5GHz, Cavity, (TE5RI01) + quantum amplifiers
#4 Updated by Daniil Frolov about 1 year ago
I have just turned on control of switch #1, that is located right after the 5 GHz cavity. I started control app and tried to read the state of the switch. It reads state correctly, and doesn't heat up. At least 10 minutes after readout command, temperature didn't change at all. A will now try to activate the switch, and turn it to different position...
#6 Updated by Daniil Frolov about 1 year ago
Some update on switching. Yesterday I forgot to use current-limiting resistor in series with the switch coil, that we planned to use when the wire of the coil will become superconducting, so the current through the coil was very high.
I tried to activate switches with current-limiting resistor this morning, and it works! Now the temperature jumps very little. Below is temperature of mixing chamber after two switches were activated, one after another.
First, temperature jumped from 12.5 to 15.5 mK, and next time it jumped from 15.5 to 18.5 mK - very small change.
Looks like that switches heat mixing chamber negligibly small.
#7 Updated by Daniil Frolov about 1 year ago
- File 6282019635.JPG 6282019635.JPG added
- File field_gui_2019_Jun_28__16h17m25s.jpg field_gui_2019_Jun_28__16h17m25s.jpg added
I'm starting to tune Yale quantum amplifier. Finished flux scan with new program. Found optimal biasing at +230 uA. Quantum amplifier resonates at 4.984 GHz with this biasing.
#9 Updated by Daniil Frolov about 1 year ago
I started to activate switches and temperature went up to 36 mK, then the gain from the quantum amplifier disappeared. I re-scanned it and found that the magnetic field vs. amplifier resonant frequency spectrogram also shifted. I foud the new parameters of the amplifier:
biasing +264 uA
pump: 14.160 GHz
pump level: +0.5 dBm
with these parameters Yale quantum amplifier gives 35 dB gain at 4.9 GHz and -70 dBm input power (to the port 8 of the fridge).
#10 Updated by Daniil Frolov about 1 year ago
- File dual_gui_2019_Jul_01__19h35m06s.jpg dual_gui_2019_Jul_01__19h35m06s.jpg added
- File twpa_disp_feature_712019736.png twpa_disp_feature_712019736.png added
Started working with TWPA. #784.
Found dispersive feature according to manual.
Made a fast spectrogram, pump frequency (qubit vertical axis in the picture) vs. TWPA transfer function ("Cavity" horizontal axis in the picture). Looks like the desired, wide gain profile is achieved when pump frequency is around 6300 MHz (1 dbm power setting on the generator).
Now will make a finer scan, to measure pump frequency more accurately.
#13 Updated by Daniil Frolov about 1 year ago
Summary on TWPA # 784:
It works, however behavior with the cavity is different from the behavior when it is connected to the source directly, without cavity. When the amplifier is tuned without cavity gain can reach 20 dB @ 5 GHz with the specific settings of pump frequency and power. However, when the cavity is connected to the input of the amplifier, gain with the same settings of pump signal disappears. With the cavity connected the amplifier requires different setting of pump frequency and power, so it moves to some different operating point and in this mode maximum gain is not achieved any more.
In the original setting without cavity, with settings of the pump= 6.287 GHz and +0.4dBm gain was around 20 dB at 5 GHz. When I connected cavity, I found maximum gain 8 dB was at pump = 6.274 GHz, - 4 dBm. So with cavity gain of the TWPA drops by at least 10 dB. This may be because of the mismatch between cavity output and TWPA input: we don't have any isolators there currently.
Also I was able to observe ~8 dB improvement in SNR with TWPA for the resonance peak measurement (see upper plot).
#15 Updated by Daniil Frolov about 1 year ago
Just finished measurements with the second TWPA # 600. Results are very similar. Pump frequency and power is slightly different: 6.317 GHz, -2 dBm for optimal gain around 5 GHz.
The same problem with low gain when connected to the cavity remains. Next run need to add idolators between cavity and TWPA.
Also see improvement in SNR when TWPA is on, about 8 dB improvement
#17 Updated by Daniil Frolov about 1 year ago
Haozhi is here. We are now starting to test NIST JPA.
We made manual scan with network analyzer and found pump frequency (bifurcation point)= 5.04 GHz at bias -7 uA, -30.8 dBm power setting on the VNA (9.2 + 40 external attenuator).
#19 Updated by Daniil Frolov about 1 year ago
Connected cavity 5GHz to the amplifier.
Slightly changed pump parameters to optimize gain. 5.0446 GHz, power= -29.34 dBm (R&S generator).
observing gain about 25 dB!
On picture: saved trace (red)- pump off, active trace (yellow)- pump on.
#20 Updated by Daniil Frolov about 1 year ago
we tried to measure noise temperature of the system with JPA and total gain of the system.
We used y-factor method with calculated quantum noise as cold source and matched load sitting at 1.002 K as a hot source (connected to JPA with superconducting wire through the switch). We studied next configurations:
1) JPA was off, its input connected to the matched load at physical temperature of 20 mK. Half-photon quantum noise temperature is 0.24 K @ 5.004 GHz. In this case total system noise should be dominated by thermal noise of the HEMT (2.6182 K). Measured output power at the spectrum analyzer (JPA off + HEMT + room temperature amplifier) was -129.46 dBm (zero span, 1 Hz bandwidth).
2) JPA was on, under the above conditions output power on the spectrum analyzer increased to -116.16 dBm.
So the difference in output power between the two conditions JPA "on" and JPA "off" is -129.46 dBm -(-116.16 dBm)= -13.3 dBm
Theoretical difference between HEMT's temperature noise power and half-photon quantum noise power is -207.145 dBm - (-194.619 dBm) = -12.52 dBm, which is only 0.78 dB off the difference that we measure.
Our estimation of total gain of the whole chain from JPA input to the spectrum analyzer is 91 dB based on the measurements above.
(See python file with calculations in today's shift folder on RF computer: c:\redmine\runs\DR1_run_5\python_measurements\2019_07_11\NIST_JPA\JPA_noise_studies
3) We tried to measure effective noise temperature with y-factor method, by connecting "hot" load @ 1.002 K to the JPA. Output power measured with the spectrum analyzer in this case was -115.06 dBm. Effective noise temperature calculated with y-factor method was 2.39 K, which is very high and seems to be wrong (should be order of 20 mK ?). To reach 20 mK one needs to increase the result of output power with the hot load by factor of ~3 which is 4.7 dB. This means that we are probably loosing 4.7 dB on the way from hot load to the input of JPA, which seems relatively high. Need to study this more.
#21 Updated by Daniil Frolov 12 months ago
Some results from low power cavity decay measurements.
Yesterday we tuned JPA and measured several decays with and without averages. Attached chart shows estimation for photon levels at the output of the cavity, based on the assumption that when no signal is applied to the input of the cavity, JPA amplifies only half-photon noise. Decay curve and noise curve on the right is measured with 20 averages.
The gain of the chain is 77.5 dB and was estimated as a ratio between the noise level measured on the spectrum analyzer when JPA was on and calculated half-photon noise. Noise of the JPA is not taken into account in this estimation.
Mixing chamber temperature was 23 mK.
Data is stored here: \\tdserver1\project\Quantum Computing\redmine\runs\DR1_run_5\manual_measurements\TE5RI001_decay_with_JPA