Tunable Bandpass Filter Questions

Hey everyone --

Right now I am designing a tunable bandpass filter from 20 MHz to 1 GHz with 10% 3dB bandwidth. I am trying to design a tunable tank circuit using pin diodes to switch a bank of capacitors, so if I have 4 capacitors with 4 pin diodes (in parallel), I can switch on different combinations. I am having a hard time realizing this with my limited knowledge (brand new to RF design). I am trying to model this in ADS but I don't know if I am doing it correctly.

Basically I have a model for forward biased and reverse biased modes:

Forward biased is a inductor and series resistor. Reverse biased is an inductor in series with a capacitor and resistor in parallel (this is from Microsemi's PIN application notes 701). Reading the data sheet (www.microsemi.com/datasheets/mmsm 7 04.pdf ) for the MPP4201 PIN diode, I see that it says there is a low series inductance (.2 nH) and low parasitic capacitance (.06 pF). It also says that Rs is equal to 2.5 ohms at 20 mA. I'm assuming I can use these values for my model (the inductance and Rs for forward, and the capacitance at .06pF for reverse biased). But what about the parallel resistance for the reverse biased mode? I've basically just put a high value resistor there. Since the impedance is largely based off the parasitic capacitance (.06 pF) it would look like an open circuit and basically be off, correct?

I am not sure if this is the correct way of approaching this problem. I'll try to upload a design in a few to give more clarity. Do you guys have any suggestions for this type of design?

Thanks,

Sam

I would start with the filter design 1st. Assume your PIN diodes are either a perfect open or a perfect short. Can you design a filter to meet your needs? If you can't do this, worrying about the PIN diodes is the least of your worries.

I hope U can upload ur design here,so I can see what's ur method correctly.By the way, I have been studying tunable filters.

sambow87 said:

Hey everyone --

from 20 MHz to 1 GHz (brand new to RF design).

Click to expand...

A realistic approach would be to try to learn how to do less demanding filter designs until you have gained some experience. Try to learn how to do a 800 to 900 MHz tunable filter, and let us know when you have figured that out.

To tune a BPF over almost 7 octaves from 20MHz up to 1GHz (and keep the 10% BW) you are going in some problems higher than actual global crisis.

I was tempted to place a similar comment, but I realized that it would be riding a dead horse. The original thread is five months old and the absurd project apparently hasn't been continued.