LeCroy DIGITAL FILTER PACKAGE User manual

DIGITAL FILTER PACKAGE

OPERATOR’S MANUAL

JULY 2001

REVISION B

LeCroy Corporation

700 Chestnut Ridge Road
Chestnut Ridge, NY 10977–6499
Tel: [phone], Fax: [phone]

Internet: www.lecroy.com

© 2001 by LeCroy Corporation. All rights reserved. Information in this publication supersedes all earlier ver­sions. Specifications subject to change.

LeCroy, ProBus and SMART Trigger are registered trademarks, and ActiveDSO, ScopeExplorer,
WaveAnalyzer and Waverunner are trademarks, of LeCroy Corporation. Centronics is a registered trademark
of Data Computer Corp. Epson is a registered trademark of Epson America Inc. Mathcad is a registered
trademark of MATHSOFT Inc. MATLAB is a registered trademark of The MathWorks, Inc. Microsoft, MS and
Microsoft Access are registered trademarks, and Windows and NT trademarks, of Microsoft Corporation.

DFP-OM-E Rev B 0701

1 Introduction 1–1

The Need 1–1
The Solution 1–1

Enhanced Solutions 1–4

Filter Types 1–5

Low-pass Filter 1–5
High-pass Filter 1–5
Band-pass Filter 1–6
Band-stop Filter 1–6

Communication Channel Filters 1–7

Raised Cosine (a low-pass filter) 1–7
Raised Root Cosine (a low-pass filter) 1–8
Gaussian 1–9

2 Operation 2–1

Setting up the scope 2–1

Running DFP 2–1

3 Remote Commands 3–1

Define Command 3–1

Examples 3–2

4 Custom Filters 4–1

Custom Filter Setup 4–1

Example 1: Using Mathcad 2000 (Visual Basic Script) 4–2
Example 2: Sending FIR Coefficients from Excel 4–5

5 Multirate Filters 5–1

Description 5–1
Example 5–1

6 Specifications 6–1

DFP-OM-E Rev B ISSUED: July 2001 iii

LeCroy Digital Filter Package

BLANK PAGE

iv ISSUED: July 2001 DFP-OM-E Rev B

THE NEED

1Introduction

In today’s complex environment, data is frequently composed of
a mixture of analog and digital components spread over a broad
range of frequencies. In many applications, the relevant data is
encoded or obscured. Capturing the right signals becomes a
challenge. Engineers find it increasingly difficulty to examine only
those parts of the data they are interested in. Traditional (or even
smart) oscilloscope triggering cannot always provide a
satisfactory answer.

For example, servo motors from disk drives add a low frequency
component to the high frequency data output. It is hard to achieve
an accurate analysis of data unless the low component is
removed.

Another common example is switched power supply units, which
inject the switching frequency component into many system
parts. Viewing digital signals mixed with this switching frequency
component could be very difficult. Filtering is definitely required.

Yet another example is in ADSL residential connectivity, where
data is transmitted over 256 narrow bands. Each band is only

4.7kHz wide, and the gap between two adjacent bands is also

4.7kHz. Examining such complex waveforms with regular DSOs
is almost impossible; filtering out unwanted frequency
components is necessary.

THE SOLUTION

At present, these needs are addressed in two ways. One way is
building analog filters and placing them in front of the
oscilloscope, providing an already filtered signal to the DSO. The
disadvantages of this approach are many. Analog filters depend
heavily on the accuracy and stability of analog components.
Although in some cases analog filters are easily implemented,
they are quite impractical for low (< 100 HZ) or high (> 100 MHz)
frequency ranges. In comparison, digital filters can provide the
desired results in those cases.

The second approach, practiced by many engineers, is using the
DSO as a digitizer. The digitized data output is then transferred to
a PC for processing. This solution frequently provides the

DFP-OM-E Rev B ISSUED: July 2001 1–1

LeCroy Digital Filter Package

required results, but it might be too slow or too limited in flexibility
for some applications.

With the Digital Filter Package, LeCroy provides a solution that
combines the best of both worlds. This package includes seven
of the most useful filter types, in addition to a custom design
feature. You can easily set the edge (or corner) frequency in
addition to the transition region width for each filter. It is possible
to use single filters or multiple (up to four) filters cascaded for
even more complex filtering. Once filtered, waveforms include
mostly relevant frequency components, undesired parts being
greatly attenuated.

If filters with special characteristics are desired, the custom
design feature allows you to design unique filters tailored to your
specific needs. The required filter can be designed with a digital
filter design or math package such as MATLAB® or Mathcad®.
Filter coefficients can be directly downloaded from the program
into the scope, using the DSOFilter utility. It is also possible to
specify the filter coefficients on an Excel spreadsheet and to use
DSOFilter to download them from the spreadsheet to the scope.

1–2 ISSUED: July 2001 DFP-OM-E Rev B

Introduction

DFP’s flexibility and ease of operation are demonstrated by the
following example:

1: a signal composed of 5 MHz pulses is filtered by:

A: 40 MHz low-pass filter with a 20 MHz transition region width
B: 30 MHz low-pass filter with a 15 MHz transition region width
C: 20 MHz low-pass filter with a 10 MHz transition region width

DFP-OM-E Rev B ISSUED: July 2001 1–3

LeCroy Digital Filter Package

Enhanced Solutions

DFP can be coupled with other LeCroy software products — such
as AORM, JTA, PMA1, or DDA — to enhance the capabilities of
these products and to provide improved solutions.

For example

• Jitter Measurement: the DFP Band-pass Filter can be

coupled with the JTA package to measure jitter over a narrow
frequency range.

• Optical Recording: In some cases equalization with a

different response is required. In such a case, the DFP
custom filter feature can be used in conjunction with AORM.
The required custom filter can be easily implemented to
provide the necessary equalization, while all other AORM
functions remain unchanged.

• Power Measurement: A Band-stop filter can be coupled with

the PMA1 package to eliminate the switching power supply
frequency component from power lines. The Band-stop filter
can be tuned to match a specific power supply switching
characteristic.

1–4 ISSUED: July 2001 DFP-OM-E Rev B

Introduction

FILTER TYPES

Low-pass Filter

High-pass Filter

1

Low-pass filters are useful for eliminating accumulated
high frequency noise and interference, and for canceling
high frequency background noise.

Sample applications are in datacom,
telecommunications, and disk drive and optical
recording analysis for accurate RF signal detection.

Band 1: Pass Band — DC to top of the transition region;
signal passes unattenuated.

Band 2: Transition Region — edge frequency to edge
frequency plus width; increasing attenuation.

Band 3: Stop Band — above end of transition region;
signal is highly attenuated.

High-pass filters are useful for eliminating DC and low
frequency components. Sample applications include
Disk Drive and Optical Recording analysis (emulation
of the SLICING function).

Band 1: Stop Band — DC to bottom of the transition
region; highly attenuated.

Band 2: Transition Region — edge frequency minus
width to edge frequency; decreasing attenuation.

Band 3: Pass Band — above edge frequency; signal
passes unattenuated.

1. Filters are optimal FIR filters of less than 1000 taps, according to the Parks-MacLellan algorithm described in Digital
Filter Design and Implementation by Parks and Burros, John Wiley & Sons, Inc., 1987

DFP-OM-E Rev B ISSUED: July 2001 1–5

LeCroy Digital Filter Package

Band-pass Filter

Band-pass filters are useful for emphasizing a selected
frequency band. Sample applications include radio
channel identification, broadband transmission, ADSL,
clock generators (i.e., eliminating the central frequency
and displaying harmonics only), and
telecommunications (Jitter measurement over a
selected frequency range).

Band 1: First Stop Band — DC to bottom of first
transition region; highly attenuated.

Band 2: First Transition Region — lower corner minus
width to lower corner; decreasing attenuation.

Band 3: Pass Band — signal passes unattenuated.
Band 4: Second Transition Region — upper corner to

upper corner plus width; increasing attenuation.
Band 5: Second Stop Band — signal highly attenuated.

Band-stop Filter

Band-stop filters are useful for eliminating a narrow
band of frequencies. Sample applications include
medical equipment, such as ECG monitors where the
dominant ripple at 50/60 Hz is rejected, leaving the low
energy biological signals intact. Digital troubleshooting:
the inherent frequency of the switched power supply is
blocked, revealing power line voltage drops and
glitches caused by the system clock generator.

Band 1: First Pass Band — DC to bottom of first
transition region; signal passes unattenuated.

Band 2: First Transition Region — lower corner minus
width to lower corner; increasing attenuation.

Band 3: Stop Band — signal is highly attenuated.
Band 4: Second Transition Region — upper corner to

upper corner plus width; decreasing attenuation.
Band 5: Second Pass Band — signal passes

unattenuated.

1–6 ISSUED: July 2001 DFP-OM-E Rev B

COMMUNICATION CHANNEL FILTERS

Raised Cosine (a low-pass filter)

These filters belong to the low-pass filter category (with
a variety of shapes). Raised cosine is one of a class of
filters used to minimize intersymbol interference: the
time domain impulse response crosses zero at all bit
time intervals except the one with the impulse.

Applying raised root cosine twice (or at the sending
and receiving end of a signal, for example) results in a
raised cosine filter effect. Sample applications include
wireless cellular communications such as WCDMA,
datacom, telecommunications, disk drive and optical
drive analysis.

Band 1: Pass Band — DC to corner frequency minus
half width; signal passes unattenuated.

Band 2: Transition Region — corner minus half width to
corner plus half width; attenuation increases with
frequency with a rolloff shape of 0.5cos(a) + 0.5, where
a ranges from 0 to ? over the transition region. This
region is determined by ?, which is specified as a
percentage of the corner frequency.

Introduction

Band 3: Stop Band — above corner frequency plus half
width; highly attenuated.

The impulse function for the raised cosine filter is:

DFP-OM-E Rev B ISSUED: July 2001 1–7

LeCroy Digital Filter Package

Raised Root Cosine (a low-pass filter)

Band 1: Pass Band — DC to corner frequency minus
half width; signal passes unattenuated.

Band 2: Transition Region — corner minus half width to
corner plus half width; attenuation increases with

frequency with a rolloff shape of 0.5[cos(a) + 0.5]½,
where a ranges from 0 to ? over the transition region.
This region is determined by ? , which is specified as a
percentage of the corner frequency.

Band 3: Stop Band: — above corner frequency plus
half width; signal is highly attenuated.

The impulse function for the square-root raised cosine
filter is:

Illustrated above are two raised root cosine filters with different beta. The diagrams
show that applying this filter twice results in a raised cosine response.

1–8 ISSUED: July 2001 DFP-OM-E Rev B

Gaussian

Introduction

Band 1: Pass Band — DC to half power bandwidth%
times modulation frequency, pass; 3 dB down at half
power bandwidth.

The shape of a Gaussian filter’s frequency response is
a Gaussian distribution centered at DC. The signal
becomes more attenuated with increasing frequency. It
is not possible to specify a transition region or a stop
band for Gaussian filters. However, the BT value, a
fraction of the symbol frequency, determines the filter’s
width, where:

B = half power bandwidth
T = bit (or modulation period)

§ § §

DFP-OM-E Rev B ISSUED: July 2001 1–9

LeCroy Digital Filter Package

BLANK PAGE

1–10 ISSUED: July 2001 DFP-OM-E Rev B

SETTING UP THE SCOPE

Running DFP

2Operation

There are five basic steps to select and set up a filter:

1. Select MATH functions.

2. Select Filter as the Math Type

3. Select the filter type.

4. Set desired values for the frequency edge and the transition
region width.

5. Select an oscilloscope channel as the filter’s input.

To select the Math function

1. Press the MATH button (MATH SE TUP or MATH TOOLS).

2. From the "ZOOM + MATH" menu, press the soft key
alongside the desired math trace (here, Redefine A)

Pressing a Zoom + Math TRACE ON/OFF (A–D) button will bring
up the next menu.

Note

The number of "max points" displayed has no
influence on filter operation.

DFP-OM-E Rev B ISSUED: July 2001 2–1

LeCroy Digital Filter Package

After you select the desired trace, the following menu shown here is
displayed.

From the "SETUP OF A" menu

1. Select Yes from "use Math?".

2. Select the Filter option from the "Math Type" menu.

3. Select a filter from the "Filter Type" menu.

4. In the example at left, the LowPass filter was selected. The
"pass until" box allows you to set the values for the frequency
edge (f) and the transition region width (w). Press the
corresponding soft key to switch between frequency and
width.

5. Turn the associated adjustment knob to set the desired
values.

6. From the "of" menu, select the channel to which the filter is to
be applied.

2–2 ISSUED: July 2001 DFP-OM-E Rev B

Operation

When you select the Band Pass filter, the menu at left appears:

Press the "MORE FILTER SETUP" soft key. The
following menu appears:

The soft key toggles between the lower (l)
and upper (u) frequency edges. Use the
associated adjustment knob to set values for
both parameters.

Edge Width is set from the lower box. Use
the soft key to increase the width, or the knob
to adjust the value.

DFP-OM-E Rev B ISSUED: July 2001 2–3

LeCroy Digital Filter Package

When you select Raised Cosine or Raised Root Cosine filters, the
menu at left appears:

Press the corresponding soft key to switch between corner
frequency (f) and beta (? ). Turn the associated knob to set values
for these parameters.

?

2–4 ISSUED: July 2001 DFP-OM-E Rev B

Operation

When you select Gaussian filters, the menu at left appears:

Press the corresponding soft key to switch between modulation
frequency (f) and BT, where B = half power bandwidth expressed
as a fraction of the modulation frequency and T = bit (or
modulation) period. Turn the associated knob to set values for
these parameters.

§ § §

DFP-OM-E Rev B ISSUED: July 2001 2–5

LeCroy Digital Filter Package

BLANK PAGE

2–6 ISSUED: July 2001 DFP-OM-E Rev B

3Remote Commands

Values associated with the above parameters are:

DEFINE COMMAND

The DFP option adds one new <equation> to the existing DEFINE remote command. The new
function available with the DFP option is FIR(<source>).

Several parameters are added to support the FIR math function. They are:
FTYPE,<ftype>

FREQ, <lfreq>
UFREQ,<ufreq>
FWIDTH,<fwidth>
FBETA,<fbeta>
MCOEFF,<memory>

<ftype>:=[ LOWPASS | HIPASS | BANDPASS | BANDSTOP | RAISEDCOS | RSDROOTCOS |
GAUSSIAN | CUSTOM ]
<lfreq> := lower corner or only corner frequency, in Hz

<ufreq>:=upper corner frequency, in Hz. Must be greater than or equal to <lfreq>
<fwidth>:=transition region width in Hz. Must be >0.3% of sample rate. <lfreq> - <fwidth> must

be greater than 0.1% of sample rate.

<fbeta>:= 0 to 100 percent. For raised cos and raise root cos, this is the % of <lfreq> + and -

over which the transition region extends. For Gaussian, this is BT, the % of <lfreq> at
which the response is 3dB down from DC response.

<memory>:=[ M1 | M2 | M3 | M4 ]

DFP-OM-E Rev B ISSUED: July 2001 3–1

LeCroy Digital Filter Package

Examples

Not all of these need to be set for any <ftype>. Here is an example of those that are actually
needed:
Low Pass filter: TA:DEF EQN,"FIR(C1)",FTYPE,LOWPASS,LFREQ,71.5E+6
HZ,FWIDTH,40E+6 HZ
High Pass filter: TA:DEF EQN,"FIR(C1)",FTYPE,HIPASS,LFREQ,71.5E+6
HZ,FWIDTH,40E+6 HZ
Band Pass filter: TA:DEF EQN,"FIR(C1)",FTYPE,BANDPASS,LFREQ,60E+6
HZ,UFREQ,70E+6 HZ,FWIDTH,40E+6 HZ
Band Stop filter: TA:DEF EQN,"FIR(C1)",FTYPE,BANDSTOP,LFREQ,60E+6
HZ,UFREQ,70E+6 HZ,FWIDTH,40E+6 HZ
Raised Cosine filter: TA:DEF EQN,"FIR(C1)",FT YPE,RAISEDCOS,LFREQ,60E+6
HZ,FBETA,30 PCT

Raised Root Cosine filter :

TA:DEF EQN,"FIR(C1)",FTYPE,RSDROOTCOS,LFREQ,60E+6 HZ,FBETA,30 PCT
Gaussian filter: TA:DEF EQN,"FIR(C1)",FTYPE,GAUSSIAN,LFREQ,60E+6 HZ,70E+6
HZ,FBETA,30 PCT
Custom filter: TA:DEF EQN,"FIR (C1)",FTYPE,CUSTOM,MCOEFF,M2

§ § §

3–2 ISSUED: July 2001 DFP-OM-E Rev B

CUSTOM FILTER SETUP

4Custom Filters

If the seven standard filters provided with DFP are not sufficient,
you can create and use filters with virtually any characteristic, up
to 1000 taps.

The required custom filter can be designed with a digital filter
design or math package such as MATLAB® or Mathcad®.

The filter coefficients can then be loaded into the scope, using the
DSOFilter utility. GPIB, LAN, or a serial RS-232 connection
between the PC and oscilloscope is required. However, if these
connections are not available, it is also possible to load the file
with a diskette.

DSOFilter is an ActiveX control that can be downloaded free of
charge from LeCroy’s web site at www.lecroy.com.

Following are two examples of how custom filters can be created
and loaded into the scope with DSOFilter. The first demonstrates
a filter design using the Mathcad program. The second shows
how to use an Excel spreadsheet. Both examples use the
DSOFilter utility for downloading coefficients in the scope.

DFP-OM-E Rev B ISSUED: July 2001 4–1

LeCroy Digital Filter Package

check

= 0.987 This is the DC gain of the filter

Example 1: Using Mathcad 2000 (Visual Basic Script)

Sending FIR coefficients to a DSO from Mathcad, using the LeCroy DSOFilter Control:
i := 0..200
sinx(x) := sin(x)/x

?

1

?

coefs

200 point sin(x)/x, a low-pass filter.
Note: Real world filters would either be windowed or made by the Remez exchange algorithm.

The point of this example is not to make a good filter, but to show how to transfer a filter to the
scope.

:

i

2

?

sin

i

?

x

?

?

coefscheck ??:

0.2

0.159

0001.100

?

2

?

?

0.1

coefs

i

0

0.035?

0.1
0 50 100 150 200

2000 i

4–2 ISSUED: July 2001 DFP-OM-E Rev B

Custom Filters

To make the connection to the ActiveX control choose Insert, Component, Scriptable Object; then

to load. Click Next.

uses the inputs from Mathcad to invoke the SetFilter method of the object.

To send it, we need to include the LeCroy DSOFilter Control, which is an ActiveX control,
as a scriptable object. Remember the name you specify for this instance of the
ActiveX control. Mathcad acts as an ActiveX scripting host. It makes the methods of
the object available using the name you specified. It invokes a VBScript interpreter
(because this example uses VBScript as the scripting language) and calls three routines:
ScriptObj_Event.Start(), ScriptObj_Event.Exec (Inputs, Outputs), and SciptObjec_Event.Stop().
Mathcad calls these three in order when calculating. Mathconnex can Start, followed by 0, 1, or
many invocations of Exec, followed by Stop.

click Next.
Check "New" and select "LeCroy DSOFilter Control" from the list. (If you don't see it,

the control has not been installed). Click Next.
Select "VB Script Language." This tells Mathcad, the scripting host, which interpreter

Give this object a Name. For this example, accept the default name "ScriptObj".
The VBScript uses this name to access the object's methods. For example: ScriptObj.Beep().
Set this object to require 2 inputs from Mathcad and to return no output values.
In this example the two inputs to the script are the coefficient array, and a value specifying the
number of points in the array. The destination is assumed to be M1. The connection is assumed
to be at GPIB address 5. These assumptions could be eliminated. Only the “_Exec” function
takes arguments, so the connect method could be called from ScriptObjEvent_Exec() instead of
ScriptObjEvent_Start(), and we could pass in an extra argument to specify the connection
method, as a string, from Mathcad.

The format of the string would be, for example, either “GPIB: 5” or “COM1: 19200,8,N,1” (RS-232
on COM1, at 19200 baud, 8 bits per character, no parity, one stop bit; the other choices for parity
are E for even or O for odd) or “IP: 128.211.87.234”. Similarly, if ScriptObjEvent_Exec() took four
arguments, the fourth could be a string specifying the destination. This fourth string would be
passed to the SetFilter method of the DSOFilter control instead of the fixed string “M1”. The script

DFP-OM-E Rev B ISSUED: July 2001 4–3

LeCroy Digital Filter Package

End Sub

DSOFilter
(coefs 201)

The script that is run to connect Mathcad to the DSOFilter Active X control for this example is:
Sub ScriptObjEvent_Start()

REM Set up - tell the DSOFilter how to talk to the scope
ScriptObj.Connect("GPIB: 5")
End Sub

Sub ScriptObjEvent_Exec(Inputs,Outputs)
REM Get the inputs
numcoefs = Inputs(1).Value
REM Need numcoefs as Long
numcoefs = CLng(numcoefs)
REM Send the coefficients to M1 - have to put them in an array
Dim coefs(999)
Dim foo
foo = inputs(0).value ' assign inputs(0).value to a local variable
For i = 0 to numcoefs-1
dtmp = foo(i)
coefs(i) = CSng(dtmp) ' they can be double, too; but single is enough
Next
retval = ScriptObj.SetFilter( "M1", numcoefs, coefs)
If (retval = False) Then
MsgBox( "SetFilter returned False - problem")
End If
REM Make the scope beep to make sure we got here
ScriptObj.Beep()
End Sub
Sub ScriptObjEvent_Stop()
REM Nothing to do for clean-up

4–4 ISSUED: July 2001 DFP-OM-E Rev B

Custom Filters

coefficients.

Example 2: Sending FIR Coefficients from Excel

When you start the DSOFilt utility, an Excel spreadsheet similar to the following one opens. (You
must enable macros beforehand in the pop-up dialogue box.) In this example, a low-pass filter
with 200 coefficients is specified on the Excel spreadsheet. The PC is connected to the
oscilloscope via a GPIB connection. Click the "Make Scope Beep" button to check this
connection. A beep will be heard after a few seconds.

After ensuring that the connection works properly, the coefficients may be sent. First, a
destination must be specified (such as memories M1–M4 in the DSO). It is also possible to send
the coefficients to a file. In that case, a filename must be specified.

Note:

If other than GPIB serial connection between the PC and the scope is used (such

as RS-232 or Ethernet 10/100 MB/s), GPIB: 5 should be set for the correct
connection type.

RS-232

?

COM1: 19200,8,N,1 (COMn: baud, bits, parity, stop)

LAN (Ethernet 10/100)

?

IP: 128.211.87.234 (IP: a,b,c,d)

Click the "Put Filter" button to send the list of coefficients. Upon completion, the beep is sounded
again.

Number of Points: If “Manual Number of Points” is set to 0, the “Automatic Number of Points” is
determined by the number of coefficients, However, if “Manual Number of Points” is other then 0,
it overrides the Automatic setting, making it possible to examine the results with only a limited
number of coefficients before employing the full filter.

Notes:

1. The chart is used only as a display tool. It shows the filter in the time domain,

but it has no influence on operation.

2. You may simply replace the coefficients in the example with your filter's

DFP-OM-E Rev B ISSUED: July 2001 4–5

LeCroy Digital Filter Package

§ § §

4–6 ISSUED: July 2001 DFP-OM-E Rev B

DESCRIPTION

5Multirate Filters

In many of today’s development environments, digital filter design
has become most challenging. Specifications typically require
higher order filters, implying increased storage capacity for filter
coefficients and higher processing power. Moreover, high-order
filters can be difficult, if not impossible, to design. In applications
such as 3G wireless systems, for example, at the receiver end
data must be filtered in large magnitude in order to be processed.

Although the LeCroy DFP option provides many filter types, the
correlation between edge frequencies and sample rate may be a
limiting factor: edge frequencies are limited from 1% to 49.5% of
the sample rate, while the minimum transition width region is 1%
of the sample rate.

Multirate, multistage filters are a practical solution for the design
and implementation of FIR filters with narrow spectral constraints.
Multirate filters change the input data rate at one or more
intermediate points within the filter itself, while maintaining an
output rate that is identical to the input rate. This approach
provides a solution with greatly reduced filter lengths, as
compared to standard single-rate filters.

This can be achieved in two or more simple steps. First, a filter
(with a relative limited edge frequency) is applied and the results
are decimated. Then, a second filter is applied to the decimated
waveform, substantially reducing the lower edge frequency limit.

EXAMPLE

A sine wave with a frequency of 3 MHz has a higher frequency
noise component. A low-pass filter is required to remove the
noise component. The sample rate of the scope is 2 GS/s. The
minimum edge frequency of the low-pass filter for this sample
rate is 20 MHz. While this filter is sufficient for removing part of
the noise, it cannot remove the high frequency component
completely. In such a case, the problem can be solved in two
stages.

DFP-OM-E Rev B ISSUED: July 2001 5–1

LeCroy Digital Filter Package

1. Channel 1 represents a noisy sine wave with a frequency of
3MHz.

2. The first low-pass filter with 20 MHz edge frequency and
30MHz transition region is applied.

3. Trace B is a decimated version of trace A. This is accom­plished by using the identity function and setting "for Math
use" in the ZOOM & MATH menu to 250 points.

4. A second low-pass filter with an edge frequency of 5 MHz
and a transition region width of 6 MHz is applied to trace B.
The result is displayed in trace C.

5. Trace C shows the zoomed signal, which was filtered by a
multistage filtering method. Notice that all high frequency
noise components were removed.

§ § §

5–2 ISSUED: July 2001 DFP-OM-E Rev B

6Specifications

• The pass-band gain of all filters (except custom) is

normalized to 1.

• The group delay of all filters is 0.

• The maximum number of coefficients for all filters is less than

1000.

• The number of input data points in memory should exceed

that of coefficients by a factor of 10.

• The minimum transition region width is 1% of the sample

rate.

• Edge frequencies for the low-pass, high-pass, band-pass,

and band-stop filters are a minimum 1% of the sample rate.
(It is possible to further reduce the lower limit by using a
multirate filter system using multiple filters and an identity —
sparse — function.)

• The grain (minimum step) for setting frequencies is 0.1% of

the sample rate; for widths it is 0.02% of the sample rate; and
for fractions (e.g., ?) it is 0.1%.

§ § §

DFP-OM-E Rev B ISSUED: July 2001 6–1

LeCroy Digital Filter Package

BLANK PAGE

6–2 ISSUED: July 2001 DFP-OM-E Rev B