I recently changed my stem and tightened up the headset in the normal way, but it ended up a little bit tighter than before. The wheel will still fall to one side under its own weight, jut a bit more slowly. I actually quite like the slightly stiffer steering, which seems to make the bike a bit less twitchy.

Is this a bad idea, and could I do any damage to bearings etc. by having the headset "firm" rather than "floppy"?

Thanks!

As long as the headset is totally free, though maybe a bit slow you're OK.

However, any binding, even very slight binding can make the bike a bit harder to handle. Riding a bike involve countless tiny steering corrections on a continuous basis. These are controlled by your cerebellum and you do them with the same level of consciousness given to tasks like walking or eating ---- namely none.

If the steering is sticky, then you have to overcompensate to make those tiny corrections, and the handling becomes jerky, as if you were drunk.

So, you want a headset that's free enough to respond immediately and precisely when you ride no hands and steer by leaning the bike. If it meets that test, all is OK, if not, back it off a hair.

.
...what you are talking about is called "preload". I generally preload all my headset bearings, and have yet to experience anything bad from the practice. That said, some bearings are more amenable to preload than others. The flop test is as good as any to judge whether you're still in the ballpark.

It's a matter of how much preload. "Preload" refers to any load applied to a bearing before locking it; bearings generally have some amount of play when uncompressed, so if you don't preload a bearing at all, it will end up loose and rattly.

You want a little bit of 'preload' but my experience has been that the times I've erred on the tight side have been when I've later gotten the 'indexed steering' problem where there are tiny depressions in the bearing race and the balls tend to fall into them - i.e. preferring the straight-ahead orientation. One explanation for the depressions is that if the lubricant is squeezed out of the ball-to-race interface and you're going straight on a rough surface the vibrations can lead to fretting of the race at the contact of the ball.

...I've heard this so many times on this forum, I'm not going to take much time to argue the point.

Long story short, I have trouble believing it as a cause and effect for indexing in a headset. I could be wrong, but my own impression is that most of these cases that I've seen are caused by "false brinelling".

There are many discussions immortalized online with regard to this business of headsets and what is or is not the correct adjustment load to prolong life. Here is one. The lubrication issues that lead to false brinelling have much more to do with the way a road bike is used, with little side to side motion to redistribute the grease, than with the tightness of your preload squeezing out the lubricant. As stated, my opinion...but it is not an unfounded one.

With regard to true brinelling, as stated by John Allen in the link, you have to hit something pretty hard to cause that in your average hardened steel headset race.

I can understand why you won't argue the point, because you and Prathmann basically agree. Fretting and false brinelling are two terms for the same process.

It's a common issue in stationary bearings, which is basically what a headset is. The slight movement when riding isn't enough to ensure good lubrication at the points of contact and the constant road vibration does it's damage.

The way to prevent it is to be a lousy rider who can't hold a line and constantly swing the front wheel side to side .as you wobble down the road. I understand alcohol helps

...where we disagree is on whether the degree of preload is a contributor. As stated, I'm doubtful that it is.

Note that your source above states: "Dimpling occurs more easily with a correctly adjusted bearing than with a loose one that rattles and clunks. Rattling replenishes lubricant between balls and races, something that would otherwise occur less easily."
Seems reasonable to me that a bearing with just a bit of preload would still allow a little movement of the balls under the vibration of a straight ride on a rough surface while one with enough preload to make it feel stiff (as in the OP's case) would be more likely to have the lubricant squeezed out at the contact point. And that's been my experience in practice - the only times I've had a problem develop was with headsets adjusted a bit tighter than usual.

This may be OCD, but I always check the headset adjustment before a ride -- turn the bars all the way back and forth through their range to check for tight spots, and then hold the front brake and rock the bike fore and aft to check for looseness. I'd like to think it helps redistribute lube.

Go ahead and believe that if it helps you feel better. Certainly the OCD headset check can't hurt.

However, any benefit gained is gone within the first few minutes, with the protection time interval depending more on the specific lube used than anything else.

Stationary bearing wear has long been a problem in the automotive industry, and modern bearing greases are formulated to help prevent fretting in these conditions.

I'm with you there. As long as preload is within a reasonable range it doesn't make a difference. The best evidence is that fretting is almost exclusively limited to the lower stack. If preload were a factor, especially excess preload, we'd expect to see more fretting in the upper.

IME - what's happening is a combination of classic fretting, plus the effects of front end weight and vibration hitting the lower stack harder because the upper only has the preload and not the other factors. I also sense that I see faster failures with loose headset, but that's only a gut feeling.

...as soon as this conversation comes up, phrases like "a bit tighter", "looser than normal", and the like inevitably make it impossible. Whatever works for you, works for you. But again, without trying to be a dick about it, I don't really see the physical mechanism you describe to be a factor.

I've already qualified this as my opinion. If anyone has anything from the official bearing engineering and science guys, I'd certainly be interested in reading it.

But, the alcohol must be applied properly. Directly to headset has the effect of removing the grease and increasing fretting. (Applying it as intended sometimes has the negative effect of causing others to fret.).

Ben

How many pounds is it, normally pre-loaded on headset bearings, just 8 or 10 right? So the difference with "slightly more" is some fraction of that. A few pounds, maybe less.

It makes me wonder if the weight we put on the bars, combined with flex in the fork and steer tube and angle thereof, cause less load on the rear top bearings than on the front bottom bearings. IF the fretting is exacerbated by preload pressure, then that might account for it being worse in the lower stack of bearings.

You may have that backward. The preload pressure is by it's very nature equal in both stacks. (Newton's Laws), while the forces generated by riding are mostly in the lower stack. (headtube flex acts like preload and loads both bearings equally, but steel head tubes don't flex much, if at all).

So, with loose headsets, you'd expect the bottom stack to suffer but not the top, while with greater preload you'd expect more equal wear.

There will also be differences in wear patterns. If the bearing has play, the fork will cam in the head tube, with load to the front of the lower stack and rear of the upper. and the wear pattern should reflect that. OTOH a straight head tube secured with proper preload, will sit square and the wear pattern will be more uniform around the entire race, though there may be some more wear at the front of the lower stack because of angular forces sloping head tubes cause.

Ha!

Very possible I do. I'm visualizing the forces due to leverage from the stem and ground, with the fork flexing because of the head tube angle, which appears to me to lessen the force against the top bearings relative to the bottom bearings. I'd need to draw a vector diagram to be sure, but visualizing it is more often right than not.

Assuming that the difference in forces is similar to the forces of the pre-load, that would account for the wear patterns being different from what you'd expect with equal preload.

of course it depends on things like whether we're looking at a steel steerer which is pretty rigid, or a more flexible CF steerer. Also, whether the stem height and length offer leverage whereby a rider may flex the fork.

However, we have to look at the time element, which is what the fretting process is about anyway.

In normal riding with a slightly loose fork, gravity would load the lower bearing and separate the upper. But the head angle would cam the fork, causing the load to be only in front on the lower stack, and back of the upper, with the rest of the bearing separated.

Larger bumps would push the fork back reversing that pattern, but again those are single events and don't account for nearly the amount of time that the normal straight riding does.

So, we might consider 3 conditions.

1- excess preload which would cause similar even load top and bottom
2- correct preload, where the lower bearing is doing the bulk of the work and we see uniform wear, but mostly or only in the lower stack.
3- inadequate preload where the fork cams and we see uneven wear on both the top/back, and lower/front due to the cammed fork.

IME - condition 2 accounts for well over 90% of the fretted bearings I've seen. That leads me to believe that headsets are mostly decently adjusted, and/or poor adjustment is not a significant factor in accelerated wear.

Yep, this is what I was talking about in the first post. Even changing "slightly loose" to normal pre-load, replace "only in" front of the lower stack etc, with "mostly in".

Now from here, I surmise that pushing down and forward on the stem, given the tube angle, has a similar affect as this. You observed that you see the indexing in the lower stack, and reasoning from that suggest that pre-load (being equal) therefore does not impact it. However, supposing that the loading is not equal in normal riding, that reasoning does not necessarily hold true.

It seems reasonable that the amount of load on the bearings does impact how quickly the bearings and balls wear. I just suggest a possible mechanism where that can be preserved while still being consistent with your observation.

Ah, the lower stack also has pressure from the crown fork doesn't it? That also might account for the difference.

Err, getting back to the original post...

I recently purchased a used bike with slightly tighter steering (it will flop on it's own when leaned a bit, but not easily, and the wheel will always track pretty strait ahead when pushing the bike by the seat), and I love it. It doesn't feel stiff at all when riding (I have some experience riding motorcycles, so bicycle steering always feels super light), but it makes everything so much more convenient day to day.

I live in the city, so I'm constantly pushing the bike through doors while holding them open, and keeping the wheel from flopping is a huge help. Likewise, when I put the bike on the stand I don't have to worry about the bars turning and smacking...

I'm now thinking about adjusting my other bikes to be the same way. And, in any case, I can't adjust this tighter one at the moment, since it has that annoying '80s Shimano 600 star nut on top.

I'm with [MENTION=465722]robertorolfo[/MENTION]. Too much debate on a subject the OP has not replied to. My take on the OP's issue is that he is describing what is happening when the bike is static and possibly in a bike stand with the wheel off the ground.

That would not be abnormal at all. Don't all front wheels flop one way or another when hung up by the tubes?

There was lots of useful content in the replies I received. My take-home is that there is some latitude in the amount of head-set tension, provided it is not actively "loose" so that the bearings can move around too freely, and get damaged.

I use the "front brake" test to make sure nothing is loose, and adjust the headset to about 2Nm torque. I moved to a longer stem on this bike, so having a fairly normal "loose" head-set tension offsets the slower steering that results from the longer stem.

So the next time I spend some time at Mcsorleys, Then ride Citibike I can say I'm helping save the bike. Awesum.

I ride with a guy who does that- maybe that's his strategy.

I figure that he covers a couple of extra miles over a double century,

but he's done over 60 of them, & his headset is probably in good shape.