Chain-Suck :
DETAILED investigations
see Main Article : Chain-Suck in a NUTSHELL

by Jonathan Levy

What is Chain-Suck ?   What causes it ?   How can it be prevented ?

How to check whether you might get it, but before that next race !!

You really want to suffer from Chain-Suck ?

Use a NEW chain with your WORN chain-ring !!
WHY ? Read on ....

From a morass of theories, counter-theories, and myths where I started on this problem, I believe that a logical mechanism, and clear solution, has been demonstrated for chain-suck.

There does not seem to be a general understanding of the issues affecting chain-suck in the general biking community. This article is intended to assist with the spread of that understanding by clear and believable demonstration of the issues.

One month after purchasing a good quality mountain bike, I experienced my first bike-and-bone jolting encounter with chain-suck. The problem recurred often during muddy or wet conditions. Some time later, I started investigating it by actual field testing using my bike, and by looking for information. I could find no adequately detailed information, despite an extensive search. It was clear that if I wanted to resolve the issue to my satisfaction, I would have to investigate thoroughly. I set up a test programme.

During the testing, an article on the Internet about wear and slippage of chains and sprockets, caught my eye. A short paragraph in it referred to chain-suck with implication of a mechanism, but that aspect was not explored further in the article. This helped confirm some of the ideas which I was investigating. The test programme was extended to include alternative bike parts, and conditions were set up which I felt might be relevant to chain-suck. The interactions between the chain and chain-rings arising from their geometry and direction of forces, was also evaluated carefully.

For a SUMMARY of the outcome of this investigation, see ...

Chain-Suck : in a NUTSHELL


Chain-suck is a failure of the chain to release from the bottom of a front chain-ring of a bike and travel smoothly back to the rear derailleur and sprocket in the normal manner. Instead the chain is carried up and around the full circumference of the front chain-ring to wind back onto itself. It brings the pedalling action to a sudden binding stop. Because it usually occurs on steep uphill trails or when otherwise pedalling hard, the bike also comes to a sudden halt and the rider, not expecting the situation, falls off.

The phenomenon usually occurs when the chain and sprockets have become muddy or wet and gritty, when the small front chain-ring ("granny-ring") is being used, and when substantial pedal effort is being exerted on steep hills or other circumstances. The problem can be continuous and tenacious, recurring whenever pedal force is re-applied, and can even recur with modest pedal force on more gradual hills. It may require thorough cleaning and lubrication of the chain and sprockets, before it goes away, thus disrupting or terminating a ride. Frequently it is necessary to walk up the hills when it occurs. Moreover, the chain is often dragged up and wedged forcibly between the chain-rings and chain-stay, which can cause serious damage to chain-stays of softer material such as aluminium or carbon fibre.

If mud initiated chain-suck occurs during a race, it is likely to scupper the rider's chances in that race.

There are two main forms of chain suck :

  • 1-ring suck : the chain hangs-up on a single ring (usually the "granny")

  • 2-ring suck : the chain hangs-up across two rings when changing gears (usually from the middle ring to the "granny")

  • Other circumstances which might appear to be chain-suck, but are NOT.


In Cape Town, where I live, we have a mix of stoney/rocky terrain with muddy areas during and after our winter rains. The type of riding which I do includes a lot of steep to very steep uphill slopes (say from 1:7 to 1:4) for say 30% of total riding distance, which is about 300 km per month

Chain-suck first occurred to my bike about 1 month after purchasing it new, when I encountered the first wet weather conditions. It occurred almost infallibly every time I rode in wet or muddy conditions thereafter, but never in dry conditions. The problem was continuous (ie not specifically brought on by changing gears) and started soon after a thin mud slurry covered the chain and sprockets. It occurred only when using the "granny-ring" (ie smallest front sprocket of 22 teeth), and did not depend on which of the rear sprockets I was using. It would usually first present itself when using reasonable pedal effort on hills. It was so tenacious that on each ride on which it started, it would recur whenever pedal effort was needed on subsequent hills, even of moderate steepness. Even after washing the chain and sprockets in streams, it would not go away readily. I had to dismount and walk up the hills. The problem would ease off only after washing the chain and riding a long distance without mud.

Lubrication was not the issue, as I lubricated the chain each time before setting out on a ride, in an attempt to prevent the problem, and tried a wide range of lubricant types. In any event, I frequently clean and lubricate the drive-train including the chain, even during dry riding conditions. "Chain-stretch" was not the issue, as the chain was quite new with negligible measured wear.

The bike, a Trek Y22, came fitted with a Shimano 4-Arm front crank-set (aluminium chain-rings and die-pressed powder-alloy granny) and a Shimano HG-70 8-speed rear cluster (steel). The chain was a Sachs PC-41.

The shop from which I bought the bike and the local Trek agent were both very helpful in assisting me with the problem; they supplied and fitted various alternative equipment types. The initial trials were with 2 alternative chain types, a Shimano HG, and a Shimano IG (which is very slightly wider and with more chamfers). This made no difference. The next trial was to use a slightly narrower crank spindle (Bottom Bracket or BB) to improve the chain-line; again no solution to the chain-suck. Removal of prominent burrs which had formed on the sides of the "granny-ring" teeth, also made no difference.

About 3 months after purchasing the bike, I visited Knysna, a popular area for MTB's about 500 km from Cape Town, and which often has muddy conditions at any time of the year; severe chain-suck occurred. An experienced cross-country rider there examined my bike and advised me to replace the aluminium "granny-ring" with a more robust steel one as he considered its teeth worn to a "hooked" shape, and that this was the cause of the problem. Consequently my bike shop fitted a Shimano STX steel sprocket, but I did only one ride in somewhat muddy conditions (no chain-suck problems occurred); and the dry season had started.

By the onset of the next wet season about 5 months later, and about 8 months after purchasing the bike, while still using the steel STX "granny-ring", I had immediate and severe chain-suck problems. By this time it was becoming too time-consuming to take my bike to the shop each time to try alternatives. The shop lent me the necessary equipment for me to test. Again I tried a new Sachs chain, as well as Shimano HG and IG chains; again they made no difference.

I then installed onto my bike, a new low-end (Shimano Acera) crankset which I had purchased for my wife's bike. I did everything I could to induce chain-suck by riding through mud puddles and even stopping a number of times to pack mud into the complete chain. I finally got chain-suck to occur, but it was "reluctant". In addition to the mud riding, it took 5 deliberate hand-packings of mud to start the chain-suck, and even then it was not tenacious. I could still pedal up steep hills, with a slight reverse pedalling to unlock the chain each time it "sucked". It then remained unlocked for some distance (say 100m) before recurring, and then gradually disappeared. Was this the solution, a cheap crank-set ??

The above experience of chain-suck was primarily the 1-ring variety.

The findings of my investigation and test programme are set out in this article. The article is mainly about 1-ring suck (see "What is Chain-suck" in section-1, above), but most of the issues identified also have bearing on 2-ring suck. Subsequent to writing an initial article, I was able to observe and evaluate 2-ring suck which affected biking colleagues, acquaintances, and respondents to the article. Some respondents provided their own valuable insights. My insights to 1-ring suck, and extension of these to understanding further issues relevant to 2-ring suck, enabled me to assist cyclists in alleviating this related problem. Although I did not undertake a formal testing programme for 2-ring suck, the understanding brought about by finding out what did, and what didn't work for these cyclists, amounted to an informal testing schedule.


I presented my findings to my bike shop, who lent me some alternative crank-sets to try to resolve the problem. Additional tests were included to check my growing belief that wear of the "granny-ring" and chain "stretch" were also key influences. The effect of lubrication and its absence was also included. The following parameters were considered in the tests:

Each of 9 set-ups were then tested on my by now familiar "chain-suck-alley", to try covering the above parameters as best as possible. All the set-ups had 22-tooth "granny-rings". A newish Shimano IG chain (0.1% "stretch) was used unless otherwise stated.

The 9 TEST set-ups were as follows:

  1. Original Shimano LX crank-set with STX RC 4-arm steel "granny"
    .... (Used, 6-months old)

    a. New HG chain : 0.0% "stretch"
    b. Newish IG chain : 0.1% "stretch"
    c. Oldish HG chain : 0.4% "stretch"

  2. Original Shimano LX crank-set with STX RC 4-arm steel "granny"
    .... but FLIPPED (& Flipped tooth faces Unused, ring 6-months old)

    a. New HG chain : 0.0% "stretch"
    b. Newish IG chain : 0.1% "stretch"
    c. Oldish HG chain : 0.4% "stretch"

  3. Cheap Shimano Acera crank-set with Acera riveted steel front rings
    .... (Fairly New, 2-weeks old)

  4. Original Shimano LX crank-set, but with the ICON 4-arm steel "granny"
    .... (New, unused)

  5. Used Shimano XT crank-set with SUGINO 5-arm unused steel "granny"
    .... (New, unused)

A) For each set-up, the chain was cleaned and lubricated before setting off. I then rode through mud, and stopped up to 5 times to pack the chain with new mud after washing off the old mud each time with muddy water. This procedure was used both to standardise the test conditions as far as possible, and to simulate a longer ride where there is greater time for any lubrication to be removed by mud, water puddles and rain. Each test lasted about 1 hour, and the same route was followed. On each occasion both the "granny" and middle rings were tested for chain-suck in combination with all reasonable rear sprockets. If severe chain-suck occurred, no further mud packing was done.

B) Each test was repeated after cleaning/de-greasing the drive-train, but without re-lubrication.

Comments from passing bikers who saw me smearing mud onto the chain varied from funny to rude !!

The following information came from examination of the "granny-rings" of each test set-up after all the testing had been completed. It was noted in case it had bearing on the investigation:

a) The used Shimano STX ring had a moderate (noticeable) degree of wear on the pressure faces of its teeth (steepening of teeth, rather than "hooking") with no burring on the sides, and no noticeable wear on the trailing tooth faces. The pressure point of the chain-rollers on the teeth was raised only slightly above the unworn position. All the other rings (new, except for the tests) had very little wear on the tooth pressure faces, but had already developed obvious burrs on the sides of their teeth.

b) The basic tooth profile of both the Shimano sprockets (used STX, and new ACERA) appeared to be fairly symmetrical and had fairly deep rounded cups between the teeth

c) The basic tooth profile of the new ICON sprocket was symmetrical and also had fairly deep rounded cups between the teeth, thus seeming to be similar to Shimano

d) The basic tooth profile of the the new SUGINO sprocket was also symmetrical; however, the cups between its teeth were slightly shallower, and their base was flatter and less rounded; this sprocket was noisier when pedalling.


The results of the TESTS were as follows (refer to the above numbering):

1a,b Severe, continuous and tenacious chain-suck within a short while of the chain first becoming covered in a thin layer of mud slurry; presence of lubrication might have delayed the onset of chain-suck to a degree, but did not seem to have a significant influence. Thereafter, chain-suck continued even on mild slopes when exerting moderate pedal force.

1c Occasional chain-suck and not tenacious, gradually disappeared; presence of lubrication might have delayed the onset of chain-suck to a degree, but did not seem to have a significant influence. Chain-suck only occurred on steep hills with high pedal force.

2a,b,c No hint of chain-suck; lubrication or the lack thereof did not seem to have any influence.

3. Occasional chain-suck and not tenacious, first appeared only after 5th hand-pack of mud, gradually disappeared; lubrication or the lack thereof did not seem to have much effect. Chain-suck only occurred on steep hills with high pedal force.

4. Incipient chain-suck, not problematic, and which did not affect the ability to ride; lubrication or the lack thereof did not seem to have an influence. No hint of chain-suck on moderate hills.

5. No hint of chain-suck; lubrication or the lack thereof did not seem to have any influence.

Key tests (1a,c 2a,c 4) were repeated a further two times; consistency of results was confirmed.

Chain-suck was independent of which rear sprocket was in use. It did not occur with the middle ring. None of the set-ups gave any indication of chain-suck when mud was not present on the drive-train, despite high pedal forces on steep hills and even when lubrication was absent. It is noted that only one of the "granny-rings" used had wear, and I judged that wear as low-to-moderate rather than severe.

No chain "jumping" over teeth of either the "granny-ring" or rear sprockets, occurred for any set-ups.

While the STX ring was the only item already worn, would the new ACERA, ICON and SUGINO rings also wear at a similar rate and give rise to chain-suck ? The obvious burring on the sides of their teeth, and a start of pressure face indentations lead me to believe that the particular rings were not made of especially wear resistant steel. From my understanding later in this article, I believe that they would probably fare worse than the STX ring (because they are not made of as tough a steel). The SUGINO ring might be an exception because its different tooth profile might (or might not) be less sensitive to chain-suck once worn; that, however, would need to be proven.

It must be telling that one individual "granny-ring" gave :
  • SEVERE chain-suck with limited exposure to mud, using its WORN tooth-faces and various NEW "unstretched" chains

  • MODERATE chain-suck with full mud exposure, using its WORN tooth-faces and a MEDIUM "stretched" chains

  • NO HINT of chain-suck, using its UN-WORN tooth-faces with OLD "stretched" or NEW "unstretched" chains, and despite throwing any amount of mud/grit at it


Although not included in these tests, my original chain-suck problem with the used and worn LX "granny-ring" was at least as severe, and probably more severe than with the used and moderately worn steel STX "granny-ring" (TESTs 1a,b,c). The original LX "granny-ring" only had about 800 km usage. Nonetheless, its die-pressed powder-alloy teeth were considerably more severely worn and "hooked" than my current steel STX "granny-ring" which has about 2300 km usage. Also, the pressure point of the chain-rollers on the teeth of the LX ring was raised considerably above the unworn position. This ring seems to be aluminium at first glance, but is actually slightly magnetic and heavier than aluminium - an unconfirmed source reported it as being a mix of tin and nickel.

Removal of substantial burrs from the sides of the worn LX "granny-ring" made no difference to alleviating chain-suck. Conversely, obvious burrs already forming on the ICON and SUGINO "granny-rings" before their pressure-faces showed much wear, did not induce chain-suck.

A further point of interest is that I also tested my wife's bike (DiamondBack Topango with Shimano Olivia 7-speed rear cluster). This gave only occasional and non-tenacious chain-suck with its new Acera crank-set and new HG chain. However when I fitted my LX crank-set to it with the used STX steel "granny-ring" and same new HG chain, the chain-suck was as severe as on my Trek Y22. Thus, the situation is not alleviated by changing from an 8 to 7-speed rear cluster, nor to a rigid framed bike.

Reference is also made to a friend's experience with a hardtail bike that I witnessed on some very steep trails (some blacktop and some dirt). Chain-suck occurred in dry dusty conditions despite a complete absence of mud, but with his drive-train very dry and inadequately lubricated; his aluminium LX "granny-ring" had fairly severe wear. Lubrication alleviated the chain-suck in this situation for the duration of two longish rides.

Subsequent to completing this investigation initially, I was able to secure a large number of worn "granny-rings" from local bike shops to examine the nature of their wear. It seemed to me there were two main categories of materials, with relevance to wear characteristics :

  • Softer materials : softer steels, aluminium, and die-pressed powder-alloy tended to have more indented wear "pockets" in the pressure-faces of their teeth. Furthermore, these "pockets" were elevated along the length of the teeth causing the chain-rollers to pressurise the teeth higher up, with significant roller clearance above the roots of the teeth when under pressure from chain tension. These softer teeth also tended to be somewhat thickened laterally and with prominent burrs on the sides.

    The rate at which tooth "wear" can occur can be surprising. The same friend whose severely worn "granny-ring" sucked badly in dry conditions replaced the ring with a new one of the same type (powder-alloy LX). On the first ride thereafter of about 20 km length and 2 hours duration (though admittedly very steep), moderate-to-severe "pockets" had formed on the pressure-faces of the teeth which had also thickened to some extent and developed side-burrs. This was a result of plastic metal failure rather than true wear. The teeth had been quite easily overloaded beyond the strength of the aluminium alloy, well before any meaningful abrasive wear could occur, and before any chain "stretch" could occur to over-stress the top teeth. Thereafter, this new "granny-ring" sucked in muddy conditions.

  • Harder materials (typically stainless steels), tended not to have actual "pockets", but rather a steepening of the pressure-faces, and an absence of thickened teeth or side burrs. The chain-rollers did not ride as far up from the tooth roots when under pressure from chain tension. Titanium "granny-rings" were not available to me.

The significance of the above aspects is discussed later in this article.


Now, what are we to make of all this, and how do we resolve a chain-suck problem on a reliable long-term basis ? The investigation was limited by my time and resources. It includes reference to the tests, my general experience, published statements and comments (no comprehensive articles were found), and discussion with others who had chain-suck. The implications drawn below are based mainly on the tests and my own experiences. Findings derived from other input sources are noted.

i) The following issues don't affect chain-suck, or are lesser influences :

  • Chain-type
  • Chain-line
  • Whether 7 or 8-speed rear clusters are used
  • Which of the rear sprockets are being used, provided pedal effort/chain-load is sufficient
  • Use of the middle-ring; chain-suck only occurred with use of the "granny"
  • Lubrication lack is not a primary cause of chain-suck. Good lubrication might alleviate the situation for a short period before it is stripped out by mud and rain. However, when there is no mud, lubrication will usually alleviate chain-suck which can occur in the dry for badly worn "grannies".

ii) Type and original tooth profile of the "granny-ring" sprocket may contribute to chain-suck, with some profiles giving rise to chain-suck more readily than others, or wearing more readily to a problematic tooth profile.

iii) Tooth pressure-face wear is the primary cause of chain-suck. This opinion has been arrived at both from the above tests and from my general experience with the STX and LX "granny-rings". I have not yet been able to determine whether the ICON and SUGINO "granny-rings" will also start showing "chain-suck" when they wear. Wear takes time (later: the ICON did give chain-suck). This finding about tooth wear is also confirmed by published comment made by others.

iv) Chain-ring material and wear resistance significantly affect the potential for chain-suck to occur. From technical articles, aluminium (even alloyed, heat-treated and with hardened surfaces) wears and deforms much worse than hard steel (there may be special aluminium alloys with acceptable strength). Titanium is far better than aluminium; it is very strong against deforming, and has good wear resistance. Tough stainless steels are also very strong and have the best wear resistance.

v) Accuracy in complying with a "correct" tooth profile gives significant benefit against chain-suck. Accurately machined teeth do not wear or deform as fast as die-stamped teeth, for a given tooth material, because the chain-rollers are correctly seated with pressure applied as close as possible to the root of the teeth, and the load is spread over a larger tooth area. I have observed that new, die-stamped steel teeth (eg Shimano STX and others), often have only half the width of the tooth's driving face in contact with the chain rollers, with the remainder of the width having been torn away at an incline by the stamping process. Later, wear extends the roller contact across the full width of the tooth which results in a worn "pocket" over half the tooth width (enough to cause a problem). Also the location of applied pressure for die-stamped teeth is not always as close to the root of the tooth, and wear "pockets" usually form even farther out along each tooth-face, even with new/"unstretched" chains; this increases the effective pitch of the teeth (being at a greater effective chain-ring diameter), and has a similar influence to a shorter chain pitch (see discussion below and later about chain "stretch").

vi) Condition of chain, ie "stretch", can have a major influence on chain-suck, both from my tests and from published comments (meaning of the term "chain-stretch", is given in Section-9). A new chain on a worn "granny-ring" gives the worst scenario. This does not imply that a new chain is a pre-requisite for chain-suck, only that if chain "stretch" progresses more slowly than wear of the "granny-ring", the situation is ripe for problems. A highly "stretched" chain usually lessens chain-suck, but it can still occur if the "granny-ring" is severely worn with deep "pockets". Later discussion deals with the full effects of chain "stretch".

vii) Other causes" of chain-suck have also been reported, but are usually less common, or less tenacious and related to design problems with the drive-train or accident damage to chain-rings (bent), teeth (bent), or the chain (twisted, or tight links). But chain-suck initiated by tooth wear and friction can occur to any drive-train eventually, even if adequately designed, manufactured and maintained.

So, key issues are :

  • How much wear to a "granny-ring" does it take before chain-suck becomes a problem ??

  • How rapidly does wear/deformation occur to the teeth of typical "granny-rings" ??

In my case, the answer was : "not much, and ..... very fast".
It became relevant to ask, and get to understand, "why ?"


The "granny" rings used in the field tests were also tested in the workshop at a later stage. The method used was to clamp the "granny" ring firmly in a strongly mounted vice, with the front, top and bottom teeth of the ring protruding clear respectively from the side, top, and lower gap of the vice.

A small hole was then drilled near the end (about 4 cm away) of a strong flat-bar about 30 cm long. The hole was then fitted with a "power-link" type of detachable link. Chains with various degrees of stretch could then be attached to this lever. The lever gave a 6:1 force advantage. Forces were applied through spring-scales with hook attachments.

Each chain being tested was then wrapped around the "granny" ring. The top run of chain was attached to the lever which had its short end placed against the far side of the vice. The bottom run of the chain was fed loosely through the gap in the vice. Varying degrees of load were then applied through the spring-scale (by tensioning it up) onto the end of the lever to exert tension on the top run of chain, and thereby load the teeth of the "granny" ring to a similar degree to pedalling.

This exercise allowed a number of interesting and informative observations :

  • Measurement of the REQUIRED disengagement force needed to pull the chain links off the bottom teeth of the ring while the chain is under tension.

  • Measurement of the AVAILABLE disengagement force actually provided by the rear derailleur was done separately on a range of bicycles.

  • Observation of the mechanism where increasing chain force causes the chain-rollers to ride up the ramped profile of worn tooth pressure-faces to contact the teeth at increasingly higher "elevations", and how the increasing "elevation" of the contact point transfers contact away from the top teeth and towards the bottom teeth, particularly if the chain has low "stretch". Use of a new unworn chain is accompanied by a marked increase in load transfer to the bottom teeth.

  • Observation of the detailed mechanism of how individual links disengage from worn teeth while under load.

In all cases except two, the chains were clean and lightly lubricated. In these two other cases, the muddied chains (new/unstretched, and old/stretched respectively) were taken directly off the bike following a field test. The results of the workshop investigation confirmed the field tests closely.

Tooth Loading Forces :

The APPLIED chain load by an 80 kg cyclist can produce tensions in the upper chain of about 300 kg and more when using the "granny-ring". For this series of tests, a chain load of only 150 kg was applied because of limitations of the equipment used. Thus the actual force required to disengage the chain could easily be double that measured (or more), assuming a simple linear elastic reponse of the whole system, which is not unreasonable.

About 10-11 teeth are seemingly involved for load bearing by a 22 tooth "granny-ring". Individual tooth loading might appear to be 30+ kg per tooth. However, there are further issues which modify this considerably. Firstly, because of the general mismatch between the pitch of chain and teeth as they wear, the load is not evenly spread over the full number of teeth. Secondly, because of the nature of chain design and subsequent wear, the pitch of inner and outer links becomes increasingly different ; the pitch between rollers spanned by inner-links remains virtually constant as they wear, while the effective pitch between rollers spanned by outer-links increases ; as a result,
every second roller applies all (or more) of the chain load to the teeth ; thus only every second tooth is loaded (or more highly loaded). This load re-distribution starts even when the amount of chain wear is very small. It is estimated that individual teeth can carry as much as 100-180 kg when wear increases, which is why wear is an accelerating phenomenon. This overloading affects both metal failure and true abrasion of the teeth. Top teeth will certainly bear high overloads when the chain is badly "stretched", but it is unlikely that the bottom teeth will be loaded quite as highly - nonetheless, they can carry considerable load, to which the high disengagement forces measured below, bear testimony.

Disengagement Forces :

The REQUIRED disengagement force was evaluated and measured. The component of lower chain tension radial to the chain-ring (ie along the line of the tooth), is relevant for the disengaging force. The bottom teeth of worn "granny" rings with new chains, became highly loaded in the workshop, and the links were difficult to disengage even though the chains were lightly lubricated. The most extreme case set up in the workshop (the
LX-ring example), had a highly worn "granny" and a new chain. With a measured tension of 150 kg in the upper chain, a force of around 4 kg was measured as necessary to disengage the bottom link. Thus, the required disengagement force could easily become 10 kg or more, for real conditions where actual chain loadings can be so much greater. If this chain had also been subjected to muddy riding, its increased friction would require an even greater disengaging force.

When a new lightly oiled chain was used on a "granny" with only a medium-worn ring (the STX-ring example), the disengaging force required was quite low (about 0.5 kg). However, when this same chain was used when muddied, the disengaging force again became high (about 3.5 kg) under the 150 kg upper chain tension. By comparison, using a highly stretched chain, there was no force on the lower teeth at all and the chain floated free, even though it was similarly muddied.

With a new "granny" and a new chain, the bottom chain links were completely free of any load under an upper chain tension of 150 kg ; the links floated free.

The magnitude of the AVAILABLE disengagement force actually present on a bicycle, was evaluated and measured next. As before, disengaging forces radial to the chain-ring are relevant. Initially, when there is no chain-suck (or at initiation of chain-suck), this force is supplied solely by the direct self-weight of the chain at the bottom tooth which is very small (about 50 gm, being about 1/6th of the total chain weight of 0.3 kg). The tension from the derailleur does not assist much with disengagement initially; firstly it is low because its spring is not loaded much initially, and secondly it is tangential to the teeth; all it does is alleviate pressure on the tooth face slightly. As the chain starts being dragged up by chain-suck, not only does the rear derailleur spring get extended thereby increasing the tension in the lower chain, but also the chain angle becomes increasingly radial thereby applying the tension at a better angle for disengagement. But even at its extreme extension, where the chain is about to crash into the chain-stay, the rear derailleur spring generates no more than about 3 kg tension. Tension from the catenary effect of the chain's weight is already accounted into the measured forces, as set out below :

Chain-Tension due to Rear Derailleur

Mountain Bike

Road Bike

Normal Position

0.4 - 0.7 kg

1.2 - 1.8 kg

Fully Extended Position

1.8 - 2.5 kg

2.7 - 3.5 kg

MTB's have lower available tension because their derailleurs have such a long arm ; this is another reason why they are more prone to chain-suck than road bikes.

The workshop simulation described above shows clearly how chain-suck can occur. A substantial disengagement force is needed (can be 10+ kg) to get the roller past the geometric (mechanical) resistance of the "bump" formed by a tooth "pocket" or "hook", and to overcome frictional resistance to rolling or sliding of the roller. And this is often greater than the force typically available from the chain's self-weight (50 gm), or its rear derailleur spring (0.5 kg - normal, 2+ kg at extreme extension).

The required disengaging force in the above circumstances seems so high that one may wonder how anyone ever manages to avoid chain-suck !! Of course, if tooth wear is low, the bottom teeth never get loaded in the first place. Even with wear, actual circumstances are dynamic ; as snagged links start moving up the rear of the chain-ring, the crank might be entering a lower load portion of its cycle with much reduced chain tension at the same time as the derailleur spring loading increases (as it starts getting pulled by the chain) - the link may then be pulled free, and this may well happen not too far above the bottom of the chain-ring [an observant cyclist will notice repetitive incipient chain-suck]. Once there is either substantial tooth wear and/or friction, then even the crank's cyclical reduction in tooth loading can remain sufficient to retain snagged rollers, and disengagement can fail completely.

Mechanism of Link Elevation and Load Transfer to Bottom teeth :

Even for unworn teeth, the meshing of a chain as it wraps over the complex shape and spacing of teeth is itself complex; the chain has a definite pitch, but the teeth may be considered to have a variable pitch which depends on the
"elevation" at which the chain effectively contacts them. Also the elevation at which contact is made, may vary along the wrap of the chain if the chain pitch does not closely match the tooth pitch near their roots. And depending on how and where it contacts the teeth, so the chain load may be distributed :
 a) favourably over a number of upper/mid-height teeth, or
 b) mostly on one or two upper teeth, or
 c) with a fair amount on the bottom teeth
The extremes for the bottom teeth are illustrated below.

NEW-teeth and/or 'stretched'-chain : Unloads bottom teeth : Chain floats free WORN-teeth and NEW-chain : Loads bottom teeth : Chain sucks

    New teeth (unworn)                                                                Worn teeth
     OLD chain (worn)                                                                 NEW chain
     [lots of clearance]                                                              [in tight contact]

The way in which a chain interacts with the shape of a worn tooth is progressive. As increasing force is applied, increasing load is transferred to the bottom teeth in an accelerating manner, and the chain-rollers progressively "ride" up the pressure-faces of the teeth. For a new chain on a worn chain-ring (or if the "stretched" chain pitch is still substantially shorter than the effective pitch of the worn teeth), the following occurs as increasing load is placed on the chain :

It can be understood from the above description why there tends to be a fairly sudden onset of chain-suck. With increasing pedal force, not only is the absolute chain load on each tooth increasing, but also an increasing proportion of the total load gets transferred to the bottom teeth. The steep or "hooked" part of the teeth occurs on all teeth at approximately the same elevation which is reached at a specific chain load ; thus mechanical resistance to disengagement starts coming into effect quite suddenly. And finally, the load amplifying effect occuring as rollers try to move past the "hooks", must be overcome in order to allow the disengaging roller to become free. This last effect is analagous to a gate with a stiff spring increasingly resisting any pull against it, being placed on all steep uphill paths ; it makes opening the chain-suck gate and getting free, suddenly much harder than just grinding up the steep hill.

Detailed description of Link Disengagement process :

Some comment is now made about the behaviour of links in the process of disengaging from the bottom of the chain-ring, as actually observed in the workshop. It is assumed that sufficient chain-suck is present to drag the chain until it is aligned to pull in a direction that is radially outward from the chain-ring at mid-height ; this occurs just before it crashes into the chain-stay.

Inner and outer chain links have very different disengagement behaviour in the face of chain-suck ; this is because the nature of force transfer and geometrics between the teeth and the chain with its links, pins, swaged bushings, and rollers, is not the same for outer and inner links. A "free protruding" inner link has its last joint clamped by the tooth/chain interaction to resist it rotating, while a "free protruding" outer link is not so clamped and can rotate freely. Their very different
wear behaviour arises for the same reasons, as already discussed :

  1. When the last two engaged rollers are on an outer-link (and an inner-link has just protruded "free"), both the last roller and its associated split-bushing are clamped between the chain's tension loading (applied through the outer-link pin) and the tooth's resisting load (applied at its pressure-face). The "free" inner-link is thus clamped to the still-engaged outer-link, and cannot rotate freely, because of the loaded friction within the joint - this friction is suprisingly high even with a lightly lubricated clean chain. Thus, the "free" link and the last-engaged-link tend to start to disengage as a semi-rigid unit without rotation at their common pin. The "free" protruding link is thus held at a tangent to the teeth of the ring, at the start of disengagement. The adjacent truly free link is only under spring loading of the derailleur (not chain pedal load), and is able to rotate - thus it pulls at right angles (ie radial to the chain-ring) on the "free" but clamped link. It is thus in a position to try freeing the last-engaged-link by pulling on its clamped, protruding partner, and rotating the whole clamped assembly.

  2. When the last two engaged rollers are on an inner-link (and an outer-link has just protruded "free"), the last roller is clamped only between its own split-bushing and the tooth, leaving the "free" protruding link able to rotate easily about its own still-engaged pin, because only the derailleur loading passes through this pin. The "free" protruding link pulls at right angles to the circumference of the ring (ie radially) to start freeing the roller of the last-engaged-link, more directly.

In both the above cases, the roller was observed to rotate off the tooth, rather than slide off it. Thus the torsional resistance to rotation by friction within the various joint interfaces, is obviously lower than the sliding friction of the tooth/roller interface.

  1. For an engaged outer-link, as its roller starts rolling off the tooth, the semi-clamped "free" protruding link also starts to rotate with it ; the roller remains clamped by friction to the split-bushing of the "free" protruding link ; so this clamped assembly simultaneously rotates around the pin of the engaged outer-link (because its pin has a lower overall torsional friction due to the geometric effect of the pin's small diameter), and also by the rollers action, carries the whole semi-clamped joint assembly past the "hook" of the tooth. All this only happens if sufficient disengaging force is available.

  2. For an engaged inner-link, as its roller starts rolling off the tooth it has no effect on the "free" protruding link as there is no significant friction clamp at the "joint" in this case.

Finally there is also the potential for rotational resistance at the joint formed by the next-in-line engaged roller ("next" roller). It is also loaded by the chain's primary tension and is thus subject to torsional friction on its rotating interfaces. Because the last-engaged-link also needs to rotate in order to allow the actual last-engaged-roller to disengage, any torsional friction at this "next" joint will also contribute to resisting disengagement. In a similar manner to the mechanisms described above, the differences between inner and outer link behaviour now have the following consequences :

  1. An engaged outer-link has its "next" roller clamped only between the associated split-bushing and the tooth pressure-face. Thus its own "inner" pin will have somewhat less resistance to rotation.

  2. An engaged inner-link has both its "next" roller and own split-bushing clamped between the associated pin and the tooth pressure-face. Thus its associated "next" pin is likely to be subjected to more rotational resistance.

As previously discussed, for a worn chain, only every 2nd roller actually loads the teeth (at all, or to a greater degree) - especially as chain wear increases. However, severe chain-suck problems often occur with new or low-wear chains, so this issue will be less pronounced. But, the difference between outer and inner links produces very different disengagement behaviour as described above ; and this, in turn, may also have a theoretical effect on the freeing forces required for inner and outer links. However, in practice in the workshop, and for conditions of strong chain-suck, the difference in force measured to disengage outer or inner link rollers, was not discernable within the variations of the fairly rough exercise undertaken.


I was aware that a range of "anti-suck" products exist, but was unwilling to attach a further device to my bike, also of questionable reliability for resolving the problem. But this beggars the question. Why do some cyclists seem to get chain-suck problems, and others don't ? What are the common elements ? And have those not experiencing chain-suck, really been subject to all the factors and conditions necessary for it to occur ? There must be a fundamental solution, rather than using an extra device.

The above effort arose because, even after investigation, I had not found any detailed articles on chain-suck with reasonably full explanation of the cause and its mechanism, so that deviations from the norm, and its resolution could be reliably understood. Surely the major manufacturers of drive-train equipment have both the resources and motivation to do such research . Perhaps they already have !

Now for some more devious thoughts :

If good research has been done, why is it not more readily available to the biking community ??

Or is the answer a silly one, such as :
all "granny-rings" will suck when they become worn ; it does not take very much wear before that point is reached for many commonly available products ; alternative tooth profiles are effective against chain-suck for longer, but they run more noisily ; tougher metal and accurate tooth machining is effective against chain-suck for a long time, but requires more expensive manufacturing processes putting up prices ;

...and large manufacturers / assemblers prefer not to point out these issues ??


Potential Causes of Chain-Suck :

From published statements and discussion, chain-suck is considered to have various possible causes :

  • Damage to chain-ring, teeth, or chain, eg bent rings or teeth and twisted chain or tight links; these are usually easy to detect and resolve.

  • Mechanical design inadequacies have also been cited as possible causes, such as :

    • spring tension on the rear derailleur which is too low
    • jamming of a narrow chain between chain-rings which are too widely spaced apart
    • middle and outer ring ramps/skirts or pickup lugs which catch and retain the chain
    • poor basic tooth profiles on the chain-rings (similar effect to worn teeth, see below)
    • and others, etc

  • The action of changing gears on the front chain-rings (or 2-ring suck); this form of chain-suck did not form part of this testing programme, but was investigated in an alternative manner.

  • Worn "granny-ring" (pressure face "pockets" or burrs on tooth sides) together with mud

This article deals with chain-suck caused by a worn "granny-ring" and friction raisers such as mud on an otherwise adequately designed and manufactured drive-train. This is probably the most tenacious, continuous, and common form of chain-suck ....... and it can occur to any drive-train.

The Fundamental Mechanism : Arises from Wear of the "granny-ring" and the effect of Friction

The basic cause of common chain-suck is "pockets" worn into the driving pressure faces of the "granny-ring's" teeth. The worn "pockets" are not always visually prominent, but they prevent easy release of the chain from the bottom of the "granny-ring". The "granny-ring" is subject to a number of influences making it more vulnerable to chain-suck than larger diameter chain-rings, as set out below.

The chain load on a "granny-ring" is very high, because of the crank-arm's leverage on this small diameter ring, and because it is used on steep hills needing high pedal forces. Because the "granny-ring" has comparatively few teeth (usually less than 28), each tooth bears a greater share of the high chain load (compared with larger rings). The small diameter of the "granny-ring" also causes the last links to resist release because they pull upward at a greater angle thus increasing the tendency for inherent chain retention. Thus, quite high load is taken even by the last teeth at the bottom of the "granny-ring", and in a less favourable direction for release of the chain. The tooth-face loading causes high friction forces internally within the chain's rotating parts and between chain-rollers and tooth faces. This increases the potential for chain retention especially as it is added to chain retention arising directly from the shape (geometry) of the worn "pockets".

Mud further increases the friction forces described above, when the chain is under load. The increased friction arises because mud/grit particles bridge the lubricant film allowing metal-particle-metal contact, and more particularly because the combination of mud and water rapidly strips out all the chain's lubricant.

In Summary : An inherent tendency for chain retention results from the "granny-ring" being subject to high chain forces and having few teeth, as well as having a small diameter. Respectively, these cause high loads on each tooth-face resulting in high friction between chain and teeth at the bottom, and an adverse direction of chain tension. Tooth "wear-pockets" worsen the situation with a poorer friction angle and mechanical (geometric) chain retention. Finally, mud causes even greater chain retention, induced by higher friction.

A point is reached where chain links are so bound by tooth-face friction, internal rotational friction, mechanical retention by "pocket" geometry, and direction of chain tension, that links at the bottom cannot rotate, nor their rollers slide or roll. In fact, the whole section of chain contacting the "granny-ring" teeth gets so bound by the friction generated by its tension loading that it becomes temporarily semi-rigid along its length. The rear-derailleur spring cannot provide enough tension nor has the chain enough self-weight to disengage from the bottom of the "granny-ring" in these circumstances. The links are carried around and up the rear of the "granny-ring" under load .... the chain "sucks" !! This is the primary mechanism of common chain-suck.

1-Ring Chain-Suck results mainly from a WORN "chain-ring"
...which loads bottom teeth and increases mechanical & frictional resistance between their pressure-faces & chain-rollers, thus preventing chain release

Only one thing will resolve it completely and reliably
..... a chain-ring with little wear

Elastic (transient) effects take place during high-load pedalling, in that the chain elongates and teeth deform slightly. Momentary elastic "dimples" form in the teeth under loading ; this can increase resistance to the chain's mechanical (and frictional) disengagement from teeth at the bottom of the "granny-ring" if they are also subjected to load there - of course, if there is no loading at the bottom, no resistance will occur, and the chain will float free of the teeth. Although this elastic effect can contribute to chain-suck to some degree, the depth of the transitory "dimples" formed by it are very small and will not have significant influence.

Formation of permanent "pockets", however, is the major issue and is caused by two contributing causes, namely permanent metal deformation and abrasive wear of the teeth. Metal deformation is caused by over-stressing at the small contact area between chain-roller and tooth-face with consequent plastic failure of the tooth metal - and can occur instantly upon over-stressing. Abrasion is caused mainly by grit between the teeth and the chain rollers, but also by direct metal-to-metal fretting - but takes time to develop. Both metal deformation and abrasive wear are increased by the high tooth loads occurring on small diameter "granny-rings".

Increased over-stressing and abrasion is more likely with :

  • Softer materials such as aluminium-alloys, die-pressed powder alloys, or unsuitable steels

  • Die-stamped teeth than accurately machined teeth; die stamping often tears away half the contact thickness of the tooth at the critical loading area, and causes irregularities in the desired tooth profile

  • "Stretched" chains which overload fewer teeth, at the top of "granny-rings"

Over-stressing leads to two effects, namely deepening of the "pockets" and displacement of that metal from the face of the tooth into thicker tooth widths or into side burrs. Stronger tooth material is less likely to be deformed permanently (ie fail), and harder tooth material is less likely to be worn by abrasion or fretting. Material for the "granny-ring" is better if both strong and hard (termed "tough").

The word "pocket" has been used to foster understanding of the effect of tooth wear, and must be taken as a relative description. In appearance, the "pockets" are not necessarily truly cupped indentations, unless wear is very severe. Indication that "pockets" are present is that the pressure faces of the "granny-ring" teeth appear slightly steeper than the trailing faces, like an ocean wave before it starts to break. In cases of more extreme wear, the driving face of the teeth may even have true "pockets" or be "hooked" in appearance, like a wave starting to break.

However, even when wear is not especially prominent, the wear can sufficiently change the overall force angle (of pressure and friction) between chain and tooth, that it has the same effect as a true "pocket" or "hook" when friction is high.

There is another major problem with wear. The wear "pockets" often cause the locations where chain pressure is applied against the tooth-face to move farther out along the teeth; this is even more prominent when the "pockets" form actual indentations. As stated earlier, this increases the effective pitch of the "granny-ring" teeth. It thus has a similar effect to a shorter chain pitch; an existing worn chain can now have the same effect as putting on a new chain. Alternatively, the effect of a new/"low-stretch" chain gets even worse (see the discussion about "chain-stretch" below).

FIG-1 : New Item : Aluminium (Soft Material) - Un-worn
  • Icon ring : 0 km usage

FIG-2: Used Item : Stainless Steel (Tough Material) - Moderately-Worn
  • Shimano STX-R&C ring : 2000 km usage
  • no side-burrs
  • no tooth thickening
  • tooth pressure-faces steeper than new
  • Roller contact a bit more elevated than for new teeth (negligible indentation)
  • SEVERE chain-suck with unstretched (new) chain and mud/grit : with intended design tooth-faces bearing chain load, first suck at 1800 km usage
  • NO chain-suck with unstretched (new) chain and mud/grit : with ring flipped, trailing tooth-faces bear chain load, still no suck after further 3000 km usage

FIG-3 : Used Item : Die-Press Powder-Alloy (Soft Material) - Severely-Worn
  • Shimano LX ring : 1000 km usage
  • has side-burrs
  • has tooth thickening
  • tooth pressure-faces much steeper than new (almost hooked)
  • Roller contact much more elevated than new teeth (substantial indentation)
  • VERY SEVERE chain-suck with unstretched (new) chain and mud/grit : with intended design tooth-faces bearing chain load, first suck at 300 km usage
  • CANNOT FLIP the ring ; it has an asymmetric cross-section

The middle and outer rings do not usually suffer chain-suck, because they have more teeth, much lower chain loading on the pressure face of each tooth, negligible/no load on the last teeth, and better chain load direction. There is less friction and easier chain release results, as well as lower stresses and less wear to the teeth.

Major additional Mechanism : Arises from various effects of "Chain-stretch"

"Chain-stretch" is not stretch in the usual sense (ie stretching of the metal chain-links). It means abrasive wear of the chain's internal bearing surfaces (by grit and by metal-to-metal fretting), over the contact area between the pins and split-sleeve bushes formed by the inner-link flanges. These bushes develop play and the whole chain elongates.

"Chain-stretch" plays a large role by accelerating wear to the pressure faces of the teeth of both the chain-rings and the rear sprockets. "Stretched" chains load fewer teeth of the rings & sprockets in turn (lesser load being borne by teeth further along); all teeth get overstressed and wear accelerates.

A further crucial issue arises from chain "stretch". When a NEW chain is used on a WORN "granny-ring", the potential for chain-suck increases dramatically, caused by a new chain being "unstretched". It has a correct pitch (spacing of links), which is slightly shorter than that of a worn chain. The pitch between the teeth (with their "pockets") of a worn "granny-ring" does not change (fundamentally) compared with a new "granny-ring". Thus "stretched" and new chains interact with the teeth very differently.

With a worn or "stretched" chain the accumulating increase in chain length because of its slightly longer individual link-lengths progressively frees the rollers from contact with the worn teeth. By the bottom of the "granny-ring" a highly "stretched" chain can be completely free of the pedalling load and even mechanically free of teeth with deeply worn "pockets". The contact force between each successive chain roller and tooth decreases as the links move away from the point of initial engagement at the top of the "granny-ring. Thus, a sufficiently worn chain might be able to disengage easily, even from markedly worn "granny-ring" teeth. The degree to which it does or doesn't "suck", depends on how much it has "stretched" relative to "granny-ring" wear. However, although a "stretched" chain might improve the chain-suck situation in the short term, it overloads the top teeth and accelerates wear substantially.

The rollers of new chains (or those with little "stretch"), on the other hand, remain tightly pulled against the worn teeth right to the bottom of the "granny-ring" because the chain's correct (shorter) pitch matches the fundamental tooth pitch closely. At the bottom of the "granny-ring", the rollers now bear their full share of the chain load causing friction and mechanical retention by the tooth "pockets" (as discussed), unlike the freer condition of a "stretched" chain.

Two interactive effects occur, with Mud, Chain-Stretch, & Tooth-Pitch, and also with Tooth-Wear & Tooth-Pitch, as follows :

  • A layer of mud/grit between the pressure zones of the teeth and the rollers is neutral in its influence on chain-stretch effects (uniform thickness layer assumed). However a layer of mud/grit can have other significant geometric effects on the chain/ring interaction; layers compressed as thin as 0.05 mm (0.002") are sufficient to cause problems. When layered on the internal bearing surfaces of the chain's contact-zones between its pins and split-bushes, it will tend to shorten the chain's effective pitch and have the effect of a "low-stretch" or even a "negative-stretch" chain. Additionally, mud/grit between the rollers etc and the roots of the teeth will elevate chain-rollers, thus increasing the effective pitch diameter at which the chain interacts with the teeth, and thereby also increase the effective tooth-pitch.

  • Wear "pockets" often elevate the chain-rollers above the roots of the teeth because the "pockets" often form farther out along the tooth length. This will also increase the effective pitch diameter of the ring, and will thereby increase the effective pitch of the teeth.

  • For both the above cases, the increase in tooth-pitch is simply the flip-side of a "low-stretch" chain (ie small chain-pitch), and will worsen the effect which a new or "low-stretch" chain can have when bearing on worn tooth faces. The effect which a "high-stretch" chain has in unloading bottom teeth will now be reversed, with the bottom teeth bearing a greater share of the chain load than the top teeth. OVERLOADING the BOTTOM TEETH has consequences of further increased friction within the chain and between chain-rollers and worn teeth, as well as more difficult mechanical disengagement from the "pockets". This is a significant additional effect (which also accelerates tooth wear). This test programme has showed that a 0.4% mismatch between chain-pitch and tooth-pitch resulted in severe chain-suck, with adversely worn tooth faces. The importance of having a tooth profile which has no tendency toward retention of the chain, becomes very clear. In its absence .......chain-suck becomes highly likely.

So why can chain-suck also occur with an existing chain which has worn simultaneously with a "granny-ring" ? Wear of "pockets" in the teeth and "stretch" of the chain are not directly coupled with each other. There is no reason for them necessarily to be geometrically compatible to avoid chain-suck. Elevated location of wear pockets has its own effect. All combinations and permutations can occur. For example, if the "granny-ring" is of softer wearing material, there is a good chance that it will wear faster than the chain will "stretch", plus the effect of wear "pockets" often forming at an elevated position on the tooth faces - leading to chain-suck. For a tough resistant "granny-ring", if the chain is not cleaned and lubricated regularly, it may "stretch" faster than the "granny-ring" wears - avoiding chain-suck. etc, etc.

please load images

Understanding of the above chain-suck mechanism has been developed assuming that all teeth are worn to a similar extent. Somewhat greater wear along 180 degree opposite sides of the chain-ring is likely, due to the non-uniform nature of pedalling loads. This is likely to increase chain-suck potential.

To complete the picture are some additional issues, not affecting chain-suck as such :

  • A "stretched" chain on a new "granny-ring" can result in "jumping" over teeth instead of chain-suck.

  • For the rear sprockets, chain-suck does not occur, irrespective of sprocket or chain wear ; the high tension in the upper chain directly from the pedal loads, forces disengagement.

  • A new chain on a worn rear sprocket can result in "jumping" over rear sprocket teeth.

Other Lesser Mechanisms :

A lesser mechanism possibly contributing to chain-suck to a small degree, is increased friction/binding between the sides of the teeth and the chain, caused :

  • directly by the mud and grit
  • by burrs on the sides of the teeth or thickened teeth due to tooth deformation

These may seem major issues. However, "pockets" on the tooth faces remain the main problem. Why do the burrs and tooth-widening, not constitute a significant cause of chain-suck ? Surely the chain (plus mud/grit) would be forced hard onto the widened teeth, causing wedging and binding ?

Actual burred and widened teeth which I have observed on many "granny-rings" (even severely worn ones), still had acceptable clearance fits within chain links. During my testing, there was no evidence of chain/tooth binding caused by wedging, even when severe chain-suck occurred; the chain always had enough lateral clearance on the teeth that it was loose the moment that upper chain tension was relaxed.

Large quantities of very stiff sticky mud adhering to the chain and rings might also conceivably cause the chain to "stick" to the teeth by the mud's cohesive and adhesive properties, and by its effect on the geometry of chain/teeth interaction (described already). However, I have not experienced this and consider it extreme.

So, although lateral-wedging of burred/thickened teeth and grit or large quantities of stiff sticky mud might be potential chain-suck mechanisms and might even occur occasionally in practice, they are not necessary conditions for common chain-suck to occur. This was confirmed by severe chain-suck occurring regularly with only a fine runny mud slurry, lacking larger particles, and on teeth which were worn but were neither burred nor thickened. No doubt tooth side-friction does contribute to common chain-suck to a small degree, but only by adding to its main cause, namely direct in-line friction, which occurs anyway

There is a very direct relationship between tooth-face loading caused by high pedal force (eg up steep hills), and in-line tooth-face/chain-roller friction; this is elementary physics. This friction is high; it inhibits chain links from both rotating and sliding off teeth, thereby binding them; it also fits all the observed phenomena. The moment that chain tension is relaxed, tooth loading ceases and consequently the in-line friction and its binding effect also ceases.

Lateral wedging, on the other hand, is indirect and reliant upon a number of factors (including the chain tension) to become strong enough to cause binding of the chain. I have also not observed it in practice, despite very careful examination of the actual situation during the occurrence of chain-suck. When chain tension is relaxed, any wedging and binding action should not disappear completely because it relies on the tooth plus large particles remaining jammed into the inner links of the chain; but I have not observed any such delayed release of tooth/chain binding.

Myths and Facts about Lubrication :

Lubrication (no matter what type) does not play much role in preventing chain-suck in the presence of mud, because it is stripped out rapidly during muddy wet conditions, and is thus not present to reduce friction. Only an unworn "granny-ring" is reliable in such conditions. This is a direct finding of a further programme of actual field testing where the drive-train was well lubricated using six alternative lubricant types (of big-name brands, and both wet and "dry" types) for a complete set of tests. The same test had also been carried out ensuring a complete absence of lubrication.

Chain-suck is possible even without muddy conditions if "granny-ring" wear is severe, if tooth loading is high from pedal effort on steep hills, and if lubrication is also inadequate at the same time. In this situation, friction between the chain and tooth-faces can again be high enough to cause chain-suck (typically in dry dusty conditions) ... but in these circumstances without mud, lubrication can alleviate chain-suck reliably, because it is not being continuously stripped out by mud and water. However, this alleviation will only work if the tooth wear is not excessive, and/or if the chain is NOT new.

Once a new "unstretched" chain is installed on a bike with very severe tooth-face wear, lubrication will no longer alleviate chain-suck, even in non-muddy conditions. This is because the deep "pockets" in the teeth or severely "hooked" teeth are able to resist mechanical disengagement of the chain with its shorter (correct) pitch, directly and continuously. Lubrication does not alleviate it, because the mechanical retention mechanism outweighs any reduction in the friction mechanism.


If you ride or race in muddy conditions, you may want to check for possible problems. OK, you could just go for a muddy ride and hope that absence of a problem is indicative. However, the particular ride might not place mud and grit in the critical places, nor perhaps in sufficient quantity, nor perhaps at key points in time. If it doesn't, how will you be sure that the problem won't occur when least wanted ?

The method described in the referred TEST document will give far more reliable confirmation as to whether chain-suck will occur or not. It has been developed to ensure conditions which are acceptably constant, repeatable, certain, and realistic to lessen these as issues of contention; it also reduces the time which might otherwise be needed to reach a conclusive result (+ve or -ve). The test takes only about 1 hour, and neither you nor your bike need get particularly dirty.

For the chain-suck TEST press the link below ...

Chain-Suck TEST


i) Use a "granny-ring" of a type which :

  • Does not wear rapidly. Such a product is likely to be as follows :

    • Strong and Hard-wearing material (tough) such as an especially tough hard steel (eg an appropriate type of stainless steel), or use titanium. Avoid all aluminium or die-pressed powder-alloys for "granny" rings. If the problem item is a larger chain-ring (say 30 or more teeth), only then is a good aluminium alloy acceptable.

    • Accurately machined by a process suitable for the material (definitely not die-stamped).

    • In checking for suitable products, company literature should show that "granny-ring" wear and chain-suck are specifically alleviated. Some apparently suitable products (wear test results not known to me), are made by the following companies:

      • Stainless Steel rings : Action Tec, Blackspire, Real, etc
      • Titanium rings : Action Tec, Cycle Dynamics, RaceFace, etc
      • Aluminium rings : Most aluminium alloys are far inferior to tough steels or titanium; but some special aluminiums are adequate for larger rings (say 30 or more teeth).

      A tough accurately machined steel, is probably the best

  • Has an initial tooth profile which is less susceptible to wear into a problematic shape

ii) If the "granny-ring" is worn, Flip it or Replace it with a new one :

  • Simply flip it over if it was completely symmetrical when new, so that the unworn trailing faces of the teeth become the pressure faces. This can provide a sprocket whose undamaged "new" pressure faces will start wearing from this time onward

  • Buy a new one if it is not symmetrical and cannot be flipped; the new one is best if made of tough material, accurately machined, and with a suitable tooth profile (see item-i)

iii) Re-establishing an acceptable tooth profile by filing worn pressure faces of the teeth (not tooth side-burrs) is a method which can be tried by the technically proficient - see separate document, Filing Tooth Profiles. This method can also be used on-the-trail. However, it should not normally be used if a new chain is to be installed.

iv) A lubricant container and mechanism attached to the bike for continuous lubrication during the ride is reported to have been used successfully by some bikers in muddy conditions; this is a messy solution and does not resolve the fundamental problem. It is unlikely to work if a new or "low-stretch" chain has been installed on badly worn chain-rings. The method is not recommended.

v) Installation of a special anti chain-suck plate is not recommended because they often do not work properly, and can then cause more damage to the bike than if they were absent.

vi) Clean and Lubricate the chain, rings, and sprockets regularly to combat "chain-stretch" and surface abrasion of teeth. Excessive "chain-stretch" causes major wear and damage to all teeth to the point of needing to replace rings and sprockets at the same time as replacing the chain. In dry conditions (provided that the teeth are not too badly worn, and a new "unstretched" chain is NOT used), chain-suck can be directly alleviated by lower friction between chain and teeth, when lubricated. However in muddy conditions, only the onset of chain-suck may be delayed to some degree, because lubrication (no matter what type) is stripped out so rapidly by mud. Minimising chain "stretch" and abrasive tooth wear remains a prime purpose of cleaning and lubrication.

vii) Replace the chain before it becomes badly "stretched". A badly "stretched" chain will accelerate wear to teeth receiving the hardest pedal effort on steep hills, namely those of the "granny-ring", and the larger rear sprockets. When a new chain is run on a worn "granny-ring", chain-suck will become worse (and "jumping" can occur on the worn larger rear sprockets). When an older "stretched" chain is run on a new "granny-ring" chain suck will improve or disappear, but the "granny-ring" will wear faster, and "jumping" over "granny-ring" teeth might occur.

A general recommendation is that a chain be replaced at or before about 0.5% "stretch". At about 1.0% "stretch" it will already have done significant damage to the teeth of both the "granny-ring" and rear sprockets. A new chain measures exactly 1 inch per link-pair. 0.5% stretch implies a 1/16th inch elongation over 12 link-pairs or 12 inches. The measurement must be done on a somewhat tensioned chain between the centrelines of the chain pins, and not using an internal measurement between rollers which is inaccurate. Measurement can be done on the bike (chain clean) over 12 link-pairs, but it is more accurate if the chain is removed and measured over 48 link-pairs.

If a "granny-ring" has been replaced or flipped, then the chain should be replaced if its wear is more than about 0.3% to prevent accelerated wear to the new "granny-ring" and the rear-sprockets.

viii) Replace any rear sprockets if the chain "jumps" over their teeth (due to their own wear), after a new chain has been installed.

ix) Check for other issues such as obvious accident damage to chain-rings (bent), to teeth (bent) and to chains (twisted or tight links), and reports about mechanical inadequacies of particular drive-trains. File off any large burrs on the sides of the teeth if these have formed (as a sensible although non-essential precaution); the presence of large burrs usually means that substantial "pockets" have also formed on the pressure faces of the teeth. All these "other" issues are either rarer as causes of chain-suck or, alternatively, are lesser influences.

To Manufacturers, Bike Assemblers, & the Biking Community

This investigation has developed understanding with minimal resources. It must be assumed that the major manufacturers of crank-sets with their chain-rings, and manufacturers/assemblers of bikes already have this understanding ; if they don't have it, they certainly should.

However, when this article was first written in 1999 it did not seem to be widely known in the general biking community. That situation seems to have changed over the intervening years, perhaps partly because of the articles published here.

That there are marketing pressures to supply bikes at the lowest possible price with components to suit average usage at a lower intensity, is a commercial reality which benefits bikers financially (in general). Clearly, such benefit does not apply to components which last no more than a single usage by intensive users before suffering from chain-suck; this happens far too regularly with "granny-rings" of soft material (eg aluminium and some steels). Thus, I believe that as soon as a crank-set is available with replaceable chain-rings, it must be regarded as a higher level product. The manufacturer must then offer an option for fitting a chain-ring which is demonstrably highly resistant to wear under heavy usage, and clearly identifiable as such, even if it is a special item at higher cost. Softer chain-rings should also be clearly identifiable (ha ha !!)

The market should also be made aware that better quality chain-rings exist. Hopefully, in future, such chain-rings will become more easily and economically available. Certainly, understanding of the chain-suck issue and development of a greater demand for more wear resistant "granny-rings" should increase their availability and decrease both the reluctance for mass manufacturing and costs.

Currently it can be difficult to find a wear resistant "granny-ring" to fit some popular crank-sets, even those on higher level bikes. It is up to the biking community to insist on having better clearly identifiable choice available !!

  Copyright Jonathan Levy, 2000. All rights reserved.
  email Jonathan Levy - must copy email address by hand

For the related articles, follow the links below ...
Chain-Suck : in a NUTSHELL ...comprehensive summary and solutions
Chain-Suck : the OVERALL picture ...details of mechanisms & solutions
Chain-Suck : the DETAILED Investigations ...field & workshop testing
Chain-Suck : the FIELD TEST ...a field testing method for identifying causes
Chain-Suck : Restoring Worn Tooth Pressure-Faces filing them
to the Link Page

: Visitors