Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.cplire.ru/html/lab234/Tarasov/reprints/SUST1991p644.pdf Дата изменения: Sat Mar 12 19:51:33 2005 Дата индексирования: Sat Dec 22 11:30:30 2007 Кодировка: Поисковые слова: http astrokuban.info astrokuban

Supercond. Sci. Technol.

4 (1991) 644-646. Printed in the UK

amplifier with external mm-wave pumping and its testing by junction noise
DC SQUID RF
M A Tarasov, V Yu Belltsky, G V Prokopenko, L V Flllppenko and V P Koshelets Institute of Radio Engineering and Electronics, USSR Academy of Sciences, Marx Avenue 18. GSP-3. Moscow 103907, USSR

SIS

1
..L..-:--,

Abstract. The power gain and the noise temperature of a oc SOUID RF amplifier at a signal frequency 400 MHz were measured in the presence of external microwave radiation at 2 cm. 8 mm and 4 mm wavelengths. We used a four-loop planar oc SOUID with an integrated input coil and very low values of inductances and stray capacitances. In our noise measurements we used an sisjunction as a precise source of input noise at high DC bias voltage V, > V,, . In such a configurationthe studied system is the Same as an sis mm-wave mixer with a SOUID IF amplifier.

.

._.__A I__ I. lllllU""lillUll

..^.

:" *l.1D

LLlb

D", Y"IL/AII'a_LIII

...-...... -..".....A

c"LII,a_LI,,

I :" LL1 177 U,=

&lrry>,car

.^,Li,,,-

According to one of the SQUID amplifier models (Zimmerman and Sullivan 1977) such an amplifier may be viewed as a peculiar type of parametric amplifier in that amplification of the signal of power Pi at frequency J is realized by up-conversion to the frequency wi + oj (where wj is the Josephson frequency) and detection (down-conversion) takes place in the same device. According to Manley-Rowe relationships for a paraPo/
+

perature, y are constants and not too high frequencies, according to Clarke and Tesche (1979), is T. N Tw(y,yI)L'z/V,. Taking into account y, = 8, yi = 5.5, yvi = 6 and V, N r/L one can obtain TN = 6.5TwL/ra2, i.e. for w = lo9, L = lo-'' H one can obtain TN= 0.04T and RPP' = a20Li(yJy,- y:Jy:)"' .r 0.3a20Li. Such low values of the noise temperature may be realized only for relatively low frequencies. At higher frequencies the SQUID amplifier is quantum noise limited
x In 3, where h is Planck's constant. Taking into account the coupling coefficient a2 = M2/LLi of SQUID

+

, by

ICSL'llC ,YO',

.nom\

the uncertainty principle, which leads to T.

anu

._a

.I_^

,,IC

_.:..

11"1SF

rCnrpc,arulC

:-

.---&.-:-.A 15 c"IIsLlalllC"

>

hfk

inductance L to the input coil Li and the noise parameters of practical SQUIDS, Tesche calculated the SQUID amplifier quantum limited noise temperature
T. = (S,w/k)[(l - a2)/a2](l 2azLV,Sv,/Sv

+

+

a4E V iSj/Sv)1'2

where S, is the energy resolution. For the energy resolution of practical coupled SQUID S, U 5h and coupling a2 -0.5 one can get T. e lOhfk, which leads to T. N 0.2 K atf= OS GHz.

2. Exlernal RF pump influence on

DC SauID

I-V curve

0953-2048/91/110&24

+

03 $03.50

0

1991 IOP Publishing Ltd

\

We studied experimentally the influence of external microwave irradiation at irequencies of i6, 26 and 75 GHz on I-V curves of integrated four-loop DC SQUIDS (Tarasov et al 1991). Figure 1 shows typical curves for different values of magnetic flux caused by input coil current bias.

\


DC SOUID RF

amplifier

10

20

10

20

10

20

I.

),A

Figure

with external irradiation at 16 GHz (a),36 GHz (6) and 75 GHz (c).Curves 1, 2, without and 3,4 with RF pumping.
1. I-V
DCSOUID

curves of

Flgure 2. Schematic layout to measure gain and noise temperature of a SOUIDamplifier with an SIS noise source.

At 16 GHz the Shapiro steps on the I-V curve are not separated, and between them the dynamic resistance is significantly higher than the normal one. This effect may be used for better impedance matching of SQUID output. In figure l(a) the I-V curve near the steps is efficiently changed by magnetic flux and the critical current modulation depth is even better than without irradiation. For more than twice the higher radiation frequency in figure l(b) the current steps are practically separated one from another, the dynamic resistance between the steps nearly achieves the normal curve and the modulation depth is lower than the previous case. -rL^ ^F " .."-",I.., iuc picrcxrrr~cvi ^..^ISUUL a .--An :" a ..-..-&:-"I,.. LLLVUC 13 paunxuiy pa'auct shift of the I-V curve between the steps under magnetic flux changes, i.e. dynamic resistance does not change, output coupling is preserved and dynamic range is wider. At 75 GHz irradiation in figure l(c) the efficiency of current drive by magnetic flux is significantly reduced and such a mode is inefficient for the SQUID amplifier. Comparison of the presented curves shows that there is no reason to increase the external RF pumping frequency higher than the self-resonant frequency of a
xp,"
I-....

varied between 1 and 20 K by changing the SISjunction bias current. The advantages of an SIS noise generator in comparison with a variable temperature load noise generator are: full compatibility with the SQUID amplifier, the possibility of noise modulation, wide noise temperature range and exclusion of the feeder cable Johnson noise. This rather easy method may be used for IF amplifier calibration in an 51s heterodyne receiver and in such a case an SIS mixer would function as the noise source (Belitsky et al 1990)
4. SQUID ampllfler nolse measurements

,uup.

,..
111

U",

^"^^

La>= ,,I=

*Le

"~,&--..",..."..+ DcII-Lc>"IIa'II

"*a..

2L-p

^^___ CUILC-

sponds to 17 GHz. The characteristic frequency of the rL circuit of the SQUID is 20 GHz. The most effective RE pumping influence for improvement of SQUID parameters was observed in the 16 GHz band, which is close to both of these frequencies. It should be mentioned that the characteristic frequency of our Josephson junctions V, = I, R, e 200 pV corresponds to Josephson frequencies of 100 GHz, and for such junctions it is reasonable to lower inductance and stray capacitances to achieve resonant frequencies of the same order.

Signal and noise characteristics were measured with a precise RF oscillator of Schlumberger 4009 type up to 550 MHg a semiconducting noise source up to 1.3 GHz and an SIS junction as a low-level noise source. Figure 3(a) shows the I-V curve, with the amplified signal at 426 MHz and the noise at the output of the next stage amplifier in the presence of external irradiation at 16 GHz. One can see that under these particular conditions the noise temperature is halved and the gain doubied when the bias poini is changed From the iirsi io the second current step. Figures 3(b) and 3(c) show the same dependences for the pump frequency 36 GHz and without RE pumping. Using an SIS noise source (figure qa)) enables one to make frequency dependences smoother and achieve a good input matching by applying SIS junctions with normal resistance equal to the SQUID amplifier optimal

3. Appllcatlon of the sis)unctlon as a cryogenlc nolse source

In our measurements of SQUID amplifier noise temperature we used a tunnel SIS junction with the bias point placed higher than the gap voltage as the noise source (figure 2). The noise voltage in this case, according to Vystavkin et al (1983), U; = (4kTR 2eIR') Af, is the accurate value. A filter-attenuator was placed between the SIS noise source and the SQUID amplifier. The noise temperature at the amplifier input may be

+

10

20

IO

20

10

20

1.p

Flgure 3.1-Vcurves (1). noise (2) and amplified signal (3) under external irradiation at 16 GHz (a),36 GHz (b) and without external pumping (c).
645


M A Tarasov et a/
a

S.

B

a

100

.
U.

0

183

.,a

r,m,

*I,

38,

411

s.mz

483

Flgure 4. Output noise frequency dependences with (1) and without (2) input signal and at zero SOUID bias (3).Curve (4) is the noise temperature of a SQUID amplifier obtained with semiconducting (a) and SIS (b) noise sources.

input impedance. For comparison, in shown the same dependences obtained ducting noise source. In these figures perature reduces to 0.4 K, which is only lated quantum limit.
5. Conclusion

figure 4(b) are with a semiconthe noise temtwice the calcu-

Acknowledgments

This work was supported by the Scientific Council on the high-T, problem and performed in the framework of projects 42 and 90463 of State Program `High Temperature Superconductivity'.

Application of external RF pumping to a DC SQUID amplifier is effective only for pump frequencies not higher than the characteristic frequency of a SQUID loop, which is determined by SQUID inductance. It is shown that a SQUID amplifier may be effectively used with an SIS mixer. Optimization of SQUID parameters allows one to achieve the SQUID amplifier noise temperature of 0.4 K, which is only twice the quantum limit for this SQUID amplifier at 400 MHz.

References

Belitsky V Yu et a1 1990 Proc. 20 EMC, Budapest pp 81620 Clarke 1 and Tesche C D 1979 J. Low Temp. Phys. 37 405-20 Tarasov M A et a1 1991 (to appear in SQUID 91 Berlin) Tesche C D 1982 Appl. Phys. Lett. 41 490-2 Vystavkin A N et a/ 1983 Sou. Phys. JETP 53 2405-8 Zimmerman J E and Sullivan D B 1977 Appl. Phys. Lett. 31
36&2

646