DIYbanter - View Single Post - Mathematical analysis of Rollie's Dad's Method

Joseph Gwinn

In article ,
wrote:

On Wed, 15 Sep 2010 23:46:28 -0400, Joseph Gwinn
wrote:

In article ,
wrote:

On Mon, 13 Sep 2010 09:12:17 -0400, Joseph Gwinn
wrote:

SNIP

From reading your original post.

Rollie Dads Method of measuring bed twist is useful and well
established.

I think it measures far more than just bed twist. Any misalignment between
spindle axis and bedway should show up.

For the commonplace case of a reasonable straight and round
test bar it gives the correct answer.

But lots of people seem to have had problems. Now, I was never sure if it
was
due to inadequate explanation, or some error in the method itself. Or maybe
both.

I'll have to try RDM on my lathe, for the experience. I think I recall
having
tried it, and having gotten nowhere, but don't know why. It may have been
that
I didn't really understand what I was doing mathematically, and so was doing
random things.

I ordered a R8 to ER25 collet chuck today, so I will soon be able to get
better
runout data.

As you point out, Rollie's "correction" for diameter is in
error. At both the near and the far measurement points, the
rotational axis is truly defined by the difference between the
highest and lowest clock readings and is independent of diameter
and diameter change.

I'll have to think about this, as it seems to me that the difference yields
the
pure runout regardless of where the cone axis might be.

However the detail content of Rollie's description is very
useful. I think it might be helpful to post to the drop box an
agreed RTM Mk2. Changes needed could be pretty small :-

I had been thinking along these lines as well. At the very least there is
an
extension to RDM, and there may be a small correction as well.

In each paras 5 and 8 delete the last sentence

I'm not sure I understand. Please quote the sentences to be deleted.

"why this method works" - delete or modify

Yes, modify.

Joe Gwinn

It's true that the method also shows up spindle alignment and
possible carriage alignment change but in practical terms,
adjustment of bed twist is normally the only available method of
correction.

RD also mentioned shimming the headstock where it rests upon the bedway,
but you are right that the adjustment options are limited. In practice
one might do both in alternation, so the lathe converges to as perfect
alignment as can be obtained given only those two "knobs".

RDM is often proposed as an alternative to the use of a precision
level which only detects bed twist. It may be worth discussing
this method in an RDM revision because the method is often
described as a series of level measurements of the bed surface
with the lathe as a whole needing to be precisely level.

This could work, but it would be necessary to separate the effects of
headstock misalignment and bed twist, or one could end up turning the
wrong knob, and making things progressively worse.

If you have precision level, it is very much simpler to mount
the level on the crossslide and observe the change in reading as
the carriage is traversed.

I do have a 6" Starrett model 98-6 precision level (0.005" per foot per
division), although I usually slide it around on the tops of the bedway
V rails, the method recommended in the Clausing manual. This will not
work for all bedway designs, but the ride-the-carriage method should
work universally.

The tops of the V rails do not wear in normal use, as the carriage rests
on and wears away the flanks of the V rails, so for older machines using
the level on the tops of the rails should be more accurate than riding
the carriage. But I'll have to think about this - I don't know how
important it will be in practice.

It is unnecessary for the lathe to be precisely level because you
are now measuring directly the effect of bed twist or distortion
on the cutting tool location. Even if the lathe were large
enough and flimsy enough for gravity induced deflections to be
significant this would be indicated directly by this method

Yes. Maybe this also explains how people adjusted lathes in ships at
sea.

Both para 5 and 8 amended to only read "Average the high and low
readings (add together and divide by two) to get the "near end
average distance".
Delete " if you suspect .........."

An alternative wording could be " Note as the reference
distance,the mid point between the two readings"

Ahh. These reference the original RDM description, not my postings.

The original description could be clearer for sure. I would start with
the why-this-works explanation of the math, then move on to some
specific applications, on the theory that the detailed method is easier
to understand if one knows how and why it all fits together.

The centre line of the bar mounted in the chuck describes a cone
whose centre line is coincident with the lathe rotational axis.
The reference distances describe a line truly parallel to that
axis.
The reference distances assume that the lathe bed is straight.
Bends or bumps would introduce their own errors.

Not to mention uneven bed wear near the headstock, often an issue with
HSM iron. But again, I wonder how important this is in practice.

Joe Gwinn