Re: http://www.automata.com/ipc/TechNet/0436.html
Hi, Jim:
I found the request for info that you posted some time back on the "average"
number of solder defects that occur on PTH assemblies. I don't know whether
you got any responses, but I thought I'd throw my comments your way. I have
talked to people who state that the solder defect rate in their factory is
in the hundreds per million (check with Jerry Rosser, Dave Hillman, Ralph
Woodgate), and years ago I visited a Hewlett Packard plant in Cupertino
which was running in the single digits per million, so I know that very low
rates are possible - not just on occasion but consistently.
Every defect has a cause - it isn't just a random event. The manufacturing
engineer who is responsible for the process must identify every cause and
eliminate it. The cause may not be the process - it may be a bad design or
acceptance of bad material. This latter category would include not only
solderability, but also plating of the through holes. There should be no
need to require solder fill to compensate for conditions like plating voids
and cracks in the knee - a board with such conditions should never get into
the factory (and the reliability of the "fix" is dubious). Many
manufacturing engineers regard problems of design and material as outside
their domain and too difficult to deal with, so they just learn to live with
chronic pain as they attempt to "make do" with the processes that are under
their control.
A defect may also be "caused" by the inspector or the acceptance
requirement, in that a condition that does not threaten the reliablity may
be identified as a defect because of a belief. "Grainy solder" is an
example of what I'm talking about. Solder that freezes slowly is not smooth
and shiny, but it's just as reliable, and it was just as reliable even back
when it was believed that solder had to be shiny to be acceptable. Each
such belief impedes progress to the intended goal; eliminating each takes a
lot of time and discussion (and may not even be successful, since some
people seem to prefer problems they understand to solutions they don't).
Some inspectors reject connections that they don't like the looks of,
regardless of whether a nonconformance to a requirement exists. In fact, of
all the processes going on in the factory, the one most likely to be out of
control is the solder inspection process, as every study that has ever been
reported has shown. Which connections get reworked is largely a matter of
who looked at them.
The defect may even be caused by the standard - you probably remember when
the Weapons Spec set a limit on how hot the top of the board could get
before the board entered the wave, limited the temperature of the solder,
and limited the time in the wave, while also requiring that the solder flow
out onto the top land. That combination of requirements artificially set a
limit on the heat capacity of an assembly that could be wave soldered, since
the top land has to get above the melting temperature of the solder in order
to wet. We've come a long way since those days, and today soldering
standards do not impose arbitrary process limits.
There is no reason why there should be any defects at all on a
properly-designed assembly if built only with parts that actually meet their
procurement specifications, and if only those conditions that reduce
reliability (which I term "risky defects") are counted, by people who have
been trained (and certified) to make good decisions.
The biggest problem I have encountered in trying to reduce the occurrence of
solder defects has been the tightly-held (and widely-held) belief that
"we've always had solder defects and we always will" - a self-fulfilling
prophecy. This fatalistic attitude regards solder defects as inevitable, as
if the most that could be hoped for is keeping them under control.
Gordon Davy
Northrop Grumman ESSD
410-993-7399
***************************************************************************
* TechNet mail list is provided as a service by IPC using SmartList v3.05 *
***************************************************************************
* To unsubscribe from this list at any time, send a message to: *
* [log in to unmask] with <subject: unsubscribe> and no text. *
***************************************************************************
* If you are having a problem with the IPC TechNet forum please contact *
* Dmitriy Sklyar at 847-509-9700 ext. 311 or email at [log in to unmask] *
***************************************************************************
|