chalmers university of technology

gan-on-Silicon Efficient mm-wave euRopean systEm iNtegration plAtform

The SERENA project has received funding from the European Union’s Horizon 2020 research

and innovation programme under grant agreement No 779305.

Christian Fager

Chalmers University of Technology

[email protected]

• Chalmers (present and past) T. Eriksson, K. Buisman, K. Rasilainen,

K. Hausmair, P. Taghikhani, E. Baptista

• Ericsson U. Gustavsson, P. Landin, K. Andersson

• OMMIC M. Marilier, R. Leblanc, A. Gasmi, M. ElKaamouch

• Fraunhofer IZM I. Ndip, U. Maaß

17 January, 2020 2

Dr. U. Gustavsson Dr. P. Landin

Prof. T. Eriksson Dr. K. Hausmair Dr. K. Buisman E. BaptistaP. Taghikhani Dr. K. Rasilainen

Dr. K. Andersson

Multi-physics simulation of mm-wave systems

• Active antenna arrays

• High integration needed to fit within l/2

3

Chip embedding

(IZM/SERENA)

Millimeter waves

Data

Heat

Energy

Vertical stacking IBM / Ericsson

17 January, 2020 Multi-physics simulation of mm-wave systems

• Multi-physical effects

Electrical

Thermal

Mechanical

…

• Efficient simulations and modeling are crucial

Before: Optimization, reliability, margins, time to market

After: Troubleshooting, performance optimization, reverse engineering

17 January, 2020 4Multi-physics simulation of mm-wave systems

Active antenna arrays

Antenna-circuit interactions

Linearity

5

MIM

O

tran

smitt

er

6

[T. Svensson, Chalmers]

𝐲 = 𝐇𝐱 +𝐰

Transistor Circuit SystemRF Sub-system


• Traditionally 50W assumed

• Single-input-single-output modeling of RF components

7

PA

RL = 50 W

Vin[n] Vout[n]Vout = fNL(Vin)

SISO model

Vout = f(Vin)


• Integrated transmitters

Mismatch and mutual coupling

Non-50W interfaces

• Dual-input models needed:

b2 = f(a1,a2)

8

GL≠50W


• “PHD” or “X-parameter®” model


( ) ( ) ( )

1 2 2

1 2 2

1

1 2 2

22 121 1 22 1

2( 1) 2( 1) 2( 2)2 *

2 1 1 2 1 2 2 1 1 2

1 1 2

P P Pp p p

p p p

p p p

T aS a S a

b a a a a a a a − − −

= = =

= + +

b2 = fNL(a1,a2)a1

b2

a2

MISO PA model

Nonlinear

50W SISO-termsNonlinear mismatch terms

[Verspecht and D. E. Root, “Polyharmonic distortion modeling,” IEEE Microw. Mag., vol. 7, no. 3, pp. 44–57, Jun. 2006.]

• “X-parameters®” with memory


( )

( )

( )

1 11

1 1

21 1

2 1 22

2 1 2

22 1

1 22

2 1 2

22 1

2( 1)

2 1, 1 1 1 1 1

0 1

2( 1)

1, 2, 2 1 1 2 2

0 0 1

2( 2)2

1, 2, 2 1 1 1 1

0 2

M Pp

m p

m p

S a

M M Pp

m m p

m m p

S a

M Pp

m m p

m m p

T a

b n a n m a n m

a n m a n m

a n m a n m

−

= =

−

= = =

−

= = =

= − − +

− − +

− −

2

*

2 2

0

M

a n m−

b2 = fNL(a1,a2)a1

b2

a2

MISO PA model

[C. Fager et al., "Prediction of Smart Antenna Transmitter Characteristics Using a New Behavioral Modeling Approach," Proc. IMS, 2014]

11

[C. Fager et al., "Prediction of Smart Antenna Transmitter Characteristics Using a New Behavioral Modeling Approach," Proc. IMS, 2014]

Input ref. plane

Outputref. plane

PA DUT

Simulation based

PA DUT

Measurement based

[S. Gustafsson et al., "A Novel Active Load-pull System with Multi-Band Capabilities," ARFTG, 2013]

PA DUT


User1 data

User2 data

UserK data

Channel state

b2 = f1(a1,a2,T)a1,1

a1,2

a2,1

b2,1

a2,2

b2,2

a2,N

b2,N

a1,N

b2 = f1(a1,a2,T)

b2 = f1(a1,a2,T)

Pre-codingDPD

...

Signal processing Circuit behavioral modeling

Antenna modeling

User1

User2

UserK

17 January, 2020 Multi-physics simulation of mm-wave systems 12

( ) ( ) ( )

1 2

1 1

*

2 21 1 1 22 1 2 22 1 2

2 2

NT

jj j

ant

a e e e =

= + +

=

a

b S a a S a a T a a

a S b

( ) ( ) ( ) * *

2 21 1 1 22 1 2 22 1 2ant ant= + +b S a a S a S b T a S b

Antenna couplingand mismatch effects

Regular 50Wnonlinear distortion

( ) ( )2,

1

, ,N

tot i i

i

E El Az n b n E El Az=

=

Etot(Elevation, Azimuth)

Far field radiationAntenna circuit interactions

Beam steering


• PA model extracted from IC CAD

• Antenna parameters from EM CAD

• Each PA perfectly linearized for 50W load (a2 = 0)

• Ideal phased array beam steering. No amplitude tapering

IC design 64 element antenna array

[C. Fager et al., "Analysis of Nonlinear Distortion in Phased Array Transmitters," Proc. INMMiC, 2017]


• Distortion highly direction dependent. Why?

User EVM vs. scan direction

Azimuth

a1(input)

b2

28 GHz Freq.

|a2|

Freq.

|a1|

Freq.

|b2|

28 GHz

28 GHz

WCDMA2BW = 3.84 MHz

a2

28 GHz GaN MMIC PA BW = 30.72 MHz

WCDMA1BW = 3.84 MHz

a1(input)

b2

28 GHz Freq.

|a2|

Freq.

|a1|

Freq.

|b2|

28 GHz

28 GHz

WCDMA2BW = 3.84 MHz

a2


WCDMA1BW = 3.84 MHz

a1(input)

b2

28 GHz Freq.|a

2|

Freq.

|a1|

Freq.

|b2|

28 GHz

28 GHz

WCDMA2BW = 3.84 MHz

a2


WCDMA1BW = 3.84 MHz

Elevation

Ele

vati

on

Azimuth

User


• Significant variation of PA load impedance vs. beam steering

Central elementGL Edge element

GL vs. scan direction


• Distortion averaging effects happening inside the array

Low distortion direction

Ga

in [d

B]

Output power [dBm] ColumnRow

Gain compression/expansionAM/AM for each of the 64 branches

DG

ain

[dB

]


• Distortion addition for some directions

• Direction dependent user distortion → Direction dependent DPD needed

High distortion direction

Gain compression/expansionAM/AM for each of the 64 branches

Ga

in [d

B]

Output power [dBm] ColumnRow

DG

ain

[dB

]


• Nonlinear effects observed also in mm-wave measurements

• 30 GHz beamforming chip

RF input

Digital

phase controlAzimuth

Elevation

AM-AM compression measurement


• Load-pull based nonlinear modeling

• 4 element patch antenna: EM simulation based modeling

• Reasonable modeling accuracy

RF input

Digital

phase controlAzimuth

Elevation

AM-AM compression modeling


Nonlinear PA–antenna interactions Far field distortion

4×Skyworks PAs

2 channel RF generator

2 channel RF generator Antenna array

1-path TX 4-path TX

[K. Hausmair et al. “Prediction of Nonlinear Distortion in Wideband Active Antenna Arrays,” IEEE T-MTT, 2017]


• Framework for efficient simulation of

active antenna systems

• Improved understanding of circuits-

antenna interactions with realistic signals

• New nonlinear effects predicted in phased

array and MIMO systems


Thermal modeling

Power dissipation modeling

mm-wave transmitter example

23

• Heat concentration in active antenna arrays

• Chip level heating effects Thermal coupling Efficient power amplifiers

• System level effects Performance degradation Reliability


• T(t) = TA + Pdiss(t)*zth

• Thermal admittance: Yth = Gth + jwCth

Cth Rth

TA

p(t)

YthT(t)

Pdiss(t)

Pdiss(t) = Pdiss(|a1(t)|,T(t))

Pin Pout

Pdiss

Thermal model RF Model


• Envelope time-stepped solution for T(t)

Cth Rth

TA

p(t)

YthT(t)

Pdiss(t)

Pdiss(t) = Pdiss(|a1(t)|,T(t))

Pin Pout

Pdiss

Thermal model RF Model


( ) ( )( )1

1 amb th th diss, th amb2 2n s n s nT f f T −

+ = + + + −T G C P C T

User1 data

User2 data

UserK data

Channel state

27

b2 = f1(a1,a2,T)Pdiss = f2(a1,a2,T)

a1,1

a1,2

a2,1

b2,1

a2,2

b2,2

a2,N

b2,N

a1,N



T1 TN

T = Zth*Pdiss+Tamb

Pre-codingDPD

...

TambHeat

dissipationgeneration dissipation

Thermal coupling

[C. Fager et al. “Analysis of Thermal Effects in Active Antenna Array Transmitters…,” Proc. INMMiC, 2015]


( ) ( ) ( )( )1

1 amb th th diss, 1, th amb2 , 2n s n n n s nT f f T −

+ = + + + −T G C P a T C T

1,n n=a Gx

( ) ( ) ( ) * *

2, 21 1, 1, 22 1, 2, 22 1, 2, 1| |, | |, ,n n n n n n ant n n n ant n n−= + + +b S a T a S a T S b T a T S b ρ

( ) ( )2,, ,T

n nE = b E



Thermal simulation results

• One GaN chip in a simplified package environment

17 January, 2020 31

Chip dissipated

power

Ambient

temperature

Thermal RC constants

Thermal response @ 7W step


• Temperature dependent gain


• Dissipated power vs. temperature


• Mismatch and mutual-

coupling neglected in model


Simulated S-parameters

• Embedded element patterns

for unity excitations


Far-field radiation patterns

3617 January, 2020

Thermal modeling RF circuit modeling Antenna modeling


• PA input-/output spectrum

for modulated signals

• PA-to-PA nonlinear

interactions


Power amplifier output

• Thermal transients

• On-off switching, e.g.

between T/R in TDD systems

• Heat spreading in arrays

17 January, 2020 38

Power amplifier temperature



[ = 90°, = 0°] [ = 90°, = -25°]

• Modulated fields in both polarizations

considering both PA, thermal, antenna and beam-

forming settings


• Thermal effects at circuit- and package

levels

• Framework for electro-thermal

transmitter simulations

• Prediction of joint electrical- and thermal

effects in mm-wave antenna arrays


• mm-wave transmitter design is cross-disciplinary Co-design between signals, circuits & antennas

• New nonlinear distortion phenomena Both circuit and array level linearization needed

Low complexity MIMO linearization proposed

• Power dissipation a great challenge Thermal coupling effects

• Understanding linearity and power dissipation effects through accurate multi-physicssimulations is critical for successful 5G system design


C E N T R E

Acknowledgments

Simulation of RF Systems

[1] C. Fager, X. Bland, K. Hausmair, J. Chani Cahuana, and T. Eriksson, "Prediction of Smart Antenna Transmitter Characteristics Using a New Behavioral Modeling Approach,"

Proc. IEEE International Microwave Symposium, Tampa, FL, 2014.

[2] K. Hausmair, S. Gustafsson, C. Sanchez-Perez, P. Landin, U. Gustavsson, T. Eriksson, C. Fager, "Prediction of Nonlinear Distortion in Wideband Active Antenna Arrays,"

IEEE Trans. Microwave Theory Tech., vol. 65, pp. 4550-4563, 2017.

[3] C. Fager, K. Hausmair, K. Buisman, D. Gustafsson, K. Andersson, and E. Sienkiewicz,

"Analysis of Nonlinear Distortion in Phased Array Transmitters," Proc. INMMIC, Graz, Austria, 2017.

[4] C. Fager, T. Eriksson, F. Barradas, K. Hausmair, T. Cunha, and J. C. Pedro, “Linearity and Efficiency in 5G Transmitters: New Techniques for Analyzing Efficiency, Linearity,

and Linearization in a 5G Active Antenna Transmitter Context,” IEEE Microw. Mag., vol. 20, no. 5, pp. 35–49, 2019.

Electro-thermal simulations

[5] C. Fager et al., "Analysis of Thermal Effects in Active Antenna Array Transmitters …," Proc. INMMiC, Oct. 2015]

[6] E. Baptista et al., "Analysis of Thermal Coupling Effects in Integrated MIMO Transmitters," Proc. IMS2017

[7] F. Besombes et al., “Electro-thermal Behavioral Modeling of RF Power Amplifier taking into account Load-pull Effects for Narrow Band Radar Application,” Microw.

Integr. Circuits Conf. (EuMIC), 2011 Eur., no. October, pp. 264–267, 2011.

[8] V. D’Alessandro, M. De Magistris, A. Magnani, N. Rinaldi, and S. Russo, “Dynamic electrothermal macromodeling: An application to signal integrity analysis in highly

integrated electronic systems,” IEEE Trans. Components, Packag. Manuf. Technol., vol. 3, no. 7, pp. 1237–1243, 2013.

[9] SERENA (gan-on-Silicon Efficient mm-wave euRopean system iNtegration platform), https://serena-h2020.eu/


https://serena-h2020.eu/

“The SERENA project has received funding from the European Union’s Horizon 2020 research and innovation programme

under grant agreement No 779305.”

If you need further information, please contact the coordinator:TECHNIKON Forschungs- und Planungsgesellschaft mbH

Burgplatz 3a, 9500 Villach, AUSTRIATel: +43 4242 233 55 Fax: +43 4242 233 55 77

E-Mail: [email protected]


The information in this document is provided “as is”, and no guarantee or warranty is given that the information is fit for any particular purpose.

The content of this document reflects only the author`s view – the European Commission is not responsible for any use that may be made of

the information it contains. The users use the information at their sole risk and liability.

mailto:[email protected]

chalmers university of technology

Documents