ROLAND,
        I'M NOT SURE IF THIS WILL BE OF ANY HELP OR NOT, BUT HERE GOES.  
     ABOUT 6 MONTHS AGO WE BEGAN HI-POT TESTING OF SOME NEW PRODUCT, (A 
     MULTILAYER BACKPLANE).  WE EXPERIENCE SHORTS ACROSS CONNECTOR PINS IN 
     VARIOUS LOCATIONS ON THE BACKPLANE.  WE REDESIGNED THE BOARD WITH 
     LARGER CLEARANCES BETWEEN PRIMARY AND SECONDARY POWERS AND GROUND.  IT 
     SHOULD HAVE WORKED, BUT IT DIDN'T.  MAJOR FUSTRATION.  WE BEGAN 
     LOOKING AT THE SYSTEM.  COME TO FIND OUT A CONTRACTED POWER SUPPLY HAD 
     HAD A BRACKET CHANGED, (WHICH DIDN'T EFFECT OUR ENGINEERING SPEC OR SO 
     WE THOUGHT).  THIS CHANGED CAUSED A POTENTIAL SHORT IN THE POWER 
     SUPPLY WHICH PASSED DOWN THE WIRING TO THE WEAKEST LINK ON THE 
     MOTHERBOARD.  I GUESS WHAT I'M SAYING IS THAT THE PROBLEM MAY NOT BE 
     WITH YOUR BOARD, BUT WITH A POWER SUPPLY OR THE AXIS MOTOR.  CONTACT 
     YOUR VENDORS AND SEE IF THEY'VE MADE ANY "IMPROVEMENTS".
                                        
     ROBERT BUTTERWORTH
     SR PCB DESIGNER
     ADC VIDEO SYSTEMS
     [log in to unmask]


______________________________ Reply Separator _________________________________
Subject: FAB:BURNS FROM SHORTS
Author:  [log in to unmask] at internet-mail
Date:    5/12/96 9:58 AM


Dear Teknets,
     
We would appreciate your help for solving a though case leading to high 
costs in the field. To our knowledge this is not very common.
     
A customer have machines that break down due to an electric short on axis
motor drivers. The shorts show up at random on the board surface, and are burns.
They are not necessarily located on holes.
     
The boards are in machines and a growing percentage is coming back from the 
field. Some are 3 years old now. Older boards from the same manufacturer don't 
have this problem. The design being the same.
     
It happens onto two different boards of different building and thicknesses. 
     
In both cases, part of the production is made pin-lam in house by the 
manufacturer,
the other is made mass-lam by a Mass-lam manufacturer.
     
REF - 125
4 layers - thickness 2.4 mm - inner + outer copper 1 oz 
FR4 core is 0.56 mm thick
80 VDC - 10 A
5 VDC - 240 mA
+- 12 VDC - 140 mA
     
REF - 126
4 layers - thickness 2.4 mm - inner + outer copper 2 oz 
FR4 core is 0.41 mm thick
200 VDC - 16.5 A
  5 VDC - 10 A
110 VAC - 5 A
124 VDC - 1 A
  5 VDC - 1 A
 18 VDC - 1 A
 24 VDC - 1 A
     
     -  -              ---------- copper outer layer 1
\/\/\]  [\/\/\/\/\/\~/\/\/\/\/\/\ FR4 Preg 7628      2 several
        ~~ ~~ ~~~~ ~ ~~~~~o~~~o~~                    x
\/\/\]  [\/\/\/\/\/\/\/\/\/\/\/\/ FR4 Preg 7628      3 several
     ]  [   --------------------- copper innerlayer  4
=====]  [======================== FR4 core           5
     ]  [   --------------------- copper innerlayer  6
\/\/\]  [\/\/\/\/\/\/\/\/\/\/\/\/ FR4 Preg 7628      7 several 
\/\/\]  [\/\/\/\/\/\/\/\/\/\/\/\/ FR4 Preg 7628      8 deveral
     -  -              ---------- copper outer layer 9
     
x = path used by the short on a long distance, 2 + 3 splitted
    and we can see dark veins (burns)containing copper residues... 
    inside the length of the board. We can see cavities + copper 
    particules. It looks like veines of a tree.
o = looks like copper residues trapped between the prepregs.
     
There is no dammage between 4 + 6.
     
An explanation could be that, the prepreg resine is removing 
oxydated copper fragments and they concentrate in the middle 
of the two prepregs. Then aging (temperature of the working 
board) the shorts are possible within those small residues.
     
I don't dare sending you all pictures in .jpg format as an 
attachment, but if you do like to have a look on the microsections 
let me know. They are stored in my PC and I will send it to you 
ASAP.
     
Thank you for your help
Roland
     
     
     
[log in to unmask]
Henri Jaquet SA
Geneva Switzerland
tel. xx41-22-794-7878
fx. xx41-22-794-3052