Appeal No. 2006-1729
Application No. 10/107,628
OR gate placed between each input and at least an output wherein
each input is coupled to the outputs differently.
Now, the question before us is what Ivanov would have taught
to one of ordinary skill in the art? To answer this question, we
find the following facts:
At page 1394, sections 1, 2.1 and 2.2, Ivanov
discloses the following:
Space compaction other than by MISRs has so far
been dominated by parity checkers, which are linear and
circuit-independent compactors [4], [5]. The
popularity of parity checkers as space compactors
follows from their generally fairly effective fault
coverage and simplicity. On the other hand, parity
checkers are quite costly to implement in terms of area
since they are usually implemented with trees of XOR
gates and the cost of an XOR gate is typically among
the highest for 2-input Boolean functions. A k: 1
parity tree compactor requires k – 1 2-input XOR gates
and thus amounts to significant area.
Consider the general case of compaction where a matrix
of test data D = [dij] of dimensions m x n bits is
transformed into a matrix C = [cij] of dimensions p x q
bits, where p < m and/or q < n. We denote the
transformation operator Φ as a matrix operator such
that C = Φ (D). We refer to the ratio m:p as the space
compaction ratio and the ratio n:q as the time
compaction ratio.
The column index of the test data matrix D is referred
to as the time dimension since it corresponds to the
output bits from a single circuit node (primary or
pseudo-output) resulting from the application of
different input test patterns to the CUT. Thus, if Φ
is such that C has its time dimension q < n, then time
8
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Next
Last modified: November 3, 2007