I did mine, with TLC
I got hold of some ABEC class 7 for both the lower and the upper set.
The spindle is constructed so that everything is a shrink fit, including the top pulley, top bearing retainer and lower bearing retainer. I used a jack type hydraulic press and judicious appliction of heat to remove the pulley and collars without causing any scoring of the shaft journals. I feel this is a project that one should not be 'rough' with the handling it, hence using heat to expand what needs expanding.
I did notice that the original lower main bearing set had apparently lost its preloading. This was evidenced as circular wear on the face of the lower retaining collar. So, I lapped that flat, using a fine grit self-adhesive sanding disk stuck down to a granite surface plate.
When I got the new set of A/C bearings, I set up a stack in the press, beginning with a freshly faced steel disk on the bottom with small raised boss to contact the inner race of the lower bearing, then placed the inner and outer spacers on that, then the top bearing, then the retaining collar and jacked it tight in the press. This applied pressure only to the fixed components as they would be on the spindle itself. No pressure was carried through the balls, except what should have occurred due to the spacers.
It became evident with this test that I was not getting any preload of the outer races as the outer spacer still slid around freely between the outer races of the bearings.
So I took and lapped the inner spacer on the surface plate. This should not be the first time that you try to lap something flat Use a figure eight motion as you move the spacer around, and then turn the spacer approximately a half turn in your hand and take another stroke. Even though you might think that you are putting even pressure on the part, you are not, so turning and lapping the part from various positions is absolutely necesssary to randomize your own systematic errors, so that they average out to zero.
Mic across the spacer frequently to see how evenly parallel you are keeping the ends. Only lap one end, leave the other end as per original, so that you do not end up with an out of square spacer, which would be totally junk at that point.
So I lapped and I retested. I achieved what I estimated to be a decent preload, I was shooting for light to medium. In my press setup, I could still, with moderate effort (light taps with a soft hammer), slide the outer spacer between the bearings with the inner stack fully compressed. I was prepared to redo this if necessary, but did not want to have to press the new bearings off the spindle if at all possible.
I think I found a happy spot in the tolerances. The reassembled spindle would show a few degrees of warmup during the run-in stages. Previously, my spindle had always remained stone cold.
During reassembly, I heated the retaining collars so that they slid on easily, again, this is a TLC moment and one does not want to induce scoring of any surfaces by cold pressing the components together.
Note that the lower collar (the really important one for preload) will shrink in length after it is installed. This will defeat the preload that you obtain in the cold test. So it will be necessary to press that collar down tight to the bearings after all has cooled down. It should 'croak' once in the press to signify that it has moved and seated. At that time, you can again test the preload feel of the outer spacer.
The top pair of bearings are just precision deep groove ball, and are not really preloaded, although there is a wave washer between the outer races. All you do is assemble the hot components so they slide on easily, then seat the top retaining collar with the press, to compress the wave washer. I've forgotten already, but I think there was significant feature on the spindle to indicate how far down the top retainer should move, so as not to overcompress the wave washer. Observe the collar seated position when taking the spindle apart, I guess.
The timing pulley is a tight fit, with a couple of thousandths interference. It takes a lot of heat to loosen it up enough and also to install it easily. Making shrink fits like those can be a tense moment. I hope you've done some before this
Now after all this, you will perhaps not have the facilities to regrind the internal taper with the spindle installed in the spindle cartridge. You should give this some thought if your taper has any damage or wear. An exchange spindle might be the way to go. Otherwise, what you get when you are done will depend on whether or not you can fix the taper up, even if you have attended to all the previous details with great care.
Be sure that your spindle oil mist is working, and run in the new spindle according to the Haas break in instructions.