The data sheet application and test circuits show the AD8331 LNA amplifier externally coupled to the VGA section with the use of two dc blocking capacitors. Is there any obvious reason why the broad band coupling capacitors could not be replaced with - for sake of example - a two pole, 10% bandwidth LC bandpass filter designed for 200 Ohms Zin and Zout assuming the DC block is maintained and good RF layout practice and component placement?
Note: the VGA input impedance is specified in the data sheet as 200 Ohms
You did the right thing and proceeded with the pcb breadboard. As you know the filter worked fine, no problems and the bandwidth came out as expected which as you know means the input and output impedances where 200 Ohms resistive. You also know there was no amplifier instability, which you never really expected but you exhibited good judgment in attempting to get a quick answer from AD before committing on a pcb.
Look at it this way you received a great if somewhat tardy answer for your question from the best analog-RF microwave engineer you ever met. And you where right in your judgement all along.
I say "Attaboy Bruce, Well done."
Gee I wonder if you can get a discount on the parts since you are friends with the AD applications engineer?
Attaboy Bruce, thank you, well done! This is similar to what I am designing and I went here to see if anyone else has tried this. I actually have a 3 pole bandpass on the LNA output centered around 40 MHz. Every thing looks good in LT spice. I actually designed the filter for 100 ohms differential in and 200 ohms differential out. I did this because the LNA output impedance is specified as 5 Ohms at 10 MHz and because the data sheet says "If an LNA output is routed to a remote PC board, it tolerates a load capacitance up to 100 pF with the addition of a 49.9 Ω series resistor or ferrite 75 Ω/100 MHz bead." I am not routing to another board, just to the Postamp but that made me decide on the 100 ohm input impedeance on the filter.
On the VGA output I have a 1st order bandpass tank across the outputs for a narrow bandwidth mode which is PIN diode switched in and out. I designed this filter for 200 Ohms in and 200 ohms out to go to the to a second 8331 VGA in the cascade with its LNA disabled. I am feeding the second filter with 100 ohm resistors off the VGA post amp outputs. I will run at high gain on post amp. One concern I have is the stability of the output at high frequency because the parallel tank has 190 pF in it which would look like a capacitive load at high frequencies. This should look like a real impedance at at the resonant frequency, the question being stability at high frequency, since the data sheet shows more ringing with increased capacitance. I would think the 100 resistors on the post amp outputs should eliminate stability issues. What are you thoughts on this? Hopefully the Spice model is reasonably accurate in modelling capcitive loads on the Preamp and Postamp outputs.
I would really appreciate if you could send me your layout because I would like to have an idea of a layout approach for the LNA to VGA filter as a guidline since we are focused on first pass success here.