A previous debate on delamination/baking boards sparked a response 
from Robert Willis from the UK, realative to pin holes/blowholes in 
solder connections.

Robert Willis stated:

"Blow holes are not caused by component terminations or poor wetting 
of the plated through hole.  If anyone thinks this they are 
misunderstanding what the term pin hole or blow hole is."  This is 
such a strong statement it forced me to add my two cents to his 2 
pence.

I will qualify my response by stating that I have performed and/or 
evaluated thousands of micro sections of plated through holes in my 
30 years in this business.  This includes both the populated and 
unpopulated board.  Let's first define pin holes and blow holes in a 
solder connection.

Pin Hole:    A pin hole is a depression in a solder connection, 
whereas, the bottom of the depression is visible after cleaning.

Blow Hole:   A blow hole is a void in a solder connection, whereas, 
the bottom of the void is not visible after cleaning.

Without a sketch, this is the best definition.  A pin hole is 
generally caused from a small plating void in the copper plating of a 
PTH.  It can also be caused from flux entrapment in the hole, whereas,
 the pre-heat did not drive all of the volitiles from the flux 
(improper control of pre-heat) prior to entering the wave.  In the 
case of copper plating voids, gasses are allowed to escape into the 
hole area.  With either flux entrapment or plating voids, a small gas 
bubble will expand with heat and contract upon cooling, thus, pulling 
the surface solder back into the hole in the form of a pin hole in 
the solder connection.

A blow hole is caused from a wetting problem with either the 
component lead or the copper hole wall.  It can also be caused from 
larger plating voids, which ultimately is also a wetting problem 
because you can't wet solder to epoxy glass.  Obviously, larger voids 
will allow more gasses into the hole area that may result in larger 
pin holes or blow holes.  However, my experience through micro 
sectioning of the holes has shown a wetting problem 90% of the time.  
The other 10% was either very large plating voids or an evaluation 
could not be made.

Mr. Willis mentions the involvement of the National Physics Lab in 
the UK, realtive to a study of outgassing.  Many years ago (10 - 20), 
I received a copy of this study before it was totally complete.  The 
study showed considerable outgassing in the hole after the 
electroless copper process.  This was a reasonable conclusion as 
electroless copper is only 75 to 100 micro inches thick and very 
porous.
The study further showed that outgassing, if any, was very minimal 
after copper electroplating to the required thickness.  They also 
mentioned outgassing from the surface of the board where the surface 
copper is only 80 - 85% covered with copper.  Certainly, this all 
makes good common sense.

Pin holes are of little concern, relative to quality issues, and 
should not be touched up with a soldering iron.  The old saying, " 
keep the the iron off the board surface unless it is absolutely 
necessary", certainly applies here.  The use of hand irons, if the 
process is not totally adhered to, will rapidly destroy good holes.  
When pin holes are evident, the source of the problem should be 
investigated and eliminated.  Though it is not a quality issue, it 
should not be ignored.

Blow holes can be a serious reliability issue, as you don't know the 
extent of the blow hole without destructive testing.  However, here 
again it must be stated that blow holes should not be touched up.  
Again, find the source of the problem and eliminate the problem.

Concave or sunken solder joints:  You will find that sunken or 
concave solder joints are generally caused from barrel cracks.  This 
is a very serious reliability issue and should be further 
investigated.

Let's assume that your fabricator plates your required copper 
thickness in the holes (.001" min.), without any voids.  I find it 
very difficult to believe that gasses can pass through a solid shield 
of .001"  thick copper in the 3 to 4 seconds that the board is 
exposed to the wave solder pot, unless there was plating defects.  
This only makes good common sense!

We all know that the board material is hydroscopic and absorbs 
moisture to some degree.  However, it is a small percentage of the 
total weight.  When viewing a micro section with blow holes in a 
microscope, either the component lead or hole wall will evidence 
absolutely no solder coating.  This is either dewetting or nonwetting.
  Though there is some moisture in the material, it would require a 
high pressure steam hose to clean the areas shown as not wetting in 
the scope.  If you don't believe this, do your own micro section the 
next time you see a blow hole.

Mr. Willis did a good job explaining his process of testing for voids 
or thin plating, with the use of oil in the hole and heating it up.  
I think most people agree that touch of solder connections with a 
soldering can create far more problems than it would cure.  To 
perform the oil test as descibed by Mr. Willis would require 500 
degrees "F" for 30 seconds on the board surface.  I submit that this 
would destroy good boards.  I may also suspect that this temperature 
and time could crack plating barrels and lead you to believe that you 
had defective boards, when it could be the test itself causing the 
problem.  Basically, whenever using a hand iron, you must get on the 
connection with the iron and off again before the epoxy surrounding 
the hole has a chance to completely expand.  This generally leaves 1 
to 1-1/2 seconds to complete a hand soldering process.  This includes 
soldering and desoldering.

It appears to me that if you are going to use Mr. Willis' oil test, 
you should ask your fabricator for scrap boards from the same lot for 
test purposes.  Don't destroy god boards!

At its best, I think you could use the oil test as an indicator of 
plating voids and very thin copper plating.  IPC-RB-276 allows voids 
in Class 2 and no voids in Class 3.  Except in very isolated cases, 
voids will seldom be an issue today.  In fact, most vendors will 
admit that if voids are evident, there is a process control problem.

Norm Einarson
PRINTED CIRCUIT TECHNOLOGY