DR1_Run_6 - Measure qubit from McDermott
#1 Updated by Daniil Frolov 10 months ago
Initiated qubit measurement. I will be following procedure suggested earlier by Chris:
1.) 1D cavity spec. From this I extract Q at high and low power, and look for the frequency shift as it moves from bright state to dressed.
2.) 2D cavity spec vs drive power sweep. This will give you a better idea of the transition power. After this, we can measure at the highest power where the system is fully in a dressed state.
3.) 2D cavity spec vs flux bias. Sweep each flux bias to see the cavity frequency change. Find the upper and lower sweet spots (USS, LSS). From now on, we usually work at the USS. At this point you can redo (1) and/or (2) if you want to see them with maximized Chi shifts. If you have multiple qubits on the same readout resonator, you will want to set one qubit to the LSS while you measure the power/flux shift of the other, so this will take some back-and-forth to get pictures that show the full picture independent of the other qubit.
4.) At this point I usually move from a VNA to DACs/ADCs, and redo the above measurements to make sure it is working.
5.) Qubit Spec
6.) Rabi oscillations
8.) Ramsey - detune by ~1MHz to see oscillations and find refine qubit frequency
#2 Updated by Daniil Frolov 10 months ago
Measured 2D spectrum vs power sweep. Found 4 resonators as in the spec, frequencies are very close to spec. Used TWPA for this measurement.
Observing shifts when power is decreased from -45 to - 60 dBm (at the VNA output):
5.2037 GHz - 5.2061 GHz = - 2.4 MHz
5.2237 GHz - 5.2204 GHz = + 3.3 MHz
#4 Updated by Daniil Frolov 10 months ago
...Back from holidays
We made some measurements during the holidays.
We tried to search for qubit with TWPA, however it looks like that qubit frequency is close to the pump frequency range of the TWPA, so it contaminates pump signal and causes very non-linear behavior of the quantum amplifier resulting in unstable gain. I disabled TWPA and connected qubit directly to the HEMT amplifier network with superconducting coax (with all necessary isolators...).
Here is what was found (horizontal is readout frequency, vertical is qubit frequency):
When qubit signal power is decreased 6.049 GHz peak disappears. Measurements were done with CW signals using network analyzer and r.f. generator. I will try to focus today on preparing pulsed setup with downconverters and digitizer.
#5 Updated by Daniil Frolov 10 months ago
- File 1057132020.JPG 1057132020.JPG added
- File 1056132020.JPG 1056132020.JPG added
- File 1059132020.JPG 1059132020.JPG added
After many efforts I finally tuned the digitizer and the downconverter to operate in pulsed mode.
Using 5000 ns pulse for readout. Below are results with power sweep of the readout resonator. Maximum frequency shift is observed at -65dBm and lower (shift is -3.3 MHz, same as in CW mode, but power level is lower than in CW mode, because of different gain in the pulse modulator.). I will use -65 dBm and 5000 ns pulse for readout. I will now start qubit search in pulsed mode. I am not using flux bias.
#6 Updated by Daniil Frolov 10 months ago
- File 1040162020.JPG 1040162020.JPG added
- File 1038162020.JPG 1038162020.JPG added
- File 1036162020.JPG 1036162020.JPG added
Update from weekend:
tried to search for qubit in pulsed mode. Found two peaks like in CW mode but slightly different in frequencies.
Then I fixed readout frequency at 5.22035 GHz @ -65dBm, 5 us PW. And repeated qubit scan in Magnitude and Phase.
6.165 GHz peak can be clearly distinguished both in magnitude and phase. I will try today to measure Rabi oscillations today using this frequency.
#8 Updated by Daniil Frolov 10 months ago
1) we measured power spectrogram with current biasing = 0.315 mA with VNA
2) then we repeated the same measurement in pulsed mode, using digitizer. From VNA we found time constant of readout resonator = 7.7 usec, so we tuned readout pulse to be longer = 25 usec. From this measurement we found suitable readout pulse level at -50 dBm (in reality it is -80 dBm, we had to add some attenuators to make trigger work, so there is 30 dB offset) and readout frequency 5.19936 GHz
3) Now using above readout settings we will search for qubit in range 5.4-6.4 GHz by applying high level (0dBm) and 500 ns pulse prior to the readout pulse. Since the readout signal is very weak each point will have 10k averages.
#9 Updated by Daniil Frolov 10 months ago
- File 516192020.JPG 516192020.JPG added
- File 0dB_and_-15dB.JPG 0dB_and_-15dB.JPG added
- File 458192020.JPG 458192020.JPG added
we scanned today qubit frequency and power and here is the result:
Two slices at 0dB and -15 dB.
We also tried various experiments to detect Rabi oscillations using 5.885 GHz as qubit frequency, one promising result is below although resolution is not good.
We were using 1000 ns qubit pulse here (also tried shorter pulse widths, but not successful) and we were changing its amplitude linearly from 0 to 100% (horizontal axis), where 100% corresponds to 0dBm power of the r.f. generator (blue curve on the second picture).
I will leave the same experiment to run again for tonight, but with more points.
#10 Updated by Silvia Zorzetti 10 months ago
- File phase_rabi_0110.png phase_rabi_0110.png added
- File magnitude_rabi_0110.png magnitude_rabi_0110.png added
- File rabi_parameters.png rabi_parameters.png added
sweeping both in qubit pulse amplitude and qubit pulse frequency
see for reference
Observing an oscillation on the magnitude axis (when observing the rdout magnitude)
and an oscillation on the pulse-length axis (when observing the rdout phase)
Repeating the same measurement with a better resolution and longer pulse length
#11 Updated by Silvia Zorzetti 10 months ago
I think we got the Rabi oscillation. Please see the following plot.
I don't observe any change in the frequency with the pulse amplitude. However the dynamic range of the oscillation is quite low. We can try to optimize the readout to increase the dynamic range, and this may also give some more information on the amplitude. Eric, please advice.
#12 Updated by Daniil Frolov 9 months ago
- File qubit_gen_off.JPG qubit_gen_off.JPG added
- File qubit_gen_off2.JPG qubit_gen_off2.JPG added
- File qubit_gen_on2.JPG qubit_gen_on2.JPG added
- File qubit_gen_on.JPG qubit_gen_on.JPG added
we were focused on hardware problems last 2 days:
1) we found that I-Q corrections for vector signal generators were not properly entered in Labber and it caused significant leakage of RF power through the modulator, so even when qubit pulse was off, there was still some significant r.f. signal driving qubit. We fixed it.
2) Later today I also tried to optimize parameters of the readout pulse. We were using 25 us, or sometimes 5 us pulse and averaging over all pulse length. I looked into the waveform and found that when a qubit pulse @5.88 GHz applied prior to the readout pulse there is change observed only in the first 2 microseconds of the readout pulse.
Pictures are below:
Qubit pulse applied prior to the readout:
And without qubit pulse:
Based on this I reduced readout pulse to ~2.3 microseconds, and tuned limits for pulse demodulator as shown in the readout waveforms below.
Qubit pulse applied:
And without qubit pulse:
Also, I increased the number of averages from 10k to 20k and repetition rate from 5 kHz to 20 kHz to speed up measurements.
I will repeat some previous measurements now to see if these new settings will improve contrast.
#13 Updated by Daniil Frolov 9 months ago
With the above new settings I think we can observe two-photon transition. I've just finished qubit scan vs. frequency and magnitude. Although the initial signal is very noisy, after some smoothing two peaks are clearly observed. Distance between them is ~133 MHz and lower frequency peak disappears when qubit pulse magnitude is reduced.
The key here is to keep magnitude of the qubit pulse very low, in these plots vertical scale is in volts. At 50 mV both peaks are already wide, in our previous measurements we ramped magnitude up to several volts, so the resolution was not enough to see change in peaks' width.
I will now make some more adjustments and leave it to run for night with more magnitude and frequency points.
#15 Updated by Daniil Frolov 9 months ago
1) Switched to upper side band +100 MHz for qubit frequency.
2) Tuned I-Q corrections for qubit upconverter for better LO and low side band suppression
3) Switched to Gaussian pulse for qubit (Although setting for this pulse shows 100ns, FWHM is ~40 ns from scope)
4) Increased amount of averages from 10k to 15k
5) Increased qubit pulse magnitude resolution to 10mV
With these settings got better Rabi oscillations.
I will set it for 3D scan now pulse width vs magnitude.
#18 Updated by Daniil Frolov 9 months ago
Increased power of the qubit oscillator to -25 dBm. This allowed to achieve same frequency of Rabi oscillations with shorter qubit pulse.
From the above scan I found pi-pulse as 55 ns, 400mV. I tried then to measure T1 with 10 ns resolution (in the below plot horizontal scale is in nanoseconds), T1~ 1us:
I then tried to measure T2 with pi/2 pulse = 55ns, 200 mV. However it decays faster than any sign of the sinusoidal behavior develops. So I decided to run Rabi scan again with even higher power of the qubit oscillator = -20dBm, to see if higher frequency of Rabi oscillations can be achieved with shorter pulses, so there will be more points for T2.
#20 Updated by Taeyoon Kim 9 months ago
- File (B,C)pulsed mode.pdf (B,C)pulsed mode.pdf added
- File (A, B, C, D)CW mode.pdf (A, B, C, D)CW mode.pdf added
- File taeyoon4_Pulsed_Charlie_Quibt_Scan_(Narrow)2.png taeyoon4_Pulsed_Charlie_Quibt_Scan_(Narrow)2.png added
- File taeyoon4_Pulsed_Charlie_Quibt_Scan_(Wide).png taeyoon4_Pulsed_Charlie_Quibt_Scan_(Wide).png added
- File taeyoon4_Pulsed_Bob(fmax)_Chevron.png taeyoon4_Pulsed_Bob(fmax)_Chevron.png added
- File taeyoon4_Pulsed_Bob(fmax)_Rabi_1.png taeyoon4_Pulsed_Bob(fmax)_Rabi_1.png added
Currently, measuring T1 of the Bob qubit in fmax.
Charlie qubit seems to have very small anharmonicity(~40 MHz), whereas the qubit frequency is around 6 GHz.
David readout resonator does not shows chi_0 shift. I think David qubit is dead.
The detailed report is in the PDF file that I attached.
1. CW measurement report for A(fmax), B(fmax), C, D qubits.
2. pulsed measurement report for B(fmax) and C qubits.
#21 Updated by Taeyoon Kim 9 months ago
- File 11472132020.JPG 11472132020.JPG added
- File 11452132020.JPG 11452132020.JPG added
- File 1142132020.JPG 1142132020.JPG added
- File 11422132020.JPG 11422132020.JPG added
- File 11402132020.JPG 11402132020.JPG added