long-term stability of an sigehbt-based active cold load.pdf

Long-term stability of an SiGe HBT-based active cold load

1

Emilie Leynia de la Jarrige1,[email protected]

L.Escotte1, E.Gonneau2, J.M.Goutoule3

1CNRS ; LAAS ; Université de Toulouse, 7 avenue du colonel Roche, F-31077 Toulouse, France2 LERISM, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, France

3 EADS- Astrium, 31 rue des Cosmonautes, 31042 Toulouse France

Introduction

2

Reference sources

External target:

- cold sky, pole…

Internal reference:

- passive termination, noise diodes …

Cold source : Low value of noise temperature (< target brightness temperature) +

Long-term stability

Alternative solution : Active Cold Load (ACL)

=> Significant improvement (> 40%) in radiometric resolution has been reported at L-band (with traditional noise injection radiometer)

Radiometer calibration onboard a satellite

3

Year author Technology Frequency Noisetemperature

1981 R.H. Frater GaAs MESFET 1.4 GHz 50 K

1997 L.P. Dunleavy 0.15 μm InP HEMT 18 GHz 126 K

2000 P.M. Buhles 0.15 μm GaAs

MHEMT

2-10 GHz

10-26 GHZ

90 K

125K

2009 N. Skou GaAs PHEMT 10.69 GHz 77 K

2010 C. Bredin 70 nm GaAs MHEMT 23.8 GHz 72 K

Introduction

Active cold loads

-> several technologies, topologies and frequency ranges : FET-based ACL

SiGe HBT : low cost, low excess noise : feasibility ?

Outline

4

I – ACL implementation

1. Topology2. Results

II – Radiometer for long-term stability measurements

1. Test-bed2. Characteristics

III – Stability results

1. Short-term2. Long-term

Conclusion

Outline

5







Conclusion

ACL implementation 1/2

Technical specifications :

• Frequency band : 1400 – 1427 MHz

• Return loss > 30 dB

• Noise temperature < 80 K

• Stability : to be evaluated

Active device :

SiGe HBT from Infineon Technologies(BFP640 – SOT343 package)

Tmin < 40 K

6

Typically for sea surfacesalinity and soil moisturemeasurements applications

0

20

40

60

80

0 2 4 6 8 10 12 14

Min

. n

ois

e te

mp

era

ture

(K)

Ic (mA)

1.8 GHz

1.4 GHz

0.9 GHz

7

Cdec

Cc

Lc 50ΩCb

Lb Le

output

Stub

Cdec

Vcc

Rc

Rb

L1 L2

Cdec

Photograph of the ACL

Electrical circuit :

- 0.635 mm alumina substrate

- Surface Mount Components

- 50 Ω microstrip lines


Bias stability :

- Resistive collector-to-base circuit

- Adjustable voltage regulator

8

-45

-40

-35

-30

-25

-20

-15

-10

-5

0

1,2 1,25 1,3 1,35 1,4 1,45 1,5 1,55 1,6Ref

lect

ion

Co

effi

cien

t (d

B)

Frequency (GHz)

50

55

60

65

70

1,39 1,4 1,41 1,42 1,43 1,44

No

ise

tem

per

atu

re(K

)

Frequency (GHz)

Sensitivity test of temperature variations : ≈ 0.3 K/°C(ex : GaAs FET-based ACL : 0.35 and 0.38 K/°C)

Reflection coefficient < -35 dBNoise temperature < 66 K

2. Results

For Ic = 5.6 mA and Vcc = 1.12 V


Outline

9







Conclusion

10

Radiometer for long-term stability measurements

Dedicated radiometer : two-load radiometer with noise injection measurements errors due to power gain variations are reduced

1. Test-bed

+

+Receiver

Tn

T2

T0

TR , G

ACL

matched

load

cold

load

A/D

Directional coupler +

Noise diode

Typical measurement cycle

11

1. Test-bed

Output voltage

Receiver

ACL

Noisediode

LNAdc amplifier

Matched load

DAQdevice

USB

Temperature-stabilized enclosure

Commandcircuit

Thermistors...

- DAQ U2351A from Agilent

Thermistors for temperature control


tt2t1t0

…

12

LNA +Filter :

Gain > 60 dB

Noise Bandwidth : 60 MHz

Noise temperature : 40 K

-10

0

10

20

30

40

50

60

70

1200 1250 1300 1350 1400 1450 1500 1550 1600

-10

0

10

20

30

40

50

60

70

1200 1250 1300 1350 1400 1450 1500 1550 1600

-10

0

10

20

30

40

50

60

70

1200 1250 1300 1350 1400 1450 1500 1550 1600

Frequency (MHz)

Tran

smis

sio

n c

oef

fici

ent

(dB

)

1. Test-bed

Coaxial switch :

Insertion loss : 0.04 dB

Isolation : > 120 dBOutput voltage

ReceiverACL

Noisediode

LNA dc amplifier

Matched load

DAQdevice

USB


Thermistors...

Commandcircuit


13

Square-law detector:

- Tunnel diode- Responsivity : 1 mV/µW à 1.4GHz- Noise : 2 nV/√Hz à 100Hz

Output voltage

ReceiverACL

Noisediode

LNA DC amplifier

Matched load

DAQdevice

USB


Thermistors...

Commandcircuit

1. Test-bed


14



Low-noise voltage amplifier:

- Gain : 80 dB- Noise floor : 1 nV/√Hz

Output voltage

ReceiverACL

Noisediode

LNA DC amplifier

Matched load

DAQdevice

USB


Thermistors...

Commandcircuit

1. Test-bed


Low-noise voltage amplifier:

- Gain : 80 dB- Noise floor : 1 nV/√Hz



15

Low-pass filter (RC circuit)

- Time constant : 5 ms

1. Test-bed

Output voltage

ReceiverACL

Noisediode

LNA DC amplifier

Matched load

DAQdevice

USB


Thermistors...

Commandcircuit


16

1. Test-bed

Output voltage

ReceiverACL

Noisediode

LNA DC amplifier

Matched load

DAQdevice

USB


Thermistors...

Commandcircuit

Noise injection :- Excess noise ratio of 15.6 dB at 1.4 GHz- Coupler : 16 dB- Noise temperature : T2 ≈ 280K


17

-> in order to limit the impact of temperature variations, we realize temperature stabilizedenclosure :

10 mm- thick aluminum enclosure +

50 mm-thick extruded polystyrene+

Temperature control circuit

1. Test-bed

ACL temperature ≈ 38.4°C

ACL temperature variations :

< 0.01 °C during 1 day< 0.03 °C during 4.5 months


1818

ACL noise temperature :

matched load Injected noise temperature

loss of the coupler

²)²(

²²)²(

²²1

²12

012

12

201

1202

VV

VVV

VV

VVV

VVVLTTn

Important characteristics :

- low value for L- V0 and V2 close- smallest variance values ; depend on integration time

20

12

10 TVV

VVLTTn

2. Characteristics

+

+ Receiver

Tn

T2

T0

TR , G

ACL

matched

load

cold

load

A/D


- Sensitivity < 31mK- Stability < 25mK for a total integration time of 95s

Optimum integration time:

τ 0 = 50 sτ 1 = 15 sτ 2 = 30 s

Allan variance:

Conditions :

- T°amb ≈ 38.4°C - Maximum deviation < 0.01°C- Period = 150 min

2

12 )(

2

1)( nny yyt

0,01

0,1

1

0,1 1 10 100

All

an

dev

iati

on

(m

V)

Integration time (s)

Matched load

Diode

ACLWhite noise

Flicker noiseRandom walk noise

19

2. Characteristics


Outline

20







Conclusion

Stability results

21

Characteristics :

- Mean : 87.48 K- Standard deviation : 32 mK- Maximum deviation : 0.205 K

87.3

87.4

87.5

87.6

0 1 2 3

Time (day)

AC

L N

ois

e T

emp

erat

ure

(K

)

87.35

87.45

87.55

87.65

1. Short-term

22

ACL noise temperature :

- Maximum deviation < 0.35 K- Slope : -2.1 mK/day

extrapolated stability < 1 K/year

2. Long-term

Stability results

87.2

87.3

87.4

87.5

87.6

87.7

0 20 40 60 80 100 120 140

Time (day)

AC

L N

ois

e T

emper

ature

(K

)

Outline

23







Conclusion

Conclusion

24

ACL

- Reflection coefficient < -35 dB- Noise temperature < 66 K

Radiometer

- Sensitivity < 31 mK- Stability < 25 mK for a total integration time of 95 s- Long-term stability (4.5 months) :

• Gain : maximum deviation < 0.03 dB• Receiver noise temperature : maximum deviation < 1 K• Physical temperature variations < 0.03°C

Long term stability (4.5months)

- Maximum deviation < 0.35 K - Slope : -2.1 mK/day extrapolated stability < 1 K/year

Author Frequency Technology Noisetemp.

Stability Duration

N. Skou 10.69 GHz GaAs PHEMT 77 K 2 K1 K

1 year8 months

L.P. Dunleavy 18 GHz 0.15 μm InP HEMT 126 K 2 K 1 yearC. Bredin 23.8 GHz 70 nm GaAs MHEMT 72 K 0.4 K 1 month

25

Thank you for your attention

long-term stability of an sigehbt-based active cold load.pdf

Business

nvhz noise

noise floor

noise bandwidth

load radiometer

low excess noise

sea surface noise temperature

lownoise voltage amplifier

resultsii radiometer