-
Tinus Stander
Going from Microwaves to mm-Waves:
What's the Same, What's Different
Electronic Warfare South Africa, Pretoria, 4 – 6 November
2019
-
The letter bands
The atmosphere
The applications
The semiconductor technologies
Components and packaging
System integration
Equipment
Agenda
-
What’s in a name?
Letter bands and mm-wave frequencies
-
Nomenclature (1)
IEEE Std 521-2002
UHF, VHF, X-band, L-band, etc.
Blanket designations > 110 GHz
“mm”: 110 – 300 GHz, “THz” 300 – 1000 GHz
Similar for ITU, NATO EW (A-M) bands
No letter designations above 100 GHz.
RADAR letter bands (Ka,Q,V,E,W) to
110 GHz… but above?
-
MIL-DTL-85/3D
“WR-XX” designation
Unique, good way to shop for components
My nomenclature
(no claim of generality!)
3 – 999.99 MHz: RF
1 – 30 GHz: microwave
30 – 300 GHz: mm-wave
Include 28 GHz
300+ GHz: Terahertz
Nomenclature (2)
Source: MI-WAVE, www.miwv.com
-
To soar through the air…
mm-Wave atmospheric properties
-
Atmospheric propagation characteristics (1)
Much higher attenuation, delay
Gaseous
Oxygen
60, 118.75 GHz
Temporally, spatially fixed
Elevation
Water vapour
22.24, 183.31, 325.5 GHz
Highly variable
Source: http://http://propagation.ece.gatech.edu
-
Atmospheric propagation characteristics (2)
Liquid Water
Clouds
Rain
Atmospheric “windows”
35 GHz
94 GHz
140 GHz
220 GHz
Source: Ancans et al, “Analysis Of Characteristics And
Requirements For 5G Mobile Communication
Systems”, Latvian Journal Of Physics And Technical Sciences,
v.54(4), 2017, pp. 69-78
-
But it is useful?
Applications of mm-waves
-
Applications: the Σ channel
Communications 5G: 26, 28, 39, (52) GHz
P2P: 71-76, 81-86, 92–95 GHz
WiGig: 60 GHz
SatCom: 26.5 - 40 GHz
Inter-satellite: 59-64 GHz
RADAR Automotive (24, 76-81 GHz)
Clouds (94 GHz)
Radio Astronomy Eg. ALMA: 31 GHz – 1 THz
Sources: www.iss.uni-stuttgart.de, www.e-band.com
-
Applications: the Δ channel
Atmospheric Radiometry
mm-Wave emission lines
Synthetic Aperture RADAR Imaging
94 GHz, UAV
Directed energy, active denial
95 GHz
Airport security screening
Active or passive, 80-100 GHz
Sources: Johannes et al, “Miniaturized high resolution Synthetic
Aperture Radar at 94 GHz for microlite aircraft
or UAV”, Proc. IEEE Sensors 2011; wikimedia.org;
www.radiometer-physics.de; www.sds.l-3com.com
-
Why the difference?
Wide available bandwidth Communication speed
RADAR resolution
Attenuation wanted! Spatial reuse, LPD, LPI
Specific physical phenomena Emission line radiometry
Stronger emissions
More compact systems
𝜆 =𝑐
𝑓, everything scales
Source: http://www.jb.man.ac.uk, Christian Wolff
(www.radartutorial.eu)
0 20 40 60 80 1000
1000
2000
3000
4000
5000
6000
7000
Location [GHz]
Ch
an
ne
l b
an
dw
idth
allo
ca
tio
n [M
Hz]
-
Best of both worlds
Multi-functional mm-Wave systems
Active protection
Surveillance
Trunking radio
Combat ID
Compact, LPI, LPD, wideband
Source: Wehling, “Multifunction millimeter-wave systems for
armored vehicle application”, IEEE Trans. MTT 53(3),
2005.
-
A spike on the Δ channel…
Wassenaar Arrangement Dual-use list
3.A.1.b.2: MMICs 31.8 – 37 GHz
> 90 GHz
3.A.1.b.4: SSPAs > 43.5 GHz.
3.A.1.b.7: Equipment Signal extenders > 90 GHz
VNAs > 110 GHz
Much stricter export control
-
What’s under the hood?
Semiconductor compounds and components
-
Semiconductor technologies: III-V
InP, InXAs: 1 THz +
HEMT and HBT
fmax > 1.5 THz
GaAs: 200+ GHz
ft: 400 GHz (mHEMT),
100 GHz (pHEMT), < 100nm
GaN HEMT: 100+ GHz
Stronger bias, higher Psat
Lower gain, less linear
ft > 100 GHz (100nm),
> 450 GHz (20nm)
Useful for:
Discrete transistors (GaN)
MMICs (circuits, RFFE)
Short production runs
Source: Micovic et al, W-Band GaN MMIC with
842 mW Output Power at 88 GHz, IMS 2010.
-
For 30 – 100 GHz: GaAs or GaN?
Power: GaN
Higher bandgap, higher
thermal conductivity: higher T
Higher breakdown: higher VDD
Higher ISAT: higher Pmax, Imax.
Noise: GaAs
Higher mobility
Gain: GaAs
Higher mobility
Longevity: GaN
Higher power failure
(no limiters!)
Higher temperatures
Higher radiation tolerance
Maturity: GaAs
More tech, foundries
GaN catching up!
-
Semiconductor technologies: Group IV
CMOS: 150-200 GHz
fmax ≈ 350 GHz (28nm)
Future: 700 GHz (10nm)?
SiGe BiCMOS: 200 – 250 GHz
Much faster for equivalent
lithography MOSFET
fmax ≈ 370 GHz (55nm)
Future: 2 THz (22nm)?
Useful for:
MMICs with some logic
Mass production
True SoC
-
Transistor amplifiers
Operation close to ft, fmax (≈ 0.3 ft) Bias, size, semi
tech.
Lower gain More gain stages
Higher NF: 1.5dB (X-band) → 5dB (E-band)
Lower power eg. W-band GaN assemblies:
2 W (W-band) vs 2 kW (X-band)
Need combiners, stacking
Lower efficiency (typ. < 10%)
Available up to 100+ GHz
Sources: Gadès et al, 100nm AlSb/InAs HEMT for ultra-low-power
consumption, low-noise applications,
The Scientific World Journal 2014(11); US Patent
20070273445.
-
Diodes
Very high bandwidth
GaAs: 130 GHz (100nm pHEMT) vs. 3 THz (1μm Schottky), UMS.
Used in
Front-end detection Bias or zero-bias
No LNA
Generation (IMPATT, Gunn)
Conversion
Mixers, harmonic mixers
MultipliersSources: Wang et al, “Design of a Low Noise
Integrated Sub-harmonic Mixer at 183GHz Using European Schottky
Diode
Technology, Proc. 4th ESA Workshop on mm-Wave Tech. App., 2006;
Rieh et al, “An Overview of Semiconductor
Technologies and Circuits for Terahertz Communication
Applications, GLOBECOM 2009.
-
Other active components
Frequency translation
Commercially available up to 2.2 THz
Multipliers, Harmonic mixers
Switches, noise sources
Readily available 180 – 220+ GHz
Phase shifters (excl. lab equipment)
MMICs published up to W-band
Exotic options up to 1 THz
Commercially up to Ka-bandSources: Rieh et al, “An Overview of
Integrated THz Electronics for Communications Applications”,
MWSCAS 2011; Shih et al, “A W-Band 4-Bit Phase Shifter in
Multilayer Scalable Array Systems”, CSIC
2007
-
Passive components
Antennas: THz
WG horns, Planar (150+ GHz)
Reflector surface roughness
Smaller aperture
Lower η, Directivity higher
Filters
Waveguide: 330+ GHz
Planar: 150+ GHz
More lossy, less tuning
Sources: www.vivatech.biz, Cheng et al, “94 GHz Substrate
Integrated Monopulse Antenna Array”, TAP 60(1), 2012; Casco,
“Surface roughness estimation of a parabolic reflector”,
arXiv:1007.4600v1; www.miwv.com; Gaskell & Stander,
“Reflection
mode mm-wave on-chip notch filters using coupled hairpin
resonators”, Electron. Lett. 52(5), 2016;
http://gam.webs.upv.es
-
Fitting the puzzle pieces
System integration and packaging
-
Non-planar system integration
Waveguide: 1 THz +
Split block / finline actives
Readily available, full-band
Limited to waveguide bands
Coaxial connectorised: 110 GHz
Standardized up to 1mm
Cables up to 110 GHz
Components < 50 GHz
Few options up to 110 GHz
Sources: Fung et al, “Power Combined Gallium Nitride Amplifier
with 3 Watt Output Power at 87 GHz”, ISSTT2011.
www.elva-1.com; www.sagemillimeter.com
-
Single chip system integration
System-on-chip (SoC): 250 GHz+
Wavelength allows for distributed
components
Digital on-chip possible (CMOS,
BiCMOS), or
Antenna & DC in, IF in/out
Antenna on chip possible
Typically wafer-level packaged
More on that laterSources: Scheytt et al, “Towards mm-wave
System-On-Chip
with integrated antennas for low-cost 122 and 245 GHz radar
sensors”, SiRF2013;
www.mwrf.com
-
Planar multi-chip system integration (1)
Thick film, MCM-C: 122+ GHz
Screen printed metal inks on ceramic
Multi-layer (LTCC or HTCC)
High resolution, DK
Readily available
Thin film, MCM-D: 100+ GHz
Deposited metal layers
Si / SiO2 substrate, BCB dielectrics
Very fine resolution
Not generally available
Sources: www.anaren.com, www.imst.de / www.ltcc.de,
www.mst.com:
Jansen at el, “Micromachined Devices for Space Telecom
Applications”, www.eetimes.com
-
Planar multi-chip system integration (2)
PCB, MCM-L
120+ GHz passives,
E-band systems
Deposited Cu etched / milled
RF Laminates + bonding layers
Low to moderate DK
Available everywhere
Low resolution
Sources: Dyadyuk et al, “Experimental evaluation of the E-band
multi-chip modules integrated using laminated
LCP substrates”, IMWSMMWIT 2011;
www.bosch-mobility-solutions.com
-
What limits the frequency?
Resolution is lower Δ10μm on λ/4 radiator is
0.023% f0 variation @ 1 GHz
2.3% f0 variation @ 100 GHz
Materials losses higher
More radiation
Addressed by: Use thinner substrates
Consider finish / plating
Use other media (eg. SIW)
Consider packaging in design
Sources: Thomson et al, “Characterization of LCP Material and
Transmission Lines on LCP Substrates from 30 to 110
GHz”, TMTT 52(4), 2004; Gold & Helmreich, “A Physical
Surface Roughness Model and Its Applications”, TMTT65(10),
2017; www.siwspace.com
-
Packaged components
40 GHz: QFN, LFCSP, LCC
Internal wirebonds, air cavities
TSLP (Infineon) : 80+ GHz
Wafer-level packaging: 120 GHz+
Ball grids
High pin count, SoC
Multi-chip integration
Bare dies typical above 40 GHz
Sources: www.analog.com, www.infineon.com; wikipedia.org
-
Die attach
Wirebonds
40 GHz broadband
Multi-wire, V-shape, ribbons
130 GHz resonant
Flip chip
Broadband up to 150 GHz
Stud bumping or wafer-scale
Tricky post-process
Sensitive to bump height
Sources: Valenta et al, “Design and experimental evaluation of
compensated bondwire interconnects above 100 GHz”,
IJMWT, 2015; www.secureidnews.com, www.twi-global.com; Heinrich,
“The Flip-Chip Approach for Millimeter-Wave
Packaging”, IEEE Microw. Mag. (2005)
-
Looking forward…
Integral
waveguide apertures
Micro-coax
Sources: L. Devlin, “The Future of mm-Wave Packaging”, Microwave
Journal 57(2)
www.ums.com, www.bridgewave.com
-
“Om te meet is om te weet”
Measurement equipment and techniques
-
Connecting to the DUT
Planar probing more common
1.1 THz on-chip, 67 GHz PCB
Coax possible
1.85mm, 1mm common
0.8mm (220 GHz) introduced
Waveguide readily used
Extenders have WG ports
-
Native frequency ranges and extenders
VNAs: 220 GHz
Spectrum Analyser: 110 GHz
Extenders quite common
1 THz (VNA)
1 THz (SA)
Limited IF bandwidth!
3 THz (CW sources)
Noise sources, attenuators, detectors, mixers, multipliers all
available to match extenders
Source: www.vadiodes.com
-
Unique antenna measurement considerations
Radiator smaller than
connector!
Need to shield
Nearfield scanning
Very high precision
“Far”field: ≈ 1.5m at 50 GHz
Fewer / smaller absorbers
Free space attenuation!
Sources: Boehm et al, “The Challenges of Mieasuring Integrated
Antennas at Millimeter-Wave Frequencies”,
AP Mag, 2017; Hunter & Stander, “A compact, low-cost
millimetre-wave anechoic chamber”, EuCAP 2016.
-
In conclusion…
-
A few good reasons to stick to microwaves…
Attenuation
Longer distance
RADAR, Comms
Rain fade, shadowing
Loss!
Skin effect, materials,
surface roughness…
Cheaper
Components, equipment
Physical phenomena
H2 observation, 1.4 GHz
Manufacturing tolerance
Wavelength dependent
components
Output power
Smaller components
-
Numerous similarities between microwave and mm-wave
Subtle, significant differences
The atmosphere is quite different
The applications are mostly similar, but with different
performance
Some notable exceptions
Component performance degrades
Packaging is a bigger issue
And, subsequently, planar system integration
Extenders on test equipment is more common
The gap is shrinking
The big picture
-
Tinus Stander
Senior Lecturer
Carl and Emily Fuchs Institute for Microelectronics
Dept. EEC Engineering
University of Pretoria
Pretoria, 0002
South Africa
+27 12 420 6704
[email protected]
