9
Propagation Delay, Pulse-Width Distortion and Propa-
gation Delay Skew
Propagation delay is a gure of merit which describes how
quickly a logic signal propagates through a system. The
propagation delay from low to high (t
PLH
) is the amount
of time required for an input signal to propagate to the
output, causing the output to change from low to high.
Similarly, the propagation delay from high to low (t
PHL
) is
the amount of time required for the input signal to propa-
gate to the output, causing the output to change from
high to low (see Figure 9).
Pulse-width distortion (PWD) results when t
PLH
and t
PHL
dier in value. PWD is dened as the dierence between
t
PLH
and t
PHL
and often PWD is dened as the dierence
between t
PLH
and t
PHL
and often determines the maxi-
mum data rate capability of a transmission system. PWD
can be expressed in percent by dividing the PWD (in ns)
by the minimum pulse width (in ns) being transmitted.
Typically, PWD on the order of 20-30% of the minimum
pulse width is tolerable; the exact gure depends on the
particular application (RS232, RS422, T-1, etc.).
Propagation delay skew, t
PSK
, is an important parameter
to consider in parallel data applications where synchroni-
zation of signals on parallel data lines is a concern.
If the parallel data is being sent through a group of opto-
couplers, dierences in propagation delays will cause the
data to arrive at the outputs of the optocouplers at dier-
ent times. If this dierence in propagation delays is large
enough, it will determine the maximum rate at which par-
allel data can be sent through the optocouplers.
Propagation delay skew is dened as the dierence be-
tween the minimum and maximum propagation delays,
either t
PLH
or t
PHL
, for any given group of optocouplers
which are operating under the same conditions (i.e., the
same supply voltage, output load, and operating temper-
ature). As illustrated in Figure 10, if the inputs of a group of
optocouplers are switched either ON or OFF at the same
time, t
PSK
is the dierence between the shortest propaga-
tion delay, either t
PLH
or t
PHL
, and the longest propagation
delay, either t
PLH
or t
PHL
. As mentioned earlier, t
PSK
can de-
termine the maximum parallel data transmission rate.
Figure 10 is the timing diagram of a typical parallel data
application with both the clock and the data lines being
sent through optocouplers. The gure shows data and
clock signals at the inputs and outputs of the optocou-
plers. To obtain the maximum data transmission rate, both
edges of the clock signal are being used to clock the data;
if only one edge were used, the clock signal would need
to be twice as fast.
Propagation delay skew represents the uncertainty of
where an edge might be after being sent through an opt-
ocoupler. Figure 10 shows that there will be uncertainty in
both the data and the clock lines. It is important that these
two areas of uncertainty not overlap, otherwise the clock
signal might arrive before all of the data outputs have
settled, or some of the data outputs may start to change
before the clock signal has arrived.
From these considerations, the absolute minimum pulse
width that can be sent through optocouplers in a parallel
application is twice t
PSK
. A cautious design should use a
slightly longer pulse width to ensure that any additional
uncertainty in the rest of the circuit does not cause a
problem.
DATA
INPUTS
CLOCK
DATA
OUTPUTS
CLOCK
t
PSK
t
PSK
50%
50%
t
PSK
I
F
V
O
I
F
V
O
50%,
CMOS
50%,
CMOS
Figure 9. Propagation delay and skew waveform
Figure 10. Parallel data transmission example