Page 89 - Building A Succesful Board-Test Strategy

Success for a particular tip style depends on both its geometry and the accom-
panying spring force. Too little force may prevent good electrical contact. Too high
a force may damage delicate board pads and other features.
Some probes rotate as they compress. This action helps the probe to pierce
through conformal coatings, corrosion, and other surface contamination to ensure
good contact.
Test engineers must carefully distribute pins across a board's surface to avoid
uneven pressure during testing. Such pressure can cause bending moments, reduc-
ing the pointing accuracy of pins near the board edge and possibly breaking solder
bonds or tombstoning surface-mounted components.
Ensuring high-quality circuit resistance measurements and digital tests
requires low interface resistance and good node-to-node isolation. According to
Mawby, fixture-loop resistance—including switching circuitry, wires, and probes—
ranges between 1O and 5£1 Although its magnitude is not much of an impediment
for shorts-and-opens and other self-learned tests that compensate for it automati-
cally, node-to-node variations must remain much less than shorts-and-opens
thresholds. In-circuit testers must subtract out this so-called "tare" resistance to
achieve necessary measurement accuracies.
A low contact resistance, usually specified as less than 500 mO, also ensures
maximum test speed and accuracy. Contact resistances tend to increase as probes
age, however, primarily from corrosion and other contamination. Cleaning helps
minimize the problem but cannot eliminate it completely.
Pointing accuracy specifies how closely a pin can hit its target. Repeatability
indicates the consistency from cycle to cycle. Keller and Cook (1985) estimated an
accuracy (1 standard deviation, or 0) of ±4 mils for 100-mil probes and ±5 mils for
50-mil probes. Of course, if pins bend as they age, these figures will get worse. The
researchers recommend surface-mount-board test pads with diameters of 30 or 4o.
At 30, approximately 3 probes in 1000 would miss their targets. At 40, less than 1
in 10,000 would miss. Based on these numbers, 100-mil probe pads should be
24 mils or 32 mils in diameter, and 50-mil pads should be 30 mils or 40 mils. Design-
ers may justifiably balk at consuming so much real estate "merely" for test pads.
Mawby refers to new probe designs that achieve accuracies of ±1.5 mils for
50-mil surface-mount boards. These probes need test pads only 9 mils or 12 mils in
diameter to meet 30 and 40 requirements. St. Onge (1993) also explores hitting
small targets with a conventional bed-of-nails.
Unfortunately, probes do not last forever. Standard models are rated for
about 1,000,000 cycles if there is no sideloading. Probe life is approximately pro-
portional to cross-sectional area. Therefore, if a 50-mil-pin barrel diameter is half
of the barrel diameter for a 100-mil pin, it will likely last only one-quarter as long.
Because pin replacement is a maintenance headache, this shortened life constitutes
a major concern. Some probe manufacturers have introduced small-center models
with larger-than-normal barrel diameters specifically to address the life issue.
Although they do not last as long as their 100-mil siblings, these models can last
twice as long as other small-center designs. To hit targets smaller than 50 mils, some
probes adopt the traditional design but completely eliminate the receptacle, mount-