Page 49 - ChipScale_Nov-Dec_2020-digital

Fo r t h e p r o d u c t io n t e s t i ng , a Summary cleaning methods negatively affect UPH,
dimensionless contact resistance ratio Advanced processes, tooling, and and consequently, increase the overall cost
(CRES RATIO ) was defined and used to practices are necessary to meet these of test.
assess the cleaning performance. The industry and consumer expectations. All major suppliers of roadmap
CRES RATIO is defined as the maximum Reliable package-level testing is critical for handlers have developed and incorporated
resistance value (CRES MAX ) divided meeting the requirements of KGD, KNB programmable ACC functionalities.
by the upper specification limit of die, and “zero-defect” manufacturing Programmed preventative cleaning using
the resistance (CRES USL ). In a stable without cost of test escalation. To maintain the handler automation for “on-demand”
testing process, the CRES RATIO should yields and consistently attain high data cleaning execution dramatically improves
have a value that is less than or equal integrity, regular socket cleaning is test cell OEE. Auto cleaning execution
to 1.0. If the CRES RATIO is greater than needed. Reactive or proactive manual is performed using customized cleaning
1.0, the testing process is considered
to be less than optimal with a high
probability for additional recovery
passes. During the qualification, the
two types of cleaning units (flat and
featured) were loaded simultaneously
into different cleaning repository
locations of the high-volume handler.
This relatively simple evaluation
strategy allowed for an iterative “on-
off-on” comparison to validate the
use of functional features for the
process stability.
I n t he test sequence, feat u red
c l e a ni n g u ni t s w e r e i ni t i a l l y
i mple me nt e d a nd t he CR ES R AT IO
remained below 1.0 during the testing.
After the first set of featured units were
cycled through, a second set of flat units
were used for the cleaning execution.
As these cleaning units were cycled
through the test cell, the trending for
the CRES RATIO values progressively
increased with a greater variance
(Figure 8). The f lat cleaning units
can remove debris from the surface
socket f loor; however, because the
crown tips of the spring pin protruded
less than 90µm above the socket
floor, these were not getting properly
cleaned. Upon reimplementation of
the featured units in the sequence, the
CRES RATIO incrementally recovered
as the crown tips and the guide holes
star ted getting properly cleaned.
After several cleaning executions, the
CRES RATIO stability was reestablished.
By creating the functional features,
it was possible to control debris from
within the floating base socket guide
holes that could not be addressed with
the flat, unfeatured units. The “on-off- RoHS
on” sequences were performed several P
times with practically identical results.
At the conclusion of the project, it was
clear that the featured cleaning units
were significantly more effective for
maintaining stable electrical contact
integrity for the floating bed socket.

47
Chip Scale Review November • December • 2020 [ChipScaleReview.com] 47