Adjusting hub: Axle compression w/ a solid axle ??
 

Shuffling the spacers so a 7-speed freewheel fits the spacing of my tank. I'm adjusting the solid-axle rear hub repeatedly, and I don't understand what I'm seeing.

Using a Stein hub axle vise; adjusting the non-drive side, meaning the drive side is bolted to the vise. What we're looking for is the point where any play in the bearings just disappears. Back off a few degrees and the play would reappear. The wheel should spin freely at that point.

Various techniques produce the same weird result. All start with the locknut backed off:

• wiggling the rim to feel for play, turn down the cone to just the right point. Hold the cone there and tighten the locknut. The bearings immediately become too tight!
• adjust the cone until there's a small amount of play - purposely too loose. Tighten the locknut and the play disappears!
• spin the wheel and take the cone down until drag appears; back off a little. Tighten the locknut and the wheel slows right down.
• repeat all the above with the wheel bolted to the outside of the dropout (so vertical instead of horizontal) and the same thing happens.

This doesn't make any sense. It's acting like a QR axle, with the axle 'compressing' when the skewer is clamped. But unlike a QR axle, there's no compression applied across the bearings.

What the heck is going on, and how does one find the "sweet spot" without repeated trial and error?

Is the lock nut tight on the drive side?

There is a small amount of play in the threads themselves between the male and female parts, so that the locknut would indeed push the cone down slightly when tightened. This is different than axle compression and is perfectly normal. The usual procedure for adjusting a solid axle bearing is to move the cone down till no play, back off slightly and then tighten the locknut. The difference is that once you have adjusted a solid axle bearing the adjustment does not change when you mount the wheel. For a trained, experienced mechanic it's still trial and error, just less of it.

Yep. And I'm watching the axle to be sure it isn't turning with the adjusting locknut.

Yes, and the cone wrench doesn't hold the cone perfectly, so I expect some variance. Maybe I'm just not expecting enough.

When I do this I end up with a bearing that's too tight - the wheel doesn't spin freely. I suppose this could be compensated by backing off a lot more that one might expect - maybe the full quarter turn that the Park site describes as "purposely too loose".

Which is essentially what I did last iteration and I now have a decently adjusted hub. But I have to go back into it because the drive side spacer is too thin, and I wanted to ask before the (hopefully) last iteration.

Yes, probably so. It all depends on the sloppiness of the thread.

No "last iteration". I held the drive-side cone still while adding spacer and it came out as good as it went it.

I still don't feel good about it, but the bike is a POS. I'll get more anal if I see something like this on a nice group.

The way I deal with this is to adjust the cone a bit loose, holding the cone with a wrench I tighten up the locknut just shy of final torque value. If the bearings tighten up I just back out the cone a bit aganist the locknut until I have proper bearing preload. YMMV