pushing the envelope of PA efficiency

ET Envelope Pathfrom digits to PA

Gerard WimpennyNujira Ltd

ARMMS Conference19th/20th November 2012

2

Agenda

• Envelope Processing• ET PA Characterisation• Isogain shaping• CFR shaping

• Envelope Amplifier Design Requirements• Sources of Impairment• Integrated Modulator• Distributed Modulator

3

ET System Anatomy

Envelope detection: mostaccurate if performed in digitaldomain

Envelope shaping: Determinesrelationship between RF powerand PA supply voltage

Envelope Amplifier: High BW,Low Noise, High efficiencyAmplifier used to generate PAsupply voltage

ET PA: ET can be applied tostandard fixed supply PA.Improved performance possible byoptimising PA for ET operation

Delay Alignment: ET requires accurate (~ns) timingalignment between envelope and RF paths. Most accurate /repeatable if performed in digital domain

4

ET PA System Principles

0

10

20

30

40

50

60

70

15 20 Pout (dBm) 25 3035

Effic

ienc

y (%

)

4.5V4.0V3.5V3.0V2.5V2.0V1.5VISO26dBFixed

Fixed Supply

In compressed region O/Ppower is determined bysupply voltage – RF inputpower has little influence

In linear region O/P poweris determined by RF inputpower – supply voltage haslittle influence

In transition region O/Ppower is determined byboth supply voltage and RFinput power

5

-1.5

-1

-0.5

0

0.5

1

1.5

Envelope Processing Basics

• Swing Range• Optimise efficiency of combined

modulator /PA• Prevent gross PA nonlinearity due

IV curve ‘knee’• Envelope ‘Shaping’

• Control envelope bandwidth• Optimise efficiency• Can be used to linearise PA

• Timing Alignment• Timing error leads to ‘memory

effect’ (AM-PM)• Fine adjustment necessary (~1ns)

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5

-1

-0.5

0

0.5

1

1.5

6

PA Characterisation Methods

Testmethodology

PA currentmeasurement

Supplyimpedance

Supplybandwidth

requirements

ETEfficiencyprediction

ET Linearityprediction

Parametersmeasured

SweptCW testing Bench PSU

Low(decouplingCapacitor)

Low (BenchPSU)

Poor, due toPA dieheating

Poor, due toPA dieheating

Gain (AM:AM),Efficiency

Pulsed RF /DCtesting

Instrumentationgrade current

probe,~5 us resolution

Low(decouplingCapacitor)

Low (BenchPSU)

Good, if shortpulses (~10

us, 10% dutycycle).

Fair(if device haslow AM/PM)

Gain (AM:AM),Efficiency

Dynamicsupply

modulation

Challenging –high BW withhigh commonmode voltagecurrent sense

Requires lowimpedance

dynamic supply(no decoupling)

High(~60 MHz BW) V. Good

V. Good(if device haslow memory

effects)

Gain (AM:AM),Phase (AM:PM),

Efficiency

Phase measurement possible in principle – but accuracy poor dueto heating effects and phase reference ‘wander’

No phasemeasurement

PA Characteristics must be known to determine Shaping table

7

AM/PM Input Surfaces

Input Gain Surface Input Phase Surface

Fixed Supply VoltageGain contours

Fixed Supply VoltagePhase contours

8

Isogain Contours

25dB IsoGainshaping contour

27dB IsoGainshaping contour

Low voltage phasecollapse

Phase peak flatteningc.f. fixed supply

Input Gain Surface Input Phase Surface

9

Isogain Shaping Functions

Isogain Shaping Functions

27dB gain 25dB gain

Input Gain Surface

10

Useful 2D Slices - Efficiency

Output Efficiency Surface Output Efficiency locus

25/27dB IsoGaincontours

Fixed Supplycontours

11

2D Slices AM/AM, AM/PM

AM/AM distortion removed Residual PM distortion

Output Gain Output Phase

12

Predicted Performance

ACPR

AM/PM

AM/AM

Predicted Efficiency = 67.7%

Waveform = HSUPA / 5.4dB PAPRShaping = Isogain 24dB

13

Measured Performance

ACPR

AM/PM

AM/AM

Predicted Efficiency = 67.7%Measured Efficiency = 67.6%

Waveform = HSUPA / 5.4dB PAPRShaping = Isogain 24dB

14

Shaping Table based CFR

Isogain shaping modified to introduce softclipping e.g using ‘Rapp’ function (VCFR)

Desired PA gain profile p p

Vpk

iso

V isoV CFR =

V1

1 ÷÷

ø

ö

çç

è

æ

÷÷ ø

öçç è

æ+ V Limiting voltage

factorSmoothnessp =

pk =

Unmodified Isogainshaping (Viso)

Envelope AmplifierMax Vout

15

0 1 2 3 4 5 6 710-4

10-3

10-2

10-1

100

dB above mean

Pro

babi

lity

0 1 2 3 4 5 6 710-4

10-3

10-2

10-1

100

dB above mean

Pro

babi

lity

Increased Pout using CFR

5 10 15 20 25 300

10

20

30

40

50

60

70

80

RF Output power (dBm)

Effi

cien

cy (%

) 6.4dB

4.9dB

Output Signal Statistics

Isogain shaping CCDF

Modified Isogainshaping CCDF

Increased Mean Power

LTE Power pdf

Controlled use of CFR allows Increased mean power andefficiency for given PA device periphery

Increased Efficiency

16

-200 1 2 3 4 x 10

7

-10

0

10

20

30

40

50

60

70

80

Offset from Carrier (Hz)

PS

D (d

B)

Reference Spectrum(No clipping)

‘Software Defined PA’RF Spectrum

0 2 4

x 107

-20

0

20

40

60

80

0 2 4

x 107

-20

0

20

40

60

80

0 2 4 x 107-20

0

20

40

60

80

0 2 4

x 107

-20

0

20

40

60

80

0 2 4

x 107

-20

0

20

40

60

80

0 2 4

x 107

-20

0

20

40

60

80

0 2 4 x 107-20

0

20

40

60

80

0 2 4

x 107

-20

0

20

40

60

80

Hard clippingSoft clipping

75% Clipping level(2.5dB CFR)

85% Clipping level(1.4dB CFR)

p=4 p=6 p=10 p=100

Shaping Table based CFR allows dynamic configuration of PA’sPower / ACPR / Efficiency characteristics

Offset from Carrier (Hz)

PS

D (d

B)

17

Agenda

• Envelope Processing• ET PA Characterisation• Isogain shaping• CFR shaping

• Envelope Amplifier Design Requirements• Sources of Impairment• Integrated Modulator• Distributed Modulator

18

Envelope AmplifierRequirements

High Bandwidth(e.g 4ch WCDMA, 20MHz LTE, 2x 10MHz WiMAX)• Envelope Bandwidth ~3x RF Bandwidth• Cannot be achieved with ‘switcher only’

architectureLow Noise / Distortion

• Required to meet ACPR specifications• Many factors to consider• Requires high Tracking Accuracy

High Efficiency• Must consider combined PA / modulator

efficiency• Linear supply would be pointless

Power• Must maintain BW and Noise at

increased power levels

Low Noise

HighBandwidth

HighEfficiency

Power

19

ET Impairment Categories

• System (Env & RF paths)• RF/Env Delay match• RF/Env Gain match• PA AM/AM and AM/PM

• RF Path• Noise

• Thermal• Quantisation

• Linearity• PA Memory effects

• Bias• Thermal

• Envelope Path

20

Envelope Path Impairments

• Shaping Accuracy• Tracking Accuracy

• Noise• DAC Quantisation• Env Amp Thermal• Switcher breakthrough• Linear Amp PSRR

• Frequency Response• Amplitude• Group Delay flatness

• Env Amp Distortion• Harmonic• Crossover

• Env Amp to PA Interaction• Env Amp Output Impedance• PA Interconnect Impedance• PA Non Linear Load Impedance

21

Tracking Accuracy Explained

• The difference between ideal and measured supply waveform after removalof DC offset, gain and timing errors

• Analogous to EVM for modulated signals• Tracking error analysis is useful diagnostic tool: RMS, Peak, Spectrum

0

10

20

30

40

50

60

V

-10-8-6-4-202468

10

V

0

10

20

30

40

50

60

V

-10-8-6-4-202468

10

V

-10-8-6-4-202468

10

V

-10-8-6-4-202468

10

V

0

10

20

30

40

50

60

V

0

10

20

30

40

50

60

V

Gross gain and timing errorGross timing errorSmall timing errorResidual modulator tracking error

Ideal and measured waveforms Tracking Error

22

Supply ‘Noise’ – RF Conversion

• PA in compression – Supply Noise & Distortion modulates RF carrier• PA can be considered as mixer

• O/P spectrum is convolution of Supply and PA input Spectra

• Conversion factor (Supply Sensitivity) for noise on supply to RFsidebands is similar to ideal AM modulator (mixer)

0

-10

-20

-30

-40

-50

-60

-70

-80

-90

-100Start: 0 Hz Stop: 200.0000 MHz

12-1

0

-10

-20

-30

-40

-50

-60

-70

-80

-90

-100Start: 1.8500 GHz Stop: 2.0500 GHz

1

2-1

3 4-35 6-5

40MHz ‘test tone’ added to Envelope Amplifier O/P(whilst amplifying 5MHz WCDMA signal)

Corresponding RF sidebands

40MHz

40MHz

Supply Spectrum RF Spectrum

23

Measured Supply Sensitivity

Average DC drain voltage 2.62VMeasured 40MHz injected tone level 17.3mV rmsCalculated RF sideband level for ideal AM modulator -49.6dBCMeasured RF sideband level -51dBCPA Supply Sensitivity (dB) -1.4dBPA Supply Sensitivity (%) 85%

An ideal AM modulator is described by:

where modulation index

This can be re-expressed in terms of carrier and LSB and USB components

where for an ideal AM modulator

env

env

rf

rf

VDVV

DV

env

env

VDV

rf

rf

VDV

)sin()]cos([)y(t = A+ M wmt wct

AMh =

)])sin(())[sin(()sin()y(t = A wct + R wc +wm t + wc -wm t

2MR =

24

Integrated Modulator Example–Coolteq.L

• Slow switching Buck converterprovides LF power

• Fast switching multilevel converterprovides HF power

• Error Amplifier ‘cleans up’ output

• Boost and Buck capable• Battery depletion resilience• Increased PA peak Power

25

Distributed Modulator Example- Coolteq.u

RF S

PLITTER

RF C

OM

BIN

ER

Coolteq.uPowerSupplyModule(PSM)

PA

HAT

PA

HAT

PA

HAT

PA

HAT

RFDriver

RF out400-475W@7.5 dB

PAPR

RF in

Envelopesignal

generation

DigitalPre

Distortion

RFUpconvert

Exciter

Modulation

Envelopeinput

DCinput

4 x Coolteq.u High Accuracy Tracking module (HAT®) • Scalable O/P Power• Allows multiple PAs per

Power Supply Module(PSM)

• Allows Envelope pathLinear Amplifier to beplaced close to PA

• PA supply impedanceminimised

26

Conclusions

• Understanding of PA characteristics key toachieving good ET performance.

• Careful selection of shaping table contentsallows optimisation key ET system performancemetrics

• ET is a simple concept, but attention must bepaid to multiple potential sources of impairmentto realise full potential

pushing the envelope of PA efficiency