Nicholas

I dug in my archive since we had the same question around January 97 in the
net.:

Blow out of PHT solder joints is not only a question of the copper
thickness but also of the moisture contents of the board and, very
important, of the quality of the drilled hole. If the PCB manufacturer is
using the drills too long or if he's feeding too much while drilling (
price pressure ) you will find ripped out glass fibres causing hollow
sections as well as glass fibres reaching out of the drilled walls. Both
cause a porous copper plating of the holes no matter how thick the final
plating is especially because the first process step of plating, where the
walls of the holes are covered electroless with a layer of copper in order
to make galvanic plating possible, will not cover all these irregularities.

Enclosed the data  from a trial for a 3D graph.You can use the data to make
a graph with any data handling program.

1. Specimen preparation:
Test boards where produced with drills of various age with various feeding
and various final plating thickness ( electroless plus electrical ).

2. Analysis:
Of each drill-feeding rate combination some samples have been drawn after
electroless plating. microsections of the holes where done without
embedding the specimen in resin. With transmission illumination the opened
holes where investigated. Not plates glass fibres where clearly visible as
bright spots.
The soldered specimen where investigated with microsections.

This work was done approx. 4 years ago at Siemens.
Due to the test set-up it is clear, that not all possible measurement
points could be found.

3. Graph
- X Axis ( classes of average final plating thickness )
- Y Axis ( classes of average amount of not metallised fibres after
electroless plating )
- Z Axis ( average amount of voids in the solder )

Data matrix as follows
- Top horizontal row = X Axis
- Left vertical row = Y Axis
- centre data = Z Axis


                        I 24-30 I 24-34 I 26-32 I 28-38 I 28-42 I 30-36 I 30-40
------------------I--------I---------I--------I---------I--------I---------I----
--
0                      I   8     I            I           I            I
4     I             I   1
1-10                 I  20    I            I           I            I
I             I
11-30               I          I            I           I   30     I
I             I   80
More than 30   I           I  172    I  165  I            I          I  95
I

You find as attachment the resulting graph as pdf file.


Best regards

Guenter Grossmann