Application Note No. 060

A pp l ic a t io n N o t e, R e v . 3. 0 , J a n. 2 00 7
A p p li c a t i o n N o t e N o . 0 6 0
A H i g h T h i r d- O r d e r I nt e r c e pt L o w N o i s e A m p l i f i e r
f o r 1 9 0 0 M H z A p pl i c a t i o n U s i n g t h e S i l i c o n Germanium BFP620 Transistor
R F & P r o t e c ti o n D e v i c e s
Edition 2007-01-09
Published by
Infineon Technologies AG
81726 München, Germany
© Infineon Technologies AG 2009.
All Rights Reserved.
LEGAL DISCLAIMER
THE INFORMATION GIVEN IN THIS APPLICATION NOTE IS GIVEN AS A HINT FOR THE IMPLEMENTATION
OF THE INFINEON TECHNOLOGIES COMPONENT ONLY AND SHALL NOT BE REGARDED AS ANY
DESCRIPTION OR WARRANTY OF A CERTAIN FUNCTIONALITY, CONDITION OR QUALITY OF THE
INFINEON TECHNOLOGIES COMPONENT. THE RECIPIENT OF THIS APPLICATION NOTE MUST VERIFY
ANY FUNCTION DESCRIBED HEREIN IN THE REAL APPLICATION. INFINEON TECHNOLOGIES HEREBY
DISCLAIMS ANY AND ALL WARRANTIES AND LIABILITIES OF ANY KIND (INCLUDING WITHOUT
LIMITATION WARRANTIES OF NON-INFRINGEMENT OF INTELLECTUAL PROPERTY RIGHTS OF ANY
THIRD PARTY) WITH RESPECT TO ANY AND ALL INFORMATION GIVEN IN THIS APPLICATION NOTE.
Information
For further information on technology, delivery terms and conditions and prices please contact your nearest
Infineon Technologies Office (www.infineon.com).
Warnings
Due to technical requirements components may contain dangerous substances. For information on the types in
question please contact your nearest Infineon Technologies Office.
Infineon Technologies Components may only be used in life-support devices or systems with the express written
approval of Infineon Technologies, if a failure of such components can reasonably be expected to cause the failure
of that life-support device or system, or to affect the safety or effectiveness of that device or system. Life support
devices or systems are intended to be implanted in the human body, or to support and/or maintain and sustain
and/or protect human life. If they fail, it is reasonable to assume that the health of the user or other persons may
be endangered.
Application Note No. 060
Application Note No. 060
Revision History: 2007-01-09, Rev. 3.0
Previous Version: 2000-10-05
Page
Subjects (major changes since last revision)
All
Document layout change
Application Note
3
Rev. 3.0, 2007-01-09
Application Note No. 060
A High Third-Order Intercept Low Noise Amplifier for 1900 MHz Application
1
•
•
•
•
•
•
A High Third-Order Intercept Low Noise Amplifier for 1900 MHz
Application Using the Silicon-Germanium BFP 620 Transistor
Gain = 14.7 dB
Very Low Noise Figure = 1.05 dB
High Input 3rd-Order Intercept Point = +10 dBm
Low Power Consumption, 8.9 mA at 3.0 V
Single Supply Voltage (no negative bias required), low parts count
Applications: Cellular Handsets for CDMA, WCDMA, UMTS
3
2
4
1
1 = B, 2 = E, 3 = C, 4 = E
1.1
Description
Infineon´s BFP620 is a high-performance, low-cost Silicon-Germanium bipolar transistor fabricated with Infineon´s
B7HF SiGe process. Housed in a 4-lead ultra-miniature SOT343 surface mount package and with a 70 GHz
transition frequency (fT), this device is ideal for high performance applications including Low Noise Amplifiers in
portable telephones and other battery operated wireless communications devices. The BFP620 offers
exceptionally low noise figure, high gain and high linearity at low power consumption levels. The BFP620 rivals
more expensive GaAs MESFET and PHEMT devices in performance without requiring a negative supply voltage.
This application note describes a low noise figure, high 3rd-order intercept 1900 MHz LNA targeted for CDMA
applications and also provides guidelines for improving 3rd-order intercept performance in amplifiers using Infineon
Transistors.
1.2
Overview of 1900 MHz High-Linearity LNA
The schematic diagram for the 1930 - 1990 MHz low noise amplifier (LNA) is shown in Figure 1. The amplifier is
intended for use in low-cost battery powered applications such as those found in PCS band or TriMode CDMA
cellular handsets. Design goals include a gain of 15 dB, unconditional stability, a noise figure under 1.1 dB, an
Input 3rd-Order Intercept Point (IIP3) of +10 dBm, reduced parts count and low PCB area requirement. Input and
output return loss values of better than 10 dB were sought in order to do easy integration with a duplexer and
image-stripping filter. The Bill Of Material and a PC Board cross-section diagram are shown in Table 2 and a more
detailed description of the LNA follows in Chapter 2. Measurement results are summarized in Table 1. These are
mean values taken from a sample lot of 10 application boards (please refer to Appendix C). The insertion losses
of the connectors and microstrip lines are not backed out for these results - if the (input) transmission line and
connector losses were subtracted, the noise figure results would be ≈ 0.2 dB lower.
Application Note
4
Rev. 3.0, 2007-01-09
Application Note No. 060
A High Third-Order Intercept Low Noise Amplifier for 1900 MHz Application
Typical Performance for BFP620 V4.3 Application at 1960 MHz, T = 25 °C, V = 3.0 V
Table 1
Parameter
Symbol
Value
Gain
dB[s21]²
14.7 dB
1.04 dB
Output 1dB compression
NF
IIP3
OIP3
OP1dB
Input Return Loss
dB[s11]²
13.3 dB
Output Return Loss
dB[s22]²
14.6 dB
Collector-Emitter Voltage
VCE
I
2.1 V
Noise Figure
rd
Input 3 Order Intercept
rd
Output 3 Order Intercept
Supply Current
+10.3 dBm
+25.0 dBm
+5.8 dBm
8.9 mA
LNA PC Board Area
≈ 45 mm2
Number of external SMT Components1)
12
1) Includes bias resistors and DC blocks
AN060_Schematic_Diagram.vsd
Figure 1
Schematic Diagram for PCS LNA, Printed Circuit Board Version 4.3
Application Note
5
Rev. 3.0, 2007-01-09
Application Note No. 060
A High Third-Order Intercept Low Noise Amplifier for 1900 MHz Application
Table 2
Bill Of Material
Reference
Designator
Value
Manufacturer / Type
Case
Size
Function
C1
10 pF
Various
0402
Input DC block
C2
1.5 pF
Various
0402
Output DC block, output RF matching
C3
0.1 µF
Various
0603
Low frequency ground at base (Input 3rd
Order Intercept improvement)
C4
10 pF
Various
0402
RF bypass / RF block
C5
10 pF
Various
0402
RF bypass / RF block
C6
0.1 µF
Various
0603
Bypass / block
C7
0.5 pF
Various
0402
High freq. stability / kill gain > 8 GHz
L1
15 nH
Murata LQG10A
(Low cost inductor)
0402
RF choke to DC bias on base
L2
3.3 nH
Murata LQG10A
(Low cost inductor)
0402
Output RF match, DC feed to collector
R1
10 Ω
Various
0402
Stability (below 2 GHz)
R2
30 kΩ
Various
0402
DC bias for base
R3
91 Ω
Various
0402
DC bias / drop voltage for collector
Q1
Infineon Technologies
SOT343 BFP620 Silicon-Germanium transistor
J1
Johnson 142-0701-841
RF input connector
J2
Johnson 142-0701-841
RF output connector
J3
AMP 5 pin header, MTA-100 series
640456-5 (Standard pin plating)
Or
641215-5 (gold plated pins)
DC connector
Pins 1, 5 = grounded
Pin 3 = VCC
Pins 2, 4 = no connection
PCB CROSS SECTION
TOP LAYER
.008 inch / 0.20 mm
.033 inch / 0.84 mm ?
INTERNAL 1
INTERNAL 2
(BOARD HAS 3 METAL LAYERS, INTERNAL 1
AND INTERNAL 2 ARE MERGED TOGETHER)
LAYER FOR MECHANICAL RIGIDITY OF PCB, THICKNESS HERE NOT CRITICAL
AS LONG AS TOTAL PCB THICKNESS DOES NOT EXCEED 0.045 INCH
(SPECIFICATION FOR TOTAL PCB THICKNESS: 0.040 + 0.005 / - 0.005 INCH)
BOTTOM LAYER
AN060_PCB_Cross_Section.vsd
Figure 2
PC Board Cross Section. Board material is standard FR4. Note spacing from top layer RF
traces to internal ground plane is 0.008 inch / 0.20 mm
Application Note
6
Rev. 3.0, 2007-01-09
Application Note No. 060
Low Noise Amplifier Design Considerations
2
Low Noise Amplifier Design Considerations
In subsequent sections the various design considerations and trade-offs for this particular LNA will be described,
including:
1.
2.
3.
4.
5.
Linearity
Stability
Noise Figure
Input / Output Match
DC Bias
2.1
Linearity
The primary design goal for this Low Noise Amplifier (LNA) was to achieve an input 3rd-order intercept point of
+10 dB, while consuming less than 10 mA of DC current. To this end, two techniques were employed: a) inductive
emitter degeneration and b) use of additional charge storage across the base-emitter junction of the transistor.
2.1.1
Some Definitions
The usual method employed for measuring 3rd order intercept point (IP3) involves injecting two equal amplitude
sinusoidal signals at frequencies f1 and f2 into the amplifier input, and observing the relative levels of the test
signals f1, f2 and third-order products 2f2 - f1 and 2f1 - f2 at the amplifier output. The third-order products, as well
as all other signals present at the amplifier output besides f1 and f2 constitute distortion and are the results of
nonlinear behaviour within the transistor, most notably across the base-emitter junction. For the testing done on
this particular LNA, each input test tone power level (Pin, input power for both f1 and f2) is set to -20 dBm, and the
spacing between the tones is nominally 1 MHz. The input 3rd-order intercept point (IIP3) is calculated as follows:
(1)
∆IM3
IIP3 = PIN + -------------2
Where ∆IM3 = the difference in amplitude between one of the two equal amplitude test tones present at the
amplifier output, and the level of the highest 3rd-order distortion product. Please refer to Figure 3.
Application Note
7
Rev. 3.0, 2007-01-09
Application Note No. 060
Low Noise Amplifier Design Considerations
∆f
∆ IM3
f 2-f1
2f1-f 2
f1
f2
2f2-f 1
Center
AN060_Output_spectrum_LNA.vsd
Figure 3
Partial presentation of output spectrum of the LNA; f1 and f2 are equal - amplitude input test
tones. For this LNA, -20 dBm powered tone is used. Other signals besides f1 and f2 are
generated as a result of non-linear device behavior
2.1.1.1
Inductive Emitter Degeneration
The process of intentionally inserting additional inductance between device emitter connections and RF ground is
a commonly employed method used for influencing device input / output match, noise match, stability and linearity.
Inductive degeneration does not seriously impact noise figure performance, as resistive degeneration does. In this
design, two microstrip lines, each with dimensions of 0.15 mm x 1.9 mm, serve as inductors between each device
emitter lead and PCB ground plane. This additional inductance provides “series-series” negative feedback,
improving amplifier 3rd-order intercept and gain compression points, at the cost of reduced gain. For this LNA
approximately 1.5 to 2 dB of gain is traded to attain improved linearity. In other words typical gain in a similar
applications board with direct emitter grounding was 16 to 16.5 dB, and this gain was reduced to 14.7 dB by the
use of inductive emitter degeneration.
2.1.1.2
Effect of Adding Additional Charge-Storage Across the Base-Emitter Junction
One of the spectral lines generated by device nonlinearity during a two-tone test is a low frequency product, (f1-f2).
For the case of a test tone spacing of 1 MHz, this 1 MHz product modulates the base-emitter and collector-emitter
voltages of the LNA at a 1 MHz rate. Recalling that for a bipolar transistor, emitter current is an exponential
function of the base-emitter voltage,
(2)
qV BE
Ie ≈ I es ⋅  ------------e  kT 
One can imagine this low frequency AC signal present at the device terminals varying the operating point of the
transistor at the rate (f2-f1), thus adversely impacting distortion product levels.
If a relatively large value capacitor is placed across the base-emitter junction in order to bypass this low frequency
product (f2-f1), the voltage fluctuation seen by the base-emitter junction of the transistor can be reduced, and the
levels of the 3rd order products minimized. Please refer to the schematic diagram in Figure 1. Capacitor C3,
0.1 µF, performs the low-frequency bypass function described here. An improvement in third-order intercept point
Application Note
8
Rev. 3.0, 2007-01-09
Application Note No. 060
Low Noise Amplifier Design Considerations
of approximately 5 to 10 dB can be expected by using this “trick”. The same effect may be seen by using extra
charge storage on the collector, but the results usually are not nearly as dramatic.
The closer together the two input test tones f1 and f2 are in frequency, the lower frequency the product or beat note
(f2-f1) is. Therefore as input test tones f1 and f2 come closer together, more capacitance is needed to achieve best
possible bypassing of the low frequency product (f2-f1). For a test tone separation of 1 MHz, 0.1 µF was found to
be more than adequate in this particular application circuit. A good physical or gut-level feeling for the efficacy of
this trick can be had by testing the LNA with and without C3 in place, and by experimenting with different test tone
spacing and capacitor values for C3.
For best results, the transistor should see a nice, low-impedance path at low frequencies between this additional
charge storage and its terminals; for this reason a coil rather than a high-value resistor is used to bring in the base
bias voltage and isolate the RF from the DC bias network. The 15 nH inductor used on the base to bring in the DC
bias has negligible impedance up to tens of megahertz, but provides enough impedance at 1.9 GHz to nearly
isolate the base of the transistor from the bias network within the LNA‘s normal operating frequency range. For
applications requiring operation over wide temperature ranges, the effect of temperature on capacitor dielectric
and capacitor performance should be investigated.
If is important to note that bypassing the (f2-f1) product as described here does not affect the compression level of
the amplifier - only the third-order intercept point, As a result, if this bypassing technique is used, the general rule
of thumb stating that there is approximately a 10 dB difference between third-order intercept and 1 dB
compression points is not longer valid. In this application, the Input IP3 is +10 dBm, and the Input P1dB is typically
-7.9 dBm, with a difference of 17.9 dB between these two points. A typical gain compression curve is shown in
Figure 4.
Gain Compression at 1960 MHz, BFP620
Application V4.2, S/N SG8, T=25C
16
Gain, dB
15
14
13
12
11
10
-6
-4
-2
0
2
4
6
8
Output Power, dBm
AN060_Gain_Compression.vsd
Figure 4
Typical Gain Compression Curve at 1960 MHz, T = 25 °C, V = 3.0 V for BFP620 Application
Version 4.2., S/N SG8
Application Note
9
Rev. 3.0, 2007-01-09
Application Note No. 060
Low Noise Amplifier Design Considerations
2.1.1.3
Test Tone Spacing vs. IIP3
The reader may wonder what happens to third-order product distortion levels at the two input test tone are “walked
apart” in frequency. A BFP620 V4.2 Applications Board was tested for Input IP3 over range of test tone spacing,
from 20 kHz to 100 MHz. In each case the “center frequency” (f1+f2) / 2 was 1960 MHz. Results are tabulated in
Table 3 and plotted in Figure 5.
As previously mentioned, one limiting factor at close tone spacing is the available charge storage (e.g. value of
C3), or “how good of a low frequency ground” one has across the base-emitter junction (and also, to a lesser
degree, at the collector). At wider tone spacing, the product (f2-f1) rises in frequency, and the impedance of the RF
chokes which isolate the DC bias circuit from the transistor begin to play a role - and this undesired impedance
begins to roll off or degrade IP3 performance. Of course the selectivity of the matching circuits at LNA input and
output also have some influence, particularly as the test tones are separated more and more widely and begin to
fall out of the normal operating frequency range of the amplifier.
Table 3
Input 3rd Order Intercept versus Test Tone Spacing1)
f1, MHz
f2, MHz
∆f
IIP3, dBm
1959.990
1960.010
20 kHz
+7.1
1959.975
1960.025
50 kHz
+8.1
1959.950
1960.050
100 kHz
+9.4
1959.900
1960.100
200 kHz
+9.6
1959.750
1960.250
500 kHz
+10.1
1959.500
1960.500
1 MHz
+10.1
1959.000
1961.000
2 MHz
+10.1
1957.500
1962.500
5 MHz
+10.0
1955.000
1965.000
10 MHz
+10.0
1950.000
1970.000
20 MHz
+9.8
1935.000
1985.000
50 MHz
+9.1
1910.000
2010.000
100 MHz
+9.0
1) For BFP620 V4.2 Application board, the “center” frequency - e.g. (f1+f2)/2 - is kept at 1960 MHz. Input power of each test
tone, f1 and f2, is -20 dBm each tone.
Application Note
10
Rev. 3.0, 2007-01-09
Application Note No. 060
Low Noise Amplifier Design Considerations
Input IP3 versus Test Tone Spacing. BFP620
Application V4.2. T=25C, V =3.0V
Input IP3, dBm
11
10
9
8
7
6
0.02 0.05 0.1
0.2
0.5
1
2
5
10
20
50
100
Tone Spacing, MHz
AN060_IIP3_versus_tone_spacing.vsd
Figure 5
Input IP3 versus test tone spacing, BFP620 Application V4.2. V = 3 V, Center Frequency =
1960 MHz, T = 25 °C
2.1.1.4
Linearity - To Sum Up
For this LNA application, two methods are used to improve third-order intercept performance: 1) inductive emitter
degeneration and 2) low-frequency bypassing of the (f2-f1) product at the base-emitter (and collector) junction.
Inductive emitter degeneration achieves linearity improvement at the cost of reduced gain. Other methods may be
employed to achieve further reduction of third-order distortion levels, but these two simple methods are less likely
to be influenced by temperature and LNA termination impedance than some of the more exotic techniques
reported in the literature [1]. Additional information on linearity topics can be found in Infineon Application Note
AN057 [4].
2.1.2
Stability Considerations
In general, for a linear two-port device characterized by S-parameters, the two necessary and sufficient conditions
to guarantee unconditional stability (e.g. no possibility of oscillation when input and output of the device are both
terminated in any passive real impedance) are
A) K > 1 and b) |∆| < 1 where
(3)
2
2
2
1 – s11 – s22 + ∆
K = ---------------------------------------------------------2 s12 ⋅ s21
(4)
∆ = s11 ⋅ s22 – s12 ⋅ s21
Application Note
11
Rev. 3.0, 2007-01-09
Application Note No. 060
Low Noise Amplifier Design Considerations
In the literature one may encounter an alternative form for these two conditions as
a) K > 1 and b) B1 > 0 where
(5)
2
2
B 1 = 1 + s11 – s22 – ∆
2
A simulation tool is of great help when trying to optimize an amplifier for stability. The Ansoft Serenade® design
environment was used for simulations in this LNA design, including evaluating the stability factors K and B1 up to
6 GHz.
2.1.2.1
Stability Below 6 GHz
At the time of this writing s-parameter were available for the BFP620 from 10 MHz to 6 GHz. Initial simulations
looked at stability factors K and B1. The use of inductive emitter degeneration for the purposes of linearity
improvement also had a major impact on the stability of the amplifier. In brief: a small amount of emitter
degeneration improves stability at lower frequencies, but as the amount of emitter inductance is increased, stability
at higher frequencies, e.g. in the range of 5 to 12 GHz - is compromised. In simulations, and later confirmed with
measurements, output resistive loading (see schematic, R1, 10 Ω) in conjunction with the additional emitter
inductance helped to bring K to a value greater than 1, particularly in the range of 800 - 1200 MHz. The RF bypass
capacitor C5, 10 pF, has less of an effect at lower frequencies as compared to at 2 GHz, hence the LNA output
“sees” this lossy 10 Ω resistor below band, improving stability in that region. Having some stability margin in this
range is especially important given that PCS band duplexers typically look very reflective in the 800 -1200 MHz
range.
2.1.2.2
Stability above 6 GHz
As stated previously, a small amount of additional emitter inductance can help stabilize the BFP620 at lower
frequencies, but larger amounts of added emitter inductance can jeopardize stability at higher frequencies - for
this particular amplifier, in the region above about 8 GHz. As the BFP620 is a very high gain device with an
extremely high transition frequency, the potential for high frequency instability exists, thus great care must be
exercised.
A large amount of inductive emitter degeneration was required in order to meet the linearity requirement target an input IP3 of +10 dBm with less than 10 mA of current. Due to this large amount of required emitter inductance,
it was found that the LNA could oscillate at 10 GHz if both LNA input and output were left open-circuited. The
10 GHz oscillation could be observed on a spectrum analyzer if the area near the LNA were “sniffed” with an RF
probe while amplifier input and output were left open-circuited. (Even a cable with a PC Board mount SMA
connector inserted into one end may be used as a makeshift RF probe, with the center pin of the SMA connector
used as an “antenna”. Be sure to DC block the spectrum analyzer input!)
Another observable symptom of the 10 GHz oscillation was a shift in DC operating current - as the input and output
of the LNA were taken from being terminated in 50 Ω to being open circuited, the DC current would increase by
0.5 to 1.5 mA.
In a real application (e.g. with a duplexer LNA input and a ceramic image-stripping filter at LNA output) the LNA
would not see a perfect open circuit at all frequencies, and a problem might not exist. Nevertheless, this situation
with marginal stability with input and output open-circuited is undesirable. As s-parameters were only available to
a 6 GHz, a careful analysis of K and B1 was not possible at 10 GHz. A more empirical approach was taken to
eliminate the potential 10 GHz oscillation.
Application Note
12
Rev. 3.0, 2007-01-09
Application Note No. 060
Low Noise Amplifier Design Considerations
If it were possible to kill the gain of the LNA above 8 GHz or so, while leaving PCS band gain largely unaffected,
it might be possible to eliminate the potential 10 GHz instability. To this end, a simple, crude, but effective fix was
employed.
Please refer to the schematic diagram Figure 1. Note the 0.5 pF capacitor C7. An approximate lumped-element
model for this capacitor is shown in Figure 6, including the short microstrip track leading to the capacitor’s selfinductance and equivalent series resistance (ESR). (In general, for best simulation accurary it is a good idea to try
and model parasitic reactance in circuit elements, e.g. modeling the parasitic capacitance / self-resonant
frequency of chip coils, or parasitic inductance and equivalent series resistance (“ESR”) of chip capacitors.)
1pF
C7
1pF
0.4nH
0.3 ohms
AN060_lumped_element_model.vs
Figure 6
Approximate lumped-element model for Capacitor C7 in 0402 package. (PCB track, chip selfinductance and Equivalent Series Resistance are accounted for)
The 0.5 pF 0402-size chip capacitor C7 has its in-circuit self-resonance at approximately 9 GHz, given by
(6)
1
fres = ------------------2π LC
Being a series resonance, C7 tends toward a short-circuit condition near 9 GHz. The main idea: LNA gain in the
vicinity of 8 to 11 GHz is reduced by “shorting-out” the LNA output in this frequency range with C7, hence killing
the potential oscillation. The effect on gain in the 1900 MHz PCS band was minimal, with a loss in gain of around
0.2 dB. Despite s-parameters being available only to 6 GHz, the simulation plot in Figure 7 gives a good
qualitative idea of how this element affects the amplifier with respect to gain above 5 or 6 GHz (the simulator
extrapolates s-parameters above 6 GHz.) In each case for the simulation, with C7 and without C7, the output is
re-matched for optimum amplifier gain and return loss. Note the reduction in gain above about 6 GHz when C7 is
added to the LNA output. This above-band gain reduction helped kill the potential 10 GHz oscillation.
Simulated and measured plots of the stability factors K and B1 for up to 6 GHz appear in Figure 8 and Figure 9
for a BFP620 Version 4.3 PC board. For Figure 9 measured s-parameter data taken from an actual BFP620 board
is imported into the circuit simulator, which then calculates and plots K and B1.
Application Note
13
Rev. 3.0, 2007-01-09
Application Note No. 060
Low Noise Amplifier Design Considerations
AN060_simulated_plot_LNA gain.vsd
Figure 7
Simulated plot of LNA gain showing effect of capacitor C7 on high frequency gain of LNA.
Note “suckout” or reduction in gain for 6 to 11 GHz region with C7 in the circuit. Gain
reduction in this range helped to kill potential 10 GHz oscillation caused by large amount of
emitter degeneration. (Sweep for this plot is from 100 MHz to 12 GHz.)
AN060_simulated_stability_factors_6GHz.vsd
Figure 8
Simulation plot of stability factors K and B1 up to 6 GHz. (Sweep is 100 MHz to 6 GHz). Note
Application Note
14
Rev. 3.0, 2007-01-09
Application Note No. 060
Low Noise Amplifier Design Considerations
K>1 and B1>0 for entire region. Minimum K value is 1.04 at 1850 MHz
Measured Stability Factors K and B1, BFP620
Application V4.2, S/N SG8, V=3.0V, T=25C
5
K, B1
4
3
2
1
6
5
4
3
2.6
2.2
1.9
1.7
1.5
1.3
1.1
0.9
0.7
0.5
0.3
0.2
0.1
0
Frequency, GHz
AN060_Measured_stability_factors_frequency.vsd
Figure 9
Measured stability factors K and B1 and B1 vs. frequency for a V4.3 BFP620 Applications
Board. Compare to simulated results in Figure 8. Measurement range is 100 MHz to 6 GHz
2.2
Noise Figure
The BFP620 is capable of excellent noise figure performance up to 6 GHz, and offers noise figure values
comparable to more expensive PHEMT and GaAs MESFET devices.
Near 2 GHz, the bias current for minimum noise figure is about 5 mA. However, the +10 dBm Input IP3
requirement could not be met with only 5 mA of current, and for this particular application a minimum current of
approximately 8 mA is required.
Table 4 gives noise parameters for the BFP620 at 1.8 GHz and a VCE of 2 V at 5, 8 and 10 mA collector current.
This should help the reader in understanding how bias current influenced noise figure in this design. If the
+10 dBM IIP3 requirement were dropped (e.g. as would be the case in a “low-desense requirement” LNA design),
less current (and less emitter degeneration) could be used - and noise figure values of just under 1.0 dB in an
application circuit should be possible.
Table 4
BFP620 Noise Parameters at 1.8 GHz for VCE = 2.0 Volts
IC
Fmin
dB
Γopt
mag
Γopt
ang
Rn/50
mA
5
0.65
0.16
71
0.14
8
0.71
0.09
97
0.08
10
0.76
0.06
111
0.11
Application Note
15
Rev. 3.0, 2007-01-09
Application Note No. 060
Low Noise Amplifier Design Considerations
The emitter degeneration used for linearity improvement also had the pleasant side-effect of moving the optimum
noise match and return loss match at bit closer together in terms of impedance. A plot of noise figure for a BFP620
V4.2 Application Board is given in Table 10.
AN060_NoiseFigure_vs_frequency.vsd
Figure 10
Plot of Noise Figure versus Frequency for BFP620 V4.2 Application, PC Board Serial Number
SG8. Span is 1900 to 2010 MHz. Temperature = 25 °C
2.3
Input / Output Match
As stated in the previous section, the use of added emitter inductance helped to move the optimum noise match
and optimum return-loss match closer together in terms of impedance.
The output matching circuit is a high-pass network consisting of L2 and C2. Due to the non-zero S12 of the device,
the output matching circuit was used to favorably influence the input return loss.
The influence of the output matching circuit on the input return loss together with the positive effect of the added
emitter inductance for both return loss and noise matching enabled the elimination of any RF matching elements
at the device input. Only a 10 pF DC blocking capacitor (C1) and a coil to bring in DC bias voltage to the base (L1)
were required at the input. This elimination of input tuning elements also permits optimizing the design for a
balance of input return loss and noise figure with fewer iterations. A good linear RF simulation tool is very helpful
in gaining insight into circuit behaviour and trends in this regard.
At the input, L1 presents a relatively high impedance at the 1900 MHz PCS band, and as such has minimal
influences on the input match of the circuit. This coil has a minimum self-resonant frequency of 2300 MHz, which
lies above the LNA’s nominal frequency range.
The input and output return loss for a typical BFP620 Version 4.3 Application at the low, middle and upper edge
of the PCS band is shown in Table 5.
Application Note
16
Rev. 3.0, 2007-01-09
Application Note No. 060
Low Noise Amplifier Design Considerations
Table 5
Input and Output Return Loss for BFP620 V4.3 Application, S/N MS6
Frequency
MHz
DB[s11]²
DB[s22]²
1930
12.2
19.5
1960
13.3
15.9
1990
14.9
13.7
2.4
DC Bias
In term of the chosen DC bias setting, a compromise of the various specification targets - gain, linearity, noise
figure and power consumption - had to be made. Higher current improves linearity and gain, but as shown in an
earlier section, noise figure increases with higher current levels. Linearity is also improved with increased
collector-emitter voltage (more voltage swing available at the amplifier output) but as one nears the collectoremitter breakdown voltage (VCEO) noise figure can begin to degrade due to the onset of voltage breakdown effects.
The BFP620’s worst case collector-emitter breakdown voltage (VCEO) is 2.3 V, but typically is higher. Balancing
and trading off the various target specification goals resulted in a nominal bias point of 2.1 V for the collectoremitter voltage and 8.9 mA for device current.
Referring to the schematic diagram in Figure 1, resistor R3 drops the supply voltage down from 3.0 volts to
approximately 2.1 V. R2 supplies bias to the base of the transistor. R1 is present for LNA stability improvement,
not for DC bias purposes.
This simple bias circuit offers a low to moderate amount of negative feedback to compensate for DC beta variation
from device to device and over temperature. If the device current increases due to temperature or device-to-device
variation in DC beta, the voltage drop across R3 increases, reducing the voltage seen by the base, thereby
providing negative feedback. The DC current shift over the -40 to +85 °C range for a Version 4.3 BFP620
application board is shown in Table 6, along with gain and input 3rd-order intercept variation over this same
temperature range. Note that the current varies by only 7.5% from cold to hot in spite of the limitations of the bias
circuit used on this application board. If a more stable bias scheme is required, an active bias circuit using either
a PNP transistor or the Infineon BCR400W active bias controller is recommended [2],[3].
Table 6
Bias Current, Gain and IIP31)
Variation, ∆%2)
Cold to Hot
Parameter
Temperature
-
-40 °C
(Cold)
+25 °C
(Room)
+85 °C
(Hot)
-
Bias Current, mA
8.8
9.3
9.5
7.5%↑
Gain, dB
15.1
14.7
14.1
6.8%↓
IIP3, dBm
+9.2
+10.2
+10.3
10.8%↑
1) Variation from -40 to +85 °C in a BFP620 V4.3 Applications Board, Serial Number MS11. Note that a simple resistive bias
scheme is used (Figure 1)
2) Variation is defined as the total swing in parameter value divided by the nominal (room temperature, 25 °C) value, with the
result subsequently multiplied by 100.
Application Note
17
Rev. 3.0, 2007-01-09
Application Note No. 060
Summary / Conclusions
3
Summary / Conclusions
Infineon’s Silicon-Germanium BFP620 bipolar transistor offers a very high performance, power-efficient and costeffective solution for high frequency low noise-amplifier (LNA) designs.
This application note describes a high-linearity LNA design for 1900 MHz PCS band applications. The BFP620 is
also a good solution for low-noise amplifiers operating in the European and Asian UMTS / WCDMA frequency
ranges. Evaluation boards for the LNA application depicted in this Application Note are available from Infineon
Technologies.
References
[1]
Vladimir Aparin, Charles Persico, “Effect of Out-Of-Band Terminations on Intermodulation Distortion in
Common-Emitter Circuits”.
[2]
“Application Considerations for the Integrated Bias Control Circuits BCR400R and BCR400W” Application
Note No. 014, Infineon Technologies, Silicon Discretes Group.
[3]
RF Design Guide, Systems, Circuits, and Equations. Peter Vizmuller, Artech House, 1995, ISBN 0-89006754-6. (Pages 76-78 show active biasing of an RF bipolar transistor.)
[4]
“A 1.9 GHz Low Noise Amplifier optimised for high IP3 using BFP540” Application Note No. 057, Infineon
Technologies, Silicon Discretes Group. (Use of a large value capacitor across the base-emitter junction of a
BFP540 transistor for IP3 improvement is discussed).
Application Note
18
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
4
Appendices
4.1
Appendix A. Scanned Images of BFP620 Application Board
Figure 11
View of Version 4.3 Printed Circuit Board
Application Note
19
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
Figure 12
Close-in shot of BFP620 V4.3 PCB showing parts placement. Microstrip lines used for
inductive emitter degeneration are visible between each of the two emitter leads and ground
vias
Application Note
20
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
4.2
Appendix B - Data Plots for a BFP620 Version 4.3 Application Board, Serial
Number MS5
CH
1
S1
1
log
MAG
5
dB/
28 Feb 2000
09:45:59 2_:-12.335
dB
1 960.000 000
MHz
REF 0
dB
PR
m
Co
r
De
Av
l
g
4
1_:-10.775
dB 1.93
GHz
3_:-14.144
dB 1.99
GHz
2
1
3
CENTER 1 960.000 000
MHz
SPAN 250.000 000
MHz
AN060_Input_Return_Loss.vsd
Figure 13
Input Return Loss, Log Mag
Application Note
21
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
CH
1
S1
1
1U
FS
28 Feb 2000
09:46:12
2_:
30.981
5.7988
PR
m
Co
r
De
l
Av
g
4
470.87
pH
1 960.000 000
MHz
1_:
28.214
1.93
6.2451
GHz
3_:
34.162
1.99
4.7402
GHz
2
1
3
CENTER 1 960.000 000
MHz
SPAN 250.000 000
MHz
AN060_Input_Return_Loss_SC.vsd
Figure 14
Input Return Loss, Smith Chart
Attention: The reference plane is normalized to the PC board RF Input Connector.
Application Note
22
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
CH1
S 21
log MAG
1 dB/
REF 15 dB
28 Feb 2000 09:46:45
2_: 14.786 dB
1 960.000 000 MHz
PRm
Cor
Del
Avg
4
1_: 14.966 dB
1.93 GHz
REFERENCE LINE POSITION
5 Div
3_: 14.64 dB
1.99 GHz
2
1
3
CENTER 1 960.000 000 MHz
SPAN 250.000 000 MHz
A N060_Forward_Gain.vsd
Figure 15
Forward Gain
Application Note
23
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
CH1
S 21
log MAG
1
10 dB/
REF 0 dB
28 Feb 2000 09:48:51
1_: 14.777 dB
1 960.000 000 MHz
PRm
Cor
Del
Smo
START
.030 000 MHz
STOP 6 000.000 000 MHz
AN060_Forward_Gain_Wide_span.vs
Figure 16
Forward Gain, Wide span (30 kHz - 6 GHz)
Application Note
24
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
CH1
S 12
log MAG
5 dB/
REF 0 dB
28 Feb 2000 09:47:12
2_:-19.597 dB
1 960.000 000 MHz
PRm
Cor
Del
Avg
1_:-19.729 dB
1.93 GHz
REFERENCE LINE POSITION
7 Div
3_:-19.621 dB
1.99 GHz
4
2
3
1
CENTER 1 960.000 000 MHz
SPAN 250.000 000 MHz
AN060_Reverse_Isolation.vsd
Figure 17
Reverse Isolation
Application Note
25
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
CH1
S 22
log MAG
5 dB/
REF 0 dB
28 Feb 2000 09:47:32
2_:-14.357 dB
1 960.000 000 MHz
PRm
1_:-15.956 dB
1.93 GHz
Cor
3_:-12.634 dB
1.99 GHz
Del
Avg
4
2
3
1
CENTER 1 960.000 000 MHz
SPAN 250.000 000 MHz
AN060_Output_Return_Loss.vsd
Figure 18
Output Return Loss, Log Mag
Application Note
26
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
CH1
S 22
1 U FS
28 Feb 2000 09:47:43
-10.381
7.8222 pF
2_: 36.803
1 960.000 000 MHz
PRm
1_: 37.518
-6.2324
1.93 GHz
Cor
3_: 34.797
-13.215
1.99 GHz
Del
Avg
4
2
1
3
CENTER 1 960.000 000 MHz
SPAN 250.000 000 MHz
AN060_Output_Return_Loss_LM.vsd
Figure 19
Output Return Loss, Smith Chart
Attention: The reference plane is normalized to the PC board RF Output Connector.
Application Note
27
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
AN060_Output_Response_LNA.vsd
Figure 20
Output Response of LNA to Two-Tone IP3 test. Input tones are f1 = 1959.5 MHz,
f2 = 1960.5 MHz, tone power = -20 dBm each tone. Input IP3 = -20+(60.98/2) = +10.5 dBm.
PC Board = V4.3, serial Number MS5
Application Note
28
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
4.3
Appendix C. Data on 10 BFP620 Version 4.3 Printed Circuit Boards, taken
randomly from a lot of 20 units. All Data taken a room temperature (25 °C)
Table 7
Board
S/N
dB[s11]²
dB[s21]²
dB[s12]²
dB[s22]²
1960
MHz
1960
MHz
1990
MHz
1930
MHz
1960
MHz
1990
MHz
1930
MHz
1960
MHz
1990
MHz
1930
MHz
1960
MHz
1990
MHz
MS1
10.6
11.9
13.9
14.7
14.6
14.5
19.5
19.4
19.2
13.9
13.0
11.9
MS2
10.7
11.7
12.9
14.9
14.8
14.6
19.6
19.1
19.1
21.6
17.6
15.1
MS3
12.2
13.3
14.9
14.9
14.7
14.5
19.2
19.2
19.1
19.5
15.9
13.7
MS4
12.8
14.3
15.5
14.9
14.7
14.5
19.6
19.5
19.4
15.7
13.5
11.9
MS5
12.3
14.0
15.6
14.9
14.7
14.5
19.6
19.5
19.5
15.2
13.4
11.8
MS6
11.3
12.7
14.4
14.8
14.7
14.5
19.5
19.3
19.3
15.9
14.6
13.1
MS7
11.7
13.1
14.6
14.8
14.7
14.5
19.4
19.3
19.2
17.2
15.3
13.5
MS8
12.5
14.0
15.5
14.8
14.7
14.5
19.3
19.3
19.3
15.4
13.6
12.2
MS9
12.4
13.7
14.9
14.9
14.8
14.6
19.5
19.3
19.3
17.8
15.0
13.3
MS10
12.6
14.0
15.2
14.9
14.7
14.5
19.5
19.5
19.4
15.7
13.6
12.1
Table 8
Board
S/N
IIP3
Noise Figure
DB
Output
P1dB
DBm
VCE
I
V
MA
DBm
1930
MHz
1960
MHz
1990
MHz
1930
MHz
1960
MHz
1990
MHz
1960
MHz
MS1
1.04
1.06
1.07
+9.7
+9.8
+10.2
+6.5
2.09
8.8
MS2
1.08
1.08
1.09
+10.2
+10.3
+10.6
+5.0
2.19
8.1
MS3
1.04
1.04
1.04
+10.5
+10.7
+10.9
+5.4
2.12
8.7
MS4
1.05
1.03
1.05
+10.5
+10.4
+10.7
+5.8
2.06
9.4
MS5
1.00
1.01
1.01
+10.1
+10.1
+10.6
+6.4
2.08
9.3
MS6
0.96
0.99
0.96
+10.0
+10.2
+10.4
+6.3
2.09
9.0
MS7
1.00
1.01
1.03
+10.1
+10.4
+10.6
+6.1
2.11
8.8
MS8
1.03
1.05
1.05
+10.2
+10.3
+10.5
+6.1
2.09
9.1
MS9
1.04
1.05
1.03
+10.3
+10.5
+10.9
+4.8
2.17
8.3
MS10
1.04
1.05
1.05
+10.2
+10.4
+10.7
+5.4
2.11
9.0
“Statistics” for n = 10 BFP620 V4.3 PC Boards at the middle of band (1960 MHz). Some parameters follow directly
from current variation (e.g. Output P1db). Note that Gain, Noise Figure and Input 3rd Order Intercept have
reasonably “tight” distributions. Standard Deviation used is population standard deviation (σn), not sample
standard deviation (σn-1).
Application Note
29
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
Input Return Loss
•
•
•
Mean value “x” = 13.3 dB
Standard deviation “σn” = 0.87
“Coefficient of variation “V”1) = 6.53
Output Return Loss
•
•
•
Mean value = 14.6 dB
Standard deviation = 1.36
Coefficient of variation = 9.32
1) Coefficient of Variation “V” is defined as V = (σn * 100) / x
Gain
•
•
•
Mean value = 14.7 dB
Standard deviation = 0.05
Coefficient of variation = 0.37
Current Consumption
•
•
•
Mean value = 8.9 dB
Standard deviation = 0.39
Coefficient of variation = 4.36
Noise Figure
•
•
•
Mean value = 1.04 dB
Standard deviation = 0.026
Coefficient of variation = 2.47
Input 3rd Order Intercept
•
•
•
Mean value = +10.3 dB
Standard deviation = 0.23
Coefficient of variation = 2.23
Output 1 dB Compression Point (P1dBout)
•
•
•
Mean value = +5.8 dB
Standard deviation = 0.57
Coefficient of variation = 9.81
Application Note
30
Rev. 3.0, 2007-01-09
Application Note No. 060
Appendices
4.4
Appendix D. Description and Photo of BFP620 AN060 Application Kit
The BFP620 AN060 Application Kit contains the following items:
1.
2.
3.
4.
5.
Populated and tested BFP620 Version 4.3 Applications Board in anti-static bag
DC wire harness
BFP620 Datasheet
Copy of Application Note AN060
Test Data Sheet (1 page) with test results for the particular PC board in kit
The entire kit is packaged in a cardboard mailer box, the PC Board contained in the anti-static bag is sandwiched
between two pieces of anti-static foam.
AN060_Application_kit.vsd
Figure 21
Photo of contents of BFT620 AN060 AN060 Application Kit
Application Note
31
Rev. 3.0, 2007-01-09