An 1368 PDF
An 1368 PDF
An 1368 PDF
com
AN-1368
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
INTRODUCTION
High resolution, high performance converters and radio frequency Figure 1 shows an example of a filtering scheme often used in
(RF) systems require a low noise design of the power supply to mixed-signal ICs to separate analog and digital supplies. The
achieve optimum performance. Good filtering of power supply scheme is suitable for any power distribution network; however,
noise and high frequency crosstalk reduction between analog and understanding its effectiveness and limitations helps users to
digital domains is required, especially on mixed-signal converters avoid characteristics that may, for example, be detrimental in high
and transceivers. performance converter applications.
Power supply noise is more evident with switching regulators, This application note discusses the important considerations that
which generate undesired output artifacts that can be harmful system designers need to be aware of when using ferrite beads
to any noise sensitive system. However, due to their high power in power supply systems, including response characteristics of
conversion efficiency, switching regulators are often found in the ferrite bead, a simplified ferrite bead model and simulation, dc
portable devices where prolonged battery life is needed and in current considerations, LC resonance effects, and damping
systems where thermal limitations exist. methods. An understanding of these elements can make the
Similar digital and analog voltage rails of mixed-signal ICs are system design approach fast and effective.
often powered from different power domains. This approach The Analog Devices, Inc., products used to demonstrate the
helps to prevent fast digital switching noise coupling onto the effects of ferrite beads as output filters are the 1.25 A, 1.2 MHz
sensitive analog supply rail and degrading the converter synchronous step-down switching regulator (ADP2120) and the
performance, but it increases system level complexity and cost. 2 A/1.2 A dc-to-dc switching regulator with independent positive
With proper high frequency isolation of supply domains, power and negative outputs (ADP5071).
supplies can be shared between analog and digital domains, ANALOG DIGITAL
SUPPLY SUPPLY
simplifying the design and reducing the cost.
An effective method for filtering high frequency power supply POWER
noise and cleanly sharing similar supply rails is the use of ferrite SUPPLY
FERRITE
beads. A ferrite bead is a passive device that filters high frequency BEAD
13359-001
noise energy over a broad frequency range. It becomes resistive
over its intended frequency range and dissipates the noise energy in
Figure 1. Sample Filter Scheme for a Mixed-Signal IC
the form of heat. The ferrite bead is connected in series with the
power supply rail and is often combined with capacitors to ground
on either side of the bead. This forms a low-pass filter network,
further reducing the high frequency power supply noise.
Rev. 0 | Page 1 of 11
Licenciado para - Claudean Ferreira soares - 07670754478 - Protegido por Eduzz.com
TABLE OF CONTENTS
Introduction ...................................................................................... 1 LC Resonance Effect .........................................................................6
Revision History ............................................................................... 2 Damping Methods.............................................................................9
Ferrite Bead Response Characteristics .......................................... 3 Conclusion....................................................................................... 11
Ferrite Bead Simplified Model and Simulation ............................ 3 References ........................................................................................ 11
DC Bias Current Considerations .................................................... 4
REVISION HISTORY
8/15—Revision 0: Initial Version
Rev. 0 | Page 2 of 11
Licenciado para - Claudean Ferreira soares - 07670754478 - Protegido por Eduzz.com
IMPEDANCE (Ω)
Z
1200
RESISTIVE
600
INDUCTIVE CAPACITIVE
Z ≈ |XC|;
1000 Z ≈ XL; LBEAD R CPAR
400
X
800
IMPEDANCE (Ω)
CROSSOVER
FREQUENCY Z 200
600
13359-004
R
400 1 10 100 1000
X FREQUENCY (MHz)
For the region on the measured ZRX plot where the bead appears
0 most inductive (Z ≈ XL; LBEAD), the bead inductance is calculated
13359-002
1 10 100 1000
FREQUENCY (MHz) by the following equation:
Figure 2. Tyco Electronics BMB2A1000LN2 ZRX Plot XL
LBEAD = (1)
At relatively low frequencies below crossover, the bead response 2×π × f
is inductive. At high frequencies, the bead response is capacitive. In where:
this case, the falling slope of the impedance vs. frequency plot is f is the frequency point anywhere in the region the bead appears
defined by the parasitic capacitance associated with the component. inductive. In this example, f = 30.7 MHz.
FERRITE BEAD SIMPLIFIED MODEL AND XL is the reactance at 30.7 MHz, which is 233 Ω.
SIMULATION Equation 1 yields an inductance value (LBEAD) of 1.208 μH.
A ferrite bead can be modeled as a simplified circuit consisting For the region where the bead appears most capacitive (Z ≈ |XC|;
of resistors, an inductor, and a capacitor, as shown in Figure 3. CPAR), the parasitic capacitance is calculated by the following
RDC corresponds to the dc resistance of the bead. CPAR, LBEAD, equation:
and RAC are the parasitic capacitance, the bead inductance, and
the ac resistance (ac core losses) associated with the bead. 1
CPAR = (2)
CPAR 2×π × f × | XC |
where:
LBEAD
RDC f is the frequency point anywhere in the region the bead appears
capacitive. In this example, f = 803 MHz.
13359-003
RAC
|XC| is the reactance at 803 MHz, which is 118.1 Ω.
Figure 3. Simplified Circuit Model Equation 2 yields a parasitic capacitance value (CPAR) of 1.678 pF.
In some cases, the simplified circuit model shown in Figure 3 can
be used to approximate the ferrite bead impedance characteristic
up to the sub-GHz range.
Rev. 0 | Page 3 of 11
Licenciado para - Claudean Ferreira soares - 07670754478 - Protegido por Eduzz.com
by up to 90%.
600 450
UP TO 90% DECREASE
400 IN BEAD INDUCTANCE
AT 50% RATED CURRENT
400
350
SIMULATION
INDUCTANCE (nH)
200 300
300mΩ 1.208µF 0
13359-006
0 1 2 3 4 5
R2 L1
AC 1A DC BIAS CURRENT (A)
I1 1.082kΩ
Figure 7. Effect of DC Bias on Bead Inductance
13359-205
R1
Rev. 0 | Page 4 of 11
Licenciado para - Claudean Ferreira soares - 07670754478 - Protegido por Eduzz.com
750mA
–40
40
–50
20
–60
0
13359-007
1 10 100 1000
–70
13359-009
FREQUENCY (MHz) 0.01 0.1 1 10 100
Figure 8. TDK MPZ1608S101A Impedance Curves with Respect to DC Bias FREQUENCY (MHz)
Current Figure 10. Measured Response of the Bead and Capacitor Low-Pass Filter vs.
120
DC Bias Current
CAPACITOR:
1500mA MURATA
13359-209
IBIAS
2000mA GRM188R61A105KA61
60 3000mA (1µF, X5R, 0603)
4000mA
5000mA Figure 11. Test Circuit Model
40
For a 3 A bead with 250 mA of bias current, the resonance
cutoff moves to the right mainly due to the inductance drop
20
described in Figure 7.
0 In addition, the effective attenuation of the network reduces by
13359-008
1 10 100 1000
as much as 15 dB at 1 MHz, where most modern switching
FREQUENCY (MHz)
Figure 9. Wurth Elektronik 742 791 510 Impedance Curves with Respect to DC
regulators operate. This reduction can cause problems if the
Bias Current system engineer relies solely on the data sheet, which shows the
impedance at zero dc bias current only. Therefore, to validate the
As shown in Figure 8 and Figure 9, the effective impedance
effectiveness of the ferrite bead filter and for accurate results, model
at 100 MHz dramatically drops from 100 Ω to 10 Ω for the
the inductance over a range of dc bias current and identify its
TDK MPZ1608S101A (100 Ω, 3 A, 0603) and from 70 Ω to
impedance characteristics under the actual operating conditions.
15 Ω for the Wurth Elektronik 742 792 510 (70 Ω, 6 A, 1812) by
applying just 50% of the rated current.
Rev. 0 | Page 5 of 11
Licenciado para - Claudean Ferreira soares - 07670754478 - Protegido por Eduzz.com
VIN
5V CIN R1
22µF 10Ω
X5R ADP2120ACPZ-2.5-R7
6.3V
1 VIN EN 10
2 PVIN SYNC/MODE 9
MPZ1608S101A R2
VOUT 10kΩ
3 SW PGOOD 8
2.5V C COUT1 COUT2 L
1µF 22µF 10µF 1.5µH
X5R X5R 4 PGND TRK 7
6.3V 6.3V
RL 5 GND FB 6
13359-010
10Ω
Figure 12. ADP2120 Application Circuit with Bead and Capacitor Low -Pass Filter Implementation for DC Bias Effect
Figure 12 shows an application circuit that uses the ADP2120 LC RESONANCE EFFECT
buck regulator running in forced pulse-width modulation (FPWM) Resonance peaking is possible when implementing a ferrite
mode with the ferrite bead filter of Figure 11. With the ADP2120 bead together with a decoupling capacitor. This commonly
in FPWM mode, the switching noise spectral output does not overlooked effect can be detrimental because it may amplify
change significantly with respect to load current. For more details ripple and noise in a given system instead of attenuating it. In
about FPWM mode, see the ADP2120 data sheet. The spectral many cases, this peaking occurs around popular switching
output is plotted in Figure 13, showing the effects of dc bias on frequencies of dc-to-dc converters.
ferrite bead response.
70
Peaking occurs when the resonant frequency of a low-pass
ADP2120 NO LOAD OUTPUT SPECTRUM
ADP2120 + UNDAMPED BEAD-C, ILOAD = 250mA filter network, formed by the ferrite bead inductance and the
ADP2120 + UNDAMPED BEAD-C, ILOAD = NO LOAD
60 high Q decoupling capacitance, is below the crossover frequency
of the bead. The resulting filter is underdamped. Figure 14
50
shows the measured impedance vs. frequency plot of the
AMPLITUDE (dBµV)
100
As the load current increases in Figure 13, the filter cutoff
CROSSOVER FREQUENCY
frequency moves to the right from approximately 180 kHz with 80
R
0
13359-012
1 10 100 1000
FREQUENCY (MHz)
Rev. 0 | Page 6 of 11
Licenciado para - Claudean Ferreira soares - 07670754478 - Protegido por Eduzz.com
–20 has been taken with a very light load (in the microampere
–30
range), but this is a realistic application in sections of circuits
that need just a few microamperes to 1 mA of load current or
–40
sections that are turned off to save power in some operating
–50 modes. This potential peaking creates additional noise in the
–60
system that can create unwanted crosstalk.
–70
As an example, Figure 17 shows an ADP5071 application circuit
13359-013
DC BIAS
CAPACITOR: CURRENT
MURATA (µA)
13359-213
GRM188R7H103KA01
(10nF, X7R, 0603)
ADP5071
RC1 SS INBK
118kΩ L1
COMP1 2.2µH D1
MBR130T1G MPZ1608S101A
CC1 +16V
820pF SW1
EN1
RFT1 C RL
68.1kΩ 10nF 3.2kΩ
VREG FB1 COUT1
RFB1 10µF
CVREG
3.57kΩ
1µF PVIN1
VIN
PVIN2 PGND
+9V CVREF
CIN1 PVINSYS
1µF
10µF
EN2 VREF
RFB2
RC2
3.09kΩ COUT2
57.6kΩ 10µF
COMP2 FB2
RFT2
CC2
64.9kΩ
2.7nF VREG SYNC/FREQ
SW2 –16V
SLEW D2
SEQ MBR130T1G
AGND
L2
13359-014
6.8µH
Figure 17. ADP5071 Application Circuit with Bead and Capacitor Low-Pass Filter Implementation on Positive Output
Rev. 0 | Page 7 of 11
Licenciado para - Claudean Ferreira soares - 07670754478 - Protegido por Eduzz.com
13359-015
0.01 0.1 1 10 100
0
13359-116
1 10 100 1k 10k 100k 1M 10M 100M
and damps for higher source resistance. However, the load FREQUENCY (Hz)
regulation degrades with this approach, making it unrealistic in Figure 21. Circuit Simulation Comparison of 10 Ω Load vs. 1 MΩ Load
practice. The dc voltage at the load droops with load current due 200fF
to the drop from the series resistance. C2
40 100mΩ 30mΩ 800nH
IN OUT
+ R3 R5 L1 C1
10Ω SOURCE 10nF
20 150Ω
0.1Ω SOURCE LOAD
AC1 L2
R4 2nH
V1
0 R2
INSERTION LOSS (dB)
R1
10mΩ
13359-316
–20
–60
–80
–100
13359-016
C2
SOURCE
30mΩ 800nH
IN OUT
+ R3 R5 L1 C1
10nF
150Ω
AC1 L2 R2
R4 2nH
V1 1MΩ
R1
10mΩ
13359-216
Rev. 0 | Page 8 of 11
Licenciado para - Claudean Ferreira soares - 07670754478 - Protegido por Eduzz.com
10
13359-019
0
C Figure 25. Parallel Resistor Across Ferrite Bead
INSERTION LOSS (dB)
–10 160
A BEAD ONLY,
DC BIAS 0mA
–20 140
R
B X
120 Z
–30
BEAD WITH
10Ω RESISTOR,
IMPEDANCE (Ω)
–40 100 DC BIAS 0mA
R
UNDAMPED X
–50 METHOD A: SERIES 10Ω RESISTOR TO DECOUPLING CAPACITOR 80 Z
METHOD B: PARALLEL 10Ω RESISTOR ACROSS FERRITE BEAD
METHOD C: DAMPED: C DAMP (1µF) + RDAMP (2Ω) 60
–60 BEAD: TDK MPZ1608S101A (100Ω, 3A, 0603)
CAPACITOR: MURATA GRM188R71H103KA01 (10nF, X7R, 0603)
40
–70
13359-017
13359-021
1M 10M 100M 1G
Method A consists of adding a series resistor to the decoupling FREQUENCY (Hz)
capacitor path (see Figure 24), which dampens the resonance of Figure 26. MPZ1608S101A Impedance Curve
the system but degrades the bypass effectiveness at high
12
frequencies.
FERRITE BEAD:
TDK MPZ1608S101A 10
(100Ω, 3A, 0603)
IN OUT
CAPACITOR: 8 BEAD ONLY,
IMPEDANCE (Ω)
4
CAPACITOR R
X
Z
Figure 24. Series Resistor to Decoupling Capacitor 2
13359-022
1M 10M 100M 1G
of the system. However, the attenuation characteristic of the filter is FREQUENCY (Hz)
reduced at high frequencies. Figure 26 and Figure 27 show the Figure 27. MPZ1608S101A Impedance Curve, Zoom View
impedance vs. frequency curve of the MPZ1608S101A with and
without a 10 Ω parallel resistor. The green dashed curve is the
overall impedance of the bead with a 10 Ω resistor in parallel. The
impedance of the bead and resistor combination is significantly
reduced and is dominated by the 10 Ω resistor. However, the
3.8 MHz crossover frequency for the bead with the 10 Ω parallel
resistor is much lower than the crossover frequency of the bead
on its own at 40.3 MHz. The bead appears resistive at a much
lower frequency range, lowering the Q for improved damped
performance.
Rev. 0 | Page 9 of 11
Licenciado para - Claudean Ferreira soares - 07670754478 - Protegido por Eduzz.com
13359-020
RC DECOUPLING FILTER 40
AMPLITUDE (dBµV)
30
Figure 28. Additional RC Decoupling Filter
13359-023
damping resistor value. The capacitor impedance must be 0.01 0.1 1 10 100
FREQUENCY (MHz)
sufficiently smaller than the damping resistance at the resonant
Figure 29. ADP5071 Spectral Output plus Bead and Capacitor Low-Pass Filter
frequency to reduce the peaking. Determine the RDAMP range by with Method C Damping
using Equation 3 and Equation 4. Using a ratio of 16 or higher
between damping capacitance (CDAMP) and decoupling capacitance Generally, Method C is the most elegant and is implemented by
(CDECOUP) provides a range of damping resistance that satisfies adding a resistor in series with a ceramic capacitor rather than
both equations. buying an expensive dedicated damping capacitor. The safest
designs always include a resistor that can be tweaked during
LBEAD prototyping and that can be eliminated if not necessary. The
RDAMP ≥ 2 (3)
C DAMP only drawbacks are the additional component cost and greater
required board space.
LBEAD
RDAMP ≤ 0.5 (4)
C DECOUP
where:
RDAMP is the damping resistance.
LBEAD is the bead inductance from Equation 1, including external
inductance such as the parasitic trace inductance of the board.
CDAMP is the damping capacitance.
CDECOUP is the decoupling capacitance.
Rev. 0 | Page 10 of 11
Licenciado para - Claudean Ferreira soares - 07670754478 - Protegido por Eduzz.com
Rev. 0 | Page 11 of 11