Leo's excellent analysis of this situation
Lyn J. Mangiameli
>From Leo Lichtman, 5/2/02 on rec.crafts.woodturning
Let's try another thought experiment--they often seem to help. Picture a
long shaft, mounted between two bearings, near the ends. Close to one of
the bearings, mount a flywheel. If you give this assembly a spin, it
revolves and slowly comes to rest in a random position, because it is in
balance. Now, take away half of the flywheel. The assembly is now way out
of balance, and the side with the half-flywheel will fall to the bottom. If
you now give it a spin, the end with the unbalance will do its best to rip
itself and its bearing loose. Now for the crucial part: Put back the
half-flywheel previously removed, but put it at the other end of the shaft,
180 degrees to the first half-flywheel. The assembly will vaguely resemble
a pair of bicycle pedals, and it will be in STATIC balance. Because the two
halves of the flywheel are on opposite sides of the shaft, the assembly will
remain stopped in any position.
Now for the excitement: get it spinning. BOTH ends are out of balance
dynamically, and BOTH ends will be trying to tear themselves loose from
their moorings. If you could watch the shape of the shaft, it would look
something like a long letter S on its side. The faster it turns, the more
centrifugal force is added to the weight, so the curves of the S become
wider. Each bearing is being jerked one way, and then the other. The
forces, which in the static situation would neutralize each other, don't
meet. They produce what in engineering terms is called a "couple."
On a headstock, with a sturdy spindle, and good solid bearing mounts, the
chances are the situation would not get out of hand. But, remember also,
that the unbalanced chuck of wood is hung out over the ways, while the
counterbalancing system (as proposed earlier) is hung outboard from the
headstock. This maximizes the distance between the unbalanced forces, which
maximizes the "couple."
If we place the disk holding the compensating weights to the same side as
the wood, we minimize the couple, reducing markedly the amount of of dynamic
unbalance. And remember, even though a spindle seems like a big hunk of
steel, it has a pretty deep thread on the outside, and the inside is reamed
out for a Morse taper. As was pointed out in an earlier thread, when Leif
Thorvaldson asked about bearing spacing, a spindle loses a lot of strength
due to these machining requirements.
This is an aside, here, but it is so relevant I will mention it. With an
unbalanced chunk on the lathe, it is easy to see why it is good to bring up
the tailstock for extra support. It takes half the load and sends it the
other way. It changes the cantilevered system to a span, which is
structurally much stronger. (and, of course, if the wood does come loose,
it can't get away.)