How do you retrieve mos parameters?

[medwatt]

In ngpsice, let's say you want to know the gm of the mosfet called m1, you can do so by writing save @m1[gm]. How do you do that in qucs?

[ra3xdh]

You have to use script simulation type if you need to extract variables that are nor voltage nor current or its dependencies. This simulation type doesn't support the save command. Use write instead. You can find some hints how to use this simulation type here: https://mos-ak.org/lausanne_2017/presentations/T03_Kuznetsov_MOS-AK_DATE_2017.pdf Sorry, there is no better documentation.

[medwatt]

Thanks for your reply. I did the simulation using Nutmeg script. This netlist that was generated is shown below. It simulated fine. I could see all the parameters in the tabular diagram and I could get the values of the node voltages and currents. However, while @m1[gm] is available, it shows no data in the diagram.

* Qucs 2.0.0 /home/medwatt/.qucs/test_prj/common_source_testing.sch
.INCLUDE "/home/medwatt/qucs-s/share/qucs-s/xspice_cmlib/include/ngspice_mathfunc.inc"
* Qucs 2.0.0  /home/medwatt/.qucs/test_prj/common_source_testing.sch
.INCLUDE "/home/medwatt/models/NMOS_VTL.inc"
V2 _net0 0 DC 1
M1 out  in  0  0 NMOS_VTL L=75n W=3.25u
V3 in _net1 DC 0 SIN(0 0.1M 1K 0 0 0) AC 0.1M ACPHASE 0
C1 0 out  5P 
R1 out _net0  5K

V1 _net1 0 DC 0.357
.control
echo "" > spice4qucs.cir.noise
echo "" > spice4qucs.cir.pz
set filetype=ascii
save all @m1[id] @m1[vth] @m1[gm]
op
write common_source_testing_custom.txt all
destroy all
reset
exit
.endc
.END

[ra3xdh]

You can also try to specify the equation dependent on device parameter as the Nutmeg equation. Keep in mind that Qucs-S may add letter (i.e M,Q,D) to the device name to ensure the SPICE compatibility. You may require to look into the netlist to find exact device reference.

[ra3xdh]

Everything is correct. You have got a set of 1x1 point plots. You may look at it only in tabular form.

[medwatt]

Maybe, my post was not clear. I am using tabular form. It shows no data.

[ra3xdh]

You may look at #277 how to assemble a set of 1x1 points into one curve.

[ra3xdh]

Yes, I understood now. The @ symbol is reserved for PlotVs() emulation. Try to assign a variable let m1_gm=@m1[gm] and plot it instead of plotting @m1[gm] directly.

[medwatt]

This doesn't work in dc sweeps. You can see in the table below that it only returns the value in the first step value. Using save @m1[vth] doesn't have this problem.

[ra3xdh]

After some experimenting I have found the following approach how to use save with the Nutmeg script and bypass the reserved @ symbol. Look at the example schematic. The result is identical to CLI simulation.