Datasheet

SZZA016B
7–247 Basic Design Considerations for Backplanes
C
t
can be calculated using the information known about the EVM and the observed Z
o(eff)
.
Assuming stripline construction with Z
o
= 95 Ω and C
o
= 2.40 pF/in, solve for C
d
using
equation 9 and assuming that the optimum R
TT
= Z
o(eff)
= 35 Ω. An interpolated R
TT
value
of 35 Ω was chosen because it produces the best incident-wave switching performance. Then,
solve for C
t
by multiplying C
d
by the separation between two transceivers which, in this case, is
0.94 in.
C
d
+ ǒ
Z
2
o
R
2
TT
* 1Ǔ C
o
+
ǒ
95
2
35
2
* 1
Ǔ
2.40 pFńin. + 15.28 pFńin.
C
t
+ C
d
d + 15.28 pFńin. 0.94 in. + 14.36 pF
Using C
d
and C
o
, the effective t
pd
and flight time can be calculated.
t
pd
for our EVM transmission line with a Z
o
of 95 Ω = 230 ps/in.
t
pd(eff)
= t
pd
×
1 )
ǒ
C
d
ńC
o
Ǔ
Ǹ
= 230 ps/in. × 2.71 = 624.3 ps/in.
The total distance traveled from slot 2 to the termination load is:
18 slots × 0.94 in. + termination stub length of 1 in. = 17.92 in.
Therefore, flight time is 17.92 in. × 624.3 ps/in., or 11.2 ns.
Round-trip flight time from slot 2 to the load, and back, is 22.4 ns.
The observed settling time is 20.8 ns.
Lightly Loaded Backplane
Figure 18 clearly shows the effect of the different termination resistors on signal integrity when
every other card is removed from the EVM and the distributed capacitive load is reduced by a
factor of two. There is still some capacitive loading (about 0.7 pF) at the empty slot position, but
the majority is removed with the female connector, stub, and device C
io
.
(13)
(14)