Interpreting RFSim99 modeling results

[ardf]

Hello,

I'm designing an ARDF (Amateur Radio Direction Finding) transmitter for 144 MHz. The work-in-progress design is published on www.easyeda.com.
The output stage of the transmitter utilises RQA0011DNS HF power MOSFET.

I've used the provided  S-parameters and the reference design (page 4 of the datasheet) for modeling the output network in RFSim99.

According to the model, there is no match to 50 Ohm and the result is lying in the unstable area (screenshot is in the attachment).

Am I missing something? Should I still use the reference design or design my own match to 50 Ohm?

Thank you!

[OwO]

The shaded region tells you what impedance presented to that port of the amplifier will make it unstable, not the output impedance looking into the amplifier. The definition of "unstable" here is making the other port's reflection coefficient >1. That means if you present an impedance in the shaded region to the output of the amplifier, its |S11| will become >1. Also don't expect the output impedance looking into the amplifier to be 50 ohms, it's not a conjugate match but rather experimentally determined for highest power efficiency. Stick with the datasheet reference design for the output matching, and match input to 50ohms. To stabilize the amplifier add some loss on the input side.

[ardf]

Thanks for the clarification on stability circles!
I knew that something is not right in my assumption, because whenever I model the reference designs of other transistors in RFSim99, the match almost always was "falling" into an unstable region.

Now for the input matching.
My amplifier stages are: Si5351A → 2SK3078 → RQA0011.
I copied (sorry) a part of Baofeng UV-3R schematics for 2SK3078 → RQA0011 interstage match (in the attach), but left the Renesas's output match. Now I think it was a bad idea, as it is quite different from the Renesas's reference design, and I will need to redesign that part of my schematic, am I right?

Thanks!

[rf-messkopf]

You do not seem have the input matching network included in your simulation, unless it is already contained in the S-parameters of the device. From experience I found that, even though there is some reverse isolation between input and output due to the transistor, a variation in the input matching can have a good deal of influence on the amplifier S22, i.e., the impedance looking into the amplifier output.

And clearly, if you want to control gain and flatness as well, you will definitely need the input matching network in your simulation.

[ardf]

Hello, rf-messkopf,

Here I included the reference design input matching into the simulation as well. Looks like the impedance isn't changed much.

[OwO]

Enable stability circles again? That 33ohm resistor on the input may be just enough to kill the instability... provided there is still enough gain of course.

[ardf]

You're right, the resistor did the job!
Sorry, my fault. I should have  been include the stability circles in my previous post.

Thanks again!

[OwO]

That's better, but if your antenna is a short circuit and your input is a capacitor it can still potentially oscillate