Meaningful SPI Production Control Through Automatic Normalization

Automated pre-placement solder paste inspection (SPI) has established itself in SMT assembly as an essential tool for optimizing the process and reducing end of line defects.  It’s been commonly accepted for many years that a majority of assembly defects can be traced back to print defects, and by using raw measurement data from printed PCBs for volume, area, height and offset, SPC histograms and control charts can quite readily be used to optimize the process.

Following optimization, SPI is used to detect print anomalies. However, the multitude of “normal” variables introduced from board characteristics, clamping methodologies, stencil variations and paste type force unacceptably wide tolerance settings.  These board-wide good/bad tolerances can easily mount to +/- 50% or more.  For most, this is unacceptable, because of course, with such wide tolerances production drift can be missed and important defects ignored.

Tolerances before and after Optimization are similar

Before and after SPI optimization

 

Aperture Release

For example, stencil pattern variations cause differences between designed (expected) and actual solder release. For large apertures, release variation is generally negligible. Small apertures can account for up to 50% release variation or even more.  Stencil thickness to aperture area ratio determines how close to 100% release efficiency is achievable.  As pads get smaller and smaller, the area of the aperture wall approaches that of the pad area.  In turn, the paste release efficiency is reduced as the adhesion forces from the walls approach that of the pad.   Thus, different pad sizes across a board will exhibit different volume performance.

A common work-around for this is possible by selecting specific areas and setting tolerances for these apertures.

Though these custom tolerance settings may be individually easy to perform, they are time consuming and do not carry over to other PCB’s due to differences in design, manufacture and finish. These differences are common practice in just in time manufacturing:

  • Solder mask thickness variations; typical thickness 25 um  +25/– 12um; Silk screen legends can add 12 um
  • Solder mask defined or non solder mask defined apertures
  • Gasketing effect variations
  • Solder pad finish
    • OSP and gold 0 % variation
    • HASL +/-20 um
  • Board Warpance

These all create variations on the “zero reference height” point on each pad for measurement.

Mek Automated Normalization

Automated Normalization has been introduced by Mek SPI which directly addresses the excessive time required for board region based tolerance tightening. This unique 2 stage system allows the user to first optimize the printing process and then once the optimization has been completed, to automatically normalise the results and then set tighter tolerances to monitor product drift, enabling meaningful production control.  The 5D measurement, combining 2D and 3D measurements ensures proper zero referencing, forming a solid base for accurate and repeatable measurements.

Tolerances after Normalization

After optimization and normalization

The normalization extraction takes user selected average area, height and volume from several previous inspections when being set. Thus, an accurate representation of the capability of the print process is incorporated. If a process change, such as stencil or print settings need be made, the normalization extraction can be readily reset and then applied once again.

www.marantz-electronics.com

Did you like this? Share it:

Leave a Reply

Your email address will not be published. Required fields are marked *

*



You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>