Transmitter architectures classification

Nomadic RF transmitter architectures, september 26th

Nomadic RF transmitter architectures: principles and limits M. Villegas, A. Diet, G. Baudoin

Short course SHS01: EuWiT - EuMC

T1


Nomadic RF transmitter architectures: principles and limits Speakers and contributors Martine Villegas : Professor at EYCOM-ESIEE Paris Geneviève Baudoin : Research director at ESIEE and general chairman of EuWiT Vaclav Valenta : PhD student ESYCOM and Brno University Martha Suarez : Assistant professor at ESYCOM-ESIEE Luis Andia : PhD student with STMicroelectronics Antoine Diet : Associate Professor at Paris Sud-11 (L2S-DRE, UMR 8506) ST Ericsson team : Fabien Robert, Fabio Epifano, Philippe Cathelin, Pascal Triaire Jacques Palicot : Professor at Supélec (IETR, UMR 6164)

T2


Nomadic RF transmitter architectures: principles and limits Why this short course ? Important evolution wireless communication systems More intelligence and more reconfigurability in the transceiver More efficient spectral resource management Better transmission quality Low energy consumption Frequency band : of interest : 600MHz – 6 GHz Different approaches : multistandard, multiradio, SDR, … Different approaches for reconfigurability and low consumption T3


Nomadic RF transmitter architectures: principles and limits M. Villegas1, A. Diet2, G. Baudoin1

Welcome (M. Villegas1) Spectrum analysis V. Valenta4,1, G. Baudoin1, R. Marsalek4, M. Villegas1 – 20 mn Possible approaches analysis for Cognitive Radio J. Palicot3, G. Baudoin1, M. Villegas1, M. Suarez1 – 30 mn Transmitter architectures classification A. Diet2, M. Suarez1, M. Villegas1, G. Baudoin1 – 30 mn (Pause)

M.

Villegas1,

High efficiency amplifiers A. Diet2, L. Andia1, F. Robert1,5 – 40 mn

Towards all-digital architectures, analysis of technical and technological locks F. Robert1,5, F. Epifano5, P. Cathelin5, P. Triaire5, G. Baudoin1, V. Valenta4,1 – 40 mn

1ESYCOM-ESIEE

T4

(EA2552), 2L2S-DRE (UMR8506), 3Supélec IETR (UMR6164), 4Brno University of Technology (DREL), 5ST-Ericsson

V.

Valenta4,1,

Spectrum analysis G. Baudoin1, R. Marsalek4, M. Villegas1 – 20 mn

Spectrum analysis

T5

1ESYCOM-ESIEE


Outline Air-interface

characteristics of mobile communication standards deployed in Europe: spectrum allocations and signal parameters Survey on spectrum utilization in Europe Goals and interests Measurement sites: Czech Republic and France Measurement equipment, measurement method and data processing Measurement results

Statistical interpretations : duty cycle, degree of the spectrum usage Spectrogram over 24 hours / 6 days

Summary

T6

Nomadic RF transmitter architectures: principles and limits Spectrum analysis V. Valenta4,1, G. Baudoin1, R. Marsalek4, M. Villegas1

Spectrum allocation of mobile communications in Europe

2170 3600 5850

2110 3400

3300

1980

1920

2010 2025 5725

2900 5470

1900

1880

1805 2690 2700

1785 5350

2483.5 2500

1710

960

915 925 2345 2360

2400

880 5150

2305 2320

800 MHz – 6 GHz

*deployment of complementary/same wireless in other radio bands may vary, depending on the country (e.g. MMDS, UMTS 900, CDMA 450, based on full licensing or light regulatory approach) T7


Air-interface characteristics and signal parameters 800 MHz – 6 GHz GSM900/1800

Frequency band [MHz]

Modulation Channel band [MHz] PAPR10 [dB] EVM12 ACPR13 [dBc] Pout MAX / Pout MIN [dBm] Sensitivity [dBm]

UMTS

802.11a/b/g

1900-2025 2110-2200

5150-5350 5470-5825 2412-2472

GMSK3/8PSK4

QPSK5/QAM6

OFDM7/DSSS8

0.2

3.84

16.25/20

0, 3 (EDGE11) Phase error < 5° RMS

6 12.5%@20dBm (16-QAM)

17

880-915 (UL) 925-960 (DL) 1710-1785 (UL) 1805-1880 (DL)

-60@400kHz

-33@5MHz -43@10MHz

5.6% -

WiMAX 802.16e 2300-2400, 2305-2320, 2469-2690, 3300-3400, 3400-3800

LTE Initially in the UTRA1 TDD/FDD2 bands

1.25, 5, 7, 8.75, 10, 20 29 3.16% (64-QAM) Variable, -60@double channel BW

OFDMA - QPSK, 16/64QAM 1.4, 3, 5, 10, 15, 20 22 12.5% @-40dBm Variable, -30-(-36) @3.84-18 MHz

OFDMA9

33, 30/5, 0

24/-50

20, 30/-

23/-50

23/-40

-99

-117 @ BER14=10-3

-82

-92

-

1UTRA

– UMTS Terrestrial Radio Access, 2TDD/FDD – Time Division Duplexing/Frequency Division Duplexing, 3GMSK - Gaussian Minimum Shift Keying, 48PSK – Phase Shift Keying, QPSK5 – Quadrature Phase Shift Keying, QAM6 – Quadrature Amplitude Modulation, OFDM7 – Orthogonal Frequency Division Multiplexing, DSSS8 – Direct Sequence Spread Spectrum, OFDMA9 – OFDM Access, PAPR10 – Peak to Average Power Ration, EDGE11 – Enhanced Data Rates for GSM Evolution, EVM 12 – Error Vector Magnitude, ACPR 13 – Adjacent Channel Power Ratio, BER 14 – Bit Error Ratio T8


Survey on spectrum utilisation in Europe

Goals and interests:

Investigate the degree of the global radio spectrum utilisation in the radio band 100 MHz – 6GHz

Compare spectrum utilisation in different regions

Determine utilisation behaviours / trends of individual communication standards in specific environments

Point out on certain poorly utilised bands that could be dynamically accessed by future opportunistic devices

To prove that the radio spectrum scarcity is an artificial product of archaic public policies rather than a reality T9


Measurement sites: Czech Republic and France Region no.1: northern suburb of Brno, Czech Republic

No.1

Radio band: 100 MHz – 3 GHz

5 km

Region no.2: eastern suburb of Paris (ESIEE Paris), France Region no.3: city of Paris, near “Place de la Nation”

No. 3 No. 2

Radio Band: 400 MHz – 6 GHz 5 km

T10


Measurement equipment, measurement method, data processing

Energy detection principle: Log.- Periodic antennas, spectrum analyzers, PC + GPIB + Matlab Instrument Control Toolbox

• 100 MHz -3 GHz

• 400 MHz -7 GHz T11


Measurement equipment, measurement method, data processing (cont’d)

The whole bandwidth divided into x 20 MHz sub-bands: Hz 0M 0 1

z H 6G

Bandwidth to be analyzed

-4 0 -6 0 -8 0 -1 0 0

…..

0.5

1

1 .5

2

2 .5

…………… up to 145/280 sub-bands •

3

SPAN 20 MHz

Power

(125/160 samples per one band making 160/125 kHz spacing) 0.74

0.745

0.75

0.755

0.76

0.765

0.77

0.775

20 MHz

0.78

• •

RBW 3 kHz / 55kHz 111.108x106 samples/6-days

(125/160 samples) T12


Measurement results Following slides will depict radio spectrum utilization as follows:

-50

-100 16:30 12:30 8:30 4:30 00:30 20:30 16:30

Duty Cycle

Time

Power [dBm]

Threshold*= -97.3 dBm

1.75

1.8

1.85

•

Radio power profile (maximum power over 6 days)

•

Power samples superior to the threshold*(considered as

1.9

occupied) 1.75

1.8

1.85

1.9

1 0.5 0

• 1.75

1.8 Frequency [GHz]

1.85

Duty Cycle =

N (P > threshold) N Total

1.9

*the decision threshold has been set in most cases to 7 dB above the level of the average background noise. T13


Power [dBm]

-40 Threshold = -94.24 dBm -60 -80 0.43

0.44

0.45

0.46

0.47 Time

0.42

0.42

0.43

0.44

0.45

0.46

0.47

0.5 0 0.41

0.42

Region

0.43 0.44 0.45 Frequency [GHz]

0.46

0.47

Utilisation in

Freq. allocation

No. of TV

400-470 MHz [%]

[MHz]

Brno

19.71

470 – 862

49

ESIEE Paris

9.83

470 – 830

45

Paris Nation

6.37

channels

Threshold = -94.37 dBm

-40 -60 -80

11:00 7:00 3:00 23:00 19:00 15:00 11:00 Duty Cycle

0.41 11:00 7:00 3:00 23:00 19:00 15:00 11:00 0.41 1

Duty Cycle

Time

Power [dBm]

410 – 860 MHz Radio Band

0.5

0.55

0.6

0.65

0.7

0.75

0.8

0.85

0.5

0.55

0.6

0.65

0.7

0.75

0.8

0.85

0.5

0.55

0.6 0.65 0.7 0.75 Frequency [GHz]

0.8

0.85

1 0.5 0

No. of “occupied”

Utilisation

Utilisation

TV channels

(Method 1)1 [%]

(Method 2)2 [%]

20

21.2

40.8

28

44.9

62.2

23

29.9

51.1

1Method 1 is based on the "thresholding method" as described in the previous slide. This method corresponds to the average duty cycle. 2Method 2 considers the whole 8MHz channel as occupied when both, the main carrier and the audio sub-carrier exceeds the threshold value (to protect analog TV). The utilization is then calculated as a ratio of occupied channels and total available channels in the band (49 and 45 in Czech Republic and France respectively).

T14


410 – 860 MHz Radio Band (cont’d)

ESIEE Paris

11:00 7:00 3:00 23:00 19:00 15:00 11:00

Paris Nation

Brno

TV 470 – 830 (862) MHz 11:00 7:00 3:00 23:00 19:00 15:00 11:00

11:00 7:00 3:00 23:00 19:00 15:00 11:00

Region

T15

dBm -60

-70 0.45

0.5

0.55

0.6

0.65

0.7

0.75

0.8

0.85 -80

-90 0.45

0.5

0.55

0.6

0.65

0.7

0.75

0.8

0.85 -100

-110 0.45

0.5

0.55

0.6 0.65 Frequency [GHz]

Utilisation in

Freq. allocation

No. of TV

400-470 MHz [%]

[MHz]

Brno

19.71

470 – 862

49

ESIEE Paris

9.83

470 – 830

45

Paris Nation

6.37

channels

0.7


0.75

0.8

0.85

Utilisation

Utilisation

TV channels

(Method 1) [%]

(Method 2) [%]

20

21.2

40.8

28

44.9

62.2

23

29.9

51.1


410 – 860 MHz Radio Band (cont’d) utilisation calculation methods in the TV band Analog TV Main carrier Audio carrier CH56

CH57

CH58

CH59

Power [dBm]

CH55

DVB-T

Threshold 0.74

0.745

0.75

0.755

0.76

0.765

0.77

0.775

Frequency [GHz]

0.78

Calculation method 1 As a ratio of the number of samples superior to the threshold level to the total number of samples in a given radio band. This value corresponds to the average duty cycle.

Calculation method 2 As a ratio of the number of TV channels considered as occupied to the number of available channels. In the case of the analog TV signal, the whole 8-MHz TV channel has been considered as occupied when both, the main carrier and the audio sub-carrier exceeded the threshold.

T16


Zoom on 410 – 470 MHz Radio Band (cont'd) CDMA 2000

ESIEE Paris

11:00 7:00 3:00 23:00 19:00 15:00 11:00

Paris Nation

Brno

UL

11:00 7:00 3:00 23:00 19:00 15:00 11:00

11:00 7:00 3:00 23:00 19:00 15:00 11:00

Region

DL

dBm

-50

-60 0.41

0.42

0.43

0.44

0.45

0.46

0.47

-70

-80

0.41

0.42

0.43

0.44

0.45

0.46

0.47

-90

-100

0.41

0.42


Utilisation in

Freq. allocation

400-470 MHz [%]

[MHz]

Brno

19.71

470 – 862

49

ESIEE Paris

9.83 470 – 830

45

Paris Nation T17

UL

DL

6.37

No. of TV channels

0.45


0.46

0.47

-110

Utilisation

Utilisation

TV channels

(Method 1) [%]

(Method 2) [%]

20

21.2

40.8

28

44.9

62.2

23

29.9

51.1


Power [dBm]

-40 Threshold = -94.28 dBm -60 -80 0.9

0.92

0.94

0.96

0.9

Time

00:00 – 8:00

0.92

0.94

0.96

0.5 0 0.88

0.9

Region

0.92 Frequency [GHz]

0.94

-40 -60 -80

1.8

1.85

1.9

1.8

1.85

1.9


1.9

00:00 – 8:00

1.75

1 0.5

0.96

Frequency allocation [MHz]

No.1 Brno No.2 ESIEE Paris


1.75

11:00 7:00 3:00 23:00 19:00 15:00 11:00 Duty Cycle

0.88 11:00 7:00 3:00 23:00 19:00 15:00 11:00 0.88 1

Duty Cycle

Time

Power [dBm]

GSM 900/1800, DECT Radio Band

0

1.75

Utilisation [%] 38.0 / 20.0

E-GSM + GSM 900: 880 – 915 (UL) / 925 – 960 (DL)

47.9 / 29.3

GSM 1800: 1710 – 1785 (UL) / 1805 – 1880 (DL) No.3 Paris Nation

T18

44.4 / 15.6


Paris Nation

ESIEE Paris

Brno

11:00 7:00 3:00 23:00 19:00 15:00 11:00 0.88 11:00 7:00 3:00 23:00 19:00 15:00 11:00 0.88 11:00 7:00 3:00 23:00 19:00 15:00 11:00 0.88

UMTS

GSM 900 Radio Band (cont'd)

0.89

0.89

0.89

Region

0.9

0.9

0.9

0.91

0.91

0.92

0.92

0.93

0.93


0.94

0.94

0.95

0.95

0.94

0.95



0.96

11:00 7:00 3:00 23:00 19:00 15:00 11:00

dBm -50

-60 -94

-93

-92

-91

11:00 7:00 3:00 23:00 19:00 15:00 11:00 0.96 -93 -92 -91 -90 -89 11:00 7:00 3:00 23:00 19:00 15:00 11:00 0.96 -94 -92 -90 Average Power [dBm]

-70

-80

-90

-100

-110


E-GSM + GSM 900: 880 – 915 (UL) / 925 – 960 (DL)

47.9 / 29.3

GSM 1800: 1710 – 1785 (UL) / 1805 – 1880 (DL) No.3 Paris Nation T19

44.4 / 15.6 Nomadic RF transmitter architectures: principles and limits Spectrum analysis V. Valenta4,1, G. Baudoin1, R. Marsalek4, M. Villegas1

Paris Nation

ESIEE Paris

Brno

GSM 1800, DECT Radio Band (cont'd) 11:00 7:00 3:00 23:00 19:00 15:00 11:00 11:00 7:00 3:00 23:00 19:00 15:00 11:00 11:00 7:00 3:00 23:00 19:00 15:00 11:00

1.72

1.72

1.72

Region

1.74

1.74

1.74

1.76

1.76

1.76

1.78

1.78

1.78

1.8

1.8

1.82

1.82


1.84

1.84

1.84

1.86

1.86

1.86

1.88

1.88

1.88



11:00 7:00 3:00 23:00 19:00 15:00 11:00 1.9 -104

dBm -50

-60 -103

-102

11:00 7:00 3:00 23:00 19:00 15:00 11:00 1.9 -100 -99 -98 -97 11:00 7:00 3:00 23:00 19:00 15:00 11:00 1.9 -99 -98 -97 Average Power [dBm]

-70

-80

-90

-100

-110


E-GSM + GSM 900: 880 – 915 (UL) / 925 – 960 (DL)

47.9 / 29.3

GSM 1800: 1710 – 1785 (UL) / 1805 – 1880 (DL) No.3 Paris Nation T20

44.4 / 15.6 Nomadic RF transmitter architectures: principles and limits Spectrum analysis V. Valenta4,1, G. Baudoin1, R. Marsalek4, M. Villegas1

T21

Power [dBm]

-40 Threshold = -96.12 dBm -60 -80 2.1

2.2

2.3

2.4

2.5

2

2.1

2.2

2.3

2.4

2.5

Time

2

0.5 0 1.9

2


2.4

2.5


-40 -60 -80

2.4 11:00 7:00 3:00 23:00 19:00 15:00 11:00 2.4 1

Duty Cycle

1.9 11:00 7:00 3:00 23:00 19:00 15:00 11:00 1.9 1

Duty Cycle

Time

Power [dBm]

UMTS & 2.4 GHz ISM Radio Band

2.42

2.44

2.46

2.48

2.5

2.42

2.44

2.46

2.48

2.5

2.48

2.5

0.5 0 2.4

2.42


Region

UMTS frequency allocation [MHz]

UMTS utilisation [%]

ISM utilisation [%]

No.1 Brno

FDD UL 1920 – 1980 ; FDD DL 2110 – 2170

2.1

0.24

No.2 ESIEE Paris

TDD 1900 – 1920, 2010 – 2025

10.8

4.47

No.3 Paris Nation

Satellite UMTS 1980 – 2010, 2170 – 2200

11.1

7.63


Paris Nation

ESIEE Paris

Brno

UMTS & 2.4 GHz ISM Radio Band (cont'd)

T22

11:00 7:00 3:00 23:00 19:00 15:00 11:00 1.9 11:00 7:00 3:00 23:00 19:00 15:00 11:00 1.9 11:00 7:00 3:00 23:00 19:00 15:00 11:00 1.9

dBm

-85

-90 1.95

2

2.05

2.1

2.15

2.2

2.25

2.3

2.35

2.4 -95

-100 1.95

2

2.05

2.1

2.15

2.2

2.25

2.3

2.35

2.4 -105

1.95

2

2.05

2.1


2.25

2.3


2.35

2.4

Region


No.1 Brno

FDD UL 1920 – 1980 ; FDD DL 2110 – 2170

2.1

0.24

No.2 ESIEE Paris

TDD 1900 – 1920, 2010 – 2025

10.8

4.47

No.3 Paris Nation

Satellite UMTS 1980 – 2010, 2170 – 2200

11.1

7.63

-110

ISM utilisation [%]


Paris Nation

ESIEE Paris

Brno

2.4 GHz ISM Radio Band (cont'd)

T23

11:00 7:00 3:00 23:00 19:00 15:00 11:00 2.4 11:00 7:00 3:00 23:00 19:00 15:00 11:00 2.4 11:00 7:00 3:00 23:00 19:00 15:00 11:00 2.4

dBm

-50

-60 2.41

2.42

2.43

2.44

2.45

2.46

2.47

2.48

2.49

2.5

-70

-80

2.41

2.42

2.43

2.44

2.45

2.46

2.47

2.48

2.49

2.5

-90

-100

2.41

2.42

2.43

2.44


2.47

2.48


2.49

2.5

Region


No.1 Brno

FDD UL 1920 – 1980 ; FDD DL 2110 – 2170

2.1

0.24

No.2 ESIEE Paris

TDD 1900 – 1920, 2010 – 2025

10.8

4.47

No.3 Paris Nation

Satellite UMTS 1980 – 2010, 2170 – 2200

11.1

7.63

-110

ISM utilisation [%]


Summary: Comparison of spectrum utilisation of wireless standards in individual locations Radio band 400-470 MHz *TV Band IV&V, 470-830(862) MHz GSM 900, 880(888)-915 MHz, 925(933)-960 MHz Radio band 960-1710 MHz GSM 1800, 1710-1785 MHz, 1805-1880 MHz DECT, 1880-1900 MHz UMTS, 1900-2025 MHz, 2110-2200 MHz ISM, 2.4-2.5 GHz Radio band 2.5-3 GHz Radio band 400 MHz - 3 GHz Radio band 400 MHz - 6 GHz 0

DREL Brno ESIEE Paris Paris Nation

10 20 30 40 Spectrum Utilization [%]

50

* Summary depicted here results from measurement campaigns that were carried out during years 2008/2009 [1] V. Valenta et al., Survey on Spectrum Utilization in Europe: Measurements, Analyses and Observations, in proceedings of CrownCom 2010. [2] Mark A. McHenry et al., Chicago spectrum occupancy measurements & analysis and a long-term studies proposal, in proceedings of Workshop on Technology and Policy for Accessing Spectrum, 2006. [3] Md Habibul Islam et al., Spectrum Survey in Singapore: Occupancy Measurements and Analyses, in proceedings of CrownCom 2008. T24


Possible approaches analysis for Cognitive Radio J. Palicot3, G. Baudoin1, M. Villegas1, M. Suarez1 – 30 mn

Possible approaches analysis for Cognitive Radio

T25

1ESYCOM-ESIEE


Outline •

Introduction to Cognitive Radio • •

General remarks A more general « View »

•Opportunistic •

Spectrum management – – –

•

The lower layer »

T26

Current situation Spectrum sharing The 5 phases of opportunistic communications

The sensors •

•

Communications

Hole detector

Conclusion Nomadic RF transmitter architectures: principles and limits Possible approaches analysis J. Palicot3, G. Baudoin1, M. Villegas1, M. Suarez1

Outline •





•

The lower layer »

T27


The sensors •

•

Communications

Hole detector


Cognitive Radio - Introduction •

Introduced by J.Mitola in 1999

Conceptualization and « theorization » of ideas and concepts fashionable in the world of Radio communications •

• • • •

•

T28

Environmental adaptation in a broad acceptation Intelligence in the network and terminal Terminal independence towards network and operator User independence towards technique

Is based on a truly Software Radio Technology

Nomadic RF transmitter architectures: principles and limits Possible approaches analysis J. Palicot3, G. Baudoin1, M. Villegas1, M. Suarez1

Cognitive Radio - Introduction • A few definitions found in the literature: • ”Cognitive radio increases the awareness that computational entities in radio have their locations, users, networks, and the larger environment”. • « Cognition tasks that might be performed range in difficulty from the goal driven choice of RF band, air interface, or protocol to higher level tasks of planning, learning, and evolving new protocols. » • “this type of learning technique makes the Software Radio trainable in a broad sense instead of just re-configurable”

• Broader than the conventional view limited to the spectrum optimization T29


sensors

"OSI layers"

User Profile: price, subscription, personal choices (ecological radio..) Sound, image,...position, speed safety

Inter-networks and intra-networks vertical handover, standards

Access, power, modulation and coding types frequency, handover…. channel estimation antennas,consumption

Concepts found in the literature

Application and IHM

“Context-Aware”

« Cross Layer Transport, Network Adaptation Connection & Optimization »

Interoperability

Physical, medium

Surrounding Networks

Link adaptation

“Middleware” and abstraction layer T30

Wide Band Software Radio Technology

Cognitive Radio - Introduction Conventional Cognitive cycle Observation (sensors)

Knowledge Base Rules

Learning Outside world

T31

Decision Action


Cognitive Radio - Introduction Decentralized view associated with a local optimization of needs and resources versus a centralized view based on the worst case scenario needs. •

• Ex : implementation of an equalizer

independently of the channel IR T32


Outline •





•

The lower layer »

T33


The sensors •

•

Communications

Hole detector


Spectrum management Integration: • in time • in space • in service

No available spectrum (scarce resource) T34

Spectrum management

BUT 2,4 GHz band occupancy 01/September/2004 in New York

TV band occupancy 01/September/2004 in New York

T35


Spectrum management

Hic et nunc (Here and now) Free spectrum Spectrum sharing • Dynamic spectrum access • • • •

Vertical and Horizontal sharing Underlay sharing Overlay sharing Opportunistic communications

Q Zhao, A Swami, “A survey of Dynamic spectrum access SP and networking perspectives, ICASSP 2007 T36


Dynamic spectrum access Command and control • Spread the signal • Power below a threshold for the primary user • Interference temperature concept • Ex UWB Secondary user

Spectrum underlay

Spectrum overlay Primary users

S. Haykin, “Cognitive radio: Brain-empowered wireless communications,”IEEE Journal on Select. Areas in Comm., Vol. 23, no. 2, pp. 201-220, Feb. 2005. T37


Dynamic spectrum access Spectrum opportunity = White spectrum = Spectrum Hole

Open access

Spectrum overlay

Opportunistic communications

Same notion = different words in the literature Spectrum opportunities identification = great challenge

T38


opportunities •Holes in the wide sense •Holes in the spectrum ( spectrum opportunities) •Holes in time (slot opportunity) •Holes in code (code free in a set of code) •Holes in other dimension…..

T39


standard 1

standard 3 free

free

free

free

free

freq.

free

free

free

free free free

ampl.

standard 2

time

Example of holes in the spectrum

T40


standard 1

standard 3 free

free

free

free

free

freq.

free

free

free

free free free

ampl.

standard 2

time

The 5 phases of opportunistic communications • • • • • T41

Filtering phase Hole detection phase Characterization phase Decision making phase Insertion phase Nomadic RF transmitter architectures: principles and limits Possible approaches analysis J. Palicot3, G. Baudoin1, M. Villegas1, M. Suarez1

Filtering phase (1/5) • In one predefine band – Ex: Check if one GSM channel is free in the GSM band

• Whatever the band is – Ex: Check if a desired Bandwidth( ex 1MHz) is free in the band [1MHz - 3 GHz]

• The solution will be different T42


Filtering phase (2/5) • FB should be able to extract channels spaced with different values. We only consider here a sequential extraction. P1 freq

P2 freq

P3 freq

T43


Filtering phase (3/5) • FB should be able to extract channels with different bandwidths simultaneously.

T44


Filtering phase (4/5) • FB must be able to be selective enough with a reasonable complexity as very sharp filtering expectations may be demanded, especially if the bandwidth of channels is small compared to the wideband acquired signal.

T45


Filtering phase (5/5) • Classical FB filters does not meet these requirements. A new scheme has been recently proposed based on FRM technique [1] Ha(z) a b

w HC(z) = 1-Ha(z)

fap

fas w

fap

fas

HC (zM)

c d

Ha(zM) w

HMA(z)

HMC(z) (m-fap)/M

fp fs

e

w

(m+1-fas)/M w

fp fs T. Hentschel, “Channelization for software defined base-stations,” Annales des Telecommunications, ISSN 0003-4347, vol. 57, pp. 386-420, no. 5-6, May-June 2002. [1] R. Mahesh,A. P. Vinod, C Moy, J Palicot, “A Low Complexity Reconfigurable Filter Bank Architecture for Spectrum Sensing in Cognitive Radios”, CROWNCOM 2008, Singapour T46


Outline •





•

The lower layer »

T47


The sensors •

•

Communications

Hole detector


Spectrum Hole detection

In each band the algorithm detection should be applied Whatever the method is • • • • T48

Energy detector Cyclostationnarity detector Covariance matrix eigenvalues detector Other…. Nomadic RF transmitter architectures: principles and limits Possible approaches analysis J. Palicot3, G. Baudoin1, M. Villegas1, M. Suarez1

H 0 : x(t ) = b(t )

H1 : x(t ) = ∑ si (t ) + b(t ), i

Spectrum Hole detection Problem positioning y (t ) x (t )

ADC

x (n.T e )

Detector

Decision

BL

• Available band or engaged band • Hypothesis test: H 0 : x (t ) = b (t ) H 1 : x (t ) = T49

∑ s (t ) + b (t ) i i


H 0 : x(t ) = b(t )

H1 : x(t ) = ∑ si (t ) + b(t ), i

Spectrum Hole detection • Radiometer: (+) simple efficient (--) requires a relevant noise level estimation

• Cyclostationary: (+) far less responsive to noise level variation (--) requires a prior knowledge of the cyclic frequencies that need testing (--) responsive to Nyquist emission filtering

Multi-cycle sensor Blind sensor

T50


Energy detector r(t)

y(t)

(• )

2

y2(t)

T

V = ∫ (•) dt 2

0

V

>
K, decides that a signal exists • Noise level threshold

* N0 is the noise spectral density

T51


Cyclostationnarity detector T=cyclic period c xx (t,τ ) = E ( x (t )x (t + τ )) = c xx (t + T, τ ) The time varying function cxx (t ,τ ) Can be developed according to the variable into the following Fourier series:

cxx (t ,τ ) = cxx (τ ) + ∑ C xx (α ,τ )ei 2πα t α ∈ψ

1 C xx (α , τ ) = lim Z →∞ Z

∫

Z /2

−Z /2

c xx (t , τ ) e − i 2 πα t dt

t

Cxx=cyclic covariance function α= cyclic frequency

• Test on a cyclic frequency (Giannakis) • Multi-cycle test1 • Blind test 1Ghozzi

M , Dohler M , Marx F , Palicot J, Cognitive radio: methods for the detection of free bands, Comptes Rendus Physique, Elsevier, pp, volume 7 September 2006 T52


Cyclostationnarity detector

Harmonics may be filtered by Nyquist filtrer

T53


Cyclostationnarity detector Digital television detection with the Multi-cycle detector

Blue=energy detector Black=Multi-cycle detector

T54

Cooperative Approach Covariance matrix eigenvalues detector Network cooperative sensing1

Secondary base stations {BS1, BS2, BS3, … , BSK} cooperatively sense the channel in order to identify a white space and exploit the spectrum. 1

L Cardoso, M Debbah, P Bianchi, J Najim ‘Cooperative Spectrum Sensing Using Random Matrix Theory », IEEE ISWPC 2008, May 2008, Santorini, Greece. T55


Cooperative Approach Covariance matrix eigenvalues detector The value of cooperation: random matrix approach

T56


Cooperative Approach Covariance matrix eigenvalues detector

T57


Cooperative Approach Covariance matrix eigenvalues detector Spectrum sensing algorithm: computing the ratios

T58


Covariance matrix eigenvalues detector

Simulations: detector performance

T59

Characterization phase • Agreement between TX and Rx of both A and B • Then the frequency band is an opportunity = other communications B

A

• To characterize this band (S/B, interference noise…) • To characterize the IR between A and B (great Challenge) T60


Decision phase • Many algorithms •Neural networks •Genetic algorithms •Statistical Signal Processing •Game theory •Example multi-armed bandit with UCB algorithm1

1Jouini

T61

W, Ernst D, Moy C, Palicot J, Upper confidence bound based decision making strategies and dynamic spectrum access International Communication Conference, ICC'10, Cape Town, South Africa, 26-29 May 2010

Insertion phase • Need of modulation with good Power Spectrum Density •OFDM/OQAM modulation •European project PHYDIAS •Increase the overall PAPR •Insertion under PAPR mitigation constraint

Palicot J, Louet Y, Hussain S, Zabre S, Frequency Domain Interpretation of Power Ratio Metric for Cognitive Radio Systems, Proceedings of IET Communications Journal, N° 2, pp 783-793, june 2008. T62


• Wireless capacity

Cooper's Law • X2 every 3 years since one century • This gain has been obtained = – – – –

X 25 using wider spectrum X 5 sharing small channels X 5 improved modulation X 1600 reduced cell sizes

We believe that the next 20 years the improvement will be given by

cognitive radio

T63

•

Vikram Chandrasekhar and Jeffrey G. Andrews, “ Femtocell Networks: A Survey “, arxiv.org/pdf/0803.0952v2.pdf

•

1

MS.Alouini and A.J.Goldsmith, “Area Spectral Efficiency of Cellular Mobile Radio Systems”, IEEE Tr on Vehicular Technology, vol 48, n°4, pp1047-1066,July 1999.

A.

Diet2,

Transmitter architectures classification M. Suarez1, M. Villegas1, G. Baudoin1 – 30 mn

Transmitter architectures classification

T64

1ESYCOM-ESIEE


Nomadic RF transmitter architectures

(Analog) RF architectures - Basics structures and theirs evolutions - New structures based on figures of merit (linearity and/or efficiency)

Digital RF architectures Conception is based on digital-RF blocs

RF architectures with "digitization" - Evolution of RF analog - New components "digitally-based"

T65

Nomadic RF transmitter architectures: principles and limits Transmitter architectures classification A. Diet2, M. Suarez1, M. Villegas1, G. Baudoin1

35

 C Eb  C = log 2 (RSB + 1) = log 2  + 1  B  B N0 

Cellulaire Tx WiMAX Tx WiFi Tx Bluetooth UWB DVB-H Rx GPS Rx T-DMB Rx

20 0 -20 -40

ASK

dBm

PSK

APSK

C/B Eb/N0

-60 -80

Modulation choice : - Data rate need + QoS - Bandwidth, coexistence, immunity - Spectral/power efficiency - Architecture design impact

-100 -120 -140 -160 -174 0

FSK

500

1000

1500

2000

2500

3000

3500

4000

5000

6000

Cellular/mobile

Connectivity UWB

GSM (900/1800)

WiMax WiFi

GSM/EDGE

TDD

T66

(H)-FDD

CSMA/CA

Bluetooth LTE

UMTS FDD/TDD

GPS DVB-H


CW

PAPR due to shape filtering

WB-CW

PAPR due to multi-carrier and shape filtering

T67

IR-UWB

Mean power depending on repetition frequency

IR

Frequency spectrum

PROS.

Power efficient

Time signal (with shaping filter for CW)

Multi-carrier case (ex. OFDM)

Power saving Spectrum efficient Spread spectrum

NB-CW

Typical modulation scheme


Signal integrity : EVM, ACPR

- WiFi (25 Mbps) - WiFi (54 Mbps) - WiMAX (64-QAM) - LTE (16-QAM) - GSM/EDGE - UMTS (QPSK)

Bi : Emitted Ai

Ei EVM

Bi

T68

rms

=

Ai : Ideal 1 N

∑

A i − C 1B i − C 0

i =1

1 N

i =1

∑ =

N

∑

N

2

Ai

2

A i − Si

15 % 5.6 % 3.1 % 12.5 % 10 % 17.5 %

QPSK eye and 16-QAM for 3% EVM (rms)

Ei : Vectorial Error

N

-16 dB -25 dB -30 dB -18 dB -20 dB -15 dB

2

i =1 N

∑

Ai

2

i =1


Signal integrity : EVM, ACPR PSD

Power level

30 dBm 25 dBm

R R0

0 dBm

P0

Padj

-25 dBm

frequency -50 dBm

ACPR =

P0 = Padj

∫

DSP (f ).df

frequency

P0

∫

DSP (f ).df

400kHz

Padj

" R = R0 + NL " ...to compare with spectral requirements Co-existence and RX sensitivity/selectivity T69

GSM/EDGE WCDMA LTE WiMAX

5MHz

10MHz

33dB 33dB 40dB

43dB 36dB 60dB

60 dB


Signal characteristics and architecture impact Peak to Average Power Ratio (PAPRRF)

PDC PRF_in

PAPR causes variations of the input power at RF frequency NL reaction of the transceiver linearization needed (EVM, ACPR)

Device or PRF_out Front-end

η AE =

η =

PRF_out - PRF_in PDC

PRF_out PDC

LINEARITY

PA

PA

BB + BB Dig + Dig NB / WB

η peak , η mean

Pin (t )

Nomadic battery lifetime

EFFICIENCY

Architecture design FLEXIBILITY (MULTI-RADIO)

T70

COMPLEXITY

SCALEABILITY (POWER)

Selectivity, linearization, linear architecture

- Power control (avoid Near-Far for ex.)

- Additional components or devices - Stability analysis (if loop) and model - Bandwidth enhancement (tech. challenge) - Increase in current consumption - Increase in size (PB process and package)

- Adaptation to the network characteristics Channel (statistic+fading)

- Ressource management


Realized functions in a transmitter architecture from the digital information towards the emitted micro-wave

Digital Information 0100110101

Coding and modulation scheme mapping

Digital to Analog Conversion (DAC) Baseband (symbol) frequency

Frequency transposition (in 1 or X steps) using IQ modulators and/or Modulated PLL

Power amplification (also Control) and Filtering (before/after the PA)

Connection: Duplexer, switch (loss... )

RX ?

- Spectrum/power efficiency - Data rate transfer and Bandwidth (BW) - Adaptation to the channel

- Modulators imperfections (imbalance, BW) - BW and stability of mod. PLL - Envelope information in mod. PLL ? - Phase noise of frequency synthesizer (SYNTH) - Flexibility of SYNTH + transposition

- Resolution and bandwidth limitation versus consumption - Adaptation to different signal characteristics (polar decomposition or cartesian) - State of the art new architectures T71

- Choice of the amplifier class (CW, SW...) - Impact of PA NL EVM, ACPR (model of the PA, power behavior) - Linearization or linear architecture - Filter required ( loss of efficiency)

- RX sensitivity/selectivity - (H-)FDD or TDD - loss efficiency - passive (power + HF) - MIMO ?

- Antenna (MIMO ?), package and CEM - Impact front-end - DDR and bandwidth - distortion + dispersion


Basics architectures LO-Antenna coupling CNA RF

RF

Homo-dyne

PLL

RF

+

0 90°

HPA

Pros: - Simple Cons: - Coupling

-

LO-HPA coupling

CNA

LO2 CNA FI

Hetero-dyne

LO1 PLL

+ 0 90°

-

RF HPA

Pros: - No Coupling Cons: - Phase Noise - Nbr components

CNA

T72


Basics architectures Polar "lite" for GSM/EDGE (small PAPR) AM detector PLL

CNA

F(p)

Modulated PLL

LO1 PLL

+

0 90°

+

VCO

RF

RF VGA

HPA

-

CNA

Angular modulation

T73

Pros: - Phase Noise/Stability Cons: - Bandwidth limited - No AM information

Bandwidth limitation :

Amplitude information :

- "in" the loop reference or feedback mod. - "over" the loop VCO input (stability ?) - "in" and "over" combination ?

- Polar lite architecture [Staszewski et al, 2005] "dynamic power supply" and polar family (w. and wo. feedback...)


Defaults impact (IQ) I = A k .cos (ϕ k ) G.cos (ωt ) Quadrature mismatch

0 90°

PLL

G.sin (ωt + θ )

+ -

G.A k (1 + a.cos [θ]).cos (ωt + ϕ k ) 2 G.A k (1 − a.cos [θ]).cos (ωt − ϕ k ) + 2 − G.A k .a.sin [θ] sin ϕ k sin [ωt ]

Smod =

Q = a.Ak .sin (ϕ k ) + Vdc Gain imbalance

[ ]

« Offset »

+ Vdc .G.sin (ωt + θ ) a = 1.2, θ = 5°, Vdc = 0

 ( 1 - a.cos [θ] ) + ( 2.a.sin [θ] ) IRR = 10.log10  ( 1 + a.cos [θ] )2  2

2

T

T

   

T74

 - sin [θ ]  + V  dc  a.cos [θ ]   cos [θ ] 0

Images of I and Q Local oscillator signal

a = 1, θ = 0°, LOR = -24 dB

 2. VDC LOR = 10.log10  Ak.[1 + a.cos(θ)]  I   1  I out  =       Q out   Q   - a.sin [θ ]

Transposed information

  

T

2

   

Defaults impact (PN) Synthesized frequency

Power θbref θref

VCO phase noise

Resulting phase noise

θbvco θout

Kφ.θbcomp

Kvco/p

F(p)

Reference phase noise

θcomp 1/N θbdiv Loop filter bandwidth

Koct KΦ θout =

Noise Floor Frequency

F (p) p

 F (p) 1 +  Kvco KΦ  Np 

   

[θ

ref

+ θbref + θbdiv + θbcomp] +

1.5

θbvco  F (p) 1 +  Kvco KΦ  Np 

   

1.0

0.5

0.0

-0.5

θout = G (p) [θref + θbref + θbdiv + θbcomp]

+

H(p) θbvco

-1.0

-1.5 -1.5

T75

-1.0

-0.5

0.0

0.5

1.0

1.5

Non-Linearities main characteristics (memoryless) ACPR HPA

EVM

ideal

ideal

+ Inter-modulation products

conversion

conversion and compression

conversion

conversion and compression T76


Vdd

AM/AM (compression)

Input AM (envelope)

AM/PM

HPA

AM/AM

Non-Linearities main characteristics (memoryless)

Input AM

AM/PM (conversion)

Vdd/AM (compression)

Input "AM/Vdd" (envelope)

Vdd/PM

HPA

Vdd/AM

Input AM

Input "Vdd"

Vdd/PM (conversion) Input "Vdd" T77


PAPR of Multi-carriers signals (ex. OFDM) 10

  C(t) = ℜ  ∑ [I n(t) + jQn(t)] e jω0t e jn∆ωt  n=− N/ 2  n= N/ 2

FFT FFT 10 PAPR CCDF [-]

Time domain

Time statistic

3

6000

2

5000

1

4000

0

10

FFT

10%

-1

FFT FFT

WiFi WiMAX LTE

-2

size size size size size

=256 =128 =512 =1024 =2048

3000

0

10

2000

-1

-3

1000 -2

I or Q OFDM signal

0

-3

-2.5 -2.0 -1.5 -1.0 -0.5 0.0

4

0.5

1.0

1.5

2.0

2.5

10

1500

3

-4

4

5

6

7

8 9 PAPR [dB]

2

1000

1 0 -1

EVMrms max/min (64/128) and spectrum (10 dB/step) in fct. of PAPR (lim. from -6 to 12 dB)

OFDM phase 500

-2 -3

0

-4

-4

2.5

2.0

-3

-2

-1

0

1

2

3

4

8000

OFDM envelope

10

11

12

13

6

4

0 -10

6000 -20

1.5

4000

2

-30

1.0 -40

2000

0.5

0.0 2.0E4

-50 -60

0 4.0E4

6.0E4

0.0

0.5

1.0

1.5

2.0

2.5

-70 4900

0 4950

5000

5050

5100

6

7

8

9

10

11

12

PAPR lim. + "HPA NL" LINEARISATION NEEDED T78


Basics/classical architectures (Narrow band signals)

Hetero-dyne PB of Synthesis Noise and nbr. components

Toward wideband

Linearization by a correction technique

+ Feed-Back + Pre-Distortion (analog/digital)

Linear architecture, using a complex decomposition and recombination

Optimization of special figures of merit (efficiency η, PAPR)

CARTESIAN

LIST+ : PWM/Σ∆ coding (PAPR)

Homo-dyne

LINC CALLUM

BASICS + Feed-Back + Feed-Forward + Pre-Distortion (analog/digital)

EER Env. Tracking

PLL mod.

DOHERTY (η)

LIST+ : PWM/Σ∆ Envelope coding (PAPR)

and dynamic biasing

POLAR FULLY DIGITAL T79


Feed-Back + PLL

HPA "A"

0 90°

-

Vdd S out A = Sin 1 + AK

|.|

K 90°

Phase is difficult to "feed-back"

Feed-Back technique : - Reduces gain and enhances linearity region - Stability of the loop (BW and TR response) - Phase is difficult to "loop" - Cartesian FB is an alternative tradeoff "nbr components Vs freedom degree (and BW)" - Thanks to ADC, digital feedback* has good performance and enable efficient adaptative algorithm

Pout (dBm)

0

Pin (dBm)

* this is also right for pre-distortion, whether the signal is in its cartesian or polar representation

T80


Feed-Forward

+ PLL

0 90°

D. and Φ.

HPA "A"

D. and Φ.

-

D. and Φ.

Delay and

Feed-Forward technique : phase alignment - Needs lots of additional components - Always Stable - Phase control is critical - Increases the consumption (crippling for "nomadic"...) - Benefits from adaptative correction (a kind of FB+FF...) - Can be very interesting for BTS [Ghannouchi et al., 1997, FF+DPD] T81

Extraction and amplification of NL


Pre-Distortion

HPA

P

20

Ideal, ACPR = −58.7836 Without PD, ACPR = −35.2099 Polyn. PD, ACPR = −57.6881

Pre-Distortion technique : - Digital / Baseband / IF / RF - Polar / Cartesian - Adaptative good results for ACPR reduction ! [Marsalek et al., 2003] - Memory effects (accuracy of models...) - The model is difficult (increase the complexity) - Bandwidth enhanced

power spectral densities

0

−20

−40

−60

−80

−100 0

1

2

3

4

5

6

7

8

frequency normalized by symbol frequency

T82


From Linearization to Linear architecture

Signal to emit

NL NL1

decomposition

NL2

The behaviour of NL components are very different Taking into account all the different parts of the transceiver : from the DAC and "digital" toward the RF/antenna

T83

correction

Towards the antenna

recombination

- Increase in complexity and Nbr components - Some new sources of defaults appears ! - keeping "nomadic" figures of merit "Linearity Vs Efficiency Vs Power Ctrl" - Flexibility ? ( multi-radio)


Avoiding AM/AM and AM/PM : LINC LInearization with Non-Linear Components, [Cox]

[

I + j.Q = R (t ).e jΦ ( t ) = r0 . e j(Φ ( t )+ θ ( t )) + e j(Φ ( t )−θ ( t )) Baseband complex information

r0 ) R(t

+θ(t)

Constant envelope signal n°1

T84

AM = Constant no AM/AM no AM/PM

+ HPA -

r0 .sin (Φ (t ) − θ (t ))

PLL

LINC : - 2x PA and IQ Mod. - LOSS of Combiner - Decomposition complexity - Bandwidth - 2x Phase noise

Constant envelope signal n°2

r0 .cos (Φ (t ) − θ (t ))

Φ(t) -θ(t)

]

r0 .cos (Φ (t ) + θ (t ))

0 90°

+ HPA -

LOSSES !

r0 .sin (Φ (t ) + θ (t ))


LINC CALLUM Combined Analogue Locked Loop Universal Modulator [Bateman] I(t)

VCO

HPA

LINC part

Q(t)

HPA

VCO -

PLL

T85

0 90°

LINC CALLUM: - 2x PA and VCO. - LOSSES of the combiner - Decomposition less complex - Bandwidth and stability (dual loop)


Avoiding AM/AM and AM/PM : EER Envelope Elimination and Restoration, [Khan] R(t) e j Φ(t)

e j Φ(t)

I (t ) + jQ (t ) = R(t ) e

R=1

Φ(t) R(t)

101010 D E C O M P O S I T I O N

T86

CNA

001101

R(t ) = I²+Q²

CNA cos (Φ) =

011001

CNA

sin (Φ) =

I I² + Q²

jΦ( t )

EER : - High efficiency HPA - Decomposition complexity - Bandwidth of Signals - Synchronisation - Envelope coding efficiency

PWM or Σ∆ coding

+ PLL

0 90°

-

H. E. HPA

Q I² + Q²


Envelope Tracking and dynamic biasing techniques ET : - High efficiency HPA + driving signal - Synchronisation - Envelope coding efficiency - PAPR and clipping

Dynamic biasing (input and/or output) "Tracking" the supplied DC power

101010 D E C O M P O S I T I O N

T87

CNA

001101

PWM or Σ∆ coding or converter

R(t ) = I²+Q²

CNA + PLL

011001

0 90°

-

H. E. HPA

CNA


Polar architecture for nomadic multi-radio - RF carrier frequency and bandwidth (high data rate) - Power scaleability (average and dynamic) - Linearity and efficiency ∀ modulation scheme and frequency (flexibility)

Slow control of the carrier frequency Frequency control Power control

Flexibility/tuning

Slow control of the average power Supply modulation Amplitude : ρ(t) Σ∆ or PWM or...

Modulated signal

Phase : φ(t)

PLL

"HPA" + Filtering NT

Front-end pass-band profile T88


BW limitation

Efficiency Sensitivity improvement to PAPR

Complexity

transfer complexity to digital

Increase in size

Frequency flexibility

Power scaleability

Main imperfections sources

FB

loop

medium

-

low

possible

low

loop

-

Stability, BW

FF

-

low

-

high

difficult

high

PA

-

Model, consumption

PD (DPD)

-

high

-

can be very high

yes

-

PD or

-

Need of adaptability (complex)

very high

difficult (decomp. possible)

high

loop and

-

Decomposition (complex) Recombination (loss)

yes

Synchronisation BW Envelope coding efficiency

yes

PA driving Envelope coding efficiency

Transposed signals LINC enhance BW combiner to mod. (CAL.) (2 loops) index

phase EER

ET

T89

Signal BW, DC/DC converter

DC/DC converter

high

PA

yes (clipping)

high

decomp. possible

high

DPD

combiner

PA

AM coding ?

high PA

yes (clipping)

high

AM coding ?

high

PA


Considerations for multi-radio: Signal PAPR and data rate ↔ EVM, ACPR , co-existence Signal BW and stability ! DAC state-of-the-art challenge Multi-radio + BW "open-loop" appreciated (ET, EER...) Output filtering must be considered more compared to NL Filtering front end ? Filtering PA ? Filtering antenna ? (flexibility ?) filters banks are currently necessary but lowers the efficiency

Observations/trends 1) Linear architectures (LINC, EER, ET,...) can highly benefit from Pre-distorsion, (digital decomposition, interfaced with the DPD algorithm) increase in complexity 2) Techniques for reducing the PAPR or its NL influence complete "linearization" but the complexity is often shifted to another figure of merit such as efficiency, selectivity... 3) Improving peak and average efficiency (cf. signal stat.) is a challenging lock for "nomadic". Often, this is not compatible with flexibility... a precise knowledge of the PA design is so mandatory (next part) T90


Improving the efficiency of a PA : Doherty 4

Ideal generator 2

Ideal generator 1

R = 1, I normalised

3

R0 = 2R

I1

I2

λ/4

I2 I1

2 1

R

0 0

200

400

600

800

1000

5

R ingen 1

4 gen 1 + λ

R in

3

 I  Ringen 1 = R 1 + 2  I1  

 I  Ringen 2 = R 1 + 1  I2  

R0 = 2R

2

R

T91

R02 = gen1 Rin

R ingen 2

1 0

gen1+ λ 4 in

4

200

400

600

800

1000

Modification of load lines for both PAs R decreases evolution from class A to AB, B or C ! (it means: improving the efficiency)


Improving the efficiency of a PA : Doherty

Main "PA" i.e. Gen. 1

+ PLL

PA

0 90°

-

λ/4

τ

Auxiliary "HPA" i.e. Gen. 2

η Main PA

Main and auxiliary PA

Pin T92

HPA

Doherty : - PA + High efficiency PA (2) - Design with load-pull - Efficiency improvement (average and peak) - Narrow band ! (λ/4)

Evolution of LIST thanks to PWM and Σ∆ (LInear amplification employing Sampling Techniques) - The use of LIST technique is known for many years.... but not at GHz ! - It provides the amplification of "coded AM" with a switched PA - Avoid AM PAPR avoid PA NL ! - The AM should be restored after amplification by filtering - The efficiency is : [ PA efficiency ] X [ coding efficiency ]

+ PLL

0 90°

-

T93

Σ∆/PWM modulator

PA

Performances are the major technical lock


Evolution of LIST thanks to PWM and Σ∆ (LInear amplification employing Sampling Techniques) Envelope

Σ∆ coder

Polar Σ∆

Cartesian Σ∆

Power reflected to the PA T94


RF Filtering need - If Σ∆ or PWM - For multi-radio - Spectrum mask

Rejection of the spectral regrowths Poor efficiency due to the AM coding Coexistence !

LC Filters : (-) Frequency (-) Size (-) Quality factor

Ceramic Filters : (+) Low IL and cost (-) Integration (-) Size

SAW* Filters : (-) Size (-) Frequency (-) Output power (-) Integration IC

LTCC** Filters : (+) Good rejection / low IL (+) Higher fmax than others. (-) Integration process

BAW*** Filters : (+) Good Rejection, low IL (+) Higher fmax than others. (+) Integrated “above IC” / Size reduction. SAW*: Surface Acoustic Wave T95

LTCC**: Low Temperature Co-Fired Ceramic

BAW***: Bulk Acoustic Wave


The antenna influence

50 Ohms

50

Antenna impedance

0

- Modification of the bandwidth and adaptation

-10

- Out of band emission (see harmonic opt. of SW class E and J PAs, for ex.)

-20

-30

- Why 50 Ohms for PA "filtering output NT" ? -40 2.5

T96

3.5

4.5

5.5

6.5


Co-design principle

ZNL

ZTX

channel

- Channel and antennas impact the front-end design - Antenna impedance modifies the NLs of the PA (+filter) - Electrical model of the antenna ? modification of the geometry under constraints...

T97

ZRX

S ( f, θ, φ, p)

Rad. characteristics DDR mod.

PA/Ant. coupling (CEM)

ZTX

PA NLs modification


Need of a complete front-end co-design

Front end blocs (PA/filter/duplexer/antenna) challenge for multi-radio

T98

- PA :

Efficiency, flexibility, power scaleability

- Filter :

Avoid ? Bank (losses) ? Enough selectivity ? Integration and cost

- Duplexer :

Losses and flexibility (?)

- Antennas :

Multi-band (relax filter selectivity) Multi-antennas losses but selective


Some References - Andia, L. et al. Specification of a Polar Sigma Delta Architecture for Mobile Multi-Radio Transmitter - Validation on IEEE 802.16e. Proc. of IEEE Radio and Wireless Symposium, pp. 159-162, Orlando, USA. Jan. 2008. - Baudoin, G. et al. Radiocommunications Numériques : Principes, Modélisation et Simulation. Dunod, EEA/Electronique, 672 pages, 2ème édition 2007. - Baudoin, G. et al. Influence of time and processing mismatches between phase and envelope signals in linearization systems using EER, application to hiperlan 2. Proc Conf. IEEE - MTT'2003 Microwave Theory and Technique, Philadelphia USA, June 2003. - Choi, J. et al. A ΣΔ digitized polar RF transmitter. IEEE Trans. on Microwave Theory and Techniques, Vol. 52, n°12, 2007, pp 2679-2690. - Cox, D. Linear amplification with non-linear components, LINC method. IEEE transactions on Communications, Vol COM-23, pp 1942-1945, December 1974. - Diet, A. et al. Flexibility of Class E HPA for Cognitive Radio. IEEE 19th symposium on Personal Indoor and Mobile Radio Communications, PIMRC 2008, 15-18 september, Cannes, France. CD-ROM ISBN 978-1-4244-2644-7. - Diet, A., Baudoin, G., Villegas, M. Influence of the EER/polar Transmitter Architecture on IQ Impairments for an OFDM Signal. International Review of Electrical Engineering, IREE Praise Worthy Prize, ISSN 1827-6660, V-3 N-2, March-April 2008, pp 410-417. - Diet, A., Baudoin,G., Villegas, M., Robert, F. Radio-Communications Architectures, pp. 1-35 , Chapter n°1 of “Radio-Communications” edited by Alessandro Bazzi, ISBN 978-953-307-091-9, INTECH (SCIYO), 712 pages, april 2010. - Dürdodt, D. et al. A low-IF Rx two-point ΣΔ-modulation Tx CMOS single-chip bluetooth solution. IEEE Trans. MTT, vol. 51, no. 9, pp. 1531–1537, Sep. 2001. - Eloranta, P., Seppinen, P., Parssinen, A. Direct-digital RF-modulator: a multi-function architecture for a system-independent radio transmitter. Com. Magazine, IEEE, V46, I4, 2008, pp 144-151. - Groe, J. A Multimode Cellular Radio. IEEE Trans. On circuits and systems—II: Express briefs, Vol. 55, No. 3, March 2008, pp. 269-273. - Groe, J. Polar Transmitters for Wireless Communications. IEEE Communications Magazine September 2007, pp. 58-63. - Hibon, I. et al. Linear transmitter architecture using a 1-bit ΔΣ. European Microwave Week 2005, Proc. Conf. ECWT, pp. 321-324, Octobre 2005. - Jardin, P., Baudoin, G. Filter Lookup Table Method for Power Amplifier Linearization. IEEE Trans. on Vehi. Tech., N° 3, Vol. 56, pp. 1076-1087, IEEE, Mai 2007. - Jeong, J., Wang, Y. A Polar Delta-Sigma Modulation (PSDM) Scheme for High Efficiency Wireless Transmitters. IEEE MTT-S Int. Microwave Symp. Dig. June 2007. - Kahn, L. Single Sideband Transmission by Envelope Elimination and Restoration. Proc. of the I.R.E., 1952, pp. 803-806. - McCune, E. Polar Modulation and Bipolar RF Power Devices. IEEE Bipolar/BiCMOS Circuits and Technology Meeting (BCTM), October 2005. - McCune, E. High efficiency, multimode, multiband terminal power amplifiers. IEEE Microwave Magazine, March 2005, Volume: 6, Issue: 1, pp: 44- 55. - Murmann, B. Digitally Assisted Analog Circuits – A Motivational Overview. IEEE International Solid-State Circuits Conf.: Special Topic Evening Session, Febr. 2007. - Nielsen, M., Larsen, T. Transmitter Architecture Based on ΔΣ Modulation and Switched Power Amplification. IEEE Trans. on Circuits and Syst. II, 2007, vol. 54, no. 8, pp. 735-739. - Robert, F. et al. Study of a polar ΔΣ transmitter associated to a high efficiency switched mode amplifier for mobile Wimax. 10th annual IEEE Wireless and Microwave Technology Conference, WAMICON, april 2009, Clearwater, FL, USA. - Rode, J., Hinrichs, J., Asbeck, P. Transmitter architecture using digital generation of RF signals. IEEE Radio and Wireless Conf., pp. 245-248, August 2003. - Sowlati, T. et al. Quad band GSM/GSM/GPRS polar loop transmitter. IEEE Journal of Solid-State Circuits, Volume 39, Issue 12, Dec. 2004 Page(s): 2179 – 2189. - Suarez Penaloza, M. et al. "Study of a Modified Polar Sigma-Delta Transmitter Architecture for Multi-Radio Applications", EuMW, 27-31 Octobre 2008, Amsterdam. - Suarez Penaloza, M. et al. "A Cartesian Sigma-Delta Transmitter Architecture", Proc. of IEEE Radio and Wireless Symposium, USA. Jan. 2009. - Villegas, M. et al. Radiocommunications Numériques : Conception de circuits intégrés RF et micro-ondes. Dunod, EEA/Electronique, 464 pages, 2ème édition 2007. - Wendell, B. et al. Polar modulator for multi-mode cell phones. Proceedings of the IEEE 2003 Custom Integrated Circuits Conference, Sept. 2003, pp: 439 – 44. T99


M.

Villegas1,


High efficiency amplifiers

T100

1ESYCOM-ESIEE


M.

Villegas1,


• Power amplifier classification

• Conventional power amplifiers

• Switch mode power amplifiers

• Technologies used for power amplifiers

T101

Power amplifier (PA) classification ? Different classifications in the literature

Conventional mode

Class A

50%

Class AB

Class B

Switching mode

Class C

Class F

78,5%

Class D, E, S

100% Theoritical Efficiency

Linear

Non linear Linearity

T102

Nomadic RF[...] High efficiency amplifiers M. Villegas1, A. Diet2, L. Andia1, F. Robert1,5

Power amplifier classification : linearity versus efficiency Why high efficiency is important ? Increased power consumption Battery cost Electrical power expenses Environmental incentive Deterioration of semiconductor reliability

the final power amplifier dominates the total power consumption Linear

100%

Linearity

Efficiency

78,5% -

A T103

AB

C

F

D ,E ,S


Conventional power amplifier principles ? Some PA characteristics :

Peak output power determined by its saturation

Peak ouput power

PA efficiency maximum close to saturation

Peak efficiency

Operating point near compression induced distortion

Power gain

PA operating point depends on the input signal PAPR

Amplifier linearity

Backoff needed for high PAPR efficiency reduction

Stability

Example : Transmitted WCDMA signal

Average efficiency

1 dB Output power (dBm)

Pout1dB

T104

40

20

RF signal power

Non linear area

RF signal power

Linear area

40

0 -20 -40 -60 2125

2130

2135 2140 2145 Frequency [MHz]

2150

2155

30

20

10

0 50

51

52 53 Time [µs]

54

Peak power: 40.3 dBm Faible Low rendement efficiency

Haut High rendement efficiency

Pin1dB Input power (dBm)

Average power: 30.0 dBm Peak-to-average: 10.3 dB Nomadic RF[...] High efficiency amplifiers M. Villegas1, A. Diet2, L. Andia1, F. Robert1,5

55

Conventional power amplifier characteristics ? Intermodulation

Efficiency

PRFin

η =

PRF ( S ) PDC

G

ηPAE =

P5thOrd_low P3rdOrd_low Pfund_low

PDC PRFout

PRF(S) − PRF(E) PDC AM-AM AM-PM

20 0 -20 -40 -60 -80 -100 -120 -140 -160 -180 -40

-35

-30 -25 -20 -15 -10 RFpower

-5

Most important conventional PA characteristics : Output power at compression point Intermodulation products AM-AM and AM-PM characteristics Efficiency

T105


Conventional power amplifier characteristics ?

Intermodulation product measurement :

Increased power

3th order intermodulation 5th order intermodulation

T106


PSD

Impact of non linearity on system characteristics System main features : ACPR and EVM

R R0

ACPR =

P0 = Padj

∫

DSP (f ).df

P0

∫

DSP (f ).df

Padj

P0

Padj frequency

Example : 16 QAM constellation Bi : Emitted Ai

Ai : Ideal

Ei

EVM Bi

T107

rms

=

Ei : Vectorial Error

1 N

N

∑

A i − C1B i − C 0

i =1

1 N

i =1

∑ =

N

∑

N

2

Ai

2

A i − Si

2

i =1 N

∑

Ai

2

i =1


Impact of non linearity on system characteristics Constant envelope signal t

HPA

f

f

Non constant envelope signal t

HPA

f

f T108


Impact of non linearity on system characteristics

1 dB ba ck -of f

Output power

Edge GSM : signal analysis after AB class amplifier

1dB compression point LINEARITY EFFICIENCY

T109

Input power


Power amplifier classification : current source versus switching mode Conventional mode (A, AB, B, C)

Supply Inductor

Supply

Switching mode (D, E, F) Inductor

charge

Output Network

T110

charge

Output Network


M.

Villegas1,


• Power amplifier classification and characteristics

• Conventional power amplifiers

• Switch mode power amplifiers

• Technologies used for power amplifiers

T111

Peak output power determined by its saturation PA efficiency maximum close to saturation Operating it into compression results in severe distortion

Pout [dBm], PAE [%], Prob. [%]

Conventional power amplifier based on current source Transistor used as variable courant source

The total PA efficiency is weighted by the signal input power probability density function Optimal choice : class AB operation linearity and efficiency trade-off most favorable

30 Pout PAE Prob.

25 20 15 10 5 0 -25

-20

-15 -10 -5 Input power [dBm]

0

5

60 Voltage Current Dissipated power

50 40

Simultaneous voltage and current Dissipation across the device Limits practical efficiency to < 50% How can the voltage×current overlap be minimized ? By using switch mode amplifier

30 20 10 0

0

0.2

0.4

0.6

0.8

1

Time

T112


Conventional power amplifier based on current source Power amplifier class definition : Vgg

Vgg

Z in Input matching network

CDC

CDC

Output matching

network *= ZS Z in

ZS

ZL . .

T113


Conventional power amplifier based on current source Power amplifier conduction angle definition : VDC

2π

2VDC max

Gate voltage

0V D G

Vpinch-off

θ

T114

RF output signal

-VRF

LOAD

RF input signal

S

Resonant circuit


Conventional power amplifier based on current source Class

Theoretical efficiency

Active device

Conduction angle

A

50%

ON 100%period

2π

AB

50 78,5%

ON > 50% period

π 2π

B

78,5%

ON 50% period

π

C

> 78,5%

ON < 50% period