'standard' pedal bearings

[Brucey]

Especially if you ride flat pedals (or cage pedals with toe clips) the chances are pretty good that you are using 'traditional' pedal bearings, i.e. cup and cone type, with indifferent (or non-existent) seals.

Some pretty reasonable pedals (eg most of the MKS range, and pretty much all of it that costs less than ~£40 a pair) are made this way. So are some of the most basic pedals too. Such pedals can last well provided they are kept lubricated and adjusted correctly. Even in unsealed pedals, a regular squirt of lube in the inboard bearings (eg aerosol SFG every time the chain gets lubed) will usually keep the bearings sweet even in fairly hard use. Some very basic pedals can last for years this way, with the bearings just getting smoother and smoother.

Also, it turns out that not only is there much similarity between pedal bearings made by any one manufacturer, but there is extensive compatibility between pedal manufacturers too. The effect of this is that the shattered remains of the most humble looking plastic-bodied pedal may contain parts that are of use in other, better pedals.

In practice there are two very common bearing formats; ~22mm dia cup inserts (using 5/32" dia balls) and ~20mm dia cup inserts (using 1/8" dia balls). To my surprise both bearing formats can use the same axles and cones. M8x1.25mm threading on the spindle and two flats (or near flats) on the spindle for the lockwasher are commonplace too. This means that cones, tab washers and locknuts often swap between pedals with minimal difficulty. The most common hex size for the locknut is 12mm. The lockwashers will usually take a fair bit of stick before they round out, so you can normally adjust the bearings using nothing more than a 12mm socket and a small screwdriver, to prod the cone round with. You just need to keep track of the cone movements vs the backlash in the tab washer, if you want to get the adjustment *just so*.

The distance from the shoulder (which bears against the crank) on the pedal spindle to the inboard bearing surface varies +/- ~1mm in spindles that don't have an obvious scallop or offset to them. Spindles also come in several different lengths, too. In any event the chances of finding a substitute spindle are not that bad.

MKS sylvan touring pedal, with some pedal spares

from L to R, top to bottom; 1) 20mm outboard cup insert, 2) 22mm outboard insert, 3) 20mm inboard insert, complete with x13 1/8" balls, 4) 22mm inboard insert, complete with x11 5/32" balls

Inboard bearing inserts typically have a ~12mm centre bore. Outboard inserts typically have a ~10mm centre bore, but otherwise are often similar. 13x 1/8" balls or 11x 5/32" balls, as pictured, are the two most common arrangements.

a medley of pedals. Some spare parts are common to all these pedals

The MKS touring pedal (from a 'very spare' set) in the first photo needed a replacement 22mm inboard bearing insert. One taken from a rather nasty plastic pedal was the same nominal size as the MKS one, and was made to fit by using epoxy resin to bond it into the pedal body. {Tolerances vary; the MKS insert is slightly thicker-walled and is precision ground on the OD to make it a consistent fit. Cheaper pedals often use slightly thinner gauge cups and different OD tolerances, hence the replacement was a loose-ish sliding fit in the MKS pedal body.]

Pedal spindles (even from quite cheap pedals) seem to be made from fairly well hardened CrMo steel, and in many cases will take a slight regrind if necessary.
Cup inserts are rarely set into the pedal body in such a way as the ID does not protrude into the pedal bore and this means they can usually be knocked out without too much difficulty. Obviously they come out of plastic pedal bodies quite easily, especially broken ones. You can enlarge (by grinding) the centre bore of an outboard insert and use it in the inboard position if you want to. If you try and use an inboard insert in the outboard position, this is usually possible, but it means the outboard bearings can drop inside the pedal body during reassembly or when the bearings get a bit slack in use.

Pedal brands which may share spindle threads, cones, lockwashers, locknuts, and (nominally) bearing cups include MKS, Wellgo, VP, Union as well as other more obscure brands. The original (and IMHO only 'good') type of DMR V8 pedals also use this architecture. Occasionally a small variation in spindle/build length may require that a substitute part (eg the spindle or cone) is ground shorter, so as to make everything fit.

Operations which can be performed include;

a) repair of 'good' pedals
b) refurbishment of a favoured pedal set (e.g. which is no longer available new)
c) conversion of BMX pedals (with 1/2" threaded spindles) to 9/16" spindles (or vice versa)
d) conversion of an odd LH pedal to a RH one, or vice versa, provided the pedal body is symmetric.

It is also (of course) possible to add lube ports and/or better seals to pedals when you have got them in bits.

In pure money terms you may be wasting your time doing any of this; you might be better off spending your time looking after the next set of pedals rather better, and avoiding whatever troubles befell the previous set. However there is satisfaction to be had from 'making good', saving the planet etc, and with the way prices are going, rescuing a set of MKS pedals from oblivion is no longer a pin-money consideration.

cheers

[JakobW]

This thread is timely: I recently bought a new pair of MKS Touring pedals, and was going to strip and re-grease them with a proper all-weather grease. I note that the instructions say to add some preload, but aren't particularly clear on how much; what's the best adjustment when new? I'm assuming I'll have to tighten them up after a few hundred miles as the bearings run in?

[PT1029]

Way back when, when I started, one of our members used to run his new bearings in. To ensure even running in (rather than more wear on the down stroke), he would clean the new bearings, reassemble with the bearing just a shade tight - no grease, just chrome cleaner. Connect to an electric drill with a flexible drive. After a period of time (5 - 10 minutes? - I forget) the bearing would be evenly run in.
Then he would strip it down, reassemble with new balls and grease. I know he did this for (cup and loose axle) bottom brackets, probably for hubs too. I can't remember if he did it for any other bearings.
It wasn't a case of making poor parts good, all this 20 or so bikes had variously TA/Campag/Shimano BB's.

[Jdsk]

Way back when, when I started, one of our members used to run his new bearings in. To ensure even running in (rather than more wear on the down stroke), he would clean the new bearings, reassemble with the bearing just a shade tight - no grease, just chrome cleaner. Connect to an electric drill with a flexible drive. After a period of time (5 - 10 minutes? - I forget) the bearing would be evenly run in.

Reminds me of one of Colin Chapman's early exploits: he had a mismatched (car) final drive so ran it in metal polish until it meshed properly, then replaced the bearings.

Jonathan

[Brucey]

This thread is timely: I recently bought a new pair of MKS Touring pedals, and was going to strip and re-grease them with a proper all-weather grease. I note that the instructions say to add some preload, but aren't particularly clear on how much; what's the best adjustment when new? I'm assuming I'll have to tighten them up after a few hundred miles as the bearings run in?

MKS pedal bearings are usually set slightly tight from new. Having seen very many sets of these pedals, I believe that this is (unlike many cup and cone hubs perhaps) deliberate, so that once the bearings are run in, they shouldn't end up slack.

If you back the adjustment off very slightly, they will run smoother out of the box, but will also require further adjustment a bit sooner, after a few months of use.

As I mentioned previously the trick when adjusting the bearings is to keep track of the relative movement of the cone vs the tab washer. When you tighten the locknut, the tab washer always swings full CW (within its range of lash). When you loosen the locknut the tab washer usually turns full ACW (within its range of lash). The lash is usually of the order of five degrees, but wear can increase this to nearer ten. When adjusting the bearing tighter, the initial movement of the cone CW usually takes the tab washer with it, but this doesn't affect the net bearing adjustment. Only once the lash is accounted for does relative movement between the cone and lockwasher occur, and thus makes a meaningful bearing adjustment . When adjusting the bearings slacker, the usual thing is that the slightest movement of the cone ACW moves it relative to the lockwasher and changes the adjustment.

Because the thread pitch is usually 1.25mm, the cone angle change for any given clearance is somewhat smaller than normal (in hubs where the thread pitch is usually 1.0mm). If you are shooting for ~10um accuracy (in terms of lateral float/preload) in the bearings then the required cone angle change between adjustments is less than three degrees. In other words the adjustment you are shooting for is (or should be) a fraction of the lash in the lockwasher. Obviously if the locknut isn't tight enough the bearings can soon go out of adjustment; note that the effects of in-service precession will tend to loosen the RH cone adjustment (but back it into the locknut) and tighten the LH cone adjustment (which loosens it against the locknut).

Worth mentioning that running-in creates wear debris, so arguably the time to service the bearings properly is after six months use or so. Prior to then you can adjust the bearings, and add better grease if you want to. Both these things can be done without taking the pedal completely apart. The outboard bearing can be greased in situ once the dustcap is unscrewed. The inboard bearing can have grease added to it by removing the dust seal ring (black plastic), wadging grease in the gap, and then replacing the seal ring. Provided you have added enough grease, replacing the dust seal ring (which BTW isn't any kind of a good seal per se) will tend to force grease into the bearings. This obviously isn't as good as a full service, but it should be good enough for a few months use.

cheers

[Brucey]

BTW 'running in' may cause two different things to occur

a) reshaping; comparatively gross changes in the shape of bearing parts. These may require 1-5um (or perhaps even more) material to be removed locally, without which the bearing parts may not be round enough to run smoothly or share the load between balls as should occur.

b) polishing of the surfaces. This ought to require less material removal, and is more or less a question of smoothing whatever texture remains on the bearing surfaces from manufacture.

In the case of super quality parts (eg campag) both cup and cone in a bearing assembly would be ground to provide a really good dimensional tolerance. Running such bearings in is therefore mainly a matter of polishing rather than reshaping.

MKS Sylvan pedal bearings are pretty good but they don't have a ground finish on the bearing surfaces, so some reshaping is required. AIUI the 'sylvan prime' models require less reshaping and mostly only require polishing. [However I don't think there is usually much difference between standard sylvan and sylvan prime models, once they have had a year's use or so; therefore the 'prime' model is arguably one for those who want (comparatively) instant gratification. MKS are discontinuing the 'prime' versions of sylvan pedals.]

If you run the bearing under load with an abrasive slurry in it (artificial running-in), then the wear mechanism is that the abrasive material is crushed between the ball and the raceway. This may be quite effective in terms of material removal for reshaping, but may not be the most effective way of polishing the bearing surfaces. Because the abrasive material degrades, the polishing action becomes finer and finer as the material breaks down. However it becomes mixed with metal particles and these may well cold-weld to the surface and cause a whole different wear mechanism to occur.

If all you want to do is to polish rather than reshape the bearing surfaces, then it may be better to use a more conventional polishing technique, eg using increasingly fine abrasive paper against a rotating part, or (in parts that cannot be spun in a fixture) using a rotary tool with a felt mop and abrasive polish. Such surfaces can be made smoother than occurs with conventional running in, or artificial running in.

cheers

[Brucey]

DMR V8 spindles are available as spare parts



eg from upgrade bikes

https://www.upgradebikes.co.uk/Catalogue/Pedals/Spares/Classic-V8-Axles

however the cost is not much less than buying a replacement set of wellgo pedals, which contain very similar parts. In any case you can see the finish of the axle extends over the bearing surface; it is this part which gets polished during 'running in' and can be polished using other methods instead/as well.

In a similar vein MKS use similar components inside their budget pedals; for example MKS PB-390



are only about £12 a pair and use rather sylvan-like bearing parts. However I don't know for sure that the spindle length is a match for other given models. if they are a match for spindle length etc, they could be a cost-effective way of getting spare parts for Sylvan models.

cheers

[NickJP]

I have some Campagnolo Nuovo Record quill pedals that must have had about a quarter of a million kilometres on them by the time I completely switched to clipless pedals about 20 years back. I still have the pedals, and the bearings in them feel just as good as when they were new. I used to service them more or less annually. Those pedals don't have seals where the spindle exits the pedal body, but the spindles are rifled at that point so that the rotation of the spindle relative to the pedal body helps to keep crud out of the inner bearings.

[JakobW]

Worth mentioning that running-in creates wear debris, so arguably the time to service the bearings properly is after six months use or so. Prior to then you can adjust the bearings, and add better grease if you want to. Both these things can be done without taking the pedal completely apart. The outboard bearing can be greased in situ once the dustcap is unscrewed. The inboard bearing can have grease added to it by removing the dust seal ring (black plastic), wadging grease in the gap, and then replacing the seal ring. Provided you have added enough grease, replacing the dust seal ring (which BTW isn't any kind of a good seal per se) will tend to force grease into the bearings. This obviously isn't as good as a full service, but it should be good enough for a few months use.

cheers

In which case I may just add grease and adjust the bearings for now, and strip the pedals completely come spring. If I've understood you correctly, the correct adjustment is to set the cones so that there's no play in the axle (finger-tight?), and then tighten the cones about a further 10um/3°, which will entail some faffing with tab washers; the factory adjustment is usually with more preload than this?

[Brucey]

With any bearing that is perfectly smooth, the only safe way of setting it is to incrementally tighten the bearing adjustment until the free play just disappears, each time testing of course with the locknut tightened in the normal way, and making it otherwise representative of the service conditions (which in the case of QR hubs means with the axles loaded via QR pressure). The reason for this is that an unacceptably high preload on the bearings may not be detectable, in that it won't at first cause much in the way of obvious roughness.

However with bearings that are not perfectly smooth, and don't have contact seals, you can use how the bearing feels as a proxy for preload. If you manually preload a slack bearing (eg by pushing axially by hand) whilst turning it you will feel about half the usual roughness (for that load), since only one bearing (inboard or outboard) sees any real load, and the other one cannot contribute any real feeling of roughness. In any event this can tell you if one bearing is rougher than the other (by loading in each direction) and whether a sensible level of preload is likely to cause a significant feeling of roughness or not.

With new pedals you can keep track of the timing of the cone vs the pedal flats (not so easily on pedals with a hexagonal spanner drive rather than two flats) and see how much preload was applied vs *just* play free adjustment. I would be very surprised if this came to more than about 15 degrees of cone turn of preload. [In fact I can easily experiment with this; I have a box with at least half a dozen new sets of MKS Sylvan pedals in it and can interrogate them to see how much (in cone turn terms) preload they leave the factory with.] Only when there has been a lot of wear (about 200um's worth) would you be in danger of being 'on a different flat' i.e. more than 1/6th of a turn of the cone away from the initial adjustment. A few photos (eg with the spindle in a bench vice) will enable you to keep track if you want to.

The worry with an absence of preload is that the bearing will run slack (initially under load only) and that this will cause uneven wear, even before there is any noticeable free play. My view is that this shouldn't be a big worry; not with 'traditional' pedal bearings using 5/32" balls, even if with smaller balls it might be. The reason for this is that load sharing between 5/32" balls is not required for service loads of the kind that you will generate in the saddle, and running for extended periods with slight free play doesn't usually seem to do any permanent damage, in that you can readjust the bearings and the worst that usually happens is that you have to go through another 'running in' cycle to get everything back to normal.

By contrast free play which appears relatively suddenly often signals that the lubrication has failed in one or other of the bearings, usually the inboard one, through water ingress. If this is not attended to promptly the result can be catastrophic wear, usually via a corrosion/erosion mechanism. You will know when this has happened because you will see rust coloured debris when the bearing is flushed (with GT85 or WD40) and once adjusted the bearing will usually make funny noises under load. The funny noises sound somewhat like those you hear as you walk on coarse shingle, and for similar reasons; balls are forced to slide and this is always noisy. The noise gets worse with thinner lubricant in place, so flushing with GT85/WD40 is a useful diagnostic tool. The wear is usually distributed over about 180 degrees of the spindle but it is typically more localised in the insert cups (esp if clips and straps are used), because these are always oriented the same way vs the downwards pressure on the pedals. The spindle can be reground/repolished so it is round again, the cups rather less easily. I have done various experiments and once pedal bearings become even slightly noisy it can take 'for ever' for them to run in again, once packed with grease correctly.

BTW I have never seen a torque figure for M8x1.25 locknuts but if the pedals 'adjust themselves' in service then it is clear that the locknuts were not tight enough. I reckon the correct torque figure is between 10 and 15Nm . The harder you push on the pedals, the more torque you need on the locknuts to keep the adjustment in place. AFAICT the worst thing that is likely to happen if you overtighten is that the locknut and/or tab washer will strip, and I think this is unlikely to damage the spindle, because it is 'all hard', and the locknuts are soft. 12mm hex M8x1.25 half-nuts are not readily found in a fastener suppliers but nearly every scrap set of pedals contains some, so they are not difficult to replace. [NB Campag pedals are a different matter BTW, they don't use such a coarse thread.]

You know how the bearings feel at present (in new, unadjusted pedals)…? If you plan to keep on top of things you should be shooting for a small fraction of that roughness initially and a very much smaller amount when the pedals are run in. Obviously you can only feel the bearings properly if the pedal is out of the crank and you are turning the spindle between your fingers.

But anyway if you keep your wits about you and/or you are prepared for a little trial and error, it isn't too difficult to do a decent job of pedal bearing adjustment.

cheers

[Brucey]

BTW just in case it isn't obvious, when you tighten the locknut you push the lower flanks of the cone threads hard against the upper flanks of the spindle threads.

By contrast when you 'finger tighten' a pedal cone, you push the upper flanks of the cone threads against the lower flanks of the spindle threads.

The two settings differ from one another by at least the amount of lash there is in the threads between the cone and the spindle. In most pedals this is I excess of 50-100um. If you finger tighten the cone and immediately tighten the locknut on top of it, you will exert several hundred kgs of load on the bearings and this can instantly damage them.

So the method (having had a pedal in bits) is to finger-tighten the cone, then back it off about 1/5th of a turn, then half-tighten the locknut. Hopefully you will find a little free play. You can get close to the correct adjustment, half-tightening the locknut each time you check, with full tightening of the locknut required during the final stages of bearing adjustment.

FWIW another thing that should be obvious (but often isn't) is that in some pedals the RH locknut can have a LH thread on it. However in so-called 'standard' pedals with large-ish cup and cone bearings, this is rarely encountered. RH cone/locknut threads on both spindles is the most common arrangement if the locknut threading is M8x1.25mm.

cheers

[rogerzilla]

New SPDs usually have a bit of "tooth" in the adjustment too. IME adjusting brand new pedals, or used ones with new balls, to be perfectly smooth means they will develop slop after 100 miles.

[bgnukem]

According to the graph of bearing preload against fatigue life, it's better to have a slight preload than slight clearance, but beyond the 'optimum' preload, fatigue life drops off more rapidly than is the case with increasing clearance, so it's probably less damaging to err on the side of slight clearance than risk over-tightening:

https://www.researchgate.net/figure/Bearing-life-vs-bearing-operating-setting-5-Rys-3-Wplyw-napiecia-wstepnego-na_fig3_321364607

I find that bearings which are pre-adjusted (hubs. pedals) are almost always over-tightened, I assume so that people don't complain about loose play as they might consider it to be wear, o rjust the desire to sell more parts.

A lot of the industrial bearings I look at have ground races rather than polished, so I wonder how much difference the surface finish makes? I always wondered if the so-called 'polished' races of the more expensive hubs, etc, was just a marketing gimmick?

Is it worth upgrading the balls to more accurate grades (e.g. Grade 10), after the races are 'run in'?

[Brucey]

...A lot of the industrial bearings I look at have ground races rather than polished, so I wonder how much difference the surface finish makes? I always wondered if the so-called 'polished' races of the more expensive hubs, etc, was just a marketing gimmick?

When it comes to what bike part manufacturers call 'polished' (for bearing surfaces), it is often similar to a good ground finish. In well-designed bearings the ball elastically deforms and so does the raceway slightly, to give a contact area which will support the load without either part seeing damagingly high stresses. Any roughness of the surfaces within this contact area will result in higher stresses than this, and consequent wear, which is why bicycle bearings that are the slightest bit less than perfect tend to 'run in'.

Industrial bearing data is interesting but often not relevant; the reason is that bicycle bearings don't usually suffer from fatigue failures per se (not unless the preload is stupid-high) and furthermore bicycle bearings turn way more slowly than most industrial bearings are either designed for or speeds they are used at. At 'optimum preload' (which is often comparable to the anticipated service loads) in cartridge bearings the balls roll and don't scuff, and at high speeds this is less damaging than reduced preload and more scuffing. Depending on the bearing design the balls scuff against the retaining clips, one another, the raceways, or all three. However at lower speeds it is less clear that reduced preload is always the worst option. In bearing application notes you will find a value called 'static load rating' which is a small fraction of the usual load rating. This is the value that applies to slow moving bearings, and by this definition all bicycle bearings are moving 'slowly'. Accordingly the optimum preload values are likely to be similarly reduced vs a high speed application. Optimal lubricants are different too.

Is it worth upgrading the balls to more accurate grades (e.g. Grade 10), after the races are 'run in'?

Yes it can be, but the original balls (which will have initially been slightly different sizes from one another and slightly out of round) will have been 'running in' too. They may be better than when they went into the pedal.

If you take a pedal with really good bearing surfaces, you can try some experiments. If you assemble the outboard bearing dry, and then finger tighten the cone, you can assess the bearing; there are two tests

a) how does the bearing feel, with increasing preload?
b) how evenly loaded are the balls?

In the former test if the balls are not perfectly spherical or vary in size the bearing won't feel perfect. In the latter test if the balls are of significantly different sizes from one another, you will find a few (often just three) are doing the lions share of the work; the others may be rarely if ever loaded, and may be rattling around uselessly until the service loads are very high indeed, and then they don't (can't) pull their weight. The rest of the time they are just rattling and scuffing uselessly. If you apply a little preload to the bearing (cone finger tight) and then prod the balls you will soon see if some are loose or not.

Years ago I foolishly let water get into my lovely campag NR pedals, and the balls came out rather matte rather than shiny. I was relieved to see that the raceways didn't seem to be badly worn or pitted. 'No problem' I thought, I'll just whack in some fresh 5/32" balls. Sure enough I had a packet of cheap 5/32" balls in my spares stash; "they are lower quality, but they have to be better than clearly damaged Campag balls..." I remember thinking to myself. "How much better are the campag balls anyway...?" I wondered.

Well the bearings felt rather rough and snatchy with the cheap new balls installed; pretty much like cheap pedals in fact. I was depressed and confused when I realised this; I thought I'd ruined my super-lovely pedals. And then it occurred to me that they probably felt worse than they did before I had opened them up. Had I imagined that...? Some reassembly showed that bafflingly, (to me, at the time) they felt much better with the old balls installed again, even though they were clearly imperfect. Reassembling dry and doing a few tests showed what was going on. The cheap balls were not sharing the load, whereas the old balls, even though they were no longer shiny, were still within a micron or two of being the exact same size as one another, still spherical, and were still sharing the load well. So I guess that a hierarchy of ball bearing quality might be (from best to worst)

1) Lightly used high quality balls
2) New high quality balls
3) slightly damaged high quality balls
4) lightly used lower quality balls
5) new lower quality balls
6) slightly damaged lower quality balls.

In other words often, in bicycle bearings, 'roundness trumps shinyness"

I reckon that when you are standing on the pedals it probably causes some slackness and scuffing in the bearings unless the preload is set incredibly high. Fortunately this kind of loading doesn't normally constitute much of the duty cycle in most cyclist's use, so the net wear is unlikely to be that bad either. The loads on the pedals when seated are single figures or low tens of kg, so an optimal preload for this use is likely to be similar. As I mentioned previously if the balls are a good size (eg 5/32") then moderate service loads don't need to be shared between balls much. If the balls are smaller (1/8" or even 3/32") then it is probably more important. Even so, a 'set to zero preload, adjust when loose' protocol seems to give a good service life to the bearings.

Edit; I guess the exception may be when the pedal spindle is shortened, and the inboard bearing is not moved outwards similarly, so as to keep the centre of pressure midway between the two bearings. In this case the outboard bearing sees much more load than the inboard bearing, and (for any given preload) the pedal pressure required to cause the balls to scuff is greatly reduced. In this case more preload is required if bearing wear is to be avoided. By and large good pedal designs avoid this design feature, but sometimes it creeps in. For example MKS have unevenly loaded bearings in such pedals as the old AR2 model, and of course some of their folding pedals too.

cheers

[Cowsham]

I have some Campagnolo Nuovo Record quill pedals that must have had about a quarter of a million kilometres on them by the time I completely switched to clipless pedals about 20 years back. I still have the pedals, and the bearings in them feel just as good as when they were new. I used to service them more or less annually. Those pedals don't have seals where the spindle exits the pedal body, but the spindles are rifled at that point so that the rotation of the spindle relative to the pedal body helps to keep crud out of the inner bearings.

Damn good idea -- my old vw is the same at the crankshaft