Slow Turn-off of MOSFET in trivial circuit

QSPICE
1

Hello,

I am attempting to learn how to manage imported models and design symbols. To that end, I’ve started with a MOSFET I’ve simulated extensively in LTSpice, the RH6L040BG.

The following circuit yields the very strange slow turn off behavior

image
image1920×408 39 KB

If I decrease R1 to 10ohms, the turn off time improves significantly. I am really unclear as to why this would be.
Here is the subckt definition for the mosfet:

.SUBCKT RH6L040BG 1 2 3
M1 11 22 3 3 MOS_N
D1 3 1 DDS1
R1 1 11 RTH 4.5653E-3
RG 2 22 10.2
D14 22 11 DGD1
.MODEL MOS_N NMOS TNOM = 25
+ LEVEL = 3
+ L = 2.0000E-6
+ W = 1
+ KP = 284.79E-6
+ RS = 0
+ RD = 0
+ VTO = 2.3551
+ TOX = 2.0000E-6
+ CGSO = 1.3386E-9
+ CGDO = 14.103E-12
+ CBD = 0
+ N = 2
+ GAMMA = 1.1618
+ ETA = 1.0000E-3
+ KAPPA = 0.2
+ NFS = 21.936E+9
+ XJ = 1.0000E-6
+ UO = 377.69
+ VMAX = 50.000E+3
.MODEL DDS1 D TNOM = 25
+ IS = 1.4839E-12
+ N = 0.9308
+ RS = 1.3025E-3
+ IKF = 0.2056
+ CJO = 1.5386E-9
+ M = 0.9495
+ VJ = 6.8496
+ XTI = 2.105
+ BV = 60
+ TT = 24.100E-9
.MODEL RTH RES TNOM = 25
+ TC1 = 6.3368E-3
+ TC2 = 0.2157E-6
.MODEL DGD1 D TNOM = 25
+ N = 10.000E+3
+ CJO = 170.50E-12
+ M = 2.9893
+ VJ = 28.732
+ FC = 0.02
+ T_ABS = 25
.ENDS
2

LTspice and Qspice yield the same result for this simulation setup. I expect that your question is not different in the simulator. There is a parasitic capacitance between the drain and source, and when the FET is OFF, the discharge path is through R1. That’s why with 1000 ohms, the discharge is much slower than with 10 ohms. If you delete CJO from the bode diode model in DDS1 and DGD1, you will speed up the turn-off speed, but that is not realistic.

3

'oof, thats a bit of an oversight on my part. I had not simulated this exact circuit in LTSpice to compare, I just blindly assumed I’d discovered an issue. Shame on me!

In any case, thank you for the check and confirmation that this is expected behavior!

1 Like