From vsundar@kingkong.ME.Berkeley.EDU Thu Sep 24 22:14 PDT 1998
To: sequin@cs.Berkeley.EDU
Subject: parts
I am sorry I could not send you the reports before.. I am sending some
problems that I encountered with the files in the next few mails
BTW, I have put all the ACIS .sat and VRML .wrl files in the
http://madmax.me/~cybercut/acisfiles.
One file - blshyu-1.dsg did not parse, probably because of some ACIS
error in my program. I shall look into the problem later tomorrow.
The rest, I think, at least went through the planner.
Sundar
------------------------------------------------------------------
From vsundar@kingkong.ME.Berkeley.EDU Thu Sep 24 22:29 PDT 1998
Subject: File - NoahFinal.dsg
1. The .dsg file was corrupted. The file has a construct that looks as
follows: ( no edges defined). I commented out the line to facilitate
parsing.
(pocket 1 (e 33653 -5)
(loop)
)
2. Corner radii are either not specified or are too large. Three of the
features have no corner radii and one of them (the crescent shaped
feature) has a corner radius of 0.67 defined on it. We use a default
corner radius of 0.25 which I think will completely wipe out the
star-shaped feature and will make the cross look very much rounded.
Of course, specifying a small corner radius will limit the depth
of the feature greatly. But it seems that some features such as the star
shaped feature actually is very shallow. So it may be possible to define
a combination of features to get the desired effect.
------------------------------------------------------------------
From vsundar@kingkong.ME.Berkeley.EDU Thu Sep 24 22:52 PDT 1998
Subject: File - drewp2.dsg
This file had no problems as far as the system is concerned. But the part
the designer gets back will probably take her/him aback. The corner
radius specified on all the features is 0.67 inches. This huge corner
radius specified makes the planner select a large tool for all features,
with the result that one of the features will not get cut at all. (when
the profile of the feature is offset by the tool radius, the geometry is
almost degenerate). Decreasing the corner radius will decrease the max
depth possible, fortunately in this part most of the features are
through, so they can be described as two features from opposite sides.
There might be some rigidity problems during fixturing. If a vise is
used, the two clamping faces would be the two legs of the "M" and that,
intiutively does not seem very rigid. It may not be possible to clamp it
down to the table, since there does not seem enough place to place toe
clamps.
We really appreciate the designers extending all the edges beyond the
extremes of the stock for the open edges.
------------------------------------------------------------------
From vsundar@kingkong.ME.Berkeley.EDU Thu Sep 24 22:59 PDT 1998
Subject: File - stacey.dsg
There seems to be an error in the design (although I am not sure about
that). The ACIS file shows a thin step along the inside of a pocket,
possibly to incorrect registration of pockets. Please look at
http://madmax.me/~cybercut/acisfiles/stacey.sat or stacey.wrl
------------------------------------------------------------------
From vsundar@kingkong.ME.Berkeley.EDU Fri Sep 25 04:18 PDT 1998
Subject: File - jaeho.dsg
No major problems, except some corner radius issues need to be
considered. There are several triangular features that have sharp
corners. In many cases, the corner gets washed out by some other
intersecting feature. In some, the corner remains.
------------------------------------------------------------------
From vsundar@kingkong.ME.Berkeley.EDU Fri Sep 25 04:32 PDT 1998
Subject: file - danyelf2.dsg
The part does not seem rigid enough to support clamping and machining
forces. If the largest dimension is regarded as the x-axis and the
shorter ones the y and the z, clamping is only possible on the y faces of
the part. The part is quite weak in the x direction.
There are several very thin sections in the part. The part being made of
plastic, will probably not withstand the machining forces. Our experience
in the past with very thin sections has been the material tends to flow
along the cutting direction. (doesn't look pretty)
The pockets that extend beyond the stock seem to extend a lot. Did the
designer have anything in mind?
------------------------------------------------------------------
From vsundar@kingkong.ME.Berkeley.EDU Fri Sep 25 04:38 PDT 1998
Subject: file - wittman1.dsg
The pockets need to overlap, otherwise they leave the corners of the
tools behind. Also if a feature touches the boundary of the stock, it
must extend beyond it, so that the corner gets cleared.
I could try running this part through the feature recognizer. That will
clean up a lot of these problems, but there are still two pockets that
don't quite touch the stock boundary. It can been seen in the ACIS .sat
file or the VRML file.
------------------------------------------------------------------
From vsundar@kingkong.ME.Berkeley.EDU Fri Sep 25 04:48 PDT 1998
Subject: file - ychen_final3.dsg
Pockets that touch the boundary of the part must extend beyond it.
There is a feature that has a corner radius of 0.025. There is no tool
available that can give such a small corner radius.
------------------------------------------------------------------
From vsundar@kingkong.ME.Berkeley.EDU Fri Sep 25 04:52 PDT 1998
Subject: General
I planned 6 files, 1 of them didn't parse properly. I guess it is some
ACIS bug in my program. I shall look into it tomorrow.
One of the recurrent problems seems to the corner radius/ extension of
the pockets beyond the stock boundary. I shall work over the weekend to
put the corner radius in the .sat and the .wrl file, so that designers
can see the corner radius, rather than just guess. I am going to make
that the top priority for the next couple of days..
I won't be able to go to your class, as I myself have class from 9 - 11.
Sundar
------------------------------------------------------------------
------------------------------------------------------------------
Subject: fixturing
Sundar,
I don't quite understand your comments on clamping / fixturing.
I thought we leave as much of the square stock as necessary
at either end so that the rod can be clamped on either side
while machining the fours sides.
Carlo