Telecommunications Technological Center of Catalonia (CTTC)

Jordi MateuResearch Associate

April - 2005

TelecommunicationsTechnological Center of Catalonia

(CTTC)

2 / 14

WPR3: Design, Modeling and Characterization of RF and MicrowaveDevices and Subsystems ( 0.4 MM )

NEWCOMNEWCOM

Nonlinear ModelingDistributed nonlinear effects

• “Nonlinear Distortion in a 8-pole Quasi-elliptic bandpass HTS filter forCDMA systems” J Mateu, C. Collado, O. Menendez, J.M. O’Callaghan. IEEE Trans.On Applied Superconductivity To be published in June 2005.

Antenna designsUltra-wideBand Antennas

• “Bow-tie shaped UWB planar antenna using a double-sided impedancematching network” P. Miskovsky, J. Mateu, A. Mollfulleda, J. Romeu. Submittedto IEEE International Conference UltraWideBand, Zurich, October 2005.

3 / 14

Nonlinear Nonlinear ModelingModeling

Antenna

Cable

RF front-end, Trec

Bandpassfilter

LNA

NONLINEARITIES

might degrade the system performance

• The performance of wireless base-station can be considerably enhancedby incorporating LNAs with HTS filters.

Broadbandmodulated signals

strong

out-of band

weak in-band

??•Spectral regrowth•Intermixing products

Microwave Nonlinearities in Superconductors

Nonlinear System Performance

• Filter Selectivity

• Nonlinear effects

• Broadband signals

4 / 14

Nonlinear Nonlinear ModelingModeling ∆L(i), ∆R(i)

• The equivalent circuit of HTSline resonators is wellestablished.

Ro dz Lo dz ∆R(i)dz ∆L(i)dz

C dz G dz v v+dv

- dv + + d vl - + dvnl - i i+di

+

-

+

-

This is used to build an equivalent circuit to makerealistic predictions of nonlinear effects in a HTS filter.

Numerical Techniques

• Harmonic Balance (HB) has been proven usefulto simulate this equivalent circuit.

• HB is sufficiently fast to adjust circuit elementsto match simulations to IMD measurements.

dz

5 / 14

Nonlinear Nonlinear ModelingModeling

Measurements

1.95 1.955 1.96 1.965 1.97 1.975-70

-60

-50

-40

-30

-20

-10

0

Frequency (GHz)

S 21 (

dB)

S11

(dB

)

With parasiticeffects

Theory

•We adjust the linear response by assuming

Gradient-based Optimization

1- We use the equivalent circuit to adjust thelinear frequency response of the filter

ResonatorCouplings

LAYOUT

Unwanted coupling

Deviation in resonant freq

Now we are ready to include nonlinearities in the equivalent circuit∆R(i), ∆L(i)

Deviation in desired couplings

6 / 14

Nonlinear Nonlinear ModelingModelingBasic resonant structure

Equivalent circuit

The most important contribution inthe nonlinear effects takes placewhere the current density is largest

We model the whole filter as a halfwave straightHTS Nonlinear lines coupled each other

Narrow band filter Weak coupling between resonatorsWe assume the profile of the currentdistribution in the resonators is almost thesame that would be in an isolated structure

[3EG02]

7 / 14

Nonlinear Nonlinear ModelingModeling2- We use the equivalent circuit to adjust the IMD measurements

by using Harmonic Balance techniques

1.95 1.955 1.96 1.965 1.97 1.975-100

-80

-60

-40

-20

0

Frequency (GHz)

Tran

sfer

func

tion

(dB

)

∆∆f f

PP22PP11

PP2121PP2121

Experiment 1: Power Sweepf0=1.96 GHz P1=P2= - 8 to 10 dBm

∆f = 100KHz Set 1∆f = 200KHz Set 2∆f = 400KHz Set 3

∆f = 1000KHz Set 4

Experiment 2: Frequency Sweepf0=1.95 to 1.975 GHz

∆f = 100KHz P1=P2= 9.5 dBm Set 1∆f = 400KHz P1=P2= 9.5 dBm Set 2∆f = 400KHz P1=P2= 0.5 dBm Set 3

∆L(i), ∆R(i)

8 / 14

Nonlinear Nonlinear ModelingModeling

Frequency Sweep Experiments

1.954 1.958 1.962 1.966 1.97-90-80-70-60-50-40-30-20-100

Frequency (GHz)

IMD (dBm)

Fundamentals (dBm)

dBm

1.954 1.958 1.962 1.966 1.97-90-80-70-60-50-40-30-20-10010

Frequency (GHz)

IMD (dBm)

Fundamentals (dBm)

dBm

1.954 1.958 1.962 1.966 1.97-90-80-70-60-50-40-30-20-10010

Frequency (GHz)

IMD (dBm)

Fundamentals (dBm)

dBm

SET1- ∆f=400KHz/ P1=0.5 dB SET2- ∆f=400KHz/ P1=9.5 dB SET3- ∆f=100KHz/ P1=9.5 dB

Power Sweep Experiments

-8 -6 -4 -2 0 2 4 6 8 10-80

-70

-60

-50

-40

-30

-20

-10

0

10

Fundamentals(dBm)

IMD (dBm)

P1=P2 (dBm)

P out

(d

Bm

)

∆f = 100KHz∆f = 200KHz∆f = 400KHz∆f = 1000KHz

∆L(i)=7.9 10-12 |i|0.2 + 2.2 10-11 |i|2 i in A and ∆L in H/m

Measurements andsimulations match

9 / 14

Nonlinear Nonlinear ModelingModelingCo-sitting of TDD and FDD UMTS base-stations

TDD FDD (Uplink)T1 T2 T3 T4 F1 F2 F3 F4

A B C

TDD-FDD

1900 MHz 1920 MHz 1980 MHz

Co-sitting TDD-FDDCo-sitting TDD-FDD

• UMTS- FDD base station filtersubject to interference of anearby TDD emitter.

1.945 1.95 1.955 1.96 1.965 1.97 1.975-220

-200-180-160-140-120-100-80-60-40-20

0

• Nonlinearities cause spectralregrowth and might block thebase station.

-60

-40

-20

0

dBm

/Hz

dB

Frequency (GHz)

10 / 14

UWB AntennasUWB Antennas

Antenna Requirements:

Low cost Easy to integrate Small size Wide-band Omnidirectional

CTTC Approach:

Planar technology

Self-Complementary Bow-tie based topology

Double-sidedTechnology

FCC definitionFractional bandwidth≥20% or>500MHz absolute bandwidth3.1GHz – 10.6 GHz

Short pulses [ns]No carrierGaussian shape and derivatesWithout DC component

11 / 14

UWB AntennasUWB Antennas

Antenna Subsystem

Radiating UnitMatching/FilteringNetwork

MatchingTechnologies

Double-SidedBow-tie Shaped

Double-SidedMatching impedanceFiltering

Double-SidedTo

Microstrip

12 / 14

UWB AntennasUWB AntennasBow-tie shaped antenna

-Bow-tie shaped antenna with acircle tangent to a 90 deg triangle

- feedline width optimized

Radiating Unit

13 / 14

UWB AntennasUWB AntennasAntenna, matching network and technology transformer

14 / 14

UWB AntennasUWB Antennas

0 1 2 3 4 5 6 7 8 9 10 11 -30

-25

-20

-15

-10

-5

0

frequency [GHz]

S11

[dB

] simulatedmeasured

Gai

n (d

B)

Frequency (GHz) 3 4 5 6 7 8 9 10 11

-10

-8

-6

-4

-2

0

2

4

62 bow-tiesbow-tie and Ridge

-Antenna gain has been measuredfirst using the same antenna as aprobe, then using a ridge antenna asprobe

- Major gain oscillation is observedprobably due to reflections from amicrostrip ground plane

-100 -80 -60 -40 -20 0 20 40 60 80 100-45

-40

-35

-30

-25

-20

-15

-10

-5

0

5

copolar polarization

angle [deg]

3 GHz5 GHz7 GHz9 GHz

- radiation pattern at low frequencies presentsundesired variations probably due to a GNDplane of a microstrip section

- at high frequencies the sidelobes becamemore important

- in general the radiation pattern is not stablewithin whole FCC-UWB frequency range