Hi all,
I'm looking to design a tunable filter using a pair of diodes with the following C/V curves:
I've opted to simplify my model in AWR's Microwave Office (13.03r build 8415 Rev 1) by representing them using the Simple Varactor Model (VRCTR). Those familiar with this model may know that you can provide it with CV data and it will attempt to approximate the curve using a polynomial of an order specified by the user (1-7).
I have attempted to do so in this case, however the data from the curve on the left, when entered into the model, will throw me a message saying the approximated curve has an error greater than 20% (which causes the simulation to fail) unless I use a 7th order polynomial. Even then, the curve is still off and causes inaccuracies in behavior when used in my desired circuit. The data from the curve on the right will always cause an error greater than 20% regardless of which order polynomial I specify.
This doesn't make sense to me, since those curves appear pretty straightforwardly logarithmic and shouldn't take more than a 3rd or 4th order polynomial to approximate accurately. Has anyone else run into this problem with the VRCTR model, and if so how did you remedy it?
Additionally, I've given some thought to other ways I can model a varactor diode in AWR. First, the DIODE2 model seems promising since I can simply enter the linear capacitance as a parameter, but I will have to use some scripting or otherwise similarly inaccurate, excel-generated equations to make it dependent on a swept voltage elsewhere in the circuit. I'm also considering the PNCAP model, although it's not exactly clear to me how to set its CJ0, VJ, and FC parameters to exhibit the curve that I want.
I'd appreciate any thoughts or wisdom community members might have regarding this type of modeling.
Thanks!