I called Ryobi and complained about the world-shaking side-to-side
vibration. Their first reaction was to send me four new flanges.

Their second reaction was to ask if the spindles were bent. Hmm. I
hadn't thought of that, but it's certainly possible. The grinder is
shipped with those heavy 8" diameter by 1" wide gray wheels in place on
those little 5/8" shafts, and if the box were dropped, it could happen.
The box doesn't look like it was dropped, but if the pallet were
dropped, the box would look OK. But the dial indicator shows less than
0.001" of runout, so that isn't the problem.

Looking at the shafts, it's apparent that the shoulder upon which the
inner flange rests isn't very large. The shaft in the motor is 0.667"
(17mm nominal), turned down to 0.627" (5/8" nominal), for a shoulder
height of 0.020" (0.5 mm). When one looks at a flange that has been
resting on this shoulder, the edge mark is halfway down the slope where
the sheet metal was pulled out of plane by the 5/8 punch that made the
center hole through which the shaft passes.

Part of the solution was to buy two precision hard steel thrust washers
(MSC 03548187), one per wheel, to rest upon this narrow shoulder and
provide a larger bearing area for the sheet-metal flanges. This cut the
wobble perhaps in half, but didn't completely solve the problem: The
mating surface of the inner flange also needs to be flatter.

Flattening was done by hand grinding with wet-dry sandpaper face up on a
granite surface plate. This almost worked. The problem is that the
wet-dry paper is too soft, allowing the edges of the ground area to be
turned down, and the thrust washer was resting upon the turned-down
region near the shaft.

So, went fully traditional (for mirror grinding anyway), using ordinary
valve-grinding compound on a piece of plate glass. This worked,
reducing the wobble to about 0.002" on the first attempt, though it's
slow going because the flange metal (mild steel) is really a bit too
soft for grinding this way. Nonetheless, I'll probably be able to get
to less than 0.001" with a little more elbow grease.

When the four new flanges arrive, I'll probably ask a friend with a
Clausing lathe to face them off, now that I know that this is the
solution.

The mystery is why Ryobi uses such lousy flanges. Adequate flanges
cannot cost that much, and the current bad flanges undermine the whole
affair. It reeks of using good fabric with bad thread while making
clothing.

Joe Gwinn.