<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
</head>
<body text="#000000" bgcolor="#ffffff">
<br>
Attendees <br>
Martin Hardwick, STEP Tools, USA <br>
Mikael Hedlind, KTH, Sweden <br>
David Loffredo, STEP Tools, USA <br>
Fred Proctor, NIST, USA <br>
Bengt Olsson, Sandvik, Sweden <br>
Fiona Zhao, NIST, USA/New Zealand <br>
Leon Xu, Boeing, USA <br>
Erik Jacobson, GE Energy, USA<b><br>
</b>Ronnie Fesperman, NIST, USA<span style="color: rgb(31, 73,
125);"></span><br>
Stephen Newman, University of Bath, UK<br>
Michael Mariani, IQL, USA<br>
Andreas Archenti, KTH, Sweden<br>
Mike Fair, Siemens, USA<br>
Thomas Charlton, ASME, USA<br>
<br>
We reviewed the minutes of the last conference call. The stacked
Homogeneous Transformations Matrices are stacked by error type, not
axis. Each matrix estimates a different type of error with respect
to positioning, straightness or some other criteria. The matrices
are stacked in the sense that they are applied in series to the
nominal position and location of the tool to determine a predicted
position and location. Each nominal position has its own matrix
stack with interpolation used to fill in for positions that have not
been measured. <br>
<br>
We discussed methods to represent this data. There is an XML format
in the B5.59 standards but the standard is a draft and the vendors
have been making extensions. <br>
<br>
We discussed how the data might be used by STEP-NC. There are
several scenarios. In the one discussed a STEP-NC program is run on
a machining simulator and the error prediction data is used to
convert the nominal tool path points into actual points and a
simulator (for volume removal or volume addition) predicts how
material will actually be added to, or removed from, the workpiece
on this machine. The final workpiece computed by the simulator is
then compared to the STEP-NC workpiece and the GD&T information
on the STEP-NC workpiece is validated by measuring the differences
between its dimensions and those of the simulated workpiece.<br>
<br>
We discussed how much machining error is due to errors in the
machining program and how much is caused by other factors such as
bending forces and thermal effects. The estimate is that on a good
machine the machining program contributes about 40% of the errors
and on a less good machine it contributes more.<br>
<br>
We discussed methods for correcting or eliminating the errors. The
first method being discussed is to develop a very accurate machining
simulator that will predict the result of machining a part on a
machine and so tell the user when a program is not suited to a
particular machine. A second method is to monitor the machining very
correctly, taking periodic measurements for critical tolerances and
correcting the machining as necessary using closed loop programming.
This was demonstrated at the Boeing Renton meeting in October. A
third method which is a hybrid of the first two is to make
experimental cuts during the machining, measure those cuts and use
the results to correct the main machining program.<br>
<br>
A recording of the call is on the ftp site at the address below. The
next call will be at the usual times on Friday May 7. <br>
<a class="moz-txt-link-freetext" href="ftp://www.steptools.com/private/Undecided/stepmanuf_telecon_20110422.wmv">ftp://www.steptools.com/private/Undecided/stepmanuf_telecon_20110422.wmv</a><br>
<br>
Action Items <br>
1. Try again to get copies of the ASME B5.59 draft standards.<br>
<br>
</body>
</html>