TECHNET Archives

1995

TechNet@IPC.ORG

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Date:
Sat, 21 Oct 95 19:52:15 EST
Content-Type:
text/plain
Parts/Attachments:
text/plain (57 lines)
          Jerry:
               You are correct in saying that most board houses don't
          do a "true" net list test based on your supplied netlist;
          rather a test fixture is created from the Gerber in most
          cases.  At least this is the case at HADCO.  HADCO takes the
          Gerber data provided and "pre-netlists" it before any
          modifications.  These modifications (teardropping, adjusting
          pad & line widths, thieving, etc.) are then performed and a
          "post-netlist" is generated.  We then do a pre/post netlist
          compare to ensure that we did not chnage the intent of the
          design in any way.
               We have embraced IPC-D-356 netlist formatting as the
          vehicle for those of our customers who prefer to supply us
          with a netlist from which to build the fixture.  We viewed
          -356 as the most universally accepted format out there.
          Unfortunately, there are usually discrepancies between
          supplied netlist and Gerber (on one job there were over 700
          discrepancies!).  For this reason, we have chosen to use a
          certification program with our customers who desire us to
          use their -356 data.  Three successive designs must be sent
          in with no discrepancies to become certified.  It's a great
          learning curve for both HADCO and our customer.  At present,
          I believe we only have two "certified" customers out of a
          customer base of over 500.  Let me share with you some of
          the common problems encountered:
               - Non-plated holes called out as a net
               - Ground areas which are physically disconnected but
                 part of the same net, which are not listed in the
                 same net
               - Naming convention errors (ex: "Signal A" later listed
                 as "Sig A" and then not tied together)
               - Manual changes to Gerber not reflected in CAE data
                 (ex: split power/ground)
               - CAE netlist not in IPC-D-356 format.  Much of the
                 data that we receive seems to have a personal touch
                 mixed in with the IPC format.  We need data that has
                 been built under the guidelines of IPC-D-356 from the
                 first column to the last (70th).
          We will gladly work with any customer who wants us to
          generate fixtures from their supplied IPC-D-356 netlist
          data.  We need the name, telephone #, and fax # of a person
          designated to work resolutions on the CAD vs. Gerber
          conflicts.
               In summary, IPC-D-356 testing is a computer-generated
          file in a specific format which describes which points are
          interconnected, or part of a common net.  Gerber-extract
          testing is taking customer Gerber image files, processing
          this information on a CAM workstation, and then
          computer-generating a netlist program for electrical test.
          Regards,
          Tom Coyle
          Field Services Engineer
          HADCO Corporation
          [log in to unmask]



ATOM RSS1 RSS2