Tool/software:
I am using the TPS7B6950 Unencrypted PSpice model, and then I have seemed a non-real phenomenon on the simulation.
Therefore, I checked the behavior model and found that node 0 (GND) is used inside this model below,
.SUBCKT TPS7B6950-Q1_TRANS VIN VOUT GND_0 GND_1
---Partially omitted
E_U1_ABM4 U1_N09035 0 VALUE { V(U1_N05348) <-- VCVS is connected to node 0
+ * (ABS(V(VOUT)) + 1e-6)
+ / (ABS(V(FB)) + 1e-6) }
R_U1_R4 U1_N09035 U1_N05382 10 TC=0,0
R_U1_R3 U1_N08164 U1_N05348 {3.333e5*SQRT(TTRN)} TC=0,0
C_U1_C1 0 U1_N05348 {1e-6*SQRT(TTRN)}
E_U1_ABM5 U1_N09745 0 VALUE { MIN(V(U1_N05382),
---
The GND_0 and GND_1 defined in the model, and also the model is using node 0 internally even though the node of GND_x are not connected to node 0. I have seen this issue in several other behavior models in TI's PSpice model.
The GND of an LDO may be connected to a floating power supply (e.g. virtual ground).
(See application note: www.tij.co.jp/.../tiduci3a.pdf)
The behavior model of the TPS7B6950 does not allow circuits with virtual ground to run properly.
Will this issue be fixed for the future?