65-nm cmos, w-band receivers for imaging applicationssorinv/papers/kt_cicc_2007_slides.pdf ·...

29
University of Toronto 2007 1 65-nm CMOS, W-band Receivers for Imaging Applications Keith Tang Mehdi Khanpour Patrice Garcia* Christophe Garnier* Sorin Voinigescu University of Toronto, *STMicroelectronics

Upload: dangtuong

Post on 02-Dec-2018

221 views

Category:

Documents


0 download

TRANSCRIPT

University of Toronto 2007 1

65-nm CMOS, W-band Receivers for Imaging Applications

Keith TangMehdi Khanpour

Patrice Garcia*Christophe Garnier*

Sorin Voinigescu

University of Toronto, *STMicroelectronics

University of Toronto 2007 2

Table of Content

• Motivation

• Circuit Schematics

• Fabrication

• Measurement Results

• Conclusion

University of Toronto 2007 3

Motivation• Investigation of W-band receivers in 65-nm GP CMOS

• CMOS might provide alternatives to III-V and SiGe technology in imaging arrays:•

Broadband (multi-GHz) •

Low noise

Low power •

Small area

• Comparison of two LNA feedback topologies

• Series-series feedback with inductor

• Shunt-series feedback with transformer

University of Toronto 2007 4

Receiver Block Diagram

University of Toronto 2007 5

• Inductive (series-series) feedback LNA

• Input matched by LG and LS

• Noise impedance matched by transistor sizing and biasingSGSTIN RRLfZ ++=ℜ π2}{

LNA – Schematic

University of Toronto 2007 6

• Transformer (shunt-series) feedback LNA

• Input matched by LP , LS and M

• Noise impedance matched by transistor sizing and biasingSP

m

PIN LLkM

MgLZ =⋅

≈ℜ ,}{

LNA – Schematic (2)

University of Toronto 2007 7

• S11 , Γopt < -10 dB from 74-100 GHz for both designs

50 60 70 80 90 100FREQUENCY [GHz]

-30 -30

-25 -25

-20 -20

-15 -15

-10 -10

-5 -5

0 0

5 5

10 10

15 15

20 20

[dB

]

S21S11Γopt

NF50NFMIN

50 60 70 80 90 100FREQUENCY [GHz]

-30 -30

-25 -25

-20 -20

-15 -15

-10 -10

-5 -5

0 0

5 5

10 10

15 15

20 20

[dB

]

S21S11Γopt

NF50NFMIN

LNA – Simulation

ind-feedback xfmr-feedback

University of Toronto 2007 8

Mixer – Schematic

• Gilbert cell mixer with inductive broad-banding

University of Toronto 2007 9

Fabrication• 65-nm GP/LP digital CMOS process

• 7 metal layers

• GP n-MOSFETs (80×60nm×1μm) with gate contacted on one side: fT /fMAX =170 GHz/200 GHz at VDS = 0.7 V

• GP MOSFETs 30% faster than LP MOSFETs and require lower VGS and VDS → lower power

• Gate leakage does not affect mm-wave performance

University of Toronto 2007 10

LNA breakouts – Die Photos

IND-feedback XFMR-feedback

IND

XFMR

490 um x 300um (pad) 120 um x 170 um (core)

University of Toronto 2007 11

Mixer breakout – Die Photo

RFIN

LOIN

IFP

IFN

470 um x 560 um (pad) 190um x 160 um (core)

University of Toronto 2007 12

Receiver – Die Photos

IND-feedback Receiver XFMR-feedback Receiver460 um x 500 um (pad) 160 um x 370 um (core)

University of Toronto 2007 13

60 65 70 75 80 85 90 95FREQUENCY (GHz)

-30 -30

-25 -25

-20 -20

-15 -15

-10 -10

-5 -5

0 0

5 5

10 10

15 15

S21

, S11

(d

B)

S21 (sim.)S21 (meas.)S11 (sim.)S11 (meas.)

60 65 70 75 80 85 90 95FREQUENCY (GHz)

-30 -30

-25 -25

-20 -20

-15 -15

-10 -10

-5 -5

0 0

5 5

10 10

15 15

S21

, S11

(d

B)

S21 (sim.) S21 (meas.)S11 (sim.)S11 (meas.)

Meas. LNA – 1st Spin

ind-feedback xfmr-feedback

• Requires 2.2 V VDD for 8 – 9 dB gain

• 4 – 5 dB below simulation

University of Toronto 2007 14

Measurements for 2nd Spin with Modified Layout

Series resistance in ground metallization of LNA was found in the first spin.

A second spin of the design was fabricated with:• Wider metal lines in ground mesh at top level

• Increased number of vias (even between M5 and M6)

• LNA inductance values adjusted to match @ 80 GHz

University of Toronto 2007 15

65 70 75 80 85 90 95FREQUENCY [GHz]

0

2

4

6

8

10

12

14

16

18

20

22

GA

IN [

dB

]

VDD=1.8V (1st

spin)

VDD=2.2V (1st

spin)

VDD=1.2V (2nd

spin)

VDD=1.5V (2nd

spin)

VDD=1.8V (2nd

spin)

65 70 75 80 85 90 95FREQUENCY [GHz]

0

2

4

6

8

10

12

14

16

18

20

22

GA

IN [

dB

]

VDD=1.8V (1st

spin)

VDD=2.2V (1st

spin)

VDD=1.2V (2nd

spin)

VDD=1.5V (2nd

spin)

VDD=1.8V (2nd

spin)

Meas. LNA – 2nd Spin

• Measured gain @ 1.5 V VDD = 13 dB

ind-feedback xfmr-feedback

University of Toronto 2007 16

65 70 75 80 85 90 95FREQUENCY (GHz)

-40 -40

-30 -30

-20 -20

-10 -10

0 0

10 10

20 20

S21

, S11

(d

B)

S21 (sim.)S11 (sim.)S21 (meas.)S11 (meas.)

65 70 75 80 85 90 95FREQUENCY (GHz)

-40 -40

-30 -30

-20 -20

-10 -10

0 0

10 10

20 20

S21

, S11

(d

B)

S21 (sim.)S11 (sim.)S21 (meas.)S11 (meas.)

2nd Spin LNA – meas. vs sims.

• Meas. gain @ VDD = 1.5 V is 1 – 2 dB below sims.

• S11 < -20 dB from 80 – 90 GHz (xfmr-feedback)

ind-feedback xfmr-feedback

University of Toronto 2007 17

50 60 70 80 90 100FREQUENCY [GHz]

-20 -20

-18 -18

-16 -16

-14 -14

-12 -12

-10 -10

-8 -8

-6 -6

-4 -4

-2 -2

0 0

[dB

]

S11S22S21MAG

Meas. Transformer S-params.

• MAG (loss) < -2 dB between 75 – 95 GHz

University of Toronto 2007 18

70 75 80 85 90 95 100RF FREQUENCY [GHz]

0

1

2

3

4

5

6

7

8

9C

ON

VE

RS

ION

GA

IN [

dB

]

VDD=1.5VVDD=1.8VVDD=1.8V (sim.)

IF = 1GHz

Meas. Mixer – Conversion Gain

• 1 – 2 dB below simulation

University of Toronto 2007 19

74 76 78 80 82 84 86 88 90 92 94 96 98RF FREQUENCY [GHz]

0

2

4

6

8

10

12

14

16

NO

ISE

FIG

UR

E [

dB

]

NFMIXER, VDD=1.5VNFMIXER, VDD=1.8V

IF = 1GHz

Meas. Mixer – NFDSB

• Includes ~2 dB transformer loss

• Lowest NFDSB mixer at 80 – 90 GHz in silicon

University of Toronto 2007 20

0 5 10 15 20IF FREQUENCY [GHz]

0 0

5 5

10 10

15 15

20 20

NF

DS

B [

dB

]

VDD=1.2VVDD=1.5VVDD=1.8V

0 5 10 15 20IF FREQUENCY [GHz]

0 0

5 5

10 10

15 15

20 20

CO

NV

ER

SIO

N G

AIN

[d

B]

VDD=1.2VVDD=1.5VVDD=1.8V

Meas. Rx Gain, NFDSB vs IF

• NFDSB ~7 – 8 dB, LO @ 89 GHzXFMR-feedback RCVR

University of Toronto 2007 21

70 75 80 85 90 95RF FREQUENCY [GHz]

0

5

10

15G

AIN

, NF

[d

B]

GAINRCVRNFDSB RCVR

-40

-30

-20

-10

0

10

20

S11

[d

B]

S11

IF = 1GHz

Meas. Rx Gain, NFDSB vs RF

• 3dB-bandwidth: 75- – 91 GHz

University of Toronto 2007 22

-30 -25 -20 -15 -10 -5 0PIN [dBm]

-15

-10

-5

0

5

10

15G

AIN

[d

B]

-15

-10

-5

0

5

10

15

PO

UT [

dB

m]

LO = 75GHzRF = 80GHz

P1dB = -16.2dBm

• LO @ 75 GHz due to equipment limitation

Receiver – P1dB

University of Toronto 2007 23

75 80 85 90FREQUENCY [GHz]

0 0

2 2

4 4

6 6

8 8

10 10

12 12

14 14

16 16

[dB

]

GAIN, VDD=1.5VGAIN, VDD=1.8VNF50, VDD=1.5VNF50, VDD=1.8V

LNA

MIXERRCVRLNA

MIXERRCVRLNA

GFFF

GGG1−

−=

−=Estimated LNA NF

XFMR-feedback LNA

• LNA gain peaks at frequency higher than measured (output pad capacitance removed)

• LNA NF50 ~6 – 7 dB

University of Toronto 2007 24

2nd

Spin

1.2 24 11.1 20 48 6.1 9 – 10< -10

(74-95+ GHz)1.5 34 13.4 30 71 13.6 7 – 8

1.8 48 14.9 45 104 17.7 6 – 7

1st

Spin

VDD

[V]

LNA IF Buffer Receiver

Pdiss

[mW]

Gain

[dB]

Pdiss

[mW]

Pdiss

[mW]

Gain

[dB]

NF

[dB]

S11

[dB]

1.8 38 5.8 47 95 11.6 9 – 10 < -10

(80-95+ GHz)2.2 57 7.8 75 150 13.5 8 – 9

Summary of Results

• Dramatic increase in performance just with better top-level ground mesh and vias

• ~ ½ of Pdiss used in IF buffer to drive 50Ω

off-chip

University of Toronto 2007 25

Conclusion• 74 – 94 GHz receiver with 8 dB NF and 13 dB gain

demonstrated in 65 nm GP CMOS technology.

• Inductive-feedback and transformer-feedback LNA topologies presented:

• Similar performance achieved by different matching procedures

• Layout style significantly affects circuit performance.

• Post-layout simulation at top-level, with ground mesh must be carried out.

University of Toronto 2007 26

Acknowledgement• Katya Laskin for measurements on the second-spin

• Alex Tomkins for inductor and transformer measurements

• Jaro Pristupa and CMC for CAD tools

• Bernard Sautreuil of STM for facilitating the technology access

• CITO for funding

• ECTI, NSERC, CFI and OIF for equipment

University of Toronto 2007 27

• S11 matched at 93 GHz for inductive-feedback LNA (increase LG )

65 70 75 80 85 90 95FREQUENCY [GHz]

-30 -30

-25 -25

-20 -20

-15 -15

-10 -10

-5 -5

0 0

5 5

10 10

15 15

20 20

S11

, S21

[d

B]

VDD=1.2VVDD=1.5VVDD=1.8V

65 70 75 80 85 90 95FREQUENCY [GHz]

-30 -30

-25 -25

-20 -20

-15 -15

-10 -10

-5 -5

0 0

5 5

10 10

15 15

20 20

S11

, S21

[d

B]

VDD=1.2VVDD=1.5VVDD=1.8V

2nd spin LNA – Meas. Gain

IND-feedback LNA XFMR-feedback LNA

University of Toronto 2007 28

-15 -10 -5 0 5

LO POWER [dBm]

-5

0

5

10

15

20

CO

NV

ER

SIO

N G

AIN

[d

B]

VDD=1.8VVDD=2.0VVDD=2.2V

LO = 85GHzRF = 84.5GHz

-15 -10 -5 0 5

LO POWER [dBm]

-5

0

5

10

15

20

CO

NV

ER

SIO

N G

AIN

[d

B]

VDD=1.8VVDD=2.0VVDD=2.2V

VDD=1.2V (2nd

spin)

VDD=1.5V (2nd

spin)

VDD=1.8V (2nd

spin)

LO = 85GHzRF = 84.5GHz

ind-feedback xfmr-feedback

Receiver – vs LO Power

• Requires 2 – 3 dBm (1st spin) and > 5 dBm (2nd spin) LO power

University of Toronto 2007 29

-30 -25 -20 -15 -10PIN [dBm]

-25

-20

-15

-10

-5

0

5

PO

UT [

dB

m]

0

2

4

6

8

10

12

GA

IN [

dB

]

LO=77GHz, RF=75GHz

-30 -25 -20 -15 -10PIN [dBm]

-25

-20

-15

-10

-5

0

5

PO

UT [

dB

m]

0

2

4

6

8

10

12

GA

IN [

dB

]

LO=77GHz, RF=75GHz

IND-feedback Receiver XFMR-feedback Receiver

Receiver – P1dB

• RF at 75 GHz due to equipment limitation