Paul W --
Here's a slightly-edited version of what I wrote for the Chaski board about five years ago:
A Starrett 98 is a fairly straight forward level instrument to repair and adjust.
From your questions, I'll assume that you have disassembled the vial housing and have cleaned the vial carrier thoroughly, removing ALL residue from the old vial and old plaster. If not, that's the first thing you need to do.
Once the vial carrier is thoroughly clean you can begin to install the new vial. I suggest checking the fit of the vial inside the carrier by simply sliding the vial through the carrier . . . it should slide freely and not hang up anyplace, but at the same time it should only be a bit smaller than the carrier's ID.
As I recall, the Mod 98 carrier is chromed, not painted. The metallic carrier ID does not provide a good background for reading the position of the bubble in the vial, and you can choose how you want to create a better background -- which is traditionally flat white although a very pale green or yellow background is sometimes seen and any very pale color would work. I prefer to use a slip of bond paper to create the background, but others simply paint the inside of the carrier. I suggest an archival-quality cotton-rag bond paper, which will keep its sparking white color for decades if kept dry.
I fit the paper liner to the carrier by simply slipping a rolled-up piece of paper into the carrier and marking the size of the carrier's opening onto the paper. Then I cut-and-try the liner IN WIDTH ONLY until the paper is just narrow enough to not extend into the carrier's cutouts.
Next, I fit the vial into the liner paper and slide the vial and paper into the carrier; adjust until the vial is centered both lengthwise and rotationally in the carrier's cutouts and the liner paper extends past the cutout but not quite to the radius between the vial's cylindrical body and its end. The vial and paper liner should be only a bit loose in the carrier; if necessary another layer or two of paper can be added to shim the vial. Once I'm satisfied with the fit, I slide the vial and liner out of the carrier, trim the liner to length, and then reinstall the vial and liner, centering them carefully.
If the vial is really loose but there is no better-fitting vial available, toothpicks or whittled-to-a-long-taper bamboo barbeque skewers can be used to GENTLY wedge the vial into the tube on one end only. The vial can be pulled against the carrier's opening with cellophane tape.
Once you're satisfied with the vial's position in the carrier tube, you can think about bedding the vial. RTV can be used, but personally I don't trust the stability of non-catalytic RTV in thick sections. Yes, I have used RTV to bed vials, but I've used a hypodermic syringe with an industrial applicator needle (square-cut end) to inject 3 or 4 half-pea sized beads bridging the vial-to-carrier gap. Then, figure on a 24-hour cure for the RTV.
Plaster bedding is traditional and I like the results. Depending on who you talk to, the suggested material is Plaster of Paris, Patching Plaster, a tooling plaster (such as US Gypsum's Hydrocal series), mixtures of a gypsum plaster and portland cement, and mixtures of a gypsum plaster and flour. Mostly I use Patching Plaster, which is available in small boxes and has a working life of 45 minutes to an hour . . . Plaster of Paris sets too rapidly, and I don't want to chase down a 50-pound sack of Hydrocal . . . but I recently tried the Patching Plaster mixed with flour at about a 1:1 ratio and I like the way the mixture works. It's stickier than the plain plaster and sets up to be very firm but not rock hard.
Even with the plaster or plaster mix, though, I prefer to use just a bit. I mix it to a soft paste consistency and inject the plaster using a needle-less syringe to form a plug of plaster that is just deep enough to cover the end of the vial. Now, check the vial to be sure that it is still centered in the carrier opening and the liner is properly in place. You can tweak the vial's position before the plaster starts to harden.
Remember that you need to put the vial carrier end plugs back in place . . . don't let the bedding block the end plugs.
Cure time, then bed the other end of the vial (after removing the wooden wedges if you needed to use them.
Cure time again. Inspect your work, and if you want you can hand-smear more plaster into the carrier opening to fill any gap between the carrier and vial.
Install one of the vial carrier end plugs, being sure that the plane of the mounting tab is oriented correctly relative to the carrier's opening.
Now install the cover spring on the end of the vial carrier away from the installed end plug, slide the cover over the sping and down to the other spring pocket, install the other spring, and slide the cover on all the way. Only now is it time to install the other end plug, and here you want to be sure that the mounting flat is oriented correctly and that both mounting flats are parallel to each other. Don't rely on "eyeballing" the parallelism of the mounting flats, set the assembled carrier on a set of parallels or 1-2-3 blocks on a surface plate.
Install the vial assembly on the level's body, leaving the adjustment nuts finger tight. (I presume that you've already checked the iron base and know that it is not warped or burred.)
You can adjust the vial assembly relative to the iron body by "reversion", essentially checking the vial against itself. To do this, you need a stable surface plate that is nearly level. How level? Ball-bearing-won't-roll-off level is close enough.
Within reasonable limits, temperature is not important but temperature stability is very important. You can not successfully adjust (or use) a precision level if it is in a beam of sunlight, near a blower vent, or if you've had your fingers on the vial recently. Got that?
Now, set your level instrument on the surface table and adjust the vial housing until the bubble is centered in the vial. Finger tight on the adjustment screws is good.
Next, mark the location of the level body on the surface table using a pencil or build a set of fence rails that will provide a nest for the level body. Go get yourself a cup of coffee or a cold soda, giving the level and surface table a half hour or so to come to temperature equilibrium.
Again examine the position of the bubble in the vial; adjust the vial assembly to nearly center the bubble if necessary. Give the bubble a few moments to settle, and then carefully note the position of the bubble in the vial. Some people note the bubble position in room coordinates ("a quarter division toward the west wall") while others use instrument coordinates ("a quarter division toward the adjusting screw"); pick a method, and call the bubble position in the vial the Initial position.
Now pick up the whole level and turn it 180 degrees. Set it down on the surface table and either nest it against the fences or adjust it to be within the pencil outline. Give the bubble a few moments to settle in the tube. Now examine the bubble to see where it is in the tube, which we’ll call the Rotated position.
If the bubble is in the same place relative to the room ("a quarter division toward the west wall" OR in the "opposite" place relative to the instrument body ("a quarter division away from the adjusting screw"), the vial-to-base adjustment is correct. In this fortunate circumstance, all you need to do is snug the vial adjustment screws.
If, more probably, the bubble is not in the right place, you'll need to adjust the vial housing. Use the adjustment screws to move the bubble all the way to its "should be" position, which is midway between the Initial and Rotated positions, finger snug the adjustment screws, and go through the whole test-and-adjust cycle again. Keep at it, and you'll get it right.
For what it's worth, more sensitive level instruments usually have a side-to-side vial adjustment in addition to the up-and-down adjustment. This adjustment allows you to rotate the vial so that its lengthwise axis is parallel to the long edge of the instrument's body. Why? Well, if the vial is not parallel to the long edge of the body the bubble will change its lengthwise position in the vial with small fore-and-aft tilts of the instrument.
OK, I'll stop for now. Post back if I can answer any other questions.
[I offer a special tip o' the hat to Chad Webb for saving my essay and E-mailing it to me when the Chaski archive copy went away.]
John