Researchers achieve quantum control of an oscillator using a Josephson circuit

Discussion in 'HardForum Tech News' started by erek, Dec 4, 2019.

  1. erek

    erek 2[H]4U

    Messages:
    3,569
    Joined:
    Dec 19, 2005
    Quantum control of an oscillator using a stimulated Josephson nonlinearity

    "Superconducting circuits extensively rely on the Josephson junction as a nonlinear electronic element for manipulating quantum information and mediating photon interactions. Despite continuing efforts in pushing the coherence of Josephson circuits, the best photon lifetimes have been demonstrated in microwave cavities. Nevertheless, architectures based on quantum memories require a qubit element for logical operations at the cost of introducing additional loss channels and limiting process fidelities. Here, we directly operate the oscillator as an isolated two-level system by tailoring its Hilbert space. Implementing a flux-tunable inductive coupling between two resonators, we can selectively Rabi drive the lowest eigenstates by dynamically activating a three-wave interaction through parametric flux modulation. Measuring the Wigner function confirms that we can prepare arbitrary states confined in the single-photon manifold, with measured coherence times limited by the oscillator intrinsic quality factor. This architectural shift in engineering oscillators with stimulated nonlinearity can be exploited for designing long-lived quantum modules and offers flexibility in studying non-equilibrium physics with photons in a field-programmable simulator."


    https://phys.org/news/2019-12-quantum-oscillator-josephson-circuit.html
     
  2. Shoganai

    Shoganai Limp Gawd

    Messages:
    416
    Joined:
    Dec 5, 2018
  3. clockdogg

    clockdogg Gawd

    Messages:
    945
    Joined:
    Dec 12, 2007
    ".. by tailoring its Hilbert space."

    Oh...not Dilbert space. Guess there's no point waiting for the nonlinear cartoon strip version...
     
    wolfofone, AceGoober and erek like this.
  4. DrDoU

    DrDoU 2[H]4U

    Messages:
    2,461
    Joined:
    Jun 4, 2007
    Can not someone speak the kings English on the subjects. Not everyone has phd in physics.
     
  5. Dan_D

    Dan_D [H]ard as it Gets

    Messages:
    55,115
    Joined:
    Feb 9, 2002
    Yep that one totally lost me.
     
  6. filip

    filip [H]ard|Gawd

    Messages:
    1,922
    Joined:
    Aug 15, 2012
    They do something better because of the title.
     
    1_rick and erek like this.
  7. nthexwn

    nthexwn [H]Lite

    Messages:
    72
    Joined:
    Feb 27, 2007
    My totally unqualified armchair analysis:

    This means we could have 300GHZ processors! Unfortunately, a modern CPU scaled to use microwave cavities in place of transistors would have to be roughly the size of the sun. Unfortunately the weight of the cooling helium alone would cause the entire thing to fuse into carbon and become a neutron star.
     
    Viper16, KazeoHin, KarsusTG and 2 others like this.
  8. tangoseal

    tangoseal [H]ardness Supreme

    Messages:
    7,985
    Joined:
    Dec 18, 2010
    Quantum tech will NEVER be used for the good of mankind. Only nefarious Nazi bullshit. Nothing the elites control is meant for the enhancement of humanity, moving towards the stars, bettering our species. Nope. Just nefarious control grid dystopia.
     
  9. blandead

    blandead Limp Gawd

    Messages:
    233
    Joined:
    Nov 6, 2010
    One day... How computers work will stop making sense to 99% of people

    Um.. LED lightbulbs are a result of "quantum tech". Also your phone, GPS, Wi-Fi. I'm just gonna stop at those.

    Another interesting fact... Einstein was German, and certainly not a Nazi.
     
    Last edited: Dec 5, 2019
  10. Mazzspeed

    Mazzspeed 2[H]4U

    Messages:
    2,069
    Joined:
    Dec 27, 2017
    I believe White Rose has already overcome this limitation as Elliot Alderson is about to find out...
     
  11. nthexwn

    nthexwn [H]Lite

    Messages:
    72
    Joined:
    Feb 27, 2007
    I'd be super proud if >1% of people ever understood how computers work!
     
    GhostCow, Red Falcon and KazeoHin like this.
  12. Nenu

    Nenu [H]ardened

    Messages:
    19,046
    Joined:
    Apr 28, 2007
    lol, this reads almost like inserted tech words on a Star Trek episode.
    I follow parts of it then some bits make me chuckle.

    I am really glad you make tech threads like this btw.
     
    Last edited: Dec 12, 2019 at 8:17 PM
    wolfofone likes this.
  13. Shoganai

    Shoganai Limp Gawd

    Messages:
    416
    Joined:
    Dec 5, 2018
    Reminds me of this joke terminology video:

     
    SFB and erek like this.
  14. KD5ZXG

    KD5ZXG Limp Gawd

    Messages:
    438
    Joined:
    Mar 24, 2017
    Still trying to bend my noodle around the decades old Quantum Flux Parametron.
    QFP's use Josephson junctions, but don't do quantum logic or interface with light.
    So, I know even less what chimeric pigmonkey they gone off about this time...
    Why exactly did I want a microwave cavity? Won't any generic resonance do?
    Was the cavity so it can interact with light maybe? Totally guessing here...

    If "parametric flux modulation" from the article is anything like parametrons of old:
    You need pump that injects power at twice the tuned frequency to keep it going.
    Maybe pump is light aimed at the cavity? But they said flux modulation, not light.
    Are they aiming flux at cavities to pump them? But they said interact with light.
    Make up you damn minds...

    Also a non-linearity (Josephson junction in this case) to boost the amplitude on
    both positive and negative swings, but no pump energy to boost zero crossings.
    Keeps zero crossings of multiple devices phase locked to 1/2 the pump freq.
    Each parametron is free to lock in 0 or 180 degrees binary phase shift key.
    Thus each amplifies weak bits (or vote of inputs) to full scale oscillation, and
    stores a 0 or 1 result in the form of phase.

    Binary majority vote logic is not "quantum" to extent of leveraging superposition
    or entanglements, even when done in quantum ways. A hard choice of phases
    that cancel when superposed resolve qubits to normal bits. I suspect whoever
    wrote or plagiarized that article had even less clue...

    Lets just throw Hilbert Space around to further confuse shit...
     
    Last edited: Dec 6, 2019
  15. Auer

    Auer [H]ard|Gawd

    Messages:
    1,032
    Joined:
    Nov 2, 2018
  16. TheBuzzer

    TheBuzzer HACK THE WORLD!

    Messages:
    12,538
    Joined:
    Aug 15, 2005
    I think they are just saying they did something in a better way.
     
  17. KD5ZXG

    KD5ZXG Limp Gawd

    Messages:
    438
    Joined:
    Mar 24, 2017
    So, how can a superconductor have a useful non-linearity?
    Useful in this case to pump an oscillation representing data.

    Does it work similar to negative resistance: no such thing.
    But special diodes can exibit negative dynamic resistance
    where the slope of its resistance curve might bend down.

    Josephson slope doesn't bend down, it just goes 0 flat.
    Always after a tiny voltage drop like a bidirectional diode.
    So the actual Volts/Amps is never zero, just the slope.
    https://en.wikipedia.org/wiki/Josephson_voltage_standard

    But Josephsons and Pi Josephsons are somehow not
    the same thing. Flat part of a Pi josephon seems to exist
    at true zero. Perhaps can superconduct with zero drop?
    https://en.wikipedia.org/wiki/Pi_Josephson_junction

    Pi also said to be AC voltage to frequency converters,
    but how you apply an AC or DC voltage to something
    that begins without a resistance, except to saturate it?
    Do I misunderstand correctly that like a capacitor of
    oppositeness, flips the phase of AC passing through?

    I'm thinking with a DC voltage bias and no DC current,
    it becomes a superconductor (and phase flipper) of AC
    on the flat part of the slope. This AC (storing the data
    as a binary phase shft key) is looped back through an
    inductor with another 180degree phase shift and also
    where the endless loop of DC current is stored.

    Not plain unmodulated DC, wouldn't be interesting...
    If you might alter the phase shift of this J junction by
    manipulating DC current in the loop, you might pump
    parametric oscillation, just as Goto did by saturating
    old fashioned inductor capacitor tanks in the 1950's.

    So, in actual operation: Maybe the junction does not
    operate at point of 180 phase shift, but -90 or +270?
    An inductor has maximum 90 degree shift, not 180...
    How else could AC circle back to 0 or 360 phase?

    Still don't make any sense, since this is at zero volts.
    How do you modulate zero? All you gots is current...
     
    Last edited: Dec 12, 2019 at 7:09 PM
  18. viivo

    viivo [H]ard|Gawd

    Messages:
    1,370
    Joined:
    Sep 7, 2005
    A josephson circuit sounds like something a shady PC repairman would make up to fleece a naive customer.
     
  19. KD5ZXG

    KD5ZXG Limp Gawd

    Messages:
    438
    Joined:
    Mar 24, 2017
    If you replace inductor with a delay line, coax or something.
    Can you work with multiple bits round robin in one device?

    BPSK could be streaming both directions simultaneously.
    Don't know if thats an actual mess, or inconsequential.
     
    Last edited: Dec 12, 2019 at 7:25 PM
  20. c3k

    c3k 2[H]4U

    Messages:
    2,104
    Joined:
    Sep 8, 2007
    "Eurekaw, eurekaw" quoth the raven.
     
  21. TordanGow

    TordanGow [H]ard|Gawd

    Messages:
    1,337
    Joined:
    May 25, 2015


    A what , what, and a what?
     
    wolfofone likes this.
  22. KD5ZXG

    KD5ZXG Limp Gawd

    Messages:
    438
    Joined:
    Mar 24, 2017
    No clue of what all that OP gibberish gone on about.
    Quantum anything doesn't yet make full sense to me.

    Tuned circuits don't spontaneously oscillate themselves.
    Nor have any reason to stay in phase lock with others.
    So, you need a non-linear element energized by clock.

    The clock pulses twice as fast as the desired oscillation.
    Ocillation peaks of either phase in time with that clock
    will be amplified. Any other noise or attempted phase
    drift is not repeatedly amplified and decays away.

    Somewhere in the machine, an oscillator is dedicated to
    maintain reference zero. Else we would not know ones
    from zeros. Both look identical except for Binary Phase
    Shift Key and that reference zero to compare against.

    Hitachi used ferrite shirt buttons as the non-linear element.
    http://museum.ipsj.or.jp/en/computer/dawn/0015.html
    Until you understand how the old ones worked, ain't much
    point complicting things with quantum.

    Inputs were mixed like audio, and whichever phase was
    in the majority would be the phase amplified to full scale.
    An answer would be held untill the clock was turned off
    and the tuned circuit allowed to ring down to a state of
    undecided, ready to take a new vote.

    Each oscillator was a complete ALU and storage bit.
    The majority of A,B,1 is the same operation as OR.
    The majority of A,B,0 is the same operation as AND.
    The majority of A,B,C exactly the same as CARRY.
    I/O was usually differential pair. Inversion was free.
    NAND and NOR are well known universal logics.
    Minority vote of A,B,C is therefore also universal.

    The undecided state was more similar to superposition
    than today's familiar tri-state. But the noise floor was
    always more than any macroscopic ancient qubits held
    in undecided superposition. Without inputs, you simply
    amplified a random vote from noise.

    Now we have ALU, Memory Bit, and RNG. Does it do
    anything else? Especially quiet ones, far more quiet
    than transistors, have been used to amplify weak radio
    (below the noise floor but correlated with the clock)
    from distant spacecraft. To resolve qubits from other
    quantum logic into normal bits. It also Juliennes fries.

    Inputs were rarely mixed strong enough to overwhelm
    any stored full scale oscillation. Once vote was decided,
    later input changes were typically ignored. Asynchronus
    gatelike logic flow is not the usual way for Parametrons
    to pass data.

    When an undecided Parametron is woken by clock, it
    listens for votes from both inputs and outputs. Those
    annoying output votes need to be put to sleep first.

    Syncronus logic flow involves a rolling wave of quiet
    forgetfulness as banks ahead of new votes are powered
    down. New logic can only appear in the wake. Requires
    a minimum of three clock banks (see the back side of
    Hitachi's Hipac module from the museum link above).
    You can make a shift register rotate in either direction or
    logic flow backward just by how you choose to clock it.

    Newfangled quantum ones seem to be using four banks.
    Not sure, but suspect a mixing manifold stage was added.
    Since superconduct obviously won't use resistors to mix.
    And here the inputs might be same strength as full scale.
    Preventing backwards flow is more complicated.

    If anyone this side of a paywall knows better, please clue me.
     
    Last edited: Dec 13, 2019 at 5:35 PM
  23. erek

    erek 2[H]4U

    Messages:
    3,569
    Joined:
    Dec 19, 2005

    What gibberish? :(
     
  24. KD5ZXG

    KD5ZXG Limp Gawd

    Messages:
    438
    Joined:
    Mar 24, 2017
    The kind of gibberish below the noise floor that doesn't correlate with any clock.

    Maybe these two pages explain it better...
    105535_Parametron_Works.png

    Only the shirt button magamp (labeled "Excitation" in diagram above) is
    replaced by a Josephson Junction. From that point, you are on your own...
     
    Last edited: Dec 13, 2019 at 5:36 PM
    erek likes this.
  25. erek

    erek 2[H]4U

    Messages:
    3,569
    Joined:
    Dec 19, 2005
    excited?
     
  26. KD5ZXG

    KD5ZXG Limp Gawd

    Messages:
    438
    Joined:
    Mar 24, 2017
    Excited? Yes, mix of correlated clock and data in nonlinearity results in amplification.
    Its been done without electricity using only light and non-linear optical crystal.
    Versatile way to make logic from whatever without transistors.
     
    erek likes this.
  27. erek

    erek 2[H]4U

    Messages:
    3,569
    Joined:
    Dec 19, 2005
    To be honest, i've been legitimately working through the notion of virtualizing the hardware clock as in getting rid of a hardware oscillator / quartz crystal. I'd like to have a purely software defined clock reference.

    my initial thoughts were to digitally sample / record a hardware clock signal to use as a reference.

    the idea is like a /dev/random *nix device


    scope this out KD5ZXG https://en.wikipedia.org/wiki/Asynchronous_circuit#Asynchronous_CPU
     
  28. KD5ZXG

    KD5ZXG Limp Gawd

    Messages:
    438
    Joined:
    Mar 24, 2017
    Yup, thats async. Gates naturally do that. Parametrons don't...

    Then again, if an async circuit needs 20 stages, doesn't get
    a free pipeline to work on multiple threads. Parametron does.

    Compartmentalized by rolling waves of sleep, could work six
    and 2/3 operations in the pipe simultaneously. Always integer
    though, like bleen: Sometimes six, sometimes seven. I would
    just stretch that pipe with a do-nothing stage so 7 always fit.

    No rule forbids using both techniques in the same circuit.
     
    Last edited: Dec 13, 2019 at 10:09 PM