Biomedical Engineering Reference
In-Depth Information
Defect site
Source
Capacitive
sensing
circuit
Sink
1234567
Te st
Figure 4.8
Example of a test outcome for a faulty array.
defect locations. Consider the example shown in Figure 4.8. The output pulse
sequence indicates a defect in the fifth column. The defect site can be pre-
cisely identified by carrying out the row tests. In some cases, it is difficult to
map test outcomes to candidate fault patterns; for example, a test outcome
missing the first pulse is the same as that missing the last pulse. To solve
this problem, the arrival time for the first test droplet is calculated before test
application; in this way, we get a reference point so that ambiguous interpre-
tation of test outcomes can be avoided.
Using the preceding method, a single faulty cell can be located as shown in
Figure 4.9; the test droplet will be stuck at the faulty cell in both the column-
test and row-test steps. If the defect is an electrode short, the test droplet will
be stuck at the short site in either the row test or column test, but not both. No
additional diagnosis steps are needed. In contrast, Euler-path-based testing
[36] relies on a binary search process to determine the exact location of the
defect. Thus, the parallel scan-like test saves time for many testing applica-
tions where only defect-type information is needed.
Defect localization is more complicated when the fault is due to a short
between two electrodes. In this case, the sensor readout indicates an error
for only one step, that is, either column test or row test
(see Figure 4.10)
.
Te st
outcomes
Source
Defect
Odd column
1
Sink
2
3
4
5
6
Even column
Odd row
Even row
123456
Figure 4.9
Diagnosis of a single-electrode defect by a “cross-parallel” scan-like test.
