Re: measuring scales and thermal compensation
Posted by
Jon Elson
on 2000-03-16 22:22:02 UTC
Ian Wright wrote:
unlikely to be good
for better than .001" on a very small machine. Ther is stretching of
the wire, slipping
of the wire on the drum, variations in how the rope lays on the drum,
variations in
the thickness of the coating on the rope, variations in the twist in the
rope strands,
chips getting under the rope as it wraps on the drum, vibrations in the
rope, thermal
variations in the wire rope tension, on and on and on.
As for thermal variations, there are several ways to think about this.
Obviously, when
the workpiece is excessively heated by the metalworking process,
accuracy suffers.
So, you need to let the workpiece cool before taking finishing cuts.
Now, if you can actually keep the workpiece and the entire machine at
the same
temperature, throughout the bulk of the machine, then you can hope that
thermal
variations will all self-correct when the work, machine, and measuring
scale
all return to 20 C. But, without tremendous effort, like bathing the
entire machine
in 20 C coolant (which is actually done, by the way) the machine is
distorted by
heat generated in the workpiece, and often by ambient temperature
changes, too.
The glass (or quartz) scale idea is that the scale is fixed at a
reference point at
one end, and allowed to sit without thermally induced stress being
applied to
it. (Only the better scales actually implement this, but many use
resilient mounting
techniques.) Since the thermal expansion of glass (and especially
quartz) is
much less than steel, the measurements of machine position will be more
accurate as temperature varies. I think, overall, that a glass scale is
more
accurate, but you have to carefully control workpiece temperature to get
any accuracy at all. But, thinking that a metal scale will heat up
exactly in
concert with the workpiece is not realistic.
In other words, something that works well for ruling measuring scales is
not
applicable to heavy machining tasks.
Jon
> I was told, was that the graduating machinesThe wire rope gizmo has so many possible sources of error, it is
> and standard scales were all made of metals having the same
> coefficient of
> expansion (iron and steel) and so, once they had been tested and
> verified as
> absolutely correct at the 20 deg C standard temperature, it didn't
> matter
> what temperature they were used at in the workshop as the raw
> materials
> would have expanded or contracted by the same amount and so would also
> be
> correct at 20 when finished. For this reason glass scales were avoided
>
> except as negatives for photoetching the cheaper range of scales. So,
> from
> this, it would seem that steel DRO encoder strip would be the best
> choice
> and that it should be intimately fixed to the machine if possible,
> with wire
> 'rope' on a rotary encoder being the next best. Comments?
unlikely to be good
for better than .001" on a very small machine. Ther is stretching of
the wire, slipping
of the wire on the drum, variations in how the rope lays on the drum,
variations in
the thickness of the coating on the rope, variations in the twist in the
rope strands,
chips getting under the rope as it wraps on the drum, vibrations in the
rope, thermal
variations in the wire rope tension, on and on and on.
As for thermal variations, there are several ways to think about this.
Obviously, when
the workpiece is excessively heated by the metalworking process,
accuracy suffers.
So, you need to let the workpiece cool before taking finishing cuts.
Now, if you can actually keep the workpiece and the entire machine at
the same
temperature, throughout the bulk of the machine, then you can hope that
thermal
variations will all self-correct when the work, machine, and measuring
scale
all return to 20 C. But, without tremendous effort, like bathing the
entire machine
in 20 C coolant (which is actually done, by the way) the machine is
distorted by
heat generated in the workpiece, and often by ambient temperature
changes, too.
The glass (or quartz) scale idea is that the scale is fixed at a
reference point at
one end, and allowed to sit without thermally induced stress being
applied to
it. (Only the better scales actually implement this, but many use
resilient mounting
techniques.) Since the thermal expansion of glass (and especially
quartz) is
much less than steel, the measurements of machine position will be more
accurate as temperature varies. I think, overall, that a glass scale is
more
accurate, but you have to carefully control workpiece temperature to get
any accuracy at all. But, thinking that a metal scale will heat up
exactly in
concert with the workpiece is not realistic.
In other words, something that works well for ruling measuring scales is
not
applicable to heavy machining tasks.
Jon