Yet Another Cycling Forum
General Category => The Knowledge => Topic started by: quixoticgeek on 31 August, 2017, 09:12:52 am
-
Last night a friend asked me to look at their bike as the steering felt a bit rough. I disassembled the headset, and discovered some of the bearings in the cages were no longer exactly round, infact, one wasn't even close. I've dispatched said friend to the local bike shop to buy some new bearings for me to fit. But this got me wondering.
I run regular bike maintenance workshops at the Amsterdam Hacker Space (techinc.nl), and I was pondering the idea of holding a stock of common ball bearings to make fixing things like this easier. Looking at the likes of sjs cycles I can get shimano branded bearings in amounts needed for a single headset, or a single hub, or for about a quid or 2 more I can get 100 balls. But if I goto ebay or ali express I can get the 100 balls for about half as much.
Does paying more get me more? Is there a false economy?
J
-
Ball bearings come in different grades, which means how much variation in size any individual ball may have. Ideally, every ball should be exactly the same diameter and roundness, so that the load is evenly carried by all of the balls, rather than just the very biggest ones. High quality equipment deserves Grade 10 or Grade 25 balls. Cheap equipment isn't aligned well enough for quality balls to make a difference.
Buying small numbers of ball bearings is kind of useless. I buy high grade balls for everything as the individual cost is so low when buying in quantity.
-
there are also different materials that are used to make balls. Some balls are carburised and others are made of (direct hardening) chrome steel, or stainless steel (which is slightly less hard).
For less expensive hubs and other parts with pressed steel races (which are somewhat flexible) the balls are either loaded one-or-two-at-a-time by design (most hubs) and/or the raceways are flexible enough (eg with many headsets) to help share the load even if the balls are not exactly the same size.
So for run-of-the-mill parts cheap balls are usually OK; e.g. use weldtite ones which are available in tubs of ~500 for not too much money.
For posher (more rigid) parts it is better to use good quality balls; get these from a bearing supplier eg 'simply bearings'.
BTW if you can assemble the component with (more) loose balls in place of (fewer) clipped balls, the result is much better load sharing and there is zero chance that the clip will break up and destroy the bearing. Cheap headsets really benefit from being fitted with loose balls instead of clipped balls.
BTW the balls wear slightly faster than the raceways in many cases; this means that (under otherwise benign conditions) cheap balls may wear to be more uniform in size and the bearing may 'improve itself' as time goes on.
cheers
-
I keep loose bearings in a sorting box, the kind with little sections. IIRC there are four sizes used on a bike. It's usually a particular size for a headset, a size for front hubs, a size for rear hubs and a size for bottom brackets. Some rear hubs are complicated by using different sizes drive-side and non-drive-side. If I'm taking bearings out, I'll measure the diameter with a vernier caliper, then label the hub so I know what size to install.
The clipped bearings, aka caged bearings, that come installed on cheaper bikes, suffer from deformation of the cage. The cage is made from thin sheet steel that's cut and folded to shape. Because it's thin, it's bendy and if any of the little prongs are bent out of shape, the cage can foul the headset surfaces*. Even if they're straight, the cages can corrode in situ, which can cause them to snap or disintegrate, causing further problems.
* A few weeks ago I assembled a headset with caged bearings - and then disassembled it because it was tight when I turned the fork. I reseated the headset in the headtube and regreased everything, but it was still catching. More measuring, adjusting, regreasing, again with no joy. As a last ditch attempt, I swapped the caged bearings for loose balls, tightened everything down and it was silky smooth. It was a 'why-didn't-I-think-of-that-before' moment.
-
Ball bearings come in different grades, which means how much variation in size any individual ball may have. Ideally, every ball should be exactly the same diameter and roundness, so that the load is evenly carried by all of the balls, rather than just the very biggest ones. High quality equipment deserves Grade 10 or Grade 25 balls. Cheap equipment isn't aligned well enough for quility balls to make a difference.
Buying small numbers of ball bearings is kind of useless. I buy high grade balls for everything as the individual cost is so low when buying in quantity.
Is there a logical setup to the grades? I notice simply bearings offer grade 100, and sjs grade 10. Which is better?
J
-
there is a nice Wikipedia article about steel balls. Gr10 is better than Gr100.
It is a waste of time and money putting Gr10 balls in a cheap, worn, headset BTW.
cheers
-
there is a nice Wikipedia article about steel balls. Gr10 is better than Gr100.
It is a waste of time and money putting Gr10 balls in a cheap, worn, headset BTW.
Ah yes, that makes sense.
Is there any reason to go with stainless over chrome steel?
J
-
Is there any reason to go with stainless over chrome steel?
not really; it just encourages the (non stainless) raceways to corrode preferentially, which you don't want to happen.
cheers
-
not really; it just encourages the (non stainless) raceways to corrode preferentially, which you don't want to happen.
That makes sense.
Will place an order from simply bearings.
I know my hubs use 1/4" (rear), and 3/16" (front). Are there any other sizes that are commonly used on bikes that it's worth me having in stock?
J
-
Headsets are, umm, smaller, but I don't know what size.
-
There was a young man of Devizes
Whose balls were of quite different sizes.
One was small
And no use at all
The other was huge and won prizes.
Sorry.
-
not really; it just encourages the (non stainless) raceways to corrode preferentially, which you don't want to happen.
That makes sense.
Will place an order from simply bearings.
I know my hubs use 1/4" (rear), and 3/16" (front). Are there any other sizes that are commonly used on bikes that it's worth me having in stock?
J
Major uses are;
3/32" for SPD, SPD-R pedals, detent balls in some shifters
1/8" for freewheels, freehub bodies, some headsets, some pedals
5/32" for pedals, headsets
3/16" for front hubs, some headsets, some IGH bearings
7/32" for some front hubs, some rear hubs, some IGH bearings (eg RH bearing in Nexus 3 hub)
1/4" for rear hubs, most IGH bearings, some front hubs, loose ball bottom brackets
There are many oddball components that use ball sizes not indicated above. 'Some' indicates a reasonably common usage but not the most common.
3/32" and 7/32" sizes are comparatively unusual so if you just want to keep the four most useful sizes 1/8", 5/32", 3/16", 1/4" will do.
This morning I rebuilt the headset in a bike with a 1-1/8" hidden-style A-headset. As per usual (I think the A-head design is a rubbish idea for bikes that see weather.... ::-)), water had run down the steerer into the bearings and ruined them. The clips were breaking up and there was rust everywhere, with some pitting in the raceways. However application of a handful of these
http://www.ebay.co.uk/itm/Weldtite-Loose-Bike-Ball-Bearings-Bulk-Tub-Sizes-1-8-5-32-3-16-7-32-1-4-/180778908700 (http://www.ebay.co.uk/itm/Weldtite-Loose-Bike-Ball-Bearings-Bulk-Tub-Sizes-1-8-5-32-3-16-7-32-1-4-/180778908700)
in 5/32" size and some decent grease restored normal function. Headsets like this are little ball-bearing piggies; I needed about eighty balls to fill the races. As per previous comments, there is no point in using better balls than these, not on cheap headsets.
Less than £1 worth of balls should enable that headset to outlast the rest of the bike....
cheers
-
And possibly as much as 5GBP if you use noticeably better balls, which is not very much anyway. That is why I buy bulk high quality bearings, regardless of what they are going into.
-
Campag hubs are supposed to come with Grade 25 balls but I've never seen these listed by Simply Bearings. You can also get ceramic bearings but I suspect they are utterly pointless at bikr component speeds.
-
This is disappointing.
-
re campag ball quality. I don't know what the current state of affairs is but the legend on every box that Nuovo Record hubs came in used to say 'con tolerenzia di un millesimo' which roughly translates to 'with tolerances of one micron'. This could be hyperbole, but if not it suggests that the balls inside are Grade 5 or better.
I let some water get into some NR pedals and I tried to replace the balls because the old ones had started to look dull; the pedals didn't run smoothly again until I put the old balls back. Investigation showed that with cheaper balls, only three were touching both raceways at once.
Re using cheap balls in cheap, knackered headsets; when the raceways are lumpy to the tune of 100 microns or more, ball quality is almost irrelevant. If it were a hub or something it'd be scrap, or due an inspection after another thousand miles, tops. But with the right grease (loaded with solid lubricants + EP additives) and a careful setup, a headset will work OK.
The reason the grease needs solid lubricants and EP additives is that in a bearing with lumpy raceways, you want the balls to slide easily if they need to. This is (at low speeds/repetitions) preferable to a pure rolling contact if the balls can't roll in the correct way, e.g. because the surfaces are so lumpy. Sliding and scuffing (as permitted by solid lubricant additives etc) are however fatal to high-speed bearings, which is why you won't find bearing greases with high quantities of these additives; a grease that is meant for something like an exposed gear train (including low speed bearings) is a better idea.
Whether bearings like hubs etc are more like low speed gear trains than high speed bearings I don't know; but I do know that nothing I have put the gear train/low speed exposed bearing grease in has unexpectedly worn out or broken yet, and it has been a long time....
There is arguably little point in spending a fiver on balls for a cheap headset; a new headset will cost about £10 or so (but may take hours of wasted time to correctly identify and source these days.... ::-) )
A Brucey top-tip; if you are converting a new/unworn headset of reasonable quality (eg a Tange one) from clipped balls to loose balls and you only have cheap balls to hand, you can do yourself a favour by
a) carefully extracting the balls from both clips
b) use those balls in the lower race
c) use the cheap balls in the upper race
The resultant headset is usually better than it would be with cheap balls throughout, and probably as durable as if it had posh balls throughout; the lower race is the one that benefits most from better balls.
cheers
-
As fretting is what kills lower races, a thin grease would also be better as it's more likely to "self repair" when displaced. I wonder if they ever went bad when they were just oiled from beneath?
-
As fretting is what kills lower races...
I know JB thought so, and argued his case accordingly, but I happen to think that
a) his circumstances were unusual
and
b) he was mistaken in his observations
If you use a thin oil as lubricant in a headset, you are doing exactly the wrong thing.
cheers
-
I think JB is right on this, because my conventional headsets on road bikes all went "indexed" very quickly yet those with 45 degree* plain bearings to accommodate fork flex, cartridge or not, have given no trouble, 1" or 1 1/8". Since the original headset design was thought to be ok for 100 years or so until this was fixed, I surmise that it may have been grease lubrication that highlighted the problem. On old steel 3-speeds you just flipped the bike over and dribbled oil into the races occasionally; those bikes never suffered from indexed steering AFAIK, although they were oily and filthy.
*some use 60 degrees
-
I think JB is right on this, because my conventional headsets on road bikes all went "indexed" very quickly yet those with 45 degree* plain bearings to accommodate fork flex, cartridge or not, have given no trouble, 1" or 1 1/8". Since the original headset design was thought to be ok for 100 years or so until this was fixed, I surmise that it may have been grease lubrication that highlighted the problem....
Plain bearings are not used on any bicycle headsets, to my knowledge. The cartridge bearings used almost invariably contain balls; very often there are not enough of them and they are too small; they brinell and fail quite easily.
JB was mistaken in his observations; damage occurs to some headsets in a far shorter period of time than would be able to cause fretting. He said that 'it was obviously not brinelling' but IMHO he didn't look carefully enough; the raised area around a noticeable brinelled indent is only a few microns high and is easily missed on an angled/curved surface.
'The problem' only really exists if you use a long/flexible steerer, in combination with too few/small balls and perhaps bad adjustment. Pre WW
-II the most common headset type had a zero degree contact angle and the races were 'drop in' type, that could articulate if the steerer flexed. Not that it would, the steerers were made of thick steel and were mostly quite short. Roll on a few decades and longer steerers that were built more flexible were the norm, and headsets that didn't have races that would articulate.
JB's circumstances were that he had a thin-gauge steerer about twice as long as normal on a bike with skinny tyres that he regularly took off-roading. As soon as I saw that I understood why he knackered every conventional 1" headset that he ever had, even of types that I know are strong enough to use on a tandem or will last 100000 miles on a normal bike....
The arguments that grease is not an adequate lubricant and that there isn't enough movement to replenish the lubricant film are completely specious; no-one rides their bike in a perfectly straight line and if they did they wouldn't find self-centring steering such a PITA.
If fretting were the sole cause of headset troubles then it would appear all round the raceways and it would appear on pretty much every bike. Neither of these things is true; draw your own conclusions...
cheers
-
The plain bearing is between the cups and the cartridges, or the cups and the loose races (à la Stronglight A9). That takes the flex and the rolling-element bearing, ball or roller, handles the rotation.
Can you "brinell" a headset with a lump hammer? Brandt says not but I've never tried it.
-
I see what you mean re your description of 'plain bearings' ; it wouldn't be the phrase that I'd use to describe the arrangement.
FWIW I've seen many A9 headsets and some of them have worn in such a way as you might think it was fretting, but it was probably corrosion-related. Also I've yet to see one which exhibited uniform wear marks on the lower race, indicating very strongly that the lower race was normally articulated off-axis in service. I think the cone angle was too shallow in such headsets, at least for short or medium length head tubes; I think headsets of this type suit longer head tubes somewhat better. I have not examined enough 45/60 degree angled seat cartridge bearings to spot any trends in uneven wear through articulation.
JB's experiment with a lump hammer is misleading; the races will articulate/flex as the blow is struck to share the load reasonably well; by contrast in service the flexing steerer ensures that the load is unevenly distributed on the balls, such that true brinelling is possible (which I have seen, displaced material and all). The load per ball is increased by preload and flexing, way over and above what is 'normal' from service loads divided by the number of balls alone would suggest.
cheers
-
This is disappointing.
Does this make it better?
https://youtu.be/z_jdiU47bFA
-
This is disappointing.
Does this make it better?
https://youtu.be/z_jdiU47bFA
:D :thumbsup:
-
Sometimes a cheap fix for brindalled headsets are slightly bigger balls.
-
The quick fix is to fit loose balls, then you get a slightly rough, but not actually indexed, headset: the balls can't all line up with the pits in the races.
The worst headsets I've used were Shimano 1055 in 1" threaded: I had a couple and they were both indexed within 500 miles. Still rideable, but annoying. The very low stack height meant not much else would fit. In the end I got a new fork.
The Cane Creek S6 on my commuter has been in place for 8 years with no attention whatsoever, and is still index-free although the outside looks awful due to road salt and (ahem) abrasion from the front brake cable housing. I think it might deserve new cartridges this year - I have them in my spares box.
-
We swapped out the ball bearings on the headset of my friends bike on Saturday.
Having extracted the old ones, I took some photos.
(http://photos.quixotic.eu/Techinc/ballbearings01_sm.jpg)
(http://photos.quixotic.eu/Techinc/ballbearings02_sm.jpg)
I'm not sure exactly how, but I'm pretty certain they aren't supposed to be that shape...
J
-
almost anything will be an improvement over those!
BTW some folk describe those balls as having 'shelled' because it looks as if a layer or shell has flaked off the outside of the balls. I have seen balls from a number of different manufacturer's hubs and BBs fail in a similar way.
It may reflect a certain method of manufacture or a failing therein (eg case-hardened balls with excessive residual stresses in them) or it could be pot luck if they fail like that.
cheers
-
Once rust has eaten away the outer hardened layer (assuming they're only case hardened), the balls can spall quite easily. Rust is no respecter of hardened steel - it ruins it just as happily as if it were mild steel. Only the chromium content slows it down.
-
there is no sign of any corrosion on QG's balls.
cheers
-
Excuse the thread necromancy, does anyone know what grade of balls are used by Shimano?
I've just done an order of stock as I realised I was out of some sizes. I couldn't find 3/32" in other than grade 100.
Is there any issue with using higher grade balls than originally intended? I went for grade 10 for everything else.
J
-
IIRC you should be using Gr25 Chrome steel bearings if you want like-for-like replacements. You can buy these on e-bay in 3/32" size, but tbh you may as well buy a bag from shimano.
FWIW I would under no circumstances
a) mix balls in SPD pedal bearings i.e. used ones from one bearing with another and/or
b) mix balls from one bag of new ones with another
If the balls don't look marked in any way, I'd vote for keeping the original ones in fact.
FWIW one of the kinds of horrible experiment I have carried out is to see what happens if I try to 'recover' knackered SPD pedal bearings. The first set I tried had oval cones and after a period of use with better quality grease (full of solid lubricants / EP additives) and better adjustment, they 'wore back' so that they were actually pretty good. In the end I used them for a further year or so and they didn't need adjusting again in that time. A good result.
Obviously those pedals weren't quite knackered enough; second experiment was to take a set of PD-M324s where the balls had run over one another in one pedal and had mangled both the cone and the raceway. These pedals were so bad that I wasn't sure if I could assemble them so they would work at all, leave alone be usable or improve with further use. That experiment is ongoing. No point in using new balls in the course of this experiment, so old balls went back in. Interestingly some of the balls that went back in were already marked. After a thousand miles or so I cleaned the grease out and sure enough there was plenty of sharpnel in the grease. However examination of the marked balls showed that there were no sharp edges on the marks; this suggested that the conditions in the bearing were stabilising. Currently the bearing is still a bit rumbly but it is a fair bit smoother than it was doesn't seem to be wearing too quickly, and it also adjusts without bad loose/tight spots. I guess the bearing is due for another examination any time soon. If the bearing surfaces are good enough I will put better ball bearings in this time.
Needless to say these bearings only work properly if the balls are able to share the load. This only happens if two conditions are met
1) that all the balls are exactly the same size and
2) there is no slack in the bearing.
IME the slightest free play in SPD pedal bearings normally causes greatly accelerated (ot to mention uneven) wear.
cheers