TECHNET Archives

January 1997

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:
[log in to unmask] (DAVY.J.G-)
Date:
Mon, 13 Jan 1997 10:19:54 -0500
Content-Type:
text/plain
Parts/Attachments:
text/plain (127 lines)
     On January 9 Thad McMillan posted eight questions on TechNet related 
     to filling through-holes with solder paste and reflowing it (instead 
     of filling by wave soldering), which is what a subcontractor does on 
     high-density double-side SMT/through-hole assemblies.  The resulting 
     connections typically have solder in the middle of the board, with 
     each end of the hole unfilled. Among the questions he asks are whether 
     such connections are reliable, what standards exist that describe a 
     "good" or "bad" connection, and how such a board might be designed to 
     get better-looking connections.
     
     I have noticed two other recent postings to TechNet that also ask 
     about hole fill.  One questioner was having difficulty filling holes 
     because of the heat capacity of the board (very thick) and the other 
     was losing solder due to blow-out from water vapor evolution. Here is 
     a modification of a response I posted previously, starting with two 
     questions of my own:
     
     1.  How hard it is worth trying to get a through-hole filled with 
     solder?
     
     2.  What is the requirement?
     
     When the word "requirement" is used, it is used in one of two ways: · 
     a "real" requirement - necessary to ensure reliability, and a "paper" 
     requirement - necessary to comply with the contract.
     
     Studies have shown that filled holes are not a real requirement. The 
     best that can be said for filling the holes was alluded to by Jim 
     Moffitt in TechNet recently: to compensate for brittle copper plating. 
     If the copper has cracks in it, solder can bridge them and help ensure 
     electrical continuity.  There was a time, about a generation ago, when 
     the plating process was not as controlled as it is today, and perhaps 
     there was some merit in using the solder as a Band-aid to save the 
     board and meet the schedule.  Perhaps.
     
     But what would be a reason for insisting that solder fill the hole 
     today?  Some would call incompletely filled holes a process indicator, 
     and in some cases it is.  It may also be a design indicator, with the 
     poor manufacturing engineer being asked to cover for the design 
     engineer who caused the condition, to save the board and meet the 
     schedule.  In the case of screening solder paste to fill the hole, it 
     can be thought of as a process indicator in the sense that it indi- 
     cates what process was used, but it doesn't indicate a _deficiency_ in 
     the process unless it is assumed that the process must fill the hole, 
     which is exactly what is at issue. In other words, if a condition 
     doesn't indicate that something went wrong, then it shouldn't be 
     called an indicator, because there is no preventive action that is to 
     be taken in response to it.
     
     The study which showed, to my satisfaction at least, that holes need 
     not be filled was paid for by the US Army (Product Assurance Directo- 
     rate, Picatinny Arsenal-Dover, NJ, Contract No. DAA21-76-C-0100) and 
     performed by the Lockheed Electronics Company in Plainfield, New 
     Jersey over twenty years ago:
     Gangemi, R. and P. Cipolleti, "The Dynamic Measurement and Functional 
     Inspection of Solder Joints", Picatinny Arsenal Technical Report No. 
     5055, December 1976.  Available as AD-A034852 from National Technical 
     Information Service, Springfield, Virginia, 703-487-4650 ($39).
     
     This study showed that connections between leads and plated-through 
     holes do not have to be very good, by prevailing visual inspection 
     standards, to be as reliable as, or even more reliable than, connec- 
     tions judged to be acceptable.  Only two causes of premature failure 
     were found:
     · Less than 55 percent of the area of land and lead on the solder side 
     showed good wetting
     · Plated-through holes were less than 25 percent filled with solder. 
     All other visible "defects" studied were found to be unrelated to 
     reliability.
     
     These results, though carefully researched and documented, fly in the 
     face of conventional wisdom almost as much today as they did then, and 
     seem to have been completely ignored.  They have never been challenged 
     with conflicting data (at least in print), but neither have they been 
     accepted.  More recently, Colin Lea of the National Physical 
     Laboratory in England has also shown that hole fill is not a real 
     requirement.  He found that filling the hole actually slightly reduced 
     reliability (as judged by thermal cycles to failure), perhaps because 
     of the reduced mechanical compliance of the leads.
     
     In 1989 a government-industry team visited a number of military repair 
     depots to look at solder connections on boards that had been returned 
     for repair.  Out of over one hundred thousand connections that the 
     team examined, a huge number of connections were found which were 
     "bad" by anyone's solder acceptance criterion, but only four were 
     found that had actually caused a failure, all due to terrible solder- 
     ability.  Ten unsoldered through-hole connections were found.  Not one 
     of them had caused an electrical problem due to opening the circuit.  
     (The report was published as EMPF-TR-008 in 1989 with the title of 
     "Tri-Service Evaluation of Field Electronic Hardware" and is still 
     available from http://empf.arl.psu.edu/library/.)
     
     The question "What is a 'good' solder connection?" has a simple 
     answer: A "good" solder connection is one that is unlikely to fail in 
     service. Of course, this answer doesn't tell how to recognize what is 
     likely to fail, but based on the above it seems that a connection can 
     look pretty bad and still not fail.  As for what the standards say is 
     "good", acceptance requirements are arrived at by people voting.  Some 
     vote on the basis of data, some vote on the basis of belief, some vote 
     on the basis of what can be sold to the customer, and some vote on the 
     basis of the probability of a condition that bad or worse actually 
     occurring.
     
     The conclusion that I draw is that a connection which includes solder 
     that wets the adjoining surfaces is unlikely to fail in service, and 
     this would be my description of a "good" solder connection.  If you 
     have a connection that you're not sure about, it's probably OK, since 
     one that is likely to fail will look so bad, you'll not think twice 
     about reworking it.  So my advice (based on the data and reasoning 
     I've presented) is, unless you are required by contract to fill it, or 
     unless you have reason to doubt the integrity of the copper plating, 
     if there is solder in the hole, leave it alone and let it go.
     
     Gordon Davy

***************************************************************************
* 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]      *
***************************************************************************



ATOM RSS1 RSS2