You are on page 1of 8

AN-837

APPLICATION NOTE
One Technology Way • P.O. Box 9106 • Norwood, MA 02062-9106, U.S.A. • Tel: 781.329.4700 • Fax: 781.461.3113 • www.analog.com

DDS-Based Clock Jitter Performance vs. DAC Reconstruction Filter Performance


by David Brandon and Ken Gentile

A reconstruction filter is an important element for creating a A companion application note, AN-823, also available from
clean, low jitter clock signal from a direct digital synthesizer Analog Devices, Inc., should be used in conjunction with this
(DDS). A reconstruction filter is used at the output of the DAC application note.
to attenuate image frequencies. However, a physical filter cannot When using a DDS for clock generation, insufficient attenuation
be implemented with ideal stop band rejection extending out to of DAC image frequencies can significantly affect system periodic
infinite frequency. This is due to component parasitic effects as jitter performance. Image frequencies theoretically extend to
well as the physical limitations of printed circuit board layout. infinity frequency with magnitudes that follow a sin(x)/x
Failing to use a reconstruction filter with sufficient stop-band response curve. Figure 1 depicts the DAC output spectrum.
rejection can degrade the performance of the DDS. The following formula calculates the theoretical magnitude
A FILTER LAYOUT EXPERIMENT (dBc) of any desired image frequency relative to fOUT, where f is
This application note describes the results obtained from a filter the frequency of the image and fc is the DAC sample rate:
layout experiment and discusses how layout and component ⎧ ⎛ ⎞⎫
selection impact stop-band rejection. This experiment focuses ⎪ ⎜ sin⎛⎜ πf ⎞⎟ ⎟⎪
⎞⎜ ⎜ ⎟
on three LC low-pass filter layouts described in this application ⎪⎛ f
dBc = 20 log ⎨⎜⎜ OUT ⎟⎜ ⎝ fC ⎠ ⎟⎟⎪
⎟ ⎬
note as Filter A, Filter B, and Filter C. All three filters are nearly ⎪⎝ f ⎠⎜ sin⎛⎜ πfOUT ⎞⎟ ⎟⎪
identical in terms of component values and desired response ⎪ ⎜⎜ ⎜ f ⎟ ⎟⎟⎪
⎩ ⎝ ⎝ C ⎠ ⎠⎭
characteristics. The purpose of the experiment is to restrict the
experimental variables to the physical layout and the physical
construction of the components.

fOUT
sin x/x ENVELOPE
x = π (f/fC)
SIGNAL AMPLITUDE

fC – fOUT
fC + fOUT 2fC – fOUT

2fC + fOUT 3fC – fOUT


fC
2 fC
3fC + fOUT
3 fC

0Hz FIRST SECOND THIRD FOURTH FIFTH SIXTH


IMAGE IMAGE IMAGE IMAGE IMAGE IMAGE
05956-001

SYSTEM CLOCK
FREQUENCY (Hz)

Figure 1. The DAC Output Spectrum

Rev. 0
Information furnished by Analog Devices is believed to be accurate and reliable. However, no
responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other
rights of third parties that may result from its use. Specifications subject to change without notice. No One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A.
license is granted by implication or otherwise under any patent or patent rights of Analog Devices. Tel: 781.329.4700 www.analog.com
Trademarks and registered trademarks are the property of their respective owners. Fax: 781.461.3113 ©2006 Analog Devices, Inc. All rights reserved.
AN-837

TABLE OF CONTENTS
A Filter Layout Experiment ........................................................ 1
The Importance of a Reconstruction Filter .............................. 3
Designing a Discrete LC Low-Pass Filter .................................. 3
Choosing Components............................................................ 3
Implementation of a Discrete LC Low-Pass Elliptic Filter...... 3
Filter Experiment Results ............................................................ 4
Filter Schematics........................................................................... 7
Bills of Material for Filters........................................................... 7
PCB Filter Layout Considerations.............................................. 8
Final Points.................................................................................... 8

Rev. 0 | Page 2 of 8
AN-837
THE IMPORTANCE OF A RECONSTRUCTION FILTER frequencies can be adequately rejected. The recommendation is
to observe stop-band rejection out to several gigahertz, or at
Figure 2 displays the measured power of the image frequencies least out to 5 fC.
in the absence of a reconstruction filter. The displayed
attenuation of the image frequencies is due solely to the sin(x)/x Off-the-shelf filters such as SAW, crystal, ceramic, or pre-
response of the DAC output. Sometimes these images are used packaged LC filters are valid candidates for reconstruction
as the desired signal, in which case a band-pass filter is required filters. However, before choosing a filter type, take care to
instead of a low-pass filter. In the spectral plot in Figure 2, the determine the stop-band rejection and insertion loss stated in
DDS sampling clock (fC) is 200 MHz and fOUT is set to 10 MHz. the filter manufacturer's data sheet. The stop-band rejection
RBW 200kHz RF ATT 30dB
specification is always given for a finite bandwidth. This
VBW 20kHz
SWT 640ms UNIT dBm bandwidth may be too narrow to determine its adequacy with
0
regards to extended stop-band rejection. In such cases, validate
–10
the filter stop-band performance before a final selection is made.
–20
DESIGNING A DISCRETE LC LOW-PASS FILTER
–30
AMPLITUDE (dBm)

A discrete LC low-pass filter is relatively inexpensive to


–40
implement and offers significant flexibility. From a component
–50
selection perspective, achieving optimal pass-band and stop-
–60 band performance begins with choosing components with a
–70 high self-resonant frequency and a high Q property. Components
–80 with a high Q rating are desirable for their lower series R value.
In a low-pass filter, this can decrease the insertion loss for the
–90
series inductors and decrease the impedance for the shunt
05956-002

–100
START 0Hz 100MHz STOP 1GHz capacitors. Consequently, the task of choosing components with
Figure 2 .Example of Image Frequencies without a Reconstruction Filter both high Q and high self-resonant frequency can be challenging.
Generally, the physics involved in the construction of capacitors
This plot (Figure 2) demonstrates the importance of using a
and inductors tends to make high Q and high self-resonant
reconstruction filter with adequate stop-band rejection. Note
frequency mutually exclusive parameters. Components must be
that there is significant energy contained in the image frequencies
carefully selected to yield optimal performance.
that must be eliminated, and the task of elimination falls
squarely on the stop band performance of the filter. Unfortu- Choosing Components
nately, the stop-band rejection of a practical filter degrades as The self-resonant frequency of each component is dictated by
frequencies extend well beyond the pass band. This is due to the component value and its physical construction. Choosing a
nonideal components and the physical layout of the components smaller size package is usually a better choice. A smaller package
on a printed circuit board (PCB). For example, real capacitors generally exhibits a higher self-resonant frequency and reduces
and inductors are not strictly capacitive or inductive. the parasitic effects associated with PCB layout. In the case of
Each can be modeled as having resistive (R), inductive (L) and shunt capacitors, the self-resonance can be pushed out to higher
capacitive (C) elements, as shown in Figure 3. From these frequencies by using two capacitors connected in parallel with
models, it is evident that there is a self-resonant component only one-half the value of the original component. The impact
made up of the L and C combination and an insertion loss of this technique is described in the Filter Experiment Results
component due to R. section.
IDEAL MODEL REAL MODEL Choosing components with less insertion loss provides more
(FIRST ORDER)
signal amplitude at the filter output. A reduction in amplitude
C C R L has a direct impact on the slew rate of the output signal. Increased
slew rate applied to the input of a receiver or squarer decreases
L
R wide band noise. Refer to AN-823 for details on the importance
L
05956-003

of signal amplitude in the context of timing jitter.


C
Figure 3. Ideal Model vs. Real Model IMPLEMENTATION OF A DISCRETE LC LOW-PASS
Filter layout can significantly affect stop-band rejection due to ELLIPTIC FILTER
parasitic capacitance between trace pads of the components, A filter of the elliptical, or Cauer, response variety is typically
which can unintentionally couple signals from input to output. recommended for an LC low-pass reconstruction filter. When
Optimal stop band rejection requires attention to both component compared to the other response types, an elliptic filter offers the
selection and layout details. To evaluate the stop band perform- fastest transition from pass band to stop band for a given level
ance of a filter, it should be measured at frequencies well of complexity (that is., filter order).
beyond its designed corner frequency to ensure that image
Rev. 0 | Page 3 of 8
AN-837
This characteristic makes the elliptic filter an attractive Figure 5 shows a shunt-connected, seventh-order, elliptical, low-
candidate as a reconstruction filter. Figure 4 displays a set of pass LC filter.
representative response curves of these four basic filter types:
Cauer, Chebyshev, Butterworth, and Bessel.

05956-005
0dB Figure 5. Shunt-Connected LC Filter
CHEBYSHEV AND CAUER PASS BAND RIPPLE
–3dB FILTER EXPERIMENT RESULTS
Figure 7 displays the results obtained from the experimental
MAGNITUDE (dB)

filter board containing Filter A, Filter B, and Filter C. The top


trace in each plot is the measured frequency response of the
filter; the bottom trace is the measured spectrum from a DDS
CAUER
evaluation board after passing through the experimental filter.
CHEBYSHEV The measurement configuration is shown in Figure 6.
BUTTERWORTH
DDS EVALUATION BOARD
BESSEL
AVDD
FREQUENCY 05956-004
PASS BAND
CORNER FREQUENCY 1:1
ADTT1-1
Figure 4. Response Curves for Four Basic Filter Types
DDS AVDD EXPERIMENTAL SPECTRUM
The tradeoff for the steep roll-off offered by the elliptic response FILTER BOARD ANALYZER
is the appearance of response ripples in the pass band and the
stop band. Note that the magnitude of the pass-band ripple

05956-009
shown in Figure 4 was arbitrarily drawn for demonstration AVDD
purposes only. Although some pass-band ripple is necessary in
Figure 6. Measurement Configuration
an elliptic filter design, it is one of the parameters that the filter
designer can control. Therefore, its magnitude is dependent on Note that the DSS evaluation board contains an output coupling
the particular filter design. transformer that has a bandwidth of ~300 MHz. Therefore, there
is additional attenuation in the measured spectrum at frequencies
Pass-band ripple can impact certain DDS applications. In single
above 300 MHz due to the response of the transformer.
tone applications, such as carrier signal generation, pass-band
ripple is not a critical parameter. However, in those applications The three LC low-pass elliptic filter implementations demon-
where the DDS generates a modulated carrier, pass-band ripple strate varying image rejection due to PCB layout. All three
must be considered as a crucial factor in the filter design. implementations are seventh-order elliptical 160 MHz low-pass
filters using discrete LC components.
LC filters are available in two general forms dictated by whether
the driving circuit is a current source or voltage source. If driven The DDS filtered output frequency for all plots is set to
by a current source, the first element is shunt connected, whereas 110.123 MHz with a DAC sample rate of 500 MSPS (REFCLK).
the first element is series connected for a voltage source. Note that a transformer (ADTT1-1 available from Mini-Circuits®)
differentially couples the DDS outputs. The transformer output
provides a single-ended connection to the filter input.

Rev. 0 | Page 4 of 8
AN-837
FILTER BOARD FILTER LAYOUT 0
–10
–20
C1 C2 –30
–40
–50 FILTER A
–60
RESPONSE
L1 C3 –70
–80
C4 C5 –90
0Hz 350MHz 3.5GHz
0
–10
L2 C6 –20
FILTER A
–30
C7 C8 –40
–50 DDS FILTERED
–60 IMAGE OUTPUT
L3 C9 –70
–80
–90
C10 C11
–100
0Hz 350MHz 3.5GHz

0
–10
C1 –20
–30
–40
–50 FILTER B
L1 C2 –60 RESPONSE
–70
C3 –80
–90
0Hz 350MHz 3.5GHz
0
L2 C4 –10
FILTER B –20
C5 –30
–40
–50 DDS FILTERED
IMAGES OUTPUT
–60
L3 C6
–70
C7 –80
–90
–100
0Hz 350MHz 3.5GHz

0
–10
–20
–30
–40
–50 FILTER C
–60 RESPONSE
C1 –70
–80
L1 C2 –90
L2 0Hz 350MHz 3.5GHz
C5 C3 0
FILTER C –10
C4
–20
L3 C6 –30
–40
C7 IMAGES
–50
DDS FILTERED
–60
OUTPUT
–70
–80
–90
–100
05956-006

0Hz 350MHz 3.5GHz

Figure 7. Experiment Results for Filter A, Filter B, and Filter C

Rev. 0 | Page 5 of 8
AN-837
The measured frequency response of the filters shown in For example, the peak-to-peak jitter associated with N spurs
Figure 7 is from 0 Hz to 3.5 GHz (the upper frequency limit of takes the form:
the measurement instrument). For a 1 GHz DDS, it is recom-
1 ⎛ N dBcn ⎞
mended that measurements extend beyond 3.5 GHz to better JitterUI peak −to − peak = arctan ⎜ ∑10 20 ⎟
π ⎜ n =1 ⎟
characterize the filter stop-band response. ⎝ ⎠
In applications requiring very low jitter, adequate suppression of Furthermore, the plot does not address random jitter, the effects
DAC image frequencies is critical. In fact, there is a direct of which must be considered separately.
relationship between the magnitude of a nonharmonic spectral 0.020
0.019
spur and the resulting periodic jitter, given by: 0.018

TIMING JITTER PEAK-TO-PEAK (dBc)


0.017
1 ⎛ dBc ⎞ 0.016
JitterUI peak −to − peak = arctan ⎜⎜10 20 ⎟⎟ 0.015
π ⎝ ⎠
0.014
0.013
0.012
where dBc is the spur magnitude relative to the fundamental 0.011
0.010
magnitude. 0.009
0.008
A plot of this formula is shown in Figure 8. The jitter is 0.007
0.006
expressed as peak-to-peak unit intervals (UI). One UI is one 0.005
0.004
period of the associated clock signal. Expressing the jitter in UI 0.003
allows the plot to apply to any arbitrary clock signal. For 0.002
0.001
example, if a clock signal exhibits a −35 dBc nonharmonic spur, 0

05956-007
–25 –30 –35 –40 –45 –50 –55 –60 –65 –70 –75 –80 –85
then one can expect about 0.006 UI of peak-to-peak jitter due SPUR MAGNITUDE (dBc)
solely to the spur. Figure 8. Timing Jitter vs. Spurious Interference
Assuming a clock frequency of 100 MHz, then 1 UI is 10 ns,
therefore 0.006 UI equates to 60 ps (0.006 × 10 ns) for a
100 MHz clock. When using this plot, keep in mind that the
jitter values are for a single spurious component. When multiple
spurs are present, their individual contributions must be
summed in the argument for the arctan function as shown in
the previous equation.

Rev. 0 | Page 6 of 8
AN-837
FILTER SCHEMATICS BILLS OF MATERIALS FOR FILTERS
FILTER A
SCHEMATIC The Bill of Materials for Filter A is provided in Table 1 while
the Bill of Materials for Filter B and Filter C are provided in
L1 L2 L3
Table 2. Note that the following points for these Bills of
C2 39nH C5 56nH C8 68nH C11 Materials:
5.6pF 12pF 15pF 10pF

C1 C4 C7 C10
• The quantity for all materials is 1.
C3 C6 C9
5.6pF 12pF 12pF 6.8pF 15pF 2.2pF 10pF
• The size of all materials is 0402.
FILTER B AND C • All materials are Murata Manufacturing Company,
SCHEMATIC
Ltd. parts.
L1 L2 L3
39nH 56nH 68nH

C1 C3 C5 C7

05956-008
C2 C4 C6
12pF 12pF 27pF 6.8pF 33pF 2.2pF 22pF

Figure 9. Schematic for Filter A Compared to Schematic for


Filter B and Filter C

Table 1. Filter A Bill of Materials


Reference Designator Value Supplier Part Number
C1 5.6 pF GRM1555C1H5R6DZ01
C2 5.6 pF GRM1555C1H5R6DZ01
C3 12 pF GRM1555C1H120JZ01
C4 12 pF GRM1555C1H120JZ01
C5 12 pF GRM1555C1H120JZ01
C6 6.8 pF GRM1555C1H6R8DZ01
C7 15 pF GRM1555C1H150JZ01
C8 15 pF GRM1555C1H150JZ01
C9 2.2 pF GRM1555C1H2R2CZ01
C10 10 pF GRM1555C1H100JZ01
C11 10 pF GRM1555C1H100JZ01
L1 39 nH LQG15HS39NJ02D
L2 56 nH LQG15HS56NJ02D
L3 68 nH LQG15HS68NJ02D

Table 2. Filter B and Filter C Bill of Materials


Reference Designator Value Supplier Part Number
C1 12 pF GRM1555C1H120JZ01
C2 12 pF GRM1555C1H120JZ01
C3 27 pF GRM1555C1H270JZ01
C4 6.8 pF GRM1555C1H6R8DZ01
C5 33 pF GRM1555C1H330JZ01
C6 2.2 pF GRM1555C1H2R2CZ01
C7 22 pF GRM1555C1H220JZ01
L1 39 nH LQG15HS39NJ02D
L2 56 nH LQG15HS56NJ02D
L3 68 nH LQG15HS68NJ02D

Rev. 0 | Page 7 of 8
AN-837
PCB FILTER LAYOUT CONSIDERATIONS FINAL POINTS
The list below offers useful layout tips for obtaining optimal A DDS is a sampled data system that produces a spectrum with
performance from the reconstruction filter. multiple copies of the fundamental arrayed around multiples of
the sampling frequency.
• Use a solid (uninterrupted) ground plane below the
components to lower the loop inductance of the circuit, The magnitude of these image frequencies can be the dominant
thus lowering the impedance for return currents. source of periodic jitter. Any spur present at the output of the
filter and within the input bandwidth of the receiver degrades
• Split shunt capacitors as shown with Filter A (see Figure 7)
performance in proportion to the spur magnitude. Therefore,
to increase the self-resonant frequency of the capacitors
before settling on a DAC reconstruction filter, measure the filter
and lower the inductive connection to ground. In addition,
performance to ensure that it meets the requirements of the
use a top ground plane alongside for the shunt components
application.
in combination with the bottom ground plane. This can
further lower the inductive connection to ground. Use Keep in mind that filter design software programs that do not
multiple vias to tie both ground planes together. support the inclusion of component and layout parameters only
simulate the ideal filter frequency response. The actual fre-
• Avoid placing the filter components in close proximity
quency response varies depending on the actual high frequency
with one another. Filter C (see Figure 7) is a good example
characteristics of the components and the printed circuit board.
of a layout where components are crowded. When
components are crowded, trace parasitics and mutual
coupling can affect the frequency response of the filter.
• Use components with both as high a self-resonant
frequency and as high a Q factor as possible.
• Maintain the impedance of the traces to match the charac-
teristic impedance of the filter. This is not as critical for
component interconnections, but it is important for the
traces that interface to circuits external to the filter, such as
the input and output connections.

©2006 Analog Devices, Inc. All rights reserved. Trademarks and


registered trademarks are the property of their respective owners.
AN05956-0-12/06(0)

Rev. 0 | Page 8 of 8

You might also like