TECHNET Archives

June 1999

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:
"Collins, Graham" <[log in to unmask]>
Reply To:
TechNet E-Mail Forum.
Date:
Mon, 28 Jun 1999 08:24:22 -0300
Content-Type:
text/plain
Parts/Attachments:
text/plain (69 lines)
Component Traceability?  Whoa, I'm having flashbacks to a previous life...

Rick, two things to decide:
1) What level of traceability do you want / need / are required to have?
What are you willing to pay for traceability, and will you actually use it
enough to justify the cost (time, $$$)?

2) Then decide how to collect the data.  The options I know of are: manual
input, barcode, OCR.

Manual input: set up on a computer a template that identifies the part
number and prompts for the date code.  A couple of problems with this
method.  Obviously the time involved is not good - imagine a couple of
benches where Kelly temps are sitting typing date codes all day...  but by
far the most accurate way to record the info from parts that are marked with
date codes (IC's, bare CCA date code, other parts large enough to be
marked).  You will have to train the typists on what to enter - I've seen
plant codes that look like date codes (e.g. 9413) on parts, you have a
problem and start to dig through the data only to find all the date codes
are entered as "9413"...  aaaggg!

Barcode: if you scan the reel of parts when they are loaded, and you know
when a particular board was run, then you have a pretty good idea what date
code of parts are on a board, plus or minus a couple of boards.  Problems
with this are:
 - Difficult to sort the data, can be very time consuming
 - you have to know what the markings on the barcode mean - format differs
widely between manufacturer.  And are they marking it with their P/N or
yours?  Do they actually mark with the info you want?
 - If you buy from distributors, they may be re-reeling parts and hence
remarking labels.  So the markings on the reel form distributor A may not be
the same as from distributor B.
 - substitute parts can get very confusing (e.g. we can use DRAM from IBM,
Hitachi, Samsung, ...)

OCR - the holy grail of traceability.  If you can make OCR work, you get the
accuracy of manual input at a fraction of the cost and time.  But I've never
seen anyone deliver on the promise...  if you do, please let the TechNet
know!  As I understand it, the big problem is that the parts are all marked
differently - different fonts, print intensities, font sizes, and usually
orientation won't be the same - try making that a design rule!

Good luck!

Graham Collins
Process Engineer,
Litton Systems Canada, Atlantic Facility
(902) 873-2000 ext 6215


-----Original Message-----
From: Rick Vernon [mailto:[log in to unmask]]
Sent: Friday, June 25, 1999 4:56 PM
To: [log in to unmask]
Subject: [TN] Traceability requirements


Hello, TechNetters!

We are starting a group to review traceability needs in our plant. I'm
looking for input on what others in the electronics manufacturing industry
are doing. We are considering traceability to component level, with options
ranging to the other extreme of no traceability.

Are computerized systems being used? Cameras? Bar coding? All input
appreciated.

Thanks. Rick Vernon (701) 277-6185

ATOM RSS1 RSS2