Defining length in an expression to be evaluated

The entered length of the device:

pv("/I7/M1" "l" ?result 'instance)

or the effective length (see "spectre -h bsim4", for example - relevant help will for the specific model you're using)

pv("/I7/M1" "leff" ?result 'output)

Regards,

Andrew.

When I perform a componenet parameter sweep with DC analysis, it is giving a single output of a specific operationg point (l=5u or leff=4.8...). In my expression I would like to see a parameter "l" which changes with different sweep points. I also created a .scs file to save each operating point for the specific device that I am extracting the waves of parameters.

Best,

Can

If you are sweeping the length of the transistor in the dc sweep, then there's no need to explicitly save the l, because it's the x-axis of any curve you get. For example:

 //

model nch bsim4 type=n

m1 (drain gate 0 0) nch w=20u l=0.8u
vgs (gate 0) vsource dc=1
vds (drain 0) vsource dc=1

dc dc dev=m1 param=l start=0.5u stop=2u step=0.1u

save drain gate m1:d

If in ViVA I plot m1:d, it will be versus l. I can always use the xval() function to get a waveform of the xaxis - e.g. xval(i("m1:d" ?result 'dc)) .

You can save the operating point info from the transistor (e.g. save m1:oppoint) - but since the length or effective length is not part of the operating point (it won't vary with the operating point), it can't be saved - the only time it would vary is as a result of an explicit sweep, in which case it's in the x-axis as I said earlier.

Regards,

Andrew.