lrPM I Memorandum

lrPM I Memorandum

To: K. Crady G. EdissR. Groves A. R. KerrD. Koller G. LauriaM. Pospieszalski S. -K. PanS. Srikanth J. Webber

From: J. Effland

Date: 18 October 2000

Subject: Recent Data Characterizing the SIS Noise Temperature Measurement System

Summary

The primary intent of this memo is to show that the chopper wheel system can now measure receiver noisetemperatures with nearly the same accuracy as the conical loads. The effective temperature assumed for thechopper's cold load is 94.5K from 200 to 260 GHz and 93K from 260-300 GHz. Figure 11 shows that there is nowa ±5% discrepancy between the conical load and chopper-based receiver temperatures.

The untuned LO tripler was suspected of degrading the receiver noise temperature, but Figure 3 shows nearlyidentical receiver noise temperatures using two different untuned tripler designs.

Also presented are mixer/preamp characteristics for the UVAV-L811A-D5-2-D2-374C-01 single-ended mixermarried to the IF4-12P.02 preamp measured over 50 frequencies across the 4-12 GHz IF band.

Finally, preliminary estimates are given of the ohmic losses for the reflector used with the chopper wheel. Figure 14shows that the ohmic losses are about 0.2 dB, and that the chopper's effective cold load temperature would beabout 85K if the reflector was lossless.

I wish to thank Ralph Groves for collecting most of the data shown in this report.

Background

Encouraging results ' ,2 have been previously obtained for receiver noise temperatures measured using the chopperwheel when compared to the lab's standard conical loads. The latest data presented in this memo includessignificantly more measurement points obtained by measuring noise temperatures over 50 IF frequencies at each LOfrequency. This high level of frequency resolution provides additional insight into the results. For example, a shortperiod ripple in the noise temperature data was quickly traced to improper positioning of the ambient load

1"Noise Temperatures Measured Manually and with Chopper Wheel", J. Effland, NRAO CDL Internal Memo, 2000-07-282 "Cold Load Measurements", J. Effland, K. Xiluri, R. Groves, NRAO CDL Internal Memo, 2000-01-10.

Printed: 18 October 2000 Page 1

MM SIS Mixer Measurement System

Measurement Setup

The measurement setup for testing mixer/preamp combinations is diagrammed in Figure 1. The 0.5 dB load in frontof the first amplifier on the warm IF plate provides a return path to ground for the first stage gate of that amplifier.Ron Harris claims this dramatically reduces amplifier failures, because the Miteq amplifiers are designed withfloating gates that provide no return path to ground. The measured noise temperature of the warm IF plate, referredto the input filter, is shown in Figure 2. This was measured using an HP 346B coaxial noise source.

The laboratory standard loads are fabricated from sheets of Eccosorb AN72 absorber formed to the shape of acone. The absorber is 9 cm in diameter inside the base of the cone and the internal height of the cone is 15 cm.The outside of the cold load' cone is enclosed in a copper case and the entire assembly is dipped in a bath of liquidnitrogen to cool it. The ambient load is a section of AN72 absorber without the copper casing and the base of itscone and is 9 cm with a height of 17 cm.

The chopper wheel is 35.6 cm in diameter and reflects the beam from the Dewar window into either an ambientload, or through a 44-cm square aluminum reflector to a cold load. The cold load is Eccosorb CV-3 absorberenclosed in a Styrofoam cylinder of dimensions 8-cm high and 14 cm in diameter. Liquid nitrogen was added to thebowl to cover the absorber tips.

An HP 436 power meter is used to measure Y-factors for all measurements. The noise powers of the mixer/ampand the warm IF system differ by as much as 30 dB. The existing HP 8484A Power head samples data at a slowrate of about 1 measurement per second when the power is less than -50 dBm (10 nW), so a new E4412A powerhead has been ordered to significantly speed up data acquisition.

Figure 3 shows that spurious signals are probably not being generated by the triplers in the LO system. MarianPospieszalski suggested that the noise temperature increase near the 5 GHz IF frequency could be caused byspurious signals generated by the untuned triplers. Figure 3 shows identical receiver noise temperatures at 210 GHzusing two different tripler designs.

Load Material Tests

The optimum load material for the two available types was determined by comparing receiver noise temperaturesmeasured with ambient loads made with AN-72 and CV-3 to the conical loads. The AN-72 is a flat sheet, and CV-3 is pyramidal. Both loads were cut from new material. Figure 4 graphs receiver noise temperatures measured at230 GHz for these three types of loads. The cold load temperature for the chopper-based measurements wasadjusted to 94.5K. The best agreement between the conical load data and chopper-based data is found when thehot load is made from CV-3. All the remaining noise temperature data was measured with CV-3 for both theambient and cold loads.

Noise Temperature Comparisons

Receiver noise temperatures measured with the conical loads and chopper wheel are show in Figure 5 to Figure 10.Also plotted in each figure is the measured noise temperature of the warm IF system, referred to the receiver input.

Discrepancies between receiver noise temperatures measured with conical load data and with the chopper wheelare shown in Figure 11. The large discrepancy at 210 GHz most likely results from the noise temperature spikethere, as shown in Figure 5. The spike is sharp, and its characteristics change rapidly with IF frequency andperhaps with time, so measurements near that frequency might exhibit large variations.


I#1 SIS Mixer Measurement System

The effective temperature assumed for the chopper's cold load is 94.5K from 200 to 260 GHz and 93K from 260-300 GHz. Figure 11 shows that there is now a +5% discrepancy between the conical load and chopper-basedreceiver temperatures, which is near the measurement uncertainty.

Mixer/Preamp Gain Measurements

Gain measurements for the combined mixer/preamp are shown in Figure 12, and were derived from noise powerand physical temperature measurements as described below.

Using an extension of the Rayleigh-Jeans law, where G is the gain of the system, k is Boltzmann's constant, T is thephysical temperature in Kelvin, and B is the bandwidth in Hz:

PNOISE = GkTB

the gain 3 is obtained from differencing noise power and physical temperature measurements. The overall system"gain" is measured from the RF input, where PRF, HOT and PRF, COLD are noise powers measured using the hot andcold loads with physical temperatures of TRF, HOT and TRF, COLD respectively:

GRF= RF,HOT - RF,COLD

TRF,HOT TRF,COLD

The system "gain" referred to the coax switch inside the Dewar is similarly:

G 'PIF,HOT PIF,COLD

TIF ,HOT - TIF,COLD

Since the gain of the IF system is included in GRF above, the true mixer/preamp gain is obtained from ratioing the RFand IF gains:

GMIXERPREAMP =10 LogO (GR /G)

Loss of Reflector Surface Used with Cold Load

To obtain insight into the high effective cold load temperature of the chopper-based cold load, Geoff Edissrecommended measuring noise temperatures with inserting two reflectors between the receiver and cold load. Anincrease in receiver noise temperature measured using these two reflectors results from scattering and lossesincurred in the second reflector.

Receiver noise temperature data were measured several times with a single reflector, as show in Figure 13. Next,the cold load was configured to use a second reflector, and another set of noise temperatures was measured, alsoshown in Figure 13.

The loss of the second reflector was determined from the receiver noise temperature increase by assuming that thesecond reflector acts like an attenuator located between the receiver input and the cold load. Given a receiver with

3This is actually the gain normalized by Boltzmann's constant and the bandwidth, which are reduced when two such "gains" areratioed.Printed: 18 October 2000 Page 3

4... SIS Mixer Measurement System

noise temperature TRX and an attenuator in front of the receiver with loss L (0<L< oo) at a physical temperature of

TPhysical,Loss, the noise temperature of the receiver (T,) becomes:

T = LT, + (L - 1Tphysicaoss

Rearranging the above provides the loss:T+T

L = TR + TPhyscaLoss

TRX+ TPhysicalLoss

The calculated ohmic losses of the second reflector are shown in Figure 14. Although the calculations assume thatthe entire noise increase from the second reflector results from ohmic losses, the 6061-T6 aluminum reflector had arough surface finish, and scattering losses may by significant. Additional tests are planned with a second reflectorpolished to a near optical finish to more closely match the finish of the single reflector that is used in themeasurement system. The reason for the apparent loss increase over such a small fractional bandwidth is unknown.Also shown in the figure is the calculated effective temperature of the chopper's cold load if a lossless reflectorwere used.

Printd: 1 Octber 000 agePrinted: 18 October 2000 Page 4

SSIS Mixer Measurement System

ChopperWheel

Clot I oadCold Ioad

& 0eLi O M PAD

arylwol one I am1nalNATIONAL RADIO ASTRONOMY OBSBRVATORY

hmN rlE II1LVA. ! W2Mixer u)iodes OVERALL DIAGRAM: MIXER/PREAMP

TESTS95 MIXER MEASUREMENT SYSTEM

Color Key a loo200-05- RF- IF \\Eagl \w-edl-ese\don \rock\McArordw

Bias Figure 1: NMeasurement Setup

Figure 1: Measurement Setup


Mixer ti-wire Bias Detail

Ij + Ij -

10K S 10OK A

S +

1K~ 0n 5Ibias 50 D

10K

Bias - Vj+

Supply

Vj-

Page 5Printed: 18 October 2000

SIS Mixer Measurement System

Warm IF Plate Noise FigureFirst Amp: AFS3-00101200-42-LN, SN 359287

Notes:Includes 12 GHz LP Filter7 dB between 2nd and 3rd stages

600

500

400

n 300t-

0z

200

100

0 1 t I 1 I

4 6 8 10 12 14

IF Freq (GHz)\\eagle\cv-cdl-sis\MeasSys\Data\UVaV-L811A-D5-2-D2-374C-01\MixerAmp\Data3.xls (IFNoiseFigure)

Figure 2: Measured Noise Temperature of Warm IF Plate


--- IF Atten =0 dB, 0.5 dB Added to Input-+-IF Atten =0 dB



Noise Temperature Data 210 GHzUVAV-L81 1A-D5-2-D2-374C-01 +1F4-12P.02

2000-09-27

140

120

100

IF Freq (GHz)\\eagle\cv-cdl-sis\MeasSys\Data\ChopperTest\Triplers.xls (chrt260)

Figure 3: Receiver Noise Temperature at 210 GHz using Two Different Triplers

50

Millitech Tripler FTT-04-1S Two different triplers used for LO chain

* Millitech Tripler FTT-04-3RS00 Chopper's cold load effective noise temperature = 94.5K 45_ Chopper data spacing =160 MHz

- -- Tif @ In ut (Right Axis

40

35

30 .

S + * * * 25 E

S1510

15

0

0


SSIS Mixer Measurement System

Noise Temperature Data 230 GHzUVAV-L811A-D5-2-D2-374C-01+IF4-12P.02

Chopper's cold load effective noise temperature = 94.5KChopper data spacing = 160 MHz

100

90

80

70

2 4 6 8 10 12 14

IF Freq (GHz)\\eagle\cv-cdl-sis\MeasSys\Data\ChopperTest\precise.xls (chrt230ANvsCV)

Figure 4: Comparison of Noise Temperatures Measured with Lab Standard and with Chopper Wheel using CV-3 and AN-72 for Hot Load

50

45

40

35

30 aC

25 E

20 'zU-

15

10

5

0



Noise Temperature Data 210 and 212 GHzUVAV-L811 A-D5-2-D2-374C-01 +1F4-12P.02

2000-09-25

140

120

100

2 4 6 8 10 12

IF Freq (GHz)\\eagle\cv-cdl-sis\MeasSysData\ChopperTestPrecise2.xls (chrt210)

Figure 5: Receiver Noise Temperature At 210 and 212 GHz

Printed: 18 October 2000

50

45

40

35

30 .

25

20 0z

15

10

5

0

Page 9


Noise Temperature Data 230 GHzUVAV-L811 A-D5-2-D2-374C-01 +IF4-12P.02

Chopper's cold load effective noise temperature = 94.5K

Chopper data spacing = 160 MHz100

90

80

70

2 4 6 8 10 12

IF Freq (GHz)\\eagle\cv-cdl-sis\MeasSys\Data\ChopperTest\Precise2.xls (chrt230)

Figure 6: Receiver Noise Temperature at 230 GHz

50

45

40

35

30 .t-

25 EI--

20 0z

15

10

5

0



Noise Temperature Data 250 GHzUVAV-L811A-D5-2-D2-374C-01 +1F4-12P.02

2000-09-25

140

120

100

2 4 6 8 10 12

IF Freq (GHz)\\eagle\cv-cdl-sis\MeasSys\Data\ChopperTestPrecise2.xls (chrt250)


UD

50

45

40

35

30 .

25 E

20 "zU.

15

10

5

0

Page 11Printed: 18 October 2000


Noise Temperature Data 260 GHzUVAV-L811A-D5-2-D2-374C-01+IF4-12P.02

2000-09-26

+ RX Temp,CV-3 Hot Load

--- RX Temp, Manual with Conical Loadsu-- Tif @ Input (Right Axis)

Chopper's cold load effective noise temperature = 93.0KChopper data spacing = 160 MHz

kXeagleiv-cdl-sisMeasSyslDaaVhopperTest recise2.xls (chrt260 (2))


140 -

120 -

100 -

-

-

IF Freq (GHz)

50

45

40

35

30

25 -

20 .-z

15

10

5

0

Tr



Noise Temperature Data 270 GHzUVAV-L811A-D5-2-D2-374C-01 +IF4-12P.02

2000-09-26

160

140

120

- 100a.E0I-

-g 80z

mC 60a

IF Freq (GHz)\\eagle\cv-cdl-sis\MeasSysDataChopperTestPrecise2.xls (chrt270)


50

* RX Temp,CV-3 Hot Load Chopper's cold load effective noise temperature = 93.0K

S,--RX Temp, Manual with Conical Loads Chopper data spacing = 160 MHz 45

- Tif Input (Right Axis

S..40

t* 35

30 ac

25E

m

105

0



Noise Temperature Data 280 GHzUVAV-L811A-D5-2-D2-374C-01 +1F4-1 2P.02

2000-09-26

50

45

40

35

30 .

25 E

I-

20 0

u.

15

10

2 4 6 8 10 12 14

IF Freq (GHz)\\eagle\cv-cdl-sis\MeasSys\Data\ChopperTest\Precise2.xs (chrt280)

Figure 10: Receiver Noise Temperatures at 280 GHz

Ez

240

220

200

180

160

140

120

100



Discrepancy between Chopper and Conical Loads

20.0%

15.0%

10.0%

5.0%

0.0%

-5.0%

-10.0%

Large value most likely from data spike --

-15.0% see graph for 210 GHz

-20.0%2 4 6 8 10 12 14

IF Frequency (eagle\cv-cdl-sis\MeasSys\Data\ChopperTestPrecise2.x (chrtDiscrep)

Figure 11: Discrepancy in Receiver Noise Temperature when Measured With Chopper and Conical Loads

U3


:iWE SIS Mixer Measurement System

IF Amplifier IF4-12P.02 Integrated withMixer (UVaV-L811A-D5-2-D2-374C-01)

2 4 6 8 10 12 14

IF Frequency (GHz)\\eagle\cvcdl-sis\MeasSys\Data\ChopperTest\Precise2.xls (chrtGmixVslF)

Figure 12: Gain of Mixer/Preamp vs. IF Frequency

m

. 30E

aS25

c 2 0

Cr



Cold Load Test DataSingle vs. Dual Reflector for Cold Load

UVAV-L811 A-D5-2-D2-374C-01 +ampl

100

90

80

70

2 4 6 8 10 12 14

IF Freq (GHz)\\EAGLE\cv-cdl-sis\MeasSys\Data\ChopperTest\DualReflectors.xls (chrt230)

Figure 13: Measured Data for Dual Reflector Between Receiver and Cold Load

50

45

40

35

30 S@

25 E

20 '8zU.

15

10

5

0


.11W 515 Mixer Measurement System

Cold Load Test DataSingle vs. Dual Reflector for Cold Load

UVAV-L811A-D5-2-D2-374C-01+IF4-12P.02

1.00

0.90

0.80

0.70

0.60oco

H 0.50

-J

0.40

0.30

0.20

0.10

0.00

100

95

90

85

0

80 Io

E75

0

7

65 W

60

55

50

IF Freq (GHz)\\EAGLE\cv-cdl-sis\MeasSys\Data\ChopperTest\DualReflectors.xls (Atten)

Figure 14: Calculated Loss of Second Reflector Between Receiver and Cold Load and Predicted Noise Temperature of Cold Load with NoReflectors

LO Freq = 230 GHz Calculated Loss of Second ReflectorChopper's cold load effective noise temperature = 94.5K

+ Effective Noise Temperature with Lossless Mirror

+ + .* * ** .4 * o * a

+ s ++

r+6

--- -- --

SIS Mixer Measurement Systemwraw


Documents

lrPM I Memorandum