Bicycle Mechanics - threaded headset is too tight, but fork still has play

I recently bought a late 80's Centurion off of eBay, and have been overhauling it in my spare time. After stripping the frame and applying frame saver, I just repacked the headset. It's a 105 threaded version with cage bearings. Pretty standard design for the time, relatively light with steel bearing races. The races themselves are smooth, not pitted at all, and the bearings look to be in fine shape - not perfectly smooth, but close with few scratches. The bearings are in retainers, and there's not any bearings missing.

After I packed the headset with grease and began adjusting it, I found that there was no happy medium between "loose with play" and "tight so won't turn easily". If it's tightened so that the fork turns comfortably, there's a fair bit of play when I hold the fork tips and alternately pull and push them with respect to the main frame. If I tighten the headset enough so that there's no play - and I can barely get it this far - then the fork doesn't turn easily. Though, in this state the fork still turns smoothly - the resistance is constant - but there's a fair bit of resistance.

Now, I don't think that the amount of resistance will make steering difficult, and riding no-handed should still be fine. However, I'm worried that
(a) the headset might be too tight, which will cause it to pit and wear out
(b) or perhaps the bearings are worn down and aren't wide enough to properly fill the open space betwen the male and female bearing races, thus resulting in the play

So, does any of this sound problematic? I've got a fair bit of mechanical experience, but haven't worked in a shop in 10 years and haven't had this issue with threaded headsets in recent experience. I overhauled the LX headset on my good bike this summer - it has cartridge bearings and I can get it adjusted so that there is no play and the fork spins 10 rotations with a good flick of the wrist. But it's a very different (and more expensive) design.

Are the cups tight in the headtube?

Is the fork crown-race on the fork securely?

Make sure that the bearing-retainers aren't in backwards...

Make sure that the bearing-retainers aren't in backwards...
Good thing to check, but they're in correctly. And if they weren't, I don't think it would spin with constant resistance.

Are the cups tight in the headtube? Is the fork crown-race on the fork securely?
I haven't tested the cups or fork-crown race, but I didn't remove them from the bike or do anything that would loosen them (unless application of frame saver would affect stuff here, which I doubt).

Good thing to check, but they're in correctly. And if they weren't, I don't think it would spin with constant resistance.

Did you use loose ball bearings or retainers and are they in correctly. If you used loose bearings and have too many in that would cause what you have described.

If the cups and the crown race are not loose, my money would also be on the bearings. Either upside down, wrong cage(Some of them will contact the race) or wrong size. The only other thing I can think of is a seal is upside down.

Even if the bearing cages are upside down they will still be suprisingly smooth as long as there is plenty of grease in there.

Did you use loose ball bearings or retainers and are they in correctly. If you used loose bearings and have too many in that would cause what you have described.
Retainers. And yeah, they're in correctly.

The bearings are the originals that came in the headset; they looked to be fine, I couldn't see any reason to not re-use them.

Also, I assume that the top and bottom races and bearings are equal (they look it), but just in case, I switched the bearings+retainers between top and bottom. Symptoms exactly the same, though.

Did you forget to put back a bearing or two?

Did you forget to put back a bearing or two?
Bearings are in retainers; they're all in there.

Do you have a finger caught in the retainer?

The Shimano retainers are weird, they're the reverse of normal kind. The solid ring of the retainer actually faces down into the cup, so that the fingers curl up at you. Easy test is to use loose bearings and see if it still does it.

Could also be they substituted another top-race from another headset and the adjustable cup doesn't quite fit it properly... or the crown-race...

I'll try flipping the retainers (or at least look at them again, to see if flipping would make any sense), but I'm pretty sure they're in correctly. I may buy new (loose) bearings after I've exchausted everything else.

But I'll have to pull both of my big toes out of the retainers first ;)


Could also be they substituted another top-race from another headset and the adjustable cup doesn't quite fit it properly... or the crown-race...
This is something I'll have to look at. The fork and the headset are both original, so I doubt this is the case, but that fits the symptoms. The play is at the bottom race, not the top race (and this bike has a 21cm head-tube).

Could also be they substituted another top-race from another headset and the adjustable cup doesn't quite fit it properly... or the crown-race...

This is something I'll have to look at. The fork and the headset are both original, so I doubt this is the case, but that fits the symptoms. The play is at the bottom race, not the top race (and this bike has a 21cm head-tube).
Actually, this seems to be the problem. Good call. There's a lot of play at the bottom race, and it's because of the crown-race being too small. As more evidence to this point, the crown-race is silver, while the race in the top of the head-tube (the one that does fit fine, and has no play) is black. So it appears that the crown-race isn't original to the headset.

So, what's the odds of finding the correct race for this through my LBS? If so, I'll do that - it's a nice headset - but it's probably easier and cheaper to just install a new headset. I've got a YST threaded 1" headsit sitting around. Not quite as nice as the 105, but it's good and not too heavy.

My guess is someone threw in the wrong size bearing retainers at some point.

I hate to be the bearer of bad news, but my bet is that your frame has an ovalized headset. If you ride a bike with a loose headset for too long, it will reshape the ends of the headtube. Get a good caliper and try measuring the top of your headtube front-to-back and side-to-side.

My guess is someone threw in the wrong size bearing retainers at some point.
I don't think this is the case; the bearing retainers are exactly the same top and bottom, and the top one fits perfectly, and the bottom one is loose. Also, the crown-race (which I think is the cause of the looseness) is silver while the race at the top of the headtube is black; the crown race looks like it's a different product than the rest of parts of the headset.
-

I hate to be the bearer of bad news, but my bet is that your frame has an ovalized headset. If you ride a bike with a loose headset for too long, it will reshape the ends of the headtube. Get a good caliper and try measuring the top of your headtube front-to-back and side-to-side.
I'll check on this. Hope it's not the case. If this comes from having a loose headset generally, would it also be caused by this "special" sort of looseness from having a crown race that's a bit small?

if the thing aint ovalized, repack it with loose balls. can't really go wrong, and will spin better.

to test the ovalized idea, if you don't have quick access to calipers, check for play front-to-back then side-to-side.

if the thing aint ovalized, repack it with loose balls. can't really go wrong, and will spin better.
Will loose balls allow me to use a race that's too small?

Edit: just talked to mechanic at my LBS, and the answer to this is No.
Also, he said that if the head-tube were ovalized, then I should be able to just pull the cup out of the bottom of the head tube... which I can't. So it looks as if this isn't ovalized.
-

to test the ovalized idea, if you don't have quick access to calipers, check for play front-to-back then side-to-side.
This isn't a perfect test because there's play to begin with, but it seems that there's a little more play front-to-back than side-to-side... but that might be just because I'm more easily able to brace the frame in front-to-back plane. Not a big difference.

Why reuse bearings when they're so cheap? Throw away those retainers and put new loose balls in it.

Prior to you disassembling and repacking the headset when you got the bike, how did the headset adjustment feel?

Your bike has the original fork, as I recall from the photos, so it seems unlikely that they put an alien crown race into it. The finish on the upper and lower crown races is not necessarily going to match. It's been a long time since I dealt with that generation of 105, so I don't recall whether it matched on 105.

I've seen SOOO many newbies get their retainers upside down and be firmly convinced that it's in the right way because it "looks right" that way. So I'm with the guys opining that's what happened there.

a bent steerer tube can cause similar results, binding in conjunction with looseness.

if the head-tube were ovalized, then I should be able to just pull the cup out of the bottom of the head tube... which I can't. So it looks as if this isn't ovalized.



Not true. It can be ovalized and still hold in there tightly. I had an ovalized tube that resulted in a lot of play front to back. It was solved by pressing a shim, cut from a soda can, into the head tube along with the cup.

I think the doomsday sayers may be a bit premature. Before you start measuring for ovalized headtubes etc, just put loose balls in there. Its the simplest test. And if you haven't already, be sure all the cups and races are secure.

@Dirtdrop and San Rensho: I'll put loose balls in, at least on the bottom. Those are the bearings that experience higher stresses. But it's still weird that the bearings with retainers aren't big enough to fill the gap between crown-race and bearing cup, when these are the bearings that came with this era of Shimano headsets.

@nick burns: good question, but I didn't carefully feel the headset and fork play before stripping the bike down. It felt fine, but there wasn't any careful inspection.

@mechBgon: you're right, the diff finish isn't necessarily an indicator that the crown-race isn't the original 105. But combined with the clear bad fat, it seems more likely.
However, the fact that all of the other parts on the bike are original, and that no bearings were pitted and the caliper brakes were susprisingly clean (I barely needed to touch them, certainly no overhauling), it does seem unlikely that the crown-race was ever replaced. That said, why is it clearly loose then?

@dafydd: I'll check on the possibility of a bent steerer tube.

@jbone: good to know that an ovalized head tube could still hold the bearing cup tightly. That was my intuition on this, but I don't have nearly the experience that this mechanic does.

I hate to be the bearer of bad news, but my bet is that your frame has an ovalized headset.

It doesn't sound ovalized to me. He said the high resistance is constant. An ovalized headset gets tighter as you turn the fork.

It doesn't sound ovalized to me. He said the high resistance is constant. An ovalized headset gets tighter as you turn the fork.

You will not feel any difference is resistance when turning the fork if the head tube is ovalized. The tube is what is ovalized therefore the headset cups have a small amount of play in them. The cups are not ovalized so you have a circle sitting inside an oval and you need to fill that gap. That is why you put a shim in there, to take up that space and prevent the cups from moving back and forth. That is why he is still getting play no matter how tight the headset is. Chris King has made headsets with a 1" depth on them specifically for this purpose. The longer depth gets the headset beyond the ovalized portion of the headtube.

Yeah, ovalization wouldn't make the resistance non-constant; it would just cause there to be resistance.
It seems, however, that the forces that caused the head tube to ovalize could have caused the upper part of the bearing cup (the part that fits in the frame) to ovalize with it. Especially if this part is aluminum. It won't cause the bearing cup itself to ovalize, because that's much wider and therefore exponentially stiffer.

That said, why is it clearly loose then?
Because one or both bearing retainers are upside down. That's where the smart money is on this one. ;)

Try this: remove the fork. Leave the top headset parts in place. In this case, the cone is in the frame, the cup is on the part that's normally on the fork. Push down hard on the cup and try to turn it. Are the balls rolling?

Now set aside the top part of the headset. Flip the bike over. Drop the fork through the head tube and do likewise. Do this without the bearing seal in place, and you can peek into the headset as you forcefully push down on the bearing and turn it. Are the balls rolling?

That said, why is it clearly loose then?Because one or both bearing retainers are upside down. That's where the smart money is on this one. ;)
Try this: remove the fork. Leave the top headset parts in place. In this case, the cone is in the frame, the cup is on the part that's normally on the fork. Push down hard on the cup and try to turn it. Are the balls rolling?
Now set aside the top part of the headset. Flip the bike over. Drop the fork through the head tube and do likewise. Do this without the bearing seal in place, and you can peek into the headset as you forcefully push down on the bearing and turn it. Are the balls rolling?
Nope, they're in correctly, as stated above. Top bearings spin smoothly and perfectly; resisntace is coming from the bottom even though the retainer is in correctly. I'll try the 2nd part of your suggestion, however, to see if the bearings are actually spinning. If they're not, though, it's not because the retainer is upside-down, but rather because either the crown race is too small, or the bearings are too small (a too-small crown-race could be rectified by larger bearings that fill up more space between crown race and bearing cup).

Nope, they're in correctly, as stated above. Top bearings spin smoothly and perfectly; resisntace is coming from the bottom even though the retainer is in correctly. I'll try the 2nd part of your suggestion, however, to see if the bearings are actually spinning. If they're not, though, it's not because the retainer is upside-down, but rather because either the crown race is too small, or the bearings are too small (a too-small crown-race could be rectified by larger bearings that fill up more space between crown race and bearing cup).Errr... don't quit the day job. ;) Using larger bearings than the races are designed for is the incorrect solution. If in doubt, get your LBS to look at it or slap the other headset in there.

edit: also, I believe the seal for that headset has an "L"-shaped cross section, so ensure that it's in the right way too. Good luck :)

Errr... don't quit the day job. ;) Using larger bearings than the races are designed for is the incorrect solution. If in doubt, get your LBS to look at it or slap the other headset in there.
Good point; I've not done actual shop mechanic work for 10 years, but should have been able to figure that out with a little 3-D mental visualization. However, I was thinking there might be some chance the bottom race of the headset originally came with loose balls that are larger than the ones in the retainer that's in there right now. That's an alternate explanation to the looseness, if the crown-race is actually the original.
I've actually got lots of physics/engineering and biomechanics training and experience, and tested protopyte running shoes for Adidas and Mizuno during college, and did a bit of consulting for Nike on this stuff too.
I still think the main thing is that the crown-race is wrong and too small, for whatever reason. That's the only plausible explanation I can think of, though it's baffling how the wrong race would have gotten on the fork crown.
I've got a few headsets sitting around that I could replace this with, but if my LBS has the correct race sitting around (unlikely for a late-80's headset, but possible) I'll just have them replace that. Tools for installing and removing headsets aren't worth having for a home-mechanic, at least not one who works at low volume, just because of economies of scale.

I've seen SOOO many newbies get their retainers upside down and be firmly convinced that it's in the right way because it "looks right" that way. So I'm with the guys opining that's what happened there.

balls facing away from cups !!

in case anyone was embarrassed to ask.

I still think the main thing is that the crown-race is wrong and too small, for whatever reason. That's the only plausible explanation I can think of, though it's baffling how the wrong race would have gotten on the fork crown.
I've got a few headsets sitting around that I could replace this with, but if my LBS has the correct race sitting around (unlikely for a late-80's headset, but possible) I'll just have them replace that. Tools for installing and removing headsets aren't worth having for a home-mechanic, at least not one who works at low volume, just because of economies of scale.You can use big screwdriver and hammer to remove cups and the fork-crown race (tap lightly and move to opposite side). A headset press can be made with a big bolt and washers.

It's very possible that the original crown-race was pitted and they replaced it with another one that was laying around. If the diameter and curvature of the race doesn't match properly, the balls may not be rolling on the right spot leading to looseness. What you can do is test where the balls are rolling on the race.

Remove fork and wipe off all the grease from the crown-race. Remove the retainer/bearings and clean off all the grease from the cup, put retainer/bearings back in and wipe most of the grease off, there should be a bare minimum. Then stick in the fork, instlal adjustable cup to hold in place and spin it by hand around a couple times. Disassemble and inspect the track of grease showing where the balls rolled. You can also wipe it all down and use prussian-blue to make the trace as well.

The movement should be in the middle of the curved part. If it's too wide, that would mean the crown-race is too small in diameter for the cup. Larger bearings would help in this case. Also be aware that there's a difference between metric vs. SAE sized bearings...

...It seems, however, that the forces that caused the head tube to ovalize could have caused the upper part of the bearing cup (the part that fits in the frame) to ovalize with it. Especially if this part is aluminum. It won't cause the bearing cup itself to ovalize, because that's much wider and therefore exponentially stiffer.

I didn't say "headtube" - I said "headset" in response to the poster who said "headset". The cone races *can* ovalize which results in increasing tightness as you turn the fork.

Also timcupery, if you could post a good close-up photo or two of the crown race...? A picture's worth a thousand words :)

Also timcupery, if you could post a good close-up photo or two of the crown race...? A picture's worth a thousand words :)
crown race is silver, and not pitted. I don't have time to take the thing apart for pictures right now. But the curvature of the crown race is slightly diff than the curvature of the race in the top of the headtube.
I'll get pictures if I get the time.

Okay all, I figured out the problem. I'd flipped the headset seals. I'm surprised that very thin plastic could cause that much resistance when the bearings weren't tightened, but it did, and it does. I just replaced the headset last night, and put an XT threaded 1" with cartridge bearings in. Having the cups and bearing races of the old headset out allowed me to mess with them easier, and I noticed that the bearings fit perfectly (smooth spin, no play) when the headset seals were out, but not when they were in. Then I tried flipping the seals, and it worked.

So, feeling a bit stupid. But I guess no one here wondered about the seals either. I'm out $10 for the headset on eBay, and it's a better, lighter headset that should last a long time. And out $20 to build my own working-on-headset tools, but I'll use those in the future. And I've got a nice headset sitting around now. Mainly I'm out lots of time.

Not quitting my day job,

Tim