nonzero ground node and noise performance of DC-SQUID

[zc-natalie]

Hello Joey, I would like to inquire about the following two questions:
(1) Does the voltage source in JOSIM have to have one terminal connected to node 0 instead of being floating? I am simulating a DC-SQUID amplifier. As shown in the figure below, the ground node of the inductance (blue) is not the same as that of the DC-SQUID (red). To distinguish them, I used non-zero nodes to represent them in JOSIM, but the simulation reported an error indicating that this is not allowed. So how should I represent the isolation of the two ground nodes in the netlist file?

(2) When I want to study the noise performance of the Josephson junction and DC-SQUID, the transient simulation settings seem to affect the simulation results. Do you have any good suggestions regarding this or any helpful example?

Thank you very much!

[JoeyDelp]

Hi,

(1) No, a voltage source can be connected between any two nodes. Using a non-zero node leads to floating node. There is no ground in that loop to anchor the circuit, which leads to a singular matrix error. To "isolate" the grounds you could use a giga-ohm resistor connected between the global reference GND/0 and an aptly named ground for your separation.

(2) The behavior you’re noticing comes from how JoSIM handles noise: it is generated at fixed time intervals (Ts) and then interpolated between updates. Because the transient solver takes its own time steps, it effectively samples that noise stream. When the solver’s steps are larger than Ts, some noise points are skipped and the spectrum is distorted; when the steps are smaller, the result is the expected band-limited noise.

To avoid this, explicitly set the third noise parameter (Ts) according to the maximum bandwidth you want to capture, and configure .tran so the time step does not exceed Ts (using exactly Ts as the fixed step gives the most consistent results). Treat the noise amplitude as a spectral density (units/√Hz). JoSIM automatically applies a 1/√(2·Ts) normalization so the power spectral density remains correct when you adjust Ts.

[zc-natalie]

Thank you for your patient explanation! I have downloaded the latest version of JOSIM from the official website and read the user manual. I noticed that the parameter Ts can be added when setting the noise source manually. However, if I need to use the .neb and .temp commands to add noise to all resistors, does this mean that the Ts parameter cannot be specified? Additionally, I have some confusion regarding the setting of the noise equivalent bandwidth (neb) for circuits containing Josephson junctions. Should this parameter be greater than the Josephson plasma frequency (fp) or the Josephson frequency (fJ) in the voltage state, even though these two frequencies may be much higher than the operating frequency band of the circuit?