Datasheet

SZZA016B
7–246
Basic Design Considerations for Backplanes
Fully Loaded Backplane
Figure 17 clearly shows the effect of the different termination resistors on signal integrity in the
fully loaded EVM. All waveforms show incident-wave switching, with upper noise margin gained
with lower termination-resistor values. The 50-Ω termination value is unacceptable. The
43.5-MHz and 11.5-MHz data waveforms are included for comparison.
V
TT
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
25-MHz Data
43.5-MHz Data
11.5-MHz Data
0.33 V/ns
0.43 V/ns
0.42 V/ns
0.28 V/ns
0.42 V/ns
0.31 V/ns
1.5 V
10 ns
10 ns
25 ns
V
REF
V
REF
V
TT
V
REF
V
TT
1.5 V
V
TT
R
TT
R
TT
= 25 Ω
R
TT
= 33 Ω
R
TT
= 38 Ω
R
TT
= 50 Ω
R
TT
= 33 Ω
R
TT
= 50 Ω
R
TT
= 38 Ω
R
TT
= 25 Ω
R
TT
= 25 Ω
R
TT
= 50 Ω
R
TT
= 33 Ω
R
TT
= 38 Ω
SN74GTLPH1655DGGR With Edge-Rate Control in Slow
Figure 17. Fully Loaded Backplane vs R
TT
(Driver in Slot 1, Receiver in Slot 2)
The V
OH
voltages at both 25-MHz and 43.5-MHz data rates never converge to the termination
voltage of 1.5 V as with 11.5-MHz data. The reason is that the reflections have not had enough
time to settle, which typically takes one round trip on the bus.
Additionally, slew rates of the optimum termination line are included for the rising and falling
edges. Typical TTL slew rates in lumped loads are from 1 V/ns to 1.4 V/ns, depending on the
capacitive load. The significantly slower GTLP edge rates result in a larger device t
pd
, but allow
for higher system frequencies because limited ringing improves signal integrity. The observed
slew rate should increase as the load is reduced.