quantcore_circuits.analysis.amplifiers

Parametric-amplifier models for the amplifier design notebooks (agent amplifiers): numpy / scipy only.

  • Linear networks — ABCD matrices over a frequency axis (series, shunt, cascade, s_params), the Bloch exponent of a periodic unit (bloch: cosh γ = (A + D)/2), ladders of JJ / KI cells (jj_cell, ki_cell), a SQUID-terminated λ/4 CPW resonator (cpw_jpa_s11) and the resonance / linewidth of a reflection (resonance: f0 at the steepest phase, κ/2π = 4 / max|dφ/df|, a lossless one-port).

  • Resonant paramps — the linearised single-mode Langevin equation in the pump frame (paramp_reflection): ȧ = −(κ/2 + iδ0) a − i g a† + √κ_e a_in, a_out = √κ_e a − a_in. Flux pumping (Yamamoto 2008; QPDK vacask_cpw_parametric_amplifier): ω0(t) = ω0 + δω cos ω_p t → g = δω/2, threshold δω = κ, peak gain ((1 + r²)/(1 − r²))², r = δω/κ (Roy & Devoret 2016). Kerr (current) pumping (Eichler & Wallraff 2014; twpa-sim twpa.jpa): pump photons n from n[(κ/2)² + (Δ − K n)²] = κ_e Φ_in, then δ0 = ω0 − ω_p + 2Kn, g = K n.

  • Travelling-wave gain — coupled-mode equations with an undepleted pump for a line whose inductance is L(I) = L0 (1 + ε I + ξ I²) (Chaudhuri, Gao & Irwin 2015; Malnou et al. 2021; O’Brien et al. 2014): 4WM (ω_s + ω_i = 2ω_p): SPM α_p = k_p ξ I_p²/8, XPM α_{s,i} = k_{s,i} ξ I_p²/4, coupling γ_{s,i} = k_{s,i} ξ I_p²/8, Δβ = (2k_p − k_s − k_i) + 2α_p − α_s − α_i; 3WM (ω_s + ω_i = ω_p): γ_{s,i} = k_{s,i} ε I_p/4, Δβ = (k_p − k_s − k_i) + α_p − α_s − α_i; G = 1 + (γ_s γ_i / g²) sinh²(g N), g² = γ_s γ_i − (Δβ/2)² (k per cell, N cells). A JJ line has ξ = 1/(2 I_c²) (QPDK: the pump’s own inductance × (1 + x/8), the signal’s × (1 + x/4), x = (I_p/I_c)²); a KI line ε = 2 I_dc/(I*² + I_dc²), ξ = 1/(I*² + I_dc²) on its kinetic share α_k.

These are first-order models: they explain trends and reproduce the sources’ headline numbers, they are not a harmonic-balance solution (JosephsonCircuits.jl / VACASK are not in this environment).

Attributes

Functions

bloch(unit) → numpy.ndarray

Complex Bloch exponent γ of a periodic unit: cosh γ = (A + D)/2 (Re γ: attenuation, Im γ: phase per unit),

cascade(*mats) → numpy.ndarray

char_impedance(unit) → numpy.ndarray

Bloch impedance of a symmetric unit: √(B/C).

cme_gain(k_s, k_i, k_p, n_cells, ip_A[, xi, eps, mode])

Signal power gain of a TWPA of n_cells cells (k in rad/cell), pump amplitude ip_A; returns

cpw_jpa_s11(f, cc_F, z0_ohm, vp_m_s, length_um, ls_H)

Reflection of port — C_c — λ/4 CPW (Z0, v_p, length_um) — L_s to ground (the SQUID), exact TL formula.

db(x)

flux_pumped_gain(detuning_Hz, kappa_Hz, r[, kappa_i_Hz])

Phase-preserving gain |S_ss|² vs signal detuning from f_p/2 (pump at 2 f0), pump ratio r = δω/κ.

flux_pumped_peak_gain(r)

ic_from_lj(lj_H) → float

jj_cell(f, lj_H, cj_F, cg_F[, resonator]) → numpy.ndarray

Symmetric JTWPA unit cell: half shunt — series (L_J ∥ C_J) — half shunt. resonator = (C_c, L_r, C_r):

jtwpa_unit(f, lj_H, cj_F, cg_F[, every, resonator, ...]) → numpy.ndarray

One repeat unit of a (resonator-loaded) JJ chain: every cells with the resonator on the last (or one

kerr_gain(omega, delta_p, kerr, kappa_e, kappa_i, n)

|S_ss|² of a Kerr paramp at pump photon number n (pump frame).

kerr_pump_photons(delta_p, kerr, kappa_e, kappa_i, flux_in) → list[tuple[float, bool]]

Intracavity pump photons n (all non-negative roots) for pump detuning Δ = ω_p − ω0, self-Kerr K (rad/s,

ki_cell(f, l_H, c_F[, l_stub_H]) → numpy.ndarray

KI-TWPA cell (twpa-sim Cell, shunt-series): two finger capacitors C/2 to ground, each through its

ki_nonlinearity(i_star_A[, i_dc_A, alpha_k, kerr_factor])

(ε, ξ) of a KI line, twpa-sim Bias: ε = 2 I_dc/(I*² + I_dc²), ξ = kerr/(I*² + I_dc²), times the kinetic

ki_supercell(f, l_H, c_u_F, c_l_F, unloaded, loaded[, ...]) → numpy.ndarray

twpa-sim supercell: unloaded/2 plain cells, loaded cells with the shorter fingers, unloaded/2 plain.

lj_from_ic(ic_A) → float

Josephson inductance L_J = Φ0 / (2π I_c), H.

paramp_reflection(omega, kappa_e, kappa_i, delta0, g)

(S_ss, S_si): signal reflection and idler conversion at offset omega (rad/s) from the pump frame of a

power(m, n) → numpy.ndarray

m (…, 2, 2) to the n-th power per frequency (repeated squaring).

resonance(f, s) → tuple[float, float]

(f0 Hz, κ/2π Hz) of a one-port resonance in a reflection trace: the steepest phase and 4 / max|dφ/df|

s_params(abcd[, z0]) → tuple[numpy.ndarray, numpy.ndarray]

(S11, S21) of a two-port between z0 ports.

series(z) → numpy.ndarray

shunt(y) → numpy.ndarray

squid_ls(ls0, flux_phi0[, asym])

L_s(Φ) of a DC SQUID (no loop inductance): L_s0 / √(cos² πΦ + d² sin² πΦ).

unwrap_bloch(phase) → numpy.ndarray

Unfold a Bloch phase per unit (bloch(unit).imag, folded into [0, π]) along an increasing frequency axis:

Module Contents

quantcore_circuits.analysis.amplifiers.bloch(unit: numpy.ndarray) → numpy.ndarray

Complex Bloch exponent γ of a periodic unit: cosh γ = (A + D)/2 (Re γ: attenuation, Im γ: phase per unit), branch with Im γ ≥ 0, unwrapped along the frequency axis.

quantcore_circuits.analysis.amplifiers.cascade(*mats) → numpy.ndarray
quantcore_circuits.analysis.amplifiers.char_impedance(unit: numpy.ndarray) → numpy.ndarray

Bloch impedance of a symmetric unit: √(B/C).

quantcore_circuits.analysis.amplifiers.cme_gain(k_s, k_i, k_p, n_cells: float, ip_A: float, xi: float = 0.0, eps: float = 0.0, mode: str = '4wm')

Signal power gain of a TWPA of n_cells cells (k in rad/cell), pump amplitude ip_A; returns (G, Δβ per cell, g per cell). See the module docstring for the coefficients.

quantcore_circuits.analysis.amplifiers.cpw_jpa_s11(f, cc_F, z0_ohm, vp_m_s, length_um, ls_H, z_port: float = 50.0, loss_tangent: float = 0.0, eps_fill: float = 1.0)

Reflection of port — C_c — λ/4 CPW (Z0, v_p, length_um) — L_s to ground (the SQUID), exact TL formula. loss_tangent adds dielectric loss (α = β tanδ / 2 · eps_fill).

quantcore_circuits.analysis.amplifiers.db(x)
quantcore_circuits.analysis.amplifiers.flux_pumped_gain(detuning_Hz, kappa_Hz, r, kappa_i_Hz: float = 0.0)

Phase-preserving gain |S_ss|² vs signal detuning from f_p/2 (pump at 2 f0), pump ratio r = δω/κ.

quantcore_circuits.analysis.amplifiers.flux_pumped_peak_gain(r)
quantcore_circuits.analysis.amplifiers.ic_from_lj(lj_H: float) → float
quantcore_circuits.analysis.amplifiers.jj_cell(f, lj_H, cj_F, cg_F, resonator: tuple[float, float, float] | None = None) → numpy.ndarray

Symmetric JTWPA unit cell: half shunt — series (L_J ∥ C_J) — half shunt. resonator = (C_c, L_r, C_r): the shunt gains C_c in series with L_r ∥ C_r (O’Brien 2014 resonant phase matching).

quantcore_circuits.analysis.amplifiers.jtwpa_unit(f, lj_H, cj_F, cg_F, every: int = 0, resonator=None, cg_loaded_F=None) → numpy.ndarray

One repeat unit of a (resonator-loaded) JJ chain: every cells with the resonator on the last (or one plain cell when every is 0).

quantcore_circuits.analysis.amplifiers.kerr_gain(omega, delta_p, kerr, kappa_e, kappa_i, n)

|S_ss|² of a Kerr paramp at pump photon number n (pump frame).

quantcore_circuits.analysis.amplifiers.kerr_pump_photons(delta_p, kerr, kappa_e, kappa_i, flux_in) → list[tuple[float, bool]]

Intracavity pump photons n (all non-negative roots) for pump detuning Δ = ω_p − ω0, self-Kerr K (rad/s, negative for a junction), incident photon flux Φ_in (1/s); each with its linear stability.

quantcore_circuits.analysis.amplifiers.ki_cell(f, l_H, c_F, l_stub_H: float = 0.0) → numpy.ndarray

KI-TWPA cell (twpa-sim Cell, shunt-series): two finger capacitors C/2 to ground, each through its stub inductance, then the series spine inductance.

quantcore_circuits.analysis.amplifiers.ki_nonlinearity(i_star_A: float, i_dc_A: float = 0.0, alpha_k: float = 1.0, kerr_factor: float = 1.0)

(ε, ξ) of a KI line, twpa-sim Bias: ε = 2 I_dc/(I*² + I_dc²), ξ = kerr/(I*² + I_dc²), times the kinetic share α_k of the line inductance (only L_k is current-dependent).

quantcore_circuits.analysis.amplifiers.ki_supercell(f, l_H, c_u_F, c_l_F, unloaded: int, loaded: int, stub_u_H=0.0, stub_l_H=0.0) → numpy.ndarray

twpa-sim supercell: unloaded/2 plain cells, loaded cells with the shorter fingers, unloaded/2 plain.

quantcore_circuits.analysis.amplifiers.lj_from_ic(ic_A: float) → float

Josephson inductance L_J = Φ0 / (2π I_c), H.

quantcore_circuits.analysis.amplifiers.paramp_reflection(omega, kappa_e, kappa_i, delta0, g)

(S_ss, S_si): signal reflection and idler conversion at offset omega (rad/s) from the pump frame of a single-mode paramp with detuning delta0 and pump coupling g (rad/s).

quantcore_circuits.analysis.amplifiers.power(m: numpy.ndarray, n: int) → numpy.ndarray

m (…, 2, 2) to the n-th power per frequency (repeated squaring).

quantcore_circuits.analysis.amplifiers.resonance(f, s) → tuple[float, float]

(f0 Hz, κ/2π Hz) of a one-port resonance in a reflection trace: the steepest phase and 4 / max|dφ/df| (exact for a lossless reflection resonator; overcoupled traces stay close).

quantcore_circuits.analysis.amplifiers.s_params(abcd: numpy.ndarray, z0: float = 50.0) → tuple[numpy.ndarray, numpy.ndarray]

(S11, S21) of a two-port between z0 ports.

quantcore_circuits.analysis.amplifiers.series(z) → numpy.ndarray
quantcore_circuits.analysis.amplifiers.shunt(y) → numpy.ndarray
quantcore_circuits.analysis.amplifiers.squid_ls(ls0, flux_phi0, asym: float = 0.0)

L_s(Φ) of a DC SQUID (no loop inductance): L_s0 / √(cos² πΦ + d² sin² πΦ).

quantcore_circuits.analysis.amplifiers.unwrap_bloch(phase) → numpy.ndarray

Unfold a Bloch phase per unit (bloch(unit).imag, folded into [0, π]) along an increasing frequency axis: after each Bragg stop band the folded phase runs back from π to 0 (or 0 to π); the true phase keeps increasing (2mπ ± φ, the smallest candidate not below the previous value). Needed for long units (a KI-TWPA supercell of 54 cells passes the zone edge inside the band).

quantcore_circuits.analysis.amplifiers.H_PLANCK = 6.62607015e-34
quantcore_circuits.analysis.amplifiers.PHI0 = 2.067833848e-15