Model component using table of impedance versus frequency

[DanielMaxwell]

I would like to model a device which I may describe as something that has a measured impedance versus frequency in a tabular format. Is there a way to do this in Multisim. I have tried researching the Analog Behavior Model (ABM) using possibly a B Source but can't come up with the syntax that works.

The following example was slightly modified for my purposes from another competing Spice Simulator user forum and it works there. This model shows a table using frequency, a real and an imaginary resistance. The resulting AC Analysis interpolates values of resistance and reactance based on this table. If I could get this to work in Multisim, or if there is another solution, it would be nice to receive help.

.subckt DeviceRI 1 2
R1 1 10  0.001
V1 10 11 0
B1 11 2 V=I(V1)  R_I FREQ=
+ 1000, 8, 0.5
+ 1100, 8.1, 0.7
+ 1200, 8.2, 0.5

+ 1300, 8.33, 0.6
.ends

Thank you.

Daniel

[OStep]

Hello Daniel,

Unfortunately the present release of the product does not support transfer function models that use frequency tables. However, this functionality has been requested in the past and is presently a work-in-progress feature.

Thanks,

[DanielMaxwell]

Well, rather than stopping in my tracks, I would think that there might be a work around solution such as trying to convert an impedance vs. frequency table like this to some type of a circuit that can be described with either a formula or spice components. Maybe someone else has some ideas. I am not well versed in creating a model but would be interested in learning.

[OStep]

Daniel,

The best potential workaround is to use the TRANSFER_FUNCTION_BLOCK component under Sources/CONTROL_FUNCTION_BLOCKS

The model that this component uses supports a frequency domain transfer function that is described using rational polynomials. If you can curve fit your complex data into an equation of this format, that should do the job.

It is theoretically possible to reproduce the response of your transfer function by created a lumped model using discrete components (by connecting together integrators, derivatives, delays, etc) but this far more challenging than to use the Transfer Function component mentioned above. 

Let me know if that helps.

Oleg