1. The principle of resistive mixers -...

IEEE MMIC seminar Bergen 2006 10 06 H Zirath

1. The principle of resistive mixers

Frequency translation by using resistive mixing

i(t)=g(t) v(t)

The current through the (timedependent) resistor is the product of the conductance and the applied voltageWe are interested in the frequencies generated by this multiplication in the time domain

g(t) can be represented by a Fourier serieswith the fundamental frequency ωp

v(t)

+

-

g( t) = Gnn =− ∞

∞

∑ ⋅ exp( j ⋅n ⋅ ω p ⋅ t)See Steven MaasNonlinear Microwave Circuitspp.114-137

g(t)


small-signal analysisConvenient representation of mixing frequencies ωn=ω0+n ωp

0 ωp−ωp 2 ωp 3 ωp−2 ωp−3 ωp

ω0 ω1 ω2 ω3ω−1ω−2ω−3

Assuming we know the variation of the resistance vs time, we can calculatethe frequency conversion properties by using the conversion matrix approach:

I = G . V G is the conversion matrix, I and V are vectors representingthe current and voltage at each mixing frequency

V(ω) I(ω)


I−N*

I−N+1*

I−N+2*

I−1*

I0

I1

IN−1

IN

⎡

⎣

⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢

⎤

⎦

⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥

=

G0 G−1 G−2 G−2N

G1 G0 G−1 G−2N+1

G2 G−1 G0 G−2N+2

G0

G0

G0

GN−1 GN−2 GN−3 G1 G0 G−1 G−N−1

GN GN−1 GN−2 G3 G2 G1 G0 G−1 G−2 G−3 G−N

GN+1 GN GN−1 G1 G0 G−1 G−N+1

G0

G0

G0

G0

G0

G2N G2N−1 G2N−2 G0

⎡

⎣

⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢

⎤

⎦

⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥

•

V−N*

V−N+1*

V−1*

V0

V1

VN−1

VN

⎡

⎣

⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢

⎤

⎦

⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥

we get:


We want to know how the mixer properties depend on the characteristic of the conductance waveform and the embedding impedance

The conversion matrix can be extended by reactive components, forming a conversion matrix in admittance form. This is a valuable tool when analysing the properties of diode and transistor mixers. We have now transfered the nonlinear problem to the frequency domain and can use linear equations for solving the problem.

port1

port2

port3

port4

port5

port6

Y

ω0

ω1

ω-1

ω-2

ω2

ω3

etc


non-square waveform, harmonic response

To illustrate the harmonic response, lets look at Gn as a function of the duty-cycle for a simple case, an ideally switched conductance

g(t) = Gnn =− ∞

∞

∑ ⋅ exp( j ⋅n ⋅ ω p ⋅ t)Gct0

T

Gc

0

We obtain: G0 = t0T

⋅GC , G1 = GC

π⋅ sin(π ⋅ t0

T)

G2 =GC

2π⋅sin(2π ⋅

t0T

) , G3 =GC

3π⋅ sin(3π ⋅

t0

T)

G4 =GC

4π⋅sin(4π⋅

t0

T) , G5 =

GC

5π⋅sin(5π⋅

t0T

)

Gn =1T

g(t )⋅ exp(−j ⋅n ⋅ω ⋅ t)−

T2

T2

∫


The most used device for the resistive mixer is the Schottky diode, it can be used up to THz-frequencies.

The FET-based resistive mixer is an interesting alternative if1. generation of intermodulation products is critical2. FET/HEMT based MMIC designs are considered3. LO and RF are close together and large bandwidth is required

+

-

VD

ID = I0 ⋅eq⋅Vd /ηkT

rB =∂ID

∂VD

⎛

⎝ ⎜ ⎞

⎠ ⎟

−1

→ηkTqID

or rB =k ⋅e−q⋅Vd / ηkT

IDS = K(VGS)⋅ 1+λ ⋅VDS( )⋅ tanh(α⋅VDS)

the diode and FET/HEMT as a variable resistance


0

0.01

0.02

0.03

0 0.5 1 1.5 2

Ids-Vgs characteristic InP-HEMT (InP15)Lg=0.1 um, Lw=100 um

Dra

in c

urre

nt (A

)

Drain-source voltage (V)

= =

The FET can be used as a variable resistance, controlled by the gate-voltageThe LO is applied to the gate, RF and IF to the drain, Vds≈0 VThe Id-Vd characteristic is approximately linear at low Vds which makes the FET-resistive mixer linear ! It's exponential for a diode !

linear resistorcontrolled by Vgs

2. FET-HEMT models for resistive mixers


The Rds-Vgs slope is important for the determination of the LO-power requirement of the mixer

10 -8

10 -7

10 -6

10 -5

0.0001

0.001

0.01

-1.5 -1 -0.5 0 0.5 1 1.5

Ids vs Vgs at small Vds -50 mV

Ids (

A)

Vgs (V)

10

100

1000

10 4

10 5

10 6

-1 -0.5 0 0.5 1

Rds vs Vgs: InP15, Lw=100 um

Rds

(ohm

)

Vgs (V)


In order to model the mixer performance, an empirical model describingthe Rds-Vgs characteristic is useful. The example shows an InP-HEMT

1

2

3

4

5

6

-1 -0.5 0 0.5 1

Model fit, model 1 and 2

Log

(Rds

)

Vgs (V)

Rmin=23, Vk=0.01alfa1=15, alfa2=0.7, alfa3=-8.0

Rmin=23.5, Vk=0.08, alfa1=-21.3

1

2

Model 1:

Model 2:

Empirical model equations :1. RDS = RMIN ⋅ (exp(α(VGS − VK )) + 1)2. RDS = RMIN ⋅ (exp(Ψ) + 1)

Ψ = α1 (VGS − VK ) + α 2 (VGS − VK )2 + α 3(VGS − VK )3

InP-HEMT


Cgd

Cgs

Rgrch

Rgs Cds

Rd

Rs

Gate

Drain

Source

Rgd

1

2

3

4

0

10

20

30

40

50

60

70

-2 -1,5 -1 -0,5 0 0,5 1

pm3, Lw=100 um

Log

(Rds

)

Cgs, C

gd (fF)

Vgs (V)

tanh-fit:Cgs=42+20 tanh(1.8(Vgs+0.41)

exp-fit

For 'small signal' simulations this equivalent circuit is usedrch and cgs,cgd are the most important bias-dependent parameters(Rds is now split into Rd+rch+Rs)

The equivalent circuit including capacitances


3. The single-ended mixer (1)

LOmatchLO

RF

IF

Vg

Performance of different FET-HEMT based resistive mixers

This mixer is simple since no hybrids/filtersare needed for the LO-injection->high BW. The LO-RF-isolation is of the order 10-20 dBdepending mainly on the transistor design.The total conversion loss is less than 6 dBThe required LO-power is typically 0-10 dBmdepending mainly on the transistor type

Q-band-mixer (50 GHz) with WG-input:


The single-ended mixer (2) comparison between different HEMTs

100-10-20-30-405

10

15

20

25

30

PM3

InP9

PM2

LO=39 GHz; RF=43 GHz

LO-power (dBm)

Con

vers

ion

loss

(dB

)

4846444240386

8

10

12

14

PM3

In9

PM2Plo=9 dBm

Freq. (GHz)

Con

vers

ion

loss

(dB

)

Three different HEMTs where compared in a 40 GHz MIC-mixer:

1. Single delta doped AlGaAs-InGaAs-GaAs HEMT. (PM2)2. Double delta doped AlGaAs-InGaAs-GaAs HEMT. (PM3)3. Single delta doped AlInAs-GaInAs-InP HEMT. (InP9)


MMIC mixer

Resistive FET-mixers have been realized up to F-band, this monolithic mixer is based on an AlInAs-GaInAs-InP HEMT-technology (including vias) and made at CTH.The conversion loss is ≈ 9 dB at 4 dBm LO-power. Frequency is 115 GHz.

LO

IF

RF

Vg-bias

Local oscillator power (dBm)

Lc (dB)

010

20

10

0-10

MEASMEAS

SIMSIM


PowersplitLO

RF

Vg

0 deg

180 deg

IF

IF

Balanced mixers

In applications where high LO-RF isolation is required, a balanced mixer can be used.The LO is splitted with 180 deg phase difference between the two ports. The residualLO at the drains will be minimized.An LO-RF isolation of better than 35 dB and a total conversion loss of better than 10 dBwas obtained for this balanced 50-60 GHz upconverting mixer with 5 dBm LO-power.

Block diagram:55 GHz Alumina MIC-mixer

Lo Rf-out

IF-in

IF-in


60 GHz resistive mixer

This mixer is based on the principle of resistive mixing i e the Local Oscillator power is applied to the gate causing the drain-source resistance to be time-dependent. The RF and IF signals are applied/ extracted to/from the drain through a diplexer. The mixer has usually good intermodulationproperties.

LO-in RF-in

IF-out

MMIC implementation Process D01PH and D01MH


Conversion Loss vs. LO power

0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

18.00

-10.00 -8.00 -6.00 -4.00 -2.00 0.00 2.00 4.00 6.00 8.00 10.00

LO Power [dBm]

CL

[dB

]

Measured results 50-62 GHz:


Conversion Loss vs. Vgs

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

50.00

-1.00 -0.80 -0.60 -0.40 -0.20 0.00

Vgs [V]

CL

[dB

]

PLO= -5 PLO= -3 PLO= -1 PLO= 1 PLO= 3 PLO= 5 PLO= 7PLO= 9PLO= 11


Balanced resistive mixer topology-LO is fed to mixers with a 180 deg phase difference->>>>-The residual LO will cancel at the drain-A Marchand balun for broadband operation...

MMIC implementation Process D01PH and D01MH


Short form data:30-60 dB LO->RF isolation, IIP3 =19dBm at 8 dBm PLO5-8 dB NF (DSB), P-1 ≈ PLO+(2-4dB)

0

2

4

6

8

10

12

14

16

18

20

-4 -2 0 2 4 6 8

Plo (dBm)

IIP3

(dBm

)

pHEMTmHEMT

0,00

5,00

10,00

15,00

20,00

25,00

30 35 40 45 50

RF Frequency (GHz)

NF_

DSB

(dB)

-0.1 -0.3

-0.5 -0.7

-0.9

NFminIIP3

0

5

10

15

20

-5 0 5 10Lo-power (dBm)

LO-RF isolation

20

25

30

35

40

45

50

55

60

65

30 35 40 45 50

Frequency (GHz)

Isol

atio

n (d

B)

LO-RF isolation

No source biasNo source bias


Image reject mixer (IRM)Three designs in WINs pHEMT/mHEMT technologies

Topology:

Chip photo, pHEMT,

3.3 × 2.8 mm2

Chip photo, mHEMT,

3.3 × 2.3 mm2

Chip photo, pHEMT,

2.0 × 2.9 mm2

Design by Tech. Lic. Dan Kuylenstierna


Image reject mixer (IRM)

World record in image rejection ratio World record in image rejection ratio (IRR) (IRR) versus versus RF frequency. URF frequency. Ultra wideband IF characteristics!ltra wideband IF characteristics!

IRM1

IRM30

5

10

15

20

25

30

35

0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5 IF frequency (GHz)

Lc; I

RR

(dB

)

Lc - IRM1Lc - IRM3IRR - IRM1IRR - IRM3

0

5

10

15

20

25

30

35

40

0 10 20 30 40 50 60 70 80 90 100RF freq ( GHz)

IRR

(dB)

[ IRM1]

[12]

[ IRM3]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

Comparison with IRMs published by others, (references

according to paper E)

Conversion loss and image rejection ratio versus IF frequency


Balanced resistive mixer (BRM)Motivation: Supresses the LO to RF leakage

RF

LO

BalunIF

Balun

RF

IFLPF

HPF

LO

Single-ended resistive mixer

RFBalun

IF

Balun

LO

Balanced resistive mixer Simple transmitter

0°

180°


Dual-Quadrature mixer (DQM)Motivation: Supresses the LO to RF leakage while

maintaining a broad IF bandwidth

Dual-Quadrature mixer

RF

LO

BalunIF

Balun

Balanced resistive mixer

0°

180°

0°

0°

90°

180°

The BRM requires an IF balun! A integrated IF balun has a typical bandwidth of a few 100 MHz.

NO IF balun required in the DQM! An IF bandwidth of many GHz is possible!!


Dual-Quadrature mixer (DQM)WIN pHEMT technology

Chip photo, 3 × 1.5 mm2Topology

The DQM topology was for the first time recognized for its The DQM topology was for the first time recognized for its potential to achieve high LO to RF isolation potential to achieve high LO to RF isolation (> 20 dB) (> 20 dB) while while

maintaining a very broad IF bandwidthmaintaining a very broad IF bandwidth (11.5 GHz!)(11.5 GHz!)


Why 60 GHz?

• 59 – 63 GHz license free band• 4 GHz (!) of bandwidth High speed wireless communication• High oxygen absorption Less interference Shorter cell re-use

distance• Today: Commercial GigaBit-Ethernet links (1.25 Gbps) • Future: Multi-GigaBit links, FWA, WLAN, WPAN, Roadside

communication…

BridgeWaveTERABEAMWIRELESS

Multifunctional MMICs

http://www.bridgewave.com/wireless_products/ge60-gigabit-wireless.html

http://nrdtech.com/2002/main/en_main.php


60 GHz activities at Chalmers – the beginning1997 – 2001 : The “60 GHz WLAN” project (CTH, Ericsson, KTH, and FOI within CHACH)

2x2x4x4x

BasebandInput

IF-Amp Mixer Power Amp

TX Filter

IF-Amp Mixer LNA RX Filter

Multipliers Buffer

PowerSplitter

BasebandOutput

PowerCombiner

orSwitch

AntennaVCO


Three main design goals were formulated:• As high level of integration as possible• As general design in terms of modulation format as possible• As high data rate as possible Transmission of broadband signals,

i.e. high data rate!!

Design group at Chalmers: Sten E. Gunnarsson, Camilla KSten E. Gunnarsson, Camilla Käärnfelt, Herbert Zirath,rnfelt, Herbert Zirath,Rumen Kozhuharov, and Dan KuylenstiernaRumen Kozhuharov, and Dan Kuylenstierna

60 GHz activities at Chalmers –continuation

2003 – 2005 : The “60 GHz broadband wireless communication systems (BRAWO)” project (CTH and Ericsson within CHACH)

Integrate all subIntegrate all sub--blocks in a 60 GHz frontblocks in a 60 GHz front--end into singleend into single--chip chip transmitter (TX) and receiver (RX) MMICstransmitter (TX) and receiver (RX) MMICs

Paper B Paper B and Cand C


Receiver (RX) and Transmitter (TX) chip designed at Chalmers – topology and chip photo

Image RejectMixer 2.5 GHz

IF

7.1875 GHz LO

X4 X2

LNA

FBA

3 stageAmp.

Buff.

60 GHzRF

RX chip

X8

Chip photo, pHEMT,5.7 x 5 mm2

Two versions in WINs pHEMT/mHEMT technologies

Bal. Res.Mixer 2.5 GHz

IF

7.1875 GHz LO

X4 X2

PA

FBA

3 stageAmp.

Buff.

60 GHzRF

TX chip

X8

Topology- RX

Topology- TX

Chip photo, pHEMT,5 x 3.5 mm2


Conversion Gain and Image Rejection Ratio versus IF Frequency for the RX chip, fLO=7.1875 × 8 =57.5 GHz.

Receiver (RX) and Transmitter (TX) chip designed at Chalmers – measured results

Maximum output power versus IF frequency for theTX chip, fLO=7.1875 GHz × 8 =57.5 GHz.

”First-shot” in pHEMT. WWorldorld record in terms of integration of a 60 GHz frontrecord in terms of integration of a 60 GHz front--endend

Redesigned and more compact chips in mHEMT

SmallerHigher gain Half of the DC power consumptionHalf of the DC power consumption

Date post:	19-May-2020
Category:	Documents
Upload:	others
View:	6 times
Download:	0 times

1. The principle of resistive mixers -...

Documents