Lumped JPAs, two couplings

The simplest Josephson parametric amplifier: one junction shunted by a capacitor, a lumped LC mode pumped near resonance. The chip has two of them, one coupled to its line through a series capacitor and one through a shared (tapped) inductance. The layout follows twpa-sim’s jpa families (MIT) on the Quantcore PDK.

This notebook compares two ways of coupling a lumped mode to a 50 Ω line, side by side on one die, each with its own pad.

The whole twpa_sim_jpa chip as the v2 notebook builds it

At a glance

Notebook

examples/designs/amplifiers/twpa_sim_jpa.ipynb (an older one-cell version is kept as twpa_sim_jpa.v1.ipynb)

Source

twpa-sim jpa and jpa_inductive (MIT)

Chip

7.5 × 7.5 mm Quantcore die, sapphire 650 µm

On the chip

2 LumpedJpa (capacitive, inductive) · 1 finger capacitor · 2 bond pads · 2 feeds

Builds today

yes (layout built on 2026-11-19 in 5 s)

How the notebook builds it

The amplifier notebooks use six Design segments (die, pads, feed, nonlinear element, routing, whole chip); see the notebook template.

The six layout segments of twpa_sim_jpa
  1. Die. The empty die with the QCT frame and the device ID LJPA.

  2. Bond pads. One pad per JPA: the capacitive one from the west edge, the inductive one from the east edge.

  3. Capacitive feed. The finger coupling capacitor (12 fingers × 80 µm) and its 50 Ω line.

  4. Nonlinear elements. The two lumped JPAs, with the junction sized to the PDK limit and tabulated against the source.

  5. Routing. The inductive JPA’s feed, ending on its tap strip.

  6. Whole chip. The two sim groups (jpa_c, jpa_l) and the saved design.

LOM estimates

Numbers are from a run of this notebook on 2026-11-19 (Setup, Design, Check and the LOM cells, 7 s). This is an amplifier, so the LOM section gives gain and phase matching, not qubit frequencies. The chip carries two JPAs, a capacitive and an inductive one:

JPA

f0 (GHz)

κ/2π (MHz)

Q_ext

K/2π (MHz)

capacitive

6.592

40.4

163

16.8

inductive

6.802

95.8

71

17.9

The single-mode Kerr model (undepleted pump) gives a peak gain of 20.4 dB at -123.6 dBm.

Gain of the two JPAs

These are lumped-model estimates (component formulas and calibration), not FEM results. The LOM page says how far to trust them.

Build spec

The rest of this page is the detailed spec the notebook was built from, with every component, port, route and length.

Source · licence: mandalsoumyajit/twpa-sim (MIT): src/twpa/layout/generate.py (families jpa, inductive_coupler), examples/layout/jpa*.json. Die: 7.5 × 7.5 mm PDK die, PDK frame, sapphire 650 µm, Nb trilayer junctions. Notebook and design: examples/designs/amplifiers/twpa_sim_jpa.ipynb / .json.

Components

instance

catalog ref

key params

notes

P_C, P_L

BondPad

west / east edge

CC

FingerCapacitorSquare

12 fingers × 80 µm

JC

LumpedJpa

capacitive, snapped to CC.b; 30 fingers/side × 150 µm IDC; PDK max junction

JL

LumpedJpa

inductive (tap strip 8 µm), rot 180; same IDC and junction

Routes

route

from → to

type

L_C

P_C.a → CC.a

CPW

L_L

P_L.a → JL.signal

CPW

LOM node structure

JC: island (C_shunt, JJ to ground), then C_c π, then pad; JL: island (C_shunt), then JJ, then tap (L_tap + C_tap to ground, pad)

Targets (from the source)

none stated (illustrative process: f0 ≈ 63 GHz)

Simplifications / PDK differences

planar IDC instead of MIM; PDK junction (0.65 µA) instead of 20 µA; galvanic tapped inductance instead of a pickup loop; junctions are flagged pdk_rn where present (amplifier junctions are far below the qubit R_n window).