Following from the EMPF HelpLine, Indianapolis. Original question posted 
10/3 (we jumped right on this question!)

Ed,

My name is Tim Crawford and as co-chairman of the IPC Ionic Conductivity 
Task Group, I will attempt to answer your question.  Your question, of 
course, is "How clean is clean?"  which has been asked millions of times by 
thousands of people.  The problem is that there is no good answer due to 3 
areas; the way we measure, what we measure, and the environment/use of the 
PCBA.

First, the way we measure.  The bulk resistivity meter you are using to 
measure cleanliness was not designed to be an analytical tool.  A study 
performed here at the EMPF in conjunction with IPC (available from the EMPF 
Tech Library by calling 317-226-5623) showed that test results varied from 
test equipment to test equipment.  Furthermore, the results could be 
influenced due to certain variables, such as test time, solvent temperature, 
or alcohol content.  This is not to say that these systems are useless, 
rather, when used consistently, they are valuable process control tools.  If 
your process typically measures PCBA's at  1.56 micrograms/cm2 every day, 
then one day the results come back  5 micrograms/cm2, then something 
happened to your process.  The cleaning solvent was not up to temperature, 
spray nozzles plugged, rinse water quality went bad, wrong flux, too much 
flux, wrong thermal profile, etc.  It could also be a problem with the way 
the test was performed, or the way the PCBA was handled.  I would not be 
concerned with the reading of  5 micrograms/cm2 (it's probably wrong anyway) 
but the significant change in test results does show a process problem.

What we measure.  These systems will only measure ionic residues, and only 
ionic residues that dissociate in 75% IPA, 25% water.  There are several 
manufacturing residues that may influence reliability that are not detected 
in this test.  Another factor in the "what we measure" category is the PCBA 
itself.  You mentioned Class 1,2 and 3, but more important is the 
technology.  An acceptable level for an old through-hole board with 
centimeters distance between components should be different than that of a 
surface mount board with millimeters distance between components.  The 
problem is that you can't establish a pass/fail criteria for every 
technology.

Finally, and probably most important, the end environment/use of the PCBA. 
 The cleanliness level of a PCBA that is going into a child's toy to be 
given away in a cereal box is different than a PCBA that controls an 
airplane's landing gear or a heart pacemaker.  Nothing against toys in 
cereal boxes, but you get the point.  In addition, the cleanliness 
requirement of a PCBA that is placed in a sealed box, in a temperature 
controlled, humidity controlled, protected from the rain, sleet, snow 
environment is probably not as critical as that of a PCBA in an unsealed box 
exposed to harsh environmental conditions.

In summary, these systems are not analytical tools.  Even if they were, 
putting a specific pass/fail level to a specific level of reliability is 
pretty hard to do.  The 1.56 micrograms/cm2 that is called out in 
ANSI/J-STD-001 came from old military specs and have been used successfully 
for about 20 years.  Not that it is a magical pass/fail level where 
reliability drops off, but rather it is an easily achievable number that 
shows good control of the process.  If you would like to discuss this 
further, you can call me at (317) 226-5634 or e-mail me at 
[log in to unmask]