Performance Verification
GTS1063 and GTS1250 GBIC Test Systems Instruction Manual
65
H
Allow the averaging of the CROSS result to settle (that is the StDev:
after several seconds, the standard deviation in the expanded CROSS
parameter window will settle below a picosecond) and record the mean
result of the crossing: this is the first edge (edge #0) and is used as the
timing reference for all subsequent nine transition edges.
NOTE
. Since the final result of this portion of the procedure is a peak-to-peak
measurement, the result will be the same no matter which of the 10 transition
edges in the 20-bit K28.5 pattern are used as the reference timing edge.
H
For each of the nine edges (rising or falling) following edge #0, adjust
the LEFT and RIGHT limits to isolate each edge for a similar CROSS
result: specify the correct slope (+ for rising, or — for falling) manually
for each CROSS measurement or else the measurement result will be an
ERROR.
7. Record the results of the CROSS measurements for all ten consecutive edges
in Table 19. The ideal crossing location for each edge is in the table to the
right of the measured values: the ideal crossing locations following reference
edge #0 are integer multiples of the UI bit period added onto the reference
edge. For example: if the first edge #0 of a GTS1250 occurs at 32.969 ns,
then the next edge, which occurs 5 UI later in this example, has an ideal
location of 32.969 ns + (5 x 800 ps) = 36.969 ns. The unit interval (UI) is the
inverse of the data rate: the GTS1063 has a data rate of 1.0625 Gb/s and
therefore its UI = 941.176 ps, the GTS1250 has a data rate of 1.2500 Gb/s
and therefore its UI = 800.000 ps. Table 20 shows example measurement
data for determining the deterministic jitter of the Tx Data output.
a. The error between each measured edge and its ideal location is deter-
mined by subtracting the exact location from the measured location.
note: since edge #0 is used as the reference, its ideal location and its
measured location are the same and hence have zero error.
b. Add the magnitude of the largest positive and largest negative error to
yield the peak-to-peak deterministic jitter in the K28.5 test pattern
(example: +5 ps error and –3 ps error yield 8 ps
p–p
). The calculated
peak-to-peak error should be
t
40 ps.
Artisan Technology Group - Quality Instrumentation ... Guaranteed | (888) 88-SOURCE | www.artisantg.com