Re: Re: test results

["Hassan S. Rabe" <RABEH-at-student.gvsu.edu>]

Hi Ray,

You are right....the fact that in both tests that were done in the beginning, there were a squareness error. Which is a machine error occurs when the axis are not perpendicular to each other. But this also can be a local error i.e. only in some regions along the axis or it can be that the whole axis is not perpendicular with the other.
 The reason why it didn't show clearly is that there was a bigger problem which is the scaling mismatch
 But now the scaling mismatch problem is eliminated...the squareness is now floating and can be seen clearly on the plot. If you remember ...that I said that the first thing that I learned in the ballbar training course was to fix the biggest problem and test again.
But even the scaling mismatch is a machine problem, but EMC is flexible enough that It lets you control the steps and by doing so, eliminating that error. 
 I guess I have to check that all the bolts are tight enough.
In fact the software produces more results than just a plot...it gives a numerical actual values and percentages of all the errors that meet the criteria’s defined within the software.

Hassan 



-----Original Message-----
From: Ray Henry <rehenry-at-up.net>
To: Multiple recipients of list <emc-at-nist.gov>
Date: Thu, 20 Feb 2003 21:00:35 -0500 (EST)
Subject: Re: test results


On Wednesday 19 February 2003 03:27 pm, Hassan wrote:
> Hi Ray,
>
> I made the changes to INI file, and tested the machine again, and
> definitely I had a better results. you can compare the current results
> with the older one that posted.
> the number of pulses that I subtracted and added to the x and y axis
> was 200.

Dave Engval sent me a couple of links to ballbar.  I see from some of the 
online literature that the kind of 45 degree squishing that we see in 
your latest plot could be attributed to x and y being non-perpendictular. 
The specific article was from Quality Magazine and I quote.

"Squareness error is shown in an oval ballbar plot, and occurs when 
normally orthogonal axes are no longer moving at 90 degrees, relative to 
one another. This may be because of a bent axis or some other 
misalignment. This squareness error oval "tilts" 45 degrees with respect 
to the two axes, and remains in the same position regardless of the 
direction of travel of the ballbar (CW or CCW)."

The address for that info is below.  You will have to put it back 
together and make a single url out of it.  Good luck

http://www.qualitymag.com/execute-vSection-Articles-vSub-Detail-vSideBar-Act
ive-vrID-32784197122C4F248698C9BA450CBE0E-vDate-7B52B3A3BB744965A5ADEDE4A45E
81BE.htm

Having said that about squareness, it does not seem to show this 
partictular feature very prominently on the plot made with the other CNC 
program.  Perhaps it is simply masked by other larger problems.

Ray