Re: how straight is straight[rails]?

Hi Carl,

I think the cost would be higher than 8K but I am very sure it can be
done. You have the posibility of adjusting every point down the line
so all you need to do is to find where your errors are. Of course
that sounds simple. And little gremlins like temperature expansion
will creep up.

And as Sven mentioned, ridigity will go way up if he wants to take
deep cuts and keep that type of accuracy.

Another course of action that I have thought about is using a laser
for absolute posistioing in both the X and Y axis, then using
software to correct for the error. This could be a physical process
on set-up or a realtime process monitoring.

yet another laser application would be to just correct the deviation
from the straight line. This would be rather simple and could be
hardwired as well as put in software. I think the Plasma folks are
doing a similar feedback adjustment. Using the arc as the thing
being sensed.

And another option might be to program the tool path errors and
correct for them in software.

The spec I just wrote was +/- a 16th for fractional sizes and +/-
0.015" for decimal, had to be held withing that over 6 inches. !
And no finish was specified, but if you drag it through the parking
lot, you must wash the gravel off before shipping.

Honest, that was in the written spec. And there were only two
fractional sizes, and one was a drilled hole.


You did mention router. is this for wood ? and is there really a
project using wood that needs that type of accuracy ? I mean doesn't
wood move that much with humidity changes ?


Dave

--- In CAD_CAM_EDM_DRO@y..., "carlcnc" <carlcnc@y...> wrote:
> I had a guy ask me if I can make a router within .001 straightness
> over 48",
> I am not sure this is possible with "normal" equipment.
> Taken literally that would mean an accuracy of .00002 per inch!
> {My ground straight edge is .002 per manf.}
> Do the "big boy" CNC builders achieve this kind of tolerance?
>
> Just to waste some time I checked my 5x8 in X.Y. best I can get
> is +/- .005 over 6',using said straight edge.
>
> The rest of his request was " flat to .005,and less than 8K"
>
> Carl