why does 2-D LUT (temperatu​re-indepen​dent) MOSFET shows following errors

Hey,

I have tried to speed-up my MOSFET simulation by using 2D-LUT based parameterization of N-Channel MOSFET. The MOSFET which I am trying to model and simulate is Infineon's IAUCN10S7L180 (100V, 39A).

I have used Output characteristic curve, and typical capacitance curve to extract test data. In N-channel MOSFET, VDS, IDS and corresponding VGS values are defined as a look-up table, whereas the data for Ciss, Crss, and Coss is from capacitance curve. I have attached the graph from datasheet and plot of extracted data.

When I am trying to run it with 3.5V VGS applied, I am getting the following errors "Estimated output characteristic gradients for data extrapolation must be greater than zero. Use linear extrapolation" and "At time 1.614255e-19, one or more assertions are triggered. Tabulated drain-source currents, Ids(Vgs,Vds) has resulted in a current in opposition to the applied voltage. Consider modifying the tabulated data to ensure that all operating points are contained within the data, and that the signs of the currents and voltages are consistent. Clipping the current to zero. The assertion comes from:

Block path: EV_Motor_Drive_version5_LUT/Subsystem/N-Channel MOSFET".

I tried to manually do linear explorate the curve data, but the curve is not being defined as per real condition in saturation region.

What am I missing here? I am attaching the data and model as well. Thank You!

Answer 1

[Yifan Yang]

Hi Hassan,

This is the warning message other than error message. There are two warning message here:

1. "Estimated output characteristic gradients for data extrapolation must be greater than zero. Use linear extrapolation." We typically use a threshold-based equation to assist with extrapolation to prevent excessively large values for the current and to ensure the MOSFET is saturated, rather than always in the linear region when the data is insufficient, as in this case for Vgs=7V or Vgs=10V. If this approach is not successful, switch back to linear extrapolation. If you would like to use linear extrapolation, you can just leave as it is, it will be used as is.

2. "At time 1.614255e-19, one or more assertions are triggered. Tabulated drain-source currents, Ids(Vgs,Vds), have resulted in a current opposing the applied voltage. Consider modifying the tabulated data to ensure that all operating points are contained within the data and that the signs of the currents and voltages are consistent. Clipping the current to zero." This warning message usually indicates that the MOSFET does not have a clear threshold showing when it is on or off. To avoid this message, it is suggested to add Vgs data near the threshold.

[Hassan Ali]

Hey Yifan,

So I added more data in LUT, including when VGS = 0, 1, 1.6V, 2, 3.5 and 4.5 and I have omitted the higher VGS values of 7 and 10V.

I tried to get the characterisitc curve through MOSFET characteristic example, but there seems to have negative current at the initial simulation time. Anyways, when i try to simulate it in my EV model, the number of times the warning comes have reduced, but it is very very slow, which I shouldn't be as the LUT is used to speed up the simulation.

I have attached the characteristic curve, model, and look -up table. Could you kindly guide whats is creating this issue?

Thank You!