I've been having some issues with tight headsets lately and am seeking some guidance.

My technique for checking headsets is to brace the bike against the floor (with my foot on a chainstay) and moderately jerk the front wheel and the top of the stem in contrary directions with my hands. I believe this allows me to exert a good amount of force on the headset and detect even small knocks/looseness. The problem is that I can't seem to eliminate the tiny knock I can almost always detect without tightening the headset down pretty tight -- tight enough that I can feel the bearings in the headset and the bars don't turn 100% freely.

So my questions are: Am I applying too much force when checking my headsets? Am I correct that there should be absolutely zero play in a proper headset adjustment? I'm assuming that it's better to err on the too tight side here, correct? Thanks.

Yes, zero play is what you want. If everything is healthy in the headset and assembled correctly, you should be able (easily) to achieve a zero play and zero binding adjustment.

Is the headset new or used? Have you inspected the bearings and cups?

For me, adjusting headsets, threaded and threadless, are the easiest to 'get right' the first time.

The simplest and quickest way to check for headset play is to lift the front end about a foot and drop it onto the floor. A loose headset will give a very characteristic "knock" or "clunk", and a tight headset won't.

The other test is to apply the front brake and rock the bike forward and back while looking and/or feeling for movement at the lower bearing.

The important indicator of a loose headset is a slight click or clunk when applying the front brake, as the fork pushes back in the bearing.

It's normal to feel the bearing somewhat when a headset is tight, but there shouldn't be any binding. If you can't get it adjusted to find a balance between no play, and no drag, there may be misalignment someplace, or if play is very difficult to resolve, sometimes the headset is OK, but one part is loose on the fork or in the head tube, ie. a crown race loose on the fork.

So it you're struggling with the adjustment, and it's not just you take it apart and see if you can spot the issue (other than adjustment). Among others I've seen retainered balls upside down, cups loose in the lower head tube, integrated headsets not seated properly, or mismatched for the frame, upper centering cones missing or damaged, and so on...

First try adjusting, then do a thorough diagnosis to find why you can't.

For me, the easiest test is to place the bike on the ground, apply the front brake enough to lock that wheel, then rock the bike forward and backward. There shouldn't be any discernible play or looseness at the headset. Sometimes there is a little play in the front brake which is a different issue but can confuse things. You can place a finger where the headset races meet to feel specifically for headset play.

I don't worry about being able to feel just a bit of drag from the headset when the bike is on the stand and I am turning the handlebars. The headset bearing isn't being asked to spin freely like a hub bearing, so that little bit of drag doesn't matter. You do want the front wheel to flop freely to one side or the other as you lean the bike.

When adjusting the headset on the stand, I tend to adjust too loose, and then have to tighten a little bit after the "rocking" test. The need for such a trial and error process always reminds me that I'm a shadetree, not professional, mechanic.

Thanks all, I appreciate the replies. My test is basically the front brake/rock bike method, but while also immoblizing the rear end of the bike, which I believe allows me to exert more force on the headset and detect even tiny knocks. Lately I seem to get tightness regardless of whether installing a new or used headset. When it's used gear my ASSumption (I know, I know) is that the races are probably somewhat worn; with new stuff I guess fork facing (I get the head tube parts pressed at the LBS.) I'm OK with a little drag (hey, I've seen Hairspray) but I'm trying to make sure I'm not just flat out over-tightening things from the get go.

You are.

You might be right. But here's the problem: in my recent experience there's either going to be a tiny knock or some drag. I suppose if I were a professional I could thoroughly diagnose, make frame/fork modifications and press a new headset into every bike. But I have to make some compromises.

PS: One of my concerns is that my special test exerts an "undue" force on headsets and is revealing tiny knocks that exist naturally in most headsets. (Or perhaps causing a tiny knock somewhere else that I can't locate.) I do believe my test applies more force to the headset than standard the "lock 'n rock" method. But if it's true that zero play is what I'm after, then recently I've had to go tight to achieve it.

My spidey sense tells me there will be skepticism on this point. That's OK -- I'd be skeptical too.

maybe this will shed some light on your problem, OP.

i've recently had a similar experience in adjusting my headset. i suspected that the crown race was not seated properly on the fork. there was always a slight clunk when hitting small holes in the road. i checked everything over the next few rides, hoping it was a clicking brake lever, or brake housing, or loose skewer. i could feel no looseness in the headset and had adjusted it a number of times. i figured if the crown race was not seated completely, a few hundred miles of bad road might make a difference. anyway, last night i was at it again, after a riding the bike, and noticed that i could create a clunk by just hitting the front tire, hard, towards the rear of the bike. i figured it had to be a loose headset. so i tightened it down to the point where there was a bit i tightness in it, then went for a ride. there was no noise on the ride and no tightness in the headset after the ride. i figure the crown race wasn't completely seated.

If you cannot eliminate play without tightening the headset so much that it no longer freely turns, there are a couple things to check: first, make sure the bearings are installed properly. If a retainer or cartridge is installed upside down it can cause adjustment problems. Second, the faces of the head tube may need to be prepped to ensure that the upper and lower races are both installed concentrically and perpendicular to the axis of the head tube. If this is not the case, the headset will appear to be tight in one spot, but loose in another. And don't forget to check that the crown race is fully seated, as hueyhoolihan notes above.

Since your "special test" is not working for you, maybe you should abandon it and go with one of the methods on the Park Tool site.

Thanks @FBinNY @hueyhoolihan and [MENTION=20548]JohnDThompson[/MENTION] -- I think you've hit the nail on the head re: troubleshooting. FWIW I think I've just had a string of difficult headsets/frames/forks. (I don't use retainers btw.) I'm going to keep working at it though. I'm not equipped to do the more involved fixes (facing, cup pressing, fork crown seating) but in cases where I just can't get it right I suppose I need to just bite the bullet and go to my LBS.

+1 to this! I've found that if I can't eliminate play by adjustment, I try to better seat the crown and cups. If that doesn't work I have the frame and fork faced and reamed to restore the fit and positioning of the cups. That has fixed it every time.

$20 spent to finish installation of a $20 to $60 headset on a frame worth $200 to $2k is a good investment if I get more years out of the headset.

+1. This is a PITA (at least for me) but when all else fails this seems the proper way to get things right -- tabula rasa so to speak.

Seems a good zombie to resurrect, as I'm in the same situation Jethin was and have a related question:

Could going from caged bearings to loose balls resolve the issue? I'm working with a '71 Campy headset. When I brought the bike home the steering was super-stiff, and I discovered the bearings had been installed upside down. Now that they're in correctly I can't get that sweet spot with zero play and no appreciable resistance.

I'm wondering if the bearing cages may have been damaged by the incorrect installation, as there's no pitting or scoring in the headset itself to suggest a problem there.

There's always the chance that they were right before and upside down now. Otherwise, I don't expect that a slightly damaged cage would cause problems with adjustment. You might hear or feel some drag, but it wouldn't change much as you adjust.

Usually trouble finding a sweet spot is caused by misalignment or an extra ball. OTOH, you don't have much to lose by giving loose balls a shot.

Right now the bearings are facing their respective cones and the cages face the cups. As far as I know that's right...?
As for "not much to lose by trying loose," that's what I was figuring: if we're good with the loose bearings in there, I know everything's aligned properly and something else was messing up. If it's still sketchy with that setup I'll know something more substantial is the issue.
I think.

I may be wrong: I believe that the cages should face the cones not the cups--- check this out.

Joe

There should be slight drag indicating preload.