on Vernier calipers won't go inside every chain link, it is necessary to articulate the chain so that the jaws can access the rollers. This makes the measurement less meaningful on half the links in a worn chain, because roller bushing wear is directional with respect to the inner side plates.

If there are no variations in this measurement it would mean that 'bad' chain checkers are actually accurate after all. If there are variations in this measurement (within and/or between chains) then it means that 'bad' chain checkers are indeed liable to be inaccurate.

In total my measurements varied (on new chains) between 19.72 and 20.27mm, i.e. a variation of 0.55mm. In any single chain the variation was typically around 0.2mm. In a few cases the measurement was noticeably different in every other link, e.g. by around 0.1mm or so.

Typical chain checkers measure over a length of about 150mm. The kind of change in that length which ought to indicate (say) a 0.5% worn chain would be around 0.75mm. However, if the checker measures in such a way as the three factors above can change the measurement, the indication is that the starting measurement can vary by up to around 0.55mm, which (in round numbers) will lead to a possible error of about 0.37% in terms of chain wear.

Maybe my measurements are unrepresentative for some reason; but if they are not, the conclusion is very clear; 'bad' chain checkers (as described in the Pardo link) are bad. To make any sense of the readings you would have to know what the same tool read when the chain was new.

But there is more; it is also clear that a small error is likely to be generated if the chain checker measures an odd number of half-links on some chains.

Finally it seems that many new chains don't have a perfectly uniform pitch when they are new. I think that it isn't at all uncommon to find that a chain that is about 100 miles old runs more quietly than one which is brand new; I think that the reason for this is that any irregularities in chain pitch (either random or systematic) are likely to be compensated for by uneven (from one link to another) roller bushing wear, caused by the uneven loading.

In well-used SRAM chains it isn't at all uncommon to find that there is about 0.25mm wear on the inside of each roller, when each pin bushing might only have worn by 0.06mm or so. This may well be deliberate on the part of the chain manufacturer, in that if the roller bushings wear fast enough, the chain will tend to wear itself into more uniform pitch than is otherwise likely to be achieved in mass production.

cheers

[MikeDee]

Here's an interesting article on the accuracy of various chain checkers http://pardo.net/bike/pic/fail-004/000.html

I have the Speedtech one. (Pardo is wrong about its accuracy because you line up the holes with the chain pins.)

[Samuel D]

(Pardo is wrong about its accuracy because you line up the holes with the chain pins.)

Why does that make him wrong? If the size of the roller on which the hook rests varies, the sighting holes move along the chain. Therefore the tool gives different results for rollers of different outer diameter.

Granted, you’d think there wouldn’t be much difference in roller size on bicycle chains. But when you’re measuring small differences to begin with, it may matter in some cases.



EDIT: I just realised there’s a hole on the left too, in your photo. Can that also be lined up with a pin? If the tool simply slides along the chain, what’s its advantage over a simple ruler? I wasn’t able to find a good explanation of this tool on the web (or even a photo of the hook).

[MikeDee]

(Pardo is wrong about its accuracy because you line up the holes with the chain pins.)

Why does that make him wrong? If the size of the roller on which the hook rests varies, the sighting holes move along the chain. Therefore the tool gives different results for rollers of different outer diameter.

Granted, you’d think there wouldn’t be much difference in roller size on bicycle chains. But when you’re measuring small differences to begin with, it may matter in some cases.



EDIT: I just realised there’s a hole on the left too, in your photo. Can that also be lined up with a pin? If it simply slides along the chain, what’s the advantage of this tool over a simple ruler? I wasn’t able to find a good explanation of this tool on the web (or even a photo of the hook).

Yes, you line up the left hole with a pin. That's why Pardo is wrong about that tool. BTW, that tool is no longer sold. It's easier to use this tool than a ruler.

[Samuel D]

What, if anything, holds the tool centred over the left-most pin while you sight through the holes at the other end?