IEEE MTT-V042-I04 (1994-04)

224 Pages·1994·20.167 MB·English

by  FANTOMASPING

IEEE T R A N S A C T I 0 N S O N

MICROWAVE THEORY
AND TECHNIQUES
A PUBLICATION OF THE IEEE MICROWAVE THEORY AND TECHNIQUES SOCIETY

APRIL 1994 VOLUME 42 NUMBER 4 IETMAB (ISSN 0018-9480)

[email protected]

PAPERS
CPW-fed active slot antennas - B.K. Kormanyos ; W. Harokopus ; L.P.B. Katehi ; G.M. Rebeiz 541 - 545
A high gain silicon AGC amplifier with a 3 dB bandwidth of 4 GHz - L.C.N. de Vreede ; A.C. Dambrine ; J.L. Tauritz ; R.G.F. Baets 546 - 552
Wideband dispersion measurements of water in reflection and transmission - D. Kralj ; L. Carin 553 - 557
Investigation of Telstar 4 spacecraft Ku-band and C-band antenna components for multipactor breakdown –
N. Rozario ; H.F. Lenzing ; K.F. Reardon ; M.S. Zarro ; C.G. Baran 558 - 564
Phased array operation of a diode grid impedance surface - L.B. Sjogren ; Hong-Xia Liu ; Xiaohui Qin ; C.W. Domier ; N.C. Luhmann 565 - 572
Temperature distribution in cylinder symmetric MM-wave devices - J.-F. Luy ; J. Schmidl 573 - 579
Calibration and normalization of time domain network analyzer measurements - T. Dhaene ; L. Martens ; D. De Zutter 580 - 589
Modeling multiport using a three-dimensional coupled analytical/finite element method application to microwave characterization of material -
D. Aregba ; J. Gay ; G. Maze-Merceur 590 - 594
Efficient computation of SAR distributions from interstitial microwave antenna arrays - K.L. Clibbon ; A. McCowen 595 - 600
TE-mode scattering from two junctions in H-plane waveguide - J.W. Lee ; H.J. Eom 601 - 606
Sensitivity analysis of lossy coupled transmission lines with nonlinear terminations - S. Lum ; M. Nakhla ; Qi-Jun Zhang 607 - 615
Analysis of twin ferrite toroidal phase shifter in grooved waveguide -
Wen Junding ; Yong-Zhong Xiong ; Mei-Juan Shi ; Guo-Fong Chen ; Ming-De Yu 616 - 621
Rigorous multimode network representation of capacitive steps - M. Guglielmi ; G. Gheri 622 - 628
New biorthogonality relations for inhomogeneous biisotropic planar waveguides - A.L. Topa ; C.R. Paiva ; A.M. Barbosa 629 - 634
A numerically efficient technique for the method of moments solution for planar periodic structures in layered media –
R.A. Kipp ; C.H. Chan 635 - 643
The origin of spurious modes in numerical solutions of electromagnetic field eigenvalue problems - W. Schroeder ; I. Wolff 644 - 653
Application of modified indirect boundary element method to electromagnetic field problems - Bin Song ; Junmei Fu 654 - 660
Applying the Exodus method to solve Poisson's equation - M.N.O. Sadiku ; S.O. Ajose ; Zhibao Fu 661 - 666
Digital signal processing of time domain field simulation results using the system identification method - W. Kumpel ; I. Wolff 667 - 671
A combined efficient approach for analysis of nonradiative dielectric (NRD) waveguide components - Ke Wu 672 - 677
Proposed expansions for the capacitance of a square centered in a circle - H.J. Riblet 678 - 680
Analysis of dominant and higher order modes for transmission lines using parallel cylinders - B.N. Das ; O.J. Vargheese 681 - 683
Partial inverse scattering method for three-dimensional heterogeneous biological bodies by using a matrix perturbation theory –
T.J. Cui ; C.H. Liang 683 - 686

( Continued on back cover)
Application of volume discretization methods to oblique scattering from high-contrast penetrable cylinders - A.F. Peterson 686 - 689
The traveling wave matching technique for cascadable MMIC amplifiers - B.J. Minnis 690 - 692
Further comments on "an analytic algorithm for unbalanced stripline impedance" - E. Costamagna ; A. Fanni 693 - 694
Two-junction tuning circuits for submillimeter SIS mixers - J. Zmuidzinas ; H.G. LeDuc ; J.A. Stern ; S.R. Cypher 698 - 706
Terahertz Shapiro steps in high temperature SNS Josephson junctions - P.A. Rosenthal ; E.N. Grossman 707 - 714
Photon induced noise in the SIS detector - N.B. Dubash ; G. Pance ; M.J. Wengler 715 - 725
An integrated superconducting sub-mm wave receiver for linewidth measurements of Josephson flux-flow oscillators -
Y.M. Zhang ; D. Winkler 726 - 733
Two-dimensional quasi-optical power-combining arrays using strongly coupled oscillators - J. Lin ; T. Itoh 734 - 741
The fabrication and performance of planar doped barrier diodes as 200 GHz subharmonically pumped mixers -
Trong-Huang Lee ; J.R. East ; Chen-Yu Chi ; G.M. Rebeiz ; R.J. Dengler ; I. Mehdi ; P.H. Siegel ; G.I. Haddad 742 - 749
Broadband quasi-optical SIS mixers with large area junctions - G. Pance ; M.J. Wengler 750 - 752
A technique for noise measurements of SIS receivers - Qing Ke ; M.J. Feldman 752 - 755
Experimental performance of a back-to-back barrier-N-N/sup +/ varactor tripler at 200 GHz -
D. Choudhury ; A.V. Raisanen ; R.P. Smith ; M.A. Frerking ; S.C. Martin ; J.K. Liu 755 - 758
Large area bolometers for THz power measurements - C.C. Ling ; J.C. Landry ; H. Davee ; G. Chin ; G.M. Rebeiz 758 - 760

(end)
EEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 42, NO. 4, APRIL 1994 54 1
CPW-Fed Active Slot Antennas
Brian K. Kormanyos, Student Member, IEEE, William Harokopus, Jr., Member, IEEE,
Linda P. B. Katehi, Senior Member, IEEE, and Gabriel M. Rebeiz, Senior Member, IEEE
Abstract-We have combined microwave oscillator design with
UT
theoretical characterization of planar antennas to build active
slot-oscillators.T he design is uniplanar, does not require via holes
and is compatible with monolithic transistor technology. The
coplanar waveguide (CPW) fed antenna impedance is calculated
using the space domain integral equation technique (SDIE), a full
wave method of moments approach. Slot-oscillators were built at
7 GHz and 20 GHz and the predicted oscillation frequencies agree
well with experiments. The 20 GHz medium power oscillator has
an output power of 17 mW and a DC to RF efficiency of 14%.
The design is easily scaled to millimeter-wave frequencies and
can be extended to power combining arrays.
I. INTRODUCTION
M
ILLIMETER-WAVE systems are becoming increas-
ingly important in many military and commercial appli-
II.
cations. Millimeter-wave receivers and transmitters have been
Fig. 1. Coplanar-waveguide (CPW) fed slot geometry.
traditionally waveguide-based systems that are expensive to
build at 60 to 200 GHz [l]. To solve this problem, several
groups have researched quasi-optical power combining topolo-
The small dimensions of the circuit allow the design of a power
gies and active antennas [2]-[6]. In this paper, we present a
combining array without triggering grating lobes.
novel active transmitter suitable for low-cost millimeter-wave
The CPW-fed oscillator (with a single element) is placed
applications. The transmitter consists of a coplanar-waveguide
on a dielectric lens. The dielectric lens synthesizes an infinite
(CPW)-fed slot antenna (or a dual-slot antenna) and a three-
dielectric substrate and therefore eliminates the excitation of
terminal device (millimeter-wave HEMT). CPW transmission
substrate modes and the associated power loss in these modes
lines have lower radiation loss and less dispersion than mi-
[7]. A slot antenna also radiates preferentially into the substrate
crostrip lines. Furthermore, the characteristic impedance and
with a ratio of over the power radiated to the air side. The
phase velocity of CPW are less dependent on the substrate
slot antenna therefore should radiate only 2% of its power to
height and more dependent on the dimensions in the plane
of the conducting surface. Also, the CPW-feed and antenna the air side when placed on a silicon substrate lens (E, = 12)
making the pattem unidirectional. The theoretical technique
are on the same side of the substrate thereby facilitating
used in this work for characterizing the input impedance of
the connection of shunt lumped elements and active devices
the slot antenna is the space domain integral equation (SDIE)
and eliminating the need for via holes. The slot-oscillator is
approach [8]-[ 121. The method has shown excellent versatility
therefore compatible with planar HEMT fabrication processes
in the study of a wide range of planar elements, and its
and can be easily scaled to higher frequencies.
accuracy will be demonstrated by comparison to measurements
The novelty of the CPW-fed oscillator is that we use the
for CPW fed slot-antennas fabricated on dielectric halfspaces.
antenna impedance, calculated by a full-wave analysis method,
as a parameter in the design of the oscillator. This results
11. CPW-FED SLOT-ANTENNIAM PEDANCE
in a more compact circuit than an approach consisting of an
oscillator with a 50 R output that is connected to a 50 R A CPW fed slot-antenna is shown in Fig. 1 and the substrate
antenna. In our design, the matching network is eliminated (or it is placed on is represented in Fig. 2. The substrate is lossless
minimized), and the circuit is much smaller than a wavelength. with infinite extent and the conducting surfaces have zero
ohmic losses. The two CPW apertures have width W and are
Manuscript received October 13, 1992; revised June 1, 1993. This work is separated by spacing S. The slot antenna has overall length
supported in part by the AF/Rome-Air Development Center and by the NASA ZA and width WA.
Center for Space Terahertz Technology at the University of Michigan.
With the use of the equivalence principle, all of the CPW
B. K. Kormanyos, L. P. B. Katehi, and G. M. Rebeiz are with NASNCenter
for Space Terahertz Technology, Electrical Engineering and Computer Science and slot apertures are covered by perfect electric conduc-
Department, University of Michigan, Ann Arbor, MI 48 109-2122. tors, and the electric field in the apertures is represented by
W. Harokopus, Jr. is with Texas Instruments, Defense Systems and Elec-
equivalent magnetic currents flowing on both sides of these
tronics Group, McKinney, Texas 75070.
IEEE Log Number 9216049. conductors with the same magnitude but opposite direction [8].
0018-9480/94$04.00 0 1994 IEEE
~
542 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 42, NO. 4, APIUL 1994
Fig. 2. Half spaces above and below ground plane (regions a, and b).
-100
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0
Frequency (GHz)
The presence of the conductors splits the original problem into
two simpler which deal with the radiation of the equiv- Fig. 3. Theoretical calculation of slot impedances as a function of frequency
for IVA/ZA = 0.02 and W’A/~A= 0.04 (w = 0.5 mm, s = 1 mm,
alent magnetic sources into a grounded dielectric half-space wA= 0.5 or 1 m, iA = 27 m).
and an air half-space respectively. In these two problems, the
magnetic fields can be written in an integral form where the
A theoretical calculation of slot-antenna impedance on an
kernel is the product of the magnetic-field Green’s function
E, = 12 half-space as a function of frequency is presented in
and the unknown equivalent magnetic current densities. This
Fig. 3 for two WA/~raAtio s. The impedance moves through
Green’s function is formulated in the form of semi-infinite
a wide range from a near short circuit at low frequencies to a
Sommerfeld integrals which, due to the uniform nature of the
high impedance near 160 R at the first resonance to a low re-
dielectric and air half spaces, do not exhibit pole singularities.
sistance value with negligible reactance at a second wideband
Extensive treatment of these integrals may be found in the
resonance. With appropriate frequency scaling, impedances
literature [9]-[ 101.
throughout the range are available for oscillator design.
The choice of the equivalent magnetic sources ensures con-
The accuracy of the theory has been verified by build-
tinuity of the electric field through the slot aperture. However,
ing microwave models of slot-antennas and measuring their
the continuity of the magnetic field is not satisfied unless it
impedance with a vector network analyzer. A slot antenna in
is enforced, resulting in a Helmholtz integral equation of the
free space and a slot antenna on a stycastl’ substrate with
second type with the equivalent magnetic current densities on
the slot-antenna and CPW apertures as the unknown quantities. t, = 12, have been measured. The infinite extent of the
substrate is simulated by using a large block of material with
The integral equation is then solved with the method of
absorber on the sides not covered by the ground plane. The
moments using a subsectional basis [12]. In this approach,
CPW-fed slot-antennas are attached to a coaxial line using a
the unknown magnetic current densities are expanded in sums
broadband coax-to-CPW transition which is normalized out of
of two-dimensional subcurrent elements each one represented
the measurements. The transition equalizes the ground planes
as a product of a vector unknown coefficient multiplied by a
at the start of the CPW feed reducing higher order modes on
roof-top function.
the line. The feed lines are short, about Xeff/4 to Xeff/2
The use of the approximate expression for the magnetic cur-
and we have not seen any benefit from additional ground
rent densities in the integral equation introduces a numerical
equalization using air bridges along the line. The geometry of
error which is minimized using Galerkin’s procedure. In this
the microwave models is shown in Fig. 4 with the theoretical
manner, the original integral equation is reduced to a linear set
and experimental impedances. The reference plane is set at the
of independent equations which is then solved using standard
input of the slot. Agreement between the measurements and the
techniques. The solution of this matrix equation provides the
theory is generally good in both cases. The variation is due to
equivalent magnetic current densities and consequently the
difficulties encountered in building the infinite substrates and
electric field on the slot and the CPW feed line. Following
ground planes and the ideal excitation of the CPW which the
the numerical solution of the integral equation, transmission
theory assumes.
line theory is utilized to find the input impedance of the slot
antenna.
As has been discussed in the literature [SI-[12], the numer- 111. OSCILLATOR DESIGNAN D MEASUREMENTS
ical results derived through the method of moments solution The oscillator design is based on the S-parameters of the
exhibit the best possible stability when the discretization is in microwave transistor using the reflection amplifier approach
the range of 30-80 subsections per guide wavelength. Other [13]. The devices are in chip form and the manufacturer
issues affecting the stability of the solution in terms of the supplied S-parameters include the effects of the bond wires
evaluation of Sommerfeld integrals have been well discussed used to connect the device to the CPW circuit. An indefinite
in the literature and will not be repeated here. However, it is scattering matrix is employed so that short circuited lengths
worth mentioning that subject to the assumptions adopted in of transmission-line may be placed at the gate and source
this solution the expected numerical error is on the order of ’
Stycast is a trademark of Emerson & Cuming Inc., 869 Washington Street,
1-5%. Canton, MA 02021.
KORMANYOS et al.: CPW-FED ACTIVE SLOT-ANTENNAS 543
-
Slot
antenna
(131-113 R)
antenna
(6.2 - 17.4 R)
(b)
Fig. 5. Equivalent circuits of the 20 GHz CPW fed slot oscillators. (a) First
design, (b) second higher power design.
plane near the slot antenna from the region of the ground
plane near the source terminals with capacitively bypassed
slits in the metallization. In our experience the slot-oscillator
does not suffer from oscillation problems at lower frequencies
because the transistor is embedded in a ground plane, and the
equivalent impedances seen by the source, gate and drain ports
at low-frequencies are zero (short-circuited) due to the short-
lengths of CPW line used. The design therefore eliminates the
need of additional resonant structures or RF chokes and results
in a compact circuit for power combining applications.
Two slot-oscillators were designed and built at 20 GHz
using commercially available hetero-junction FET' s. The first
design is based on the NEC NE32100, a low noise small signal
device with a gate length of 0.3 pm and a unity current gain
(b) frequency of 56 GHz. To obtain more output power, a second
Fig. 4. Theoretical and experimental slot impedances on (a) tr = 1 and design was based on the Fujitsu FLR016XV, a K-band power
(b) t, = 12. transistor. Equivalent circuits of the active slot oscillators are
shown in Fig. 5. The circuit layouts are shown in Fig. 6 with
terminals. The source connections (two of them) are DC ground equalizing bond wire air bridges installed around the
short-circuited to the ground plane and a metal-insulator-metal gate, drain, and the two source terminals to insure that the odd
capacitor is integrated at the gate end for applying the gate bias mode is excited on the CPW lines.
voltage. Computer optimization is applied to the lengths of The circuits oscillated when placed at the focus of an elliptic
the CPW transmission-line at the source and gate to maximize silicon substrate lens with a diameter of 2.6 cm. The measured
the reflection coefficient at the drain of the device. In this operating frequencies are 5-10% less than the small signal
way a reflection magnitude greater than one is obtained at design frequencies due to changes in the S-parameters as
the design frequency without the use of an external feedback large signal conditions are reached. The elliptical lens acts
network and its associated complications. A slot antenna is as an infinite dielectric medium and collimates the radiation
designed so that its reflection coefficient through a length of pattern from the slot antenna into a pattern that is diffraction
transmission line has a phase angle opposite in sign to that limited by the diameter of the lens. The measured radiation
of the reflection coefficient at the drain of the device. The patterns for a 20 GHz oscillator are shown in Fig. 7. The
drain bias is applied by DC isolating the region of the ground radiation patterns for both 20 GHz oscillators are very similar.
544 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 42, NO. 4. APRIL 1994
Capadtlve
overlay
-30-60 -40 -20 0 20 40 BO
ground plane Angle (degrees)
for draln bias
Fig. 7. Measured radiation patterns of a 20 GHz slot-oscillator on a 2.6 cm
(a)
silicon substrate lens.
r
ffoorr ddrrnniinn bbiiaass MSV34-60-E28 s? G
I
223 pm Capacitive
overlay
(b)
Fig. 6. CPW circuit layouts for the 20 GHz oscillators showing the capacitive
bypassing to allow application of DC bias. (a) First design, (b) second higher
power design.
The slot antenna provides a uniform E-plane feed pattern that
is transformed by the elliptical lens to a narrow diffraction-
-
limited pattern with high sidelobes. On the other hand the final
Dlmenrlon, mm
H-plane pattern is wider and with very low sidelobes due to antenna
(b)
the tapered (nearly sin2@ H-plane pattern of the slot antenna.
it is possible to result in a symmetrical E and H-plane pattern Fig. 8. (a) Equivalent circuit and (b) CPW layout for the 7 GHz VCO.
with the use of a double slot-antenna design [14]. The pattern
directivity (explained above). The total measured radiated
directivity is estimated from a geometric mean of the front
power is 3 mW at 22.45 GHz for the first design and 17 mW
side E-plane, H-plane, and 45-degree plane patterns with an
at 20.07 GHz for the design based on the power transistor.
appropriate reduction due to the power radiated to the back
side. Measured back side patterns indicate that about 10% of The DC to RF efficiencies are 3.8% and 14% respectively.
the power is going into the air behind the lens. This differs These power measurements are accurate to about f5% with
from the expected 2% back side power loss given by the most of the error due to uncertainty in the estimation of the
rule since there is no matching layer on the lens front surface pattern directivity. These numbers are roughly consistent with
to eliminate multiple reflections. the gain compression data supplied by the manufacturers of
The total oscillator power is calculated from the absolute the transistors.
power received by a standard gain horn at the pattern peak, A 7 GHz VCO (Fig. 8) was also designed using the above
the Friis transmission equation, and the measured pattern method with the incorporation of varactor diodes (Metelics
KORMANYOS et al.: CPW-FED ACTIVE SLOT-ANTENNAS 545
[6] K. Chang, K. A. Hummer, and J. L. Klein, “Experiments on injection
locking of active antenna elements for active phased arrays and spatial
-il__’i power combiners,” IEEE Trans. Microwave Theory Tech., vol. 37, July
1989.
[7] D. B. Rutledge, D. P. Neikirk, and D. P. Kasilingam, in K. J. Button,
Ed. Integrated circuit antennas, in Infrared and Millimeter- Waves, vol.
10. New York: Academic Press, 1983, pp. 1-90.
-20 [8] N. I. Dib, P. B. Katehi, G. E. Ponchak, and R. N. Simons, “Modeling of
shielded CPW discontinuities using the space domain integral equation
--29506 .7 6.8 6.9 7.0 7.1 7.2 7.3 7.4
method (SDIE),” J. Electromagnet. Waves Appl., to appear.
[9] N. I. Dib, W. P. Harokopus, Jr., G. E. Ponchak, and L. P. B. Katehi,
“A comparative study between shielded and open coplanar waveguide
Frequency (CHz) discontinuities,” Int. J. Microwave Millimeter-Wave Computer-Aided
Eng., vol. 2, no. 4, pp. 331-341, 1992.
Fig. 9. Relative power as a function of frequency for the 7 GHz VCO with
[lo] P. B. Katehi, “A space domain integral equation approach in the analysis
less than 2 dB power variation over a 400 MHz range.
of dielectric-covered slots,’’ Radio Sci., vol. 24, Apr. 1989.
[11] P. B. Katehi and N. G. Alexopoulis, “Real axis integration of Som-
merfeld integrals with applications to printed circuit antennas,” J. Math.
MSV34-60-E28) as the source terminals of the ET. The FET Phys., vol. 24, Mar. 1983.
is an NE32100 in a plastic package (NE32184). The oscillator [12] R. F. Harrington, Field Computation by Moment Methods. New York:
Macmillan, 1968.
was placed on a large stycast block (E, = 12) and no pattern [13] J. W. Boyles, “The oscillator as a reflection amplifier: an intuitive
measurements were made. The output power was sampled approach to oscillator design,” Micmwave J., June 1986.
with a standard gain horn pointed at the back (air side) of the [I41 G. Gauthier, T. P. Budka, W. Y. Ali-Ahmad, and G. M. Rebeiz, “A
low noise 86-90 GHz uniplanar Schottky-receiver,” ZEEE MTT-S Int.
antenna. This provides a measure of the relative output power
Microwave Symp., session OF-1-29, Atlanta, GA, June 14-18, 1993.
of the oscillator as its frequency is electronically varied. There
is less than 2 dB variation in output power over a 400 MHz
range from 6.85 GHz to 7.25 GHz (Fig. 9). This shows that
electronically tunable slot-oscillators are possible for phase Brian K. Kormanyos (S’93) was born in Ann Arbor, MI, in 1967. He received
the B.S. in electrical engineering from the University of Washington, Seattle
locked loops or other applications.
and the Ph.D. degree from the University of Michigan, Ann Arbor, in applied
electromagnetics and solid-state devices, in 1989 and 1994, respectively.
His research interests are in microwave and millimeter-wave circuits and in
IV. CONCLUSION high-frequency (VHF, UHF, Cellular Communication) analog circuit design.
We have successfully demonstrated a uniplanar medium-
power (17 mW) quasi-optical oscillator with high DC to
RF efficiency (14%) at 20 GHz. By using the slot-antenna
William Harokopus, Jr. (S’87-M’91) received the B.S.E.E., M.S.E.E.,a nd
impedance as a parameter in the design we obtain a very Ph.D. degrees from the University of Michigan, Ann Arbor in 1985, 1986,
compact circuit. No via holes are required, and the circuit is and 1991, respectively.
compatible with monolithic transistor technology. The design From 1987 to 1991, he worked as a Research Assistant in the Radiation
Lab at the University of Michigan. His research consisted of the development
can be easily extended to a double slot-antenna oscillator to of numerical techniques to study the behavior of microstrip and coplanar
yield a symmetric far field pattern. Furthermore, small 2 x 2 or waveguide circuits and antennas. In 1991, he joined the Advanced Technology
3 x 3 power combining arrays can be integrated on a silicon and Components Division of Texas Instruments. He is currently working as
an antenna engineer in the AntennaNon-Metallics Departement, McKinney,
substrate lens to result in a high power microwave source. Texas. Since joining TI, he has been involved in the design and modeling of
The silicon substrate lens should be much larger than the electrically scanned arrays and radomes.
array so that it acts as an infinite dielectric medium for the Dr. Harokopus currently serves as vice-chairman of the Dallas Antenna and
Propagation Society, and is a member of Sigma Xi.
edge antennas. The oscillator element can be used as a self
mixing receivedtransmitter in an inexpensive doppler system.
VCO designs are also possible to provide electronic tuning
for frequency control and use in phase locked systems or FM Linda P. B. Katehi, (S’81-M’84-SM‘89) for a biography, see page 83 of
radar applications. the January issue of this TRANSACTIONS.
REFERENCES
[l] P. H. Siegel, et al. “A 200 GHz planar diode subharmonically pumped Gabriel M. Rebeiz, (S’86-M’88-SM’93)was born in December 1964 in
waveguide mixer with state-of-the-art performance,” IEEE MTT-S Int. Beirut, Lebanon. He graduated in 1982 from the American University in
Microwave Symp. Dig., vol. 2, Albuquerque, NM, June 1-5, 1992, pp. Beirut with a B.E. (Honors) in electrical engineering. In September 1082,
595-598. he joined the Califonia Institute of Technology, and earned the Ph.D. in
[2] J. W. Mink, “Quasi-optical power combining of solid state millimeter electrical engineering in June 1988.
wave sources,” ZEEE Trans. Microwave Theory Tech., vol. 34, pp. He joined the faculty of the University of Michigan in September 1988
273-279, Feb. 1986. where he is now an Assistant Professor in the Electrical Engineering and
[3] Z. B. Popovic, R. M. Weikle, M. Kim, and D. B. Rutledge, “A 100 Computer Science Department.
MESFET planar grid oscillator,” IEEE Trans. Microwave Theory Tech., Dr. Rebeiz has been awarded a NASA-Certificate of Recognition Award
vol. 39, pp. 193-199, Feb. 1991. for his contribution to the millimeter-wave space program (March 1990) and
[4] R. A. York, and R. C. Compton, “Quasi-optical power combining the Best Paper Award at the 1990 International Confemece on Antennas,
using mutually synchronized oscillator arrays,” IEEE Trans. Microwave Nice, France. He received an NSF Predential Young Investigator Award in
Theory Tech, vol. 39, 1991. 1991. His research interests lie in planar millimeter-wave antennas, receivers
[5] N. Camilleri and T. Itoh, “A quasi-optical multiplying slot-array,” ZEEE and transmitters, and fabrication and measurements of novel millimeter-wave
Trans. Microwave Theory Tech., vol. 33, pp. 1189-1 195, Nov. 1985. transmission-lines and devices.
546 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 42, NO. 4, APRIL 1994
A High Gain Silicon AGC Amplifier
with a 3 dB Bandwidth of 4 GHz
L. C. N. de Vreede, A. C. Dambrine, J. L.Tauritz, Member, IEEE, and R. G. F. Baets, Member, IEEE
Abstract- In this paper, the design and realization of an
integrated high frequency AGC amplifier in BiCMOS technology
are discussed. The amplifier has 36 dB voltage gain, 4 GHz
bandwidth, dynamic range exceeding 50 dB, low spectral dis-
tortion and low power consumption. The amplifier is suitable for
application in wide-band optical telecommunication systems.
I. INTRODUCTION pol t
S
ILICON is the material of choice for medium or large
scale integration of system blocks in many telecommu-
nication applications. In the near future, however, bit rates vee
of 5 to 10 GBit/s will become common. The design and Fig. 1. Top level of the AGC amplifier.
use of silicon MMIC's for these applications is of growing
interest and importance to the microwave community. This Furthermore, in all the simulations the influence of bondwires,
paper addresses the hierarchical design of an integrated AGC process tolerances etc. have been considered. The top level of
amplifier in silicon using commercial microwave design soft- the AGC amplifier design, including packaging parasitics, is
ware. The amplifier chip has been realized in QUBiC, a Philips given in Fig. 1. A block diagram of the silicon chip is depicted
Semiconductors BiCMOS foundry process, featuring 1 micron in Fig. 2.
geometry and encompasses more than eighty active devices. The signal travels from the left to right passing respectively
The AGC amplifier was designed for use in a 2.5 GBit/s the matching input buffer (IB), two gain controllable amplifier
coherent optical receiver. The amplifier was required to have cells (A1 and A2) each with a maximum of 12 dB gain, a
a 3 dB bandwidth of at least 4 GHz and a minimum voltage fixed 12 dB gain amplifier cell (A3) and the output buffer
gain of 30 dB (equivalent to a S,, of 24 dB). Additional (OB). Since dc coupling between stages is used, differential
requirements were a gain control range larger than 30 dB and operation is required. The gain control signals are related to
the use of a standard 32 pin quad flat-pack ceramic package. the differential voltage coming from the AC component peak
detector PD1 and the DC reference peak detector PD2. The
11. THE CIRCUIT DESIGN peak detectors feed their signals to the off-chip integrator
circuit. The harmonic distortion of the high frequency signal is
The ordered design of complex integrated analog circuits
lowered by using an offset control circuit to reduce unbalance
is predicated on limited interaction between the constituent
in the dc operating points of the amplifier stages.
circuit blocks. One way to satisfy this, is the use of cascaded
amplifier cells with a large inter-cell impedance mismatch,
leading to potentially large bandwidth [2]. Inter-connects be- A. The Input Buffer
tween the amplifier stages must be kept short with respect to The implementation of the input buffer together with an
the minimum wavelength involved. The work of [5] which unbalanced 50 ohm external source is shown in Fig. 3. It
described an AGC amplifier design with a 3 dB bandwidth of should be noted that in combination with an external matching
2.5 GHz has been used as a starting point for this study. circuit the impedance level can be chosen to be 50 or 100
The design of the AGC amplifier included additional speci- ohms. The extra resistor (marked with an asterisk in Fig. 3)
fications related to gain control, temperature stabilization and is necessary in this input circuit to avoid common mode to
balanced as well as unbalanced application of the circuit. differential mode conversion of the supply voltage disturbance
component on chip. The bondwire inductance in combination
Manuscript received February 19, 1993; revised June 21, 1993.
L. C. N. de Vreede, J. L. Tauritz, and R. G. Baets are with Delft with the input impedance of the emitter follower input buffer
University of Technology, Dept. of Electrical Engineering, Laboratory for can lead to unwanted resonances. Considerable effort has been
Telecommunication and Remote Sensing Technology, P.O. Box 5031, 2600
expended on the development of a multi-purpose broadband
GA Delft, The Netherlands.
A. C. Dambrine was with Delft University of Technology, Dept. of Elec- input buffer to circumvent this problem. A solution has been
trical Engineering, Laboratory for Telecommunication and Remote Sensing found in a configuration, using a series RC-network in parallel
Technology, P.O. Box 5031, 2600 GA Delft, The Netherlands. She is now
with the input transistor to compensate for the negative input
with Dassault Electronique, 92214 St. Cloud, France.
IEEE Log Number 92 16061 . impedance. This yields an input buffer transfer characteristic
0018-9480/94$04.00 0 1994 IEEE
DE VREEDE ef ab: A HIGH GAIN SILICON AGC AMPLIFIER WITH A 3 dB BANDWIDTH OF 4 GHz 541
gain control OUT double stage en1 tter follower
(Cherry and Hooper) stages
{g
control ldGC
CONTROL
'
IB A1 A2 A3
::I:}
.. ,
Pic Pic
M gain control LIN slgnaYl level
Fig. 2. Block diagram of the AGC amplifier.
Fig. 4. Principle circuit of the differential amplifier cells.
out B lems posed by process tolerances and variation of the bondwire
Q
inductances. The cells A1 and A3 have been provided with
this peaking facility. Using this externally controllable tuning
facility the slope of the gain frequency characteristic can
be modified, canceling out hard to control parasitic effects,
yielding a flat overall gain-frequency characteristic.
out
Consistent with the dynamic range requirement the first two
amplifier cells (Al,A2) have variable gain. Gain control of
these cells is based on the four-quadrant multiplier principle
which offers large dynamic range and high linearity for large
Fig. 3. The input buffer with unbalanced input circuit. input voltage swings. In principal differential signals are
converted to common mode signals at low gain levels. In
(including package effects) that is flat within 2 dB up to at least
the case of total conversion of the differential signal into a
4 GHz. An additional advantage of the input compensation is
common mode signal no amplification will occur. A Cherry
very low power consumption for this circuit block.
and Hooper stage with a four-quadrant multiplier is shown
in Fig. 5. This proved to be well suited for the present task
B. The Amplijier Cells [5].W e used the circuit shown in Fig. 5 in the first amplifier
cell (Al). In this configuration four (equal) series feedback
Wide-band amplifier cells based on the circuit principles
resistors and two shunt capacitors are required in the two upper
first proposed by Cherry and Hooper 29 years ago [2] have
differential stages. The schematic for the second amplifier cell
been designed. The combination of alternating transadmittance
(A2) shown in Fig. 6 is similar to that of cell Al, with the
(TAS) and transimpedance stages (TIS) results in substantial
exception of the reversed gain-control and signal input. This
impedance mismatch between succeeding stages leading to
has two advantages:
excellent wide-band performance. In Fig. 4, a dc-coupled
Two emitter-follower stages are sufficient for shifting
differential amplifier cell based on this principle is given. -
the signal level between the first and second cell.
This cell consists of a Cherry and Hooper stage and two
emitter-follower stages. The emitter followers provide dc level - This approach leads to compensation of the gain-
frequency characteristics of cell A1 and A2 for different
shift and increased impedance transformation. Proper transistor
gain levels. This is made possible since cell A1 the gain
dimensioning and biasing are essential for obtaining 4 GHz
control is part of the ac signal path in contrast with the
wide-band performance. The base resistance of the transistors
situation in cell A2.
involved is one of the bandwidth limiting factors in cascaded
Cherry and Hooper stages. Through careful design low base The current sources in Fig. 4 are implemented as simple
resistance in combination with an acceptable bias current resistors. A current mirror at this point is undesirable due to
can be obtained. To further increase bandwidth a decrease the dominant parasitic capacitances of the transistors, which
in the TAS feedback with frequency is desired. This can be would lead to a decrease in impedance with frequency.
implemented by adding a simple capacitor or a tunable peaking
C. The Output Buffer
impedance in parallel to the series feedback emitter impedance
(see Fig. 4). The output buffer should fulfil the following requirements
This peaking impedance changes the local feedback in the - good impedance match at the output (to avoid instabil-
Cherry and Hooper stage and can be used to overcome prob- ity),
548 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 42, NO. 4, APRIL 1994
doubla stage
(Cherry and thoper)
trans1 npedance
stage (TIS)
output
four-quadrant -;1:-
lultipl ier
- - -
gain Fig. 7. The output buffer.
control
0 V
0
-
Fig. 5. A Cherry and Hooper stage with a four-quadrant multiplier (used NT-u 1
in cell Al). v,)
m I
75
OC output buffer
double stage with campensatian-
0 . .
transinpedance 0Y
stage (TIS) I w h
I ]output 10.0 MHz f req 5.0 GHz
II ’ Fig. 8. Simulated output match for two different types of output buffer.
I
I high output impedance. However, sensitivity to bond wire
I
2 variation remains a problem. Feedback contributed by the
parasitic capacitances Cbc of the output transistors, results in
additional mismatch in the output coupled with some peaking
in Szl. This leads in turn to reduction in the stability factor
K. Simulations have shown that a modified open collector
buffer using the extemal output circuit depicted in Fig. 7 is
less sensitive to this problem. The compensation circuitry has
transadmittance a positive influence on the stability factor K as well as the
output match. This is illustrated in Figs. 8 and 9.
D. The Gain Control
The gain control-voltage generation is based on a two Peak
Fig. 6. A Cherry and Hooper stage with a four-quadrant multiplier (used Detector (PD) structure as shown in Fig. 2. We have chosen
in cell A2). two PD’s to cancel out the effects of temperature variation.
One PD detects the maximum value of the output signal (dc
and ac component). The second PD detects only the value of
- a flat frequency characteristic, the dc component. The difference between the output signals
- insensitive to variation in the bond wire inductance, of the PD’s is directly related to the ac amplitude. Temperature
- extemal output impedance (open collector output). variations which affect the dc level of the output signal will
This list of requirements on the output buffer is nontrivial. not be of influence on the differential voltage between the PD
In particular the matching problem (which leads to a 25 ohm outputs.
load, as seen from the chip) in combination with the inductance The differential voltage output of the peak detector structure
of the bond wires ranging from 2 to 6 nH is problematical. is fed to an external integrator. The output of the integrator
Paralleling pins will reduce the inductance. One commonly circuit is connected to the AGC controller (on chip). This AGC
used solution has a balanced open collector output to provide controller supplies the control voltages for the gain cells A1