Joseph Gwinn wrote:

This is about three times what the 5914 weighs. If one assumes that
this is accomplished by tripling the thickness of everything, the
stiffness will increase by the cube of three, or a factor of 27, so I'm
not surprised the a few hundred pounds on the bed has little effect.

Yes, this was a move up from a 12" Atlas/Craftsman. Wow, what a
difference! The cross slide is beefier than the carriage ways
on the Atlas, and the carriage spans 25" along the bed.
What are the dimensions of your Sheldon? One reason I didn't go that
heavy was space limitations, and the difficulty of getting such a heavy
machine into my basement.

It is about 80" long, and I had to take off the nice door over
the back end of the headstock, as there's a bench right there.
The bed itself is 6 feet, exactly, giving a conservative 42"
between centers. I'm sure I could turn a 48" long part with no
trouble.

One of the reasons I bought this house was the grade-level door
to the back yard from the basement. I had to take a couple
handles off the lathe to slide it through the door. The biggest
problem is the soft terrain and the immense weight of fork lift
trucks.

See http://jelinux.pico-systems.com/sheldon.html
for the story and some pictures of the lathe.



Now, another comment is that you may have a poor "fit up"
between your carriage and bed. I totally rescraped my bed and
carriage to get the best fit between these surfaces. In my
case, due to the hardened bed, I had to grind it with Cratex
polishing wheels and then lap it smooth with bench stones.
(A carbide scraper blade wouldn't even SCRATCH the bed.)
I then used Moglice castable way liner to make the underside of
the carriage conform to the bed.


There is some bed wear which prevents the hold-down plates from being
too tight on the underside of the bed way, but the slop is about 0.001".
I did take one 0.002" shim out from under both hold-down plates, which
both had visible wear.

Depending on the shape of this wear, it may allow rocking of the
carriage on the convex curvature of the bed. This is "bad"!
It allows external forces to easily move the cutting point.
The larger the lathe, and the more distance between bed and
spindle centerline, the more this rocking is magnified.
One reason that self-feeding is less of a problem when the tool moves
perpendicular to the spindle axis than when the tool motion is parallel
to the axis may be that the carriage is narrower along the way than
across the way, by a factor of about two.

The ways were hardened and ground, but not scraped. But I doubt that I
will ever have the time or energy to re-scrape them.

Well, you can't scrape them, I suspect. Hardenable beds are not
grey or white cast iron, and when hardened, even without flame
spraying, they get immensely hard. I have developed a system
for "hand scraping" hardened beds, using a die grinder and the
muslin-bonded Cratex polishing wheels. If hand scraping is
slow, this "hand polishing" is even slower! It took me 14
months to do my Sheldon bed, but that involved several detours
down wrong paths.

I think that it does rock slightly, as I can feel slight movement if I
put a finger spanning the gap between carriage and way. But the skin is
very sensitive to motion in such setups, and this is probably the
~0.001" slop I already know of. The problem is that this was observed
while the lathe was shaking with chatter, and everything was bouncing,
and it wasn't clear what was cause and what was effect.

A dial indicator between carriage and way and an iron bar in the boring
bar holder (to apply force while the lathe is stopped) may be useful
here, though.

More generally, I've been finding and fixing many problems, few of which
were huge, but they do add up.
I think it really doesn't take much rocking to set up the
vibrations, then the natural resonant frequency of the machine
and workpiece take over. I have noticed with the Atlas machines
that the setup of tool overhang has a large effect on this.

Jon