I've seen that sprockets were sold separately (or by "groups/spiders"). Given the small sprockets wear faster than the big ones, it could make sense to replace them at higher interval than than the large ones, rather than replacing the whole cassette "every 3 chains".

Is there guidance to define these intervals? Is there some prerequisite to have this kind of maintenance strategy (like more frequent chain replacement)? Or is the idea bad to to start with?

Also, are there significant differences between sprockets of different brands/kinds? For instance, road and MTBs are using the same chains (maybe not true anymore for 12-speed though). So are there differences between road and MTB sprockets with the same "speed count" and number of teeth?

EDIT: the assumption of 'wearing faster' is not linked to the time spent "in the sprocket", but to the force is transmitted by the chain is spread over a smaller number of teeth. It's an argument I've read to justify the use of softer alloys for the large sprockets, and also to replace the chainrings at lower intervals.

  • 1
    Personally I mostly use the bigger half of the cassette and never had problems with premature wear of the smaller sprockets.
    – Michael
    Jun 30 at 11:53
  • I was always told to just check the shape of the teeth. If they looked worn, they were worn. Estimating a replacement interval is impossible even with knowing the cleanliness/conditions you ride in, how much use each cog sees, what the cross-chaining is when you use them, etc. -- I don't think there's any scientific data available on cog use, even in controlled (and clean) laboratory conditions. I say you're worrying too much. Some of my bicycles have never even had their chains replaced, and the cogs don't look like they'll need changing any time soon.
    – jayded-bee
    Jun 30 at 13:51
  • Is it that the largest-tooth-count sprockets don't get used ? If so, maybe a smaller-range cassette could suit your riding, and spread the wear across more cogs in the range you ride most. Dropping from a 32 to a 28 tooth large cog could help a lot with reducing cassette wear in the 11-19 area
    – Criggie
    Jul 1 at 2:35
  • Also consider a 11 tooth will go around 3x as often as a 33 tooth would, so there is more wear from triple the mating and unmating friction.
    – Criggie
    Jul 1 at 2:36
  • 2
    @Criggie that's indeed another question. I prioritize range over having 'optimal ratios'. It's not because because I don't use often a sprocket that I don't appreciate having it when riding — or the other way around: having a wide range allows me to plan for rides that I wouldn't do otherwise. And I only have room for 1 non-utility bike, so I can't really afford to have 1 bike per riding style.
    – Renaud
    Jul 1 at 4:59

4 Answers 4


To gauge whether they need replacing, a good way is use a Rohloff HG-Check. There's not yet a 12-speed version, but here you could make that not matter because if you really wanted, you could take the replaceable cog off and mount it on the freehub body by itself, then take your measurement without the other cogs to get in the way. Nominally it's for Shimano only, but in practice it works well enough on anything. The Rohloff tool detects whether the effective tooth pitch that the chain sees under load has drifted away from new enough to matter.

Barring a special tool, do it by observing how the upper left corner of the cog teeth look. Compare it to your new one. Worn cogs usually begin to get noticeably rounded or folded in that area. I believe the physics are something along the lines that that area shouldn't get loaded when the pitch of the cog and the chain are still a match, but it begins to when they aren't. You can also look at how thick the teeth are viewed from the side. For practical purposes you could also just come to your own assessment based on how much you use your smallest cogs that you should replace them every X chains. If you have a big money cassette you're trying to preserve wear on, the marginal increase in cost of being aggressive about this and just doing it every chain or whatever is almost certainly worth it in a monetary sense.

First-position cogs are usually inset into the second-position cog by some arbitrary amount. Therefore they have to be the right match for each other for the spacing of the cassette to be correct.

The shift aids on the whole thing won't necessarily be "as-intended" by the manufacturer if any mixing and matching is done. However, spacing is the only thing that matters for it to work in a basic functionality sense. If the replacement cogs you can find result in the same spacing, which is mostly a function of their thickness, it will generally work even if the ramping and entry/exit point profiling aren't clocked as intended. Whether one likes the shifting that much less smooth and considers it a succcessful or reasonable thing is a seperate question.

Hypothetically it's possible for the tooth centerline to differ from the centerline of the cog such that spacing could be thrown off that way, which for example can be a problem when mixing and matching certain chainrings that have an offset to the teeth even on a flat chainring. I don't think there are any cogs like that, but if there were then that would throw off the above guideline that cog thickness determines spacing.


The sprockets have ramps and such that assist in gear changes. Each brand is different, likely significant difference within brands. Changing to a sproket that is not the exact same part number will almost certainly compromise shifting. Wear also occurs on the ramps, that enable smooth and precise gear changes. Eventually replacing sprockets will lead to a cassette that shifts poorly.

Smaller sprockets are often steel and larger one aluminum (depends on the cassette). Its night be a larger sproket is more worn because of this.

You state "Given the small sprockets wear faster than the big ones," which is not universally true. It largely depends what gears people are riding in. While a few people live in top gear (on the small sprockets), most ride on the larger sprockets around the middle of the cassette most of the time. The drive line is more efficient on larger sprockets so it better to gear a bike to use them for your typical ride.

From an economics perspective, while cassettes are not cheap, a mechanics labor is not either. Paying someone to replace sprockets would almost certainly be uneconomic compared to paying them to replace the entire cassette. Most of the time would be used working which sprockets to replace, then discussing it with the customer. It would also be high risk work in terms of brand reputation and customer satisfaction. For DIY these are not considerations (but still worth thinking about if you value your time ), and it might be something an individual would benefit from.

From a logistics perspective it becomes easier - fewer parts to manage inventory and shipping on. Some would argue getting people to throw away a partly used item and buy a new one regularly is better business.

Given all this, the reasons to do are worth considering -less waste and environment impact. Even if you send old cassettes to the metal recycler, there waste in manufacturing that is reduced. Cost saving would be hard to work out but might be there.

Without a quantitative way to work out if a cassette sprocket needs replacing how would you when to replace them? We have the same problem with the cassette, but we use the chain as a proxy. Perhaps this is the problem to solve - come up with a tool that quantifiably (like a chain gauge) tells us when to replace a sproket.

  • Thanks for the reply. Do you have then an idea about the reason why manufacturers offer this possibility if it doesn't make sense both from an economical point of view and technical — I would assume that a manufacturer knows very well that a repair shop considers that replacing the cassette is "safer"? I don't think of cases where it's possible to 'break' one sprocket.
    – Renaud
    Jun 30 at 21:36
  • 1
    @Renaud there's some people out there that will tell you a bike is simply unridable if it doesnt have a 16t sprocket. I suspect the sale of individual sprockets is for individuals (and perhaps pro teams) that want to 'build' cassettes with custom ratios
    – Andy P
    Jun 30 at 21:45
  • 1
    @AndyP from I've seen on bikester, it's really spare parts. They sell the for example the M8000 cassette in parts (the two spiders + the individual smaller sprockets). So it's would exclude a use to "customize" an existing cassette.
    – Renaud
    Jun 30 at 21:48
  • Having just looked through the Shimano MTB cassette lineup it seems you are correct, someone must be buying these as spare parts because there's no alternative options to 'customise' with
    – Andy P
    Jun 30 at 22:00
  • 1
    @Renaud You could rip a tooth off, but that’s equally likely to be a manufacturing defect as it is user error.
    – MaplePanda
    Jul 1 at 8:24

Or is the idea bad to to start with?

Definitely not a bad idea, but some difficulties with implementation. @mattnz mentions in his answer that we don't really have a good way to determine exactly when a sprocket is worn and needs replacing.

But lets imagine that we could. Lets say that I wear out my 19T sprocket first and replace it. Well chances are i'm probably going to wear out my 17T a month after that, my 21T a month after that, and my 15T a month after that. I worry that we'd end up creating a situation where we are in an endless loop of taking off cassettes and replacing sprockets rather than performing the task once.

One place where this idea definitely has merit is on a cassette used exclusively for ERG mode on an indoor trainer. In this situation you can ensure that all the wear is on a single sprocket.


If some sprockets are skipping when others still work well. Assuming the road is the same and terrain is flat most of the way, likely one sprocket that fits would see much more usage.

This started to happen for me as soon as after 400 km on my first E-bike where I was using mostly the smallest sprocket for everything save climbing. I was much more careful with my second bicycle, using the second smallest sprocket for gravel and even larger sprockets where it looked reasonable, but after 1000 km the fastest sprocket seems already "tired" - it can sustain the ride but skips when accelerating. Previously I could not believe and spent lots of time trying to adjust the cable tension instead, but no position worked and all problems were gone after replacing the sprocket. Maybe I need a bigger chainring at front to use the bigger sprockets for the same speed, did not investigate this so far.

Big sprockets may be riveted together so difficult to replace. The first two are often not riveted so easy to do if you find the matching parts.

  • 1
    Your experience somehow confirms what was discussed in the chat: e-commuters are riding with much lower cadences than sporty riders. They would actually benefit the most from replacing only the smallest sprockets.
    – Renaud
    Jul 1 at 8:05
  • 1
    There is the cadence indicator on the display. My typical cadence was about 55.
    – nightrider
    Jul 1 at 9:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.