Bill Leach: Open_loop Tools

[FROM Dennis Delprato (950627)]

Bill Leach:

In message of 950626.22:27 U.S. Eastern Time, you made a comment
about open-loop "control" and negative-feedback control
machine tools. I understand the main point, but have no
experience with machine tools. Perhaps you could say a bit
more about the two classes of machine tools in terms of
specifics. I hear much about machine tooling around S. E.
Michigan, and you may provide some information that would
help me better "relate" to those familiar with mchine tools vis-a-
vis control.


Dennis Delprato

<[Bill Leach 950628.01:07 U.S. Eastern Time Zone]

[FROM Dennis Delprato (950627)]

machine tools

The best understood (by most people) machine tool is the lathe.

Early lathes did nothing more than spin the work. Any "machining"
performed was accomplished by the machinest holding the tool against the
work as necessary.

Sometime later, a "carriage" with a "longitudinal" and a "cross" "feed"
were added.

The "carriage" is a unit that slides along the "ways" adjacent to and
parallel to the axis of rotation of the piece being worked.

Longitudinal" travel then is travel parallel to the spin axis of the
work and "cross-feed" (or "transvers feed") is at right angle to the spin
axis. Note that many lathes have a "compound feed" that does not have to
be at 90 degrees.

Now a "constant" diameter could be machined onto a piece of work by
setting the tool position for the desired cut using the manual adjust of
the cross-feed and then engaging the longitudinal drive. Of course the
diameter was not constant.

The tool wears as it cuts and therefore the diameter would increase with
wear. The work is somewhat elastic and will actually "bow away" from the
tool. This problem is affected by the nature of the material, the
diameter of the material, the length of the material, how the spinning
material is supported and where the tool is along the length of the

In general a skilled machinest could achieve the sort of accuracy
demanded by making many "incremental" cuts until the correct diameter was

Every different diameter required careful set up by the machinest.
Non-linear diameter changes could only be approximated to the extent that
the skill of the machinest permitted (the machinest had to turn the
cross-feed by hand as the longitudinal feed moved the tool along the

The pattern machine was developed to allow the machine to handle multiple
diameter cuts and non-linear cuts. Basically the machine had a
"controller" added.

The original controller actually had a "pattern" that was a flat plate
representing half the desired finished cross section for the work. A
"follower" cam ran against the pattern and was used to set the cross feed

The "controller" then was only slightly more than a unit that would
change the cross-feed position as the longitudinal position changed.
Early controllers were designed to limit the cross-feed change (in feed
only) for each pass (that is each cycle of the longitudinal feed from one
end to the other).

As the demand increased for both more complex pieces and more precise
pieces the electronics industry was creating some pretty nifty devices
suited for control purposes.

The "pattern controller" gave way to the "tape" machine where the
specifications for the pattern were fed to an electronic controller.
These units were not quite "open loop" in that the controller did measure
the position of the cross head as opposed to just turn the feed screw.
However, they were "open loop" with respect to the work. Some attempt at
"compensation" for disturbance was attempted. For example, some
prediction can be made for tool wear based upon amount of use and the
engineer (not the machinest) could sometimes compensate somewhat for the
"springing of the work".

Most modern precision machine tools now have sensors actually measuring
the work and then adjusting the tool. In the extreme, some of these
machine tools can handle a piece of work that may weigh as much as your
house and machine it to accuracies measured in angstroms. Some machines
measure the temperature of the work (multipoint) and even stress to
compensate the cutting tool position reference.

More that you ever wanted to hear? :slight_smile:


<[Bill Leach 950628.18:26 U.S. Eastern Time Zone]

<[Bill Leach 950628.01:07 U.S. Eastern Time Zone]

>[FROM Dennis Delprato (950627)]

Before anyone deservedly jumps on my ahh... err... message:

The control systems do not measure Stress but rather measure "Strain" and
attempt to compute stress.