Lecroy 93XXC-OM-E22 User Manual

'

RIGHT CURSOR

50 % (Mesial)

Upper Threshold

(90 % Amplitude)

(10 % Amplitude)

$SSHQGL['3DUDPHWHU0HDVXUHPHQW

3DUDPHWHUVDQG+RZ7KH\:RUN

In this Appendix, a general explanation of how the instrument’s
standard parameters are computed (
table listing, defining and describing those parameters (

).

D–5

'HWHUPLQLQJ7RSDQG
%DVH/LQHV

Proper determination of the
fundamental for ensuring correct parameter calculations. The
analysis begins by computing a histogram of the waveform data over
the time interval spanned by the left and right time cursors. For
example, the histogram of a waveform trans itioning in two states will
contain two peaks (
two clusters that contain the largest data density. Then the most
probable state (centroids) associated with these two clusters will be
computed to determine the
line corresponds to the top and the

Fig. D–1

top

and

). The analysis will attempt to identify the

top

and

base

base

see below

) is followed by a

page

base

reference lines is

reference levels: the

line to the bottom centroid.

top

maximum

top

LEFT CURSOR

rise

ampl

'²

*not to scale

pkpk

HISTOGRAM*

Lower Threshold

base
minimum

fall

width

Figure D–1

$SSHQGL['

'HWHUPLQLQJ5LVHDQG
)DOO7LPHV

Once

top

and

base

are estimated, calculation of the

times is easily done (

Fig.1

). The 90 % and 10 % threshold levels

rise

and

fall

are automatically determined by the oscilloscope, using the

ampl

amplitude (
Threshold levels for

absolute or relative settings (
are chosen, the

) parameter.

rise

or

rise

or

fall

fall

time can also be se lected using

r@level, f@level

). If absolute settings

time is measur ed as the time interval
separating the two crossing points on a rising or falling edge. But
when relative settings are chosen, the vertical interval spanned

base

and

top

between the

lines is subdivided into a percentile
scale (base = 0 %, top = 100 %) to determine the vertical position
of the crossing points.

The time interval separating the points on the rising or falling
edges is then estimated to yield the rise or fall time. These
results are averaged over the number of transition edges that
occur within the observation window.

Mr

1

Rising Edge Duration

Falling Edge Duration

Where Mr is the number of leading edges found, Mf the number of
trailing edges found,

level, and

x

Tf

the time when falling edge i crosses the x % level.

i

x

Tr

the time when rising edge i crosses the x %

i

()

∑

Mr

i

=

1

Mf

1

()

∑

Mf

i

=

1

1090

−

TrTr

ii

9010

−

TfTf

ii

'HWHUPLQLQJ7LPH
3DUDPHWHUV

Time param eter measurements such as
are carried out with respect to the mesial reference level (

), located halfway (50 %) between the top and base

D–2

width, period

and

delay

Fig.

reference lines.
Time-parameter estimation depends on the number of cycles

included within the observation window. If the number of cycles is

rms

or

not an integer, parameter meas urements such as

mean

will be biased.

'²

3DUDPHWHU0HDVXUHPHQW

width

delay

50 %

(Mesial)

first

LEFT CURSOR

width

PERIOD PERIOD

freq period

= 1/

TWO FULL PERIODS: = 2

cmean, cmedian, crms, csdev

computed on interval periods

area, points, data

computed between cursors

width

duty width/period

=

cycles

Figure D–2

last

TRIGGER

POINT

RIGHT CURSOR

To avoid these bias effects, the instrument uses cyclic
parameters, including

crms

and

cmean

, that restrict the

calculation to an integer number of cycles.

'HWHUPLQLQJ'LIIHUHQWLDO
7LPH0HDVXUHPHQWV

The oscilloscope enables accurate differential time

measurements between two traces — for example, propagation,
setup and hold delays (

Parameters such as

Fig. D–3

∆

c2d

).

±

require the transition polarity of the

clock and data signals to be specified.

'²

RIGHT CURSOR

DATA (1)

CLK (2)

HYSTERESIS

Noisy spikes ignored due
to Hysteresis band

∆−

c2d (1, 2)

∆

c2d+(1, 2)

$SSHQGL['

THRESHOLD

LEFT CURSOR

CLOCK EDGE = Positive Transition

DATA EDGE = Negative Transition

TRIGGER POINT

Figure D–3

Moreover, a hysteresis range may be specified to ignore any
spurious transition that does not exceed the boundaries of the

∆

c2d

−

hysteresis interval. In Figure 3,

(1, 2) measures the time

interval separating the rising edge of the clock (tr igger) from the

∆

c2d

+

first negative transition of the data signal. Sim ilarly,

(1, 2)

measures the time interval between the trigger and the next
transition of the data signal.

'²