QUCS RF Simulation and impedance matching

[aht2000]

I am trying to use QUCS to simulate the input and output impedance of a simple common emitter amplifier based on BFP420. As this is new to me, I would like to verify with the experts if the steps I performed are the right ones.

• I created a schematic with the amplifier I want, and updated all the parameters of the transistor with the one from Infineon (the built in QUCS spice mode was not giving correct DC bias and current results as per actual physical implementation).
• Performed a DC analysis to get the VCE and IC at which the transistor is operating. I verified the same with LTSpice and got similar results for DC bias. See attached "DC Bias.JPG".
• Based on the VCE and IC, I selected the corresponding S2P file from Infineon website. See attached "BFP420_w_noise_VCE_2.5V_IC_1.5mA.s2p".
• I replaced the transistor in the schematic with an s-parameter file component, and edit the component to point to the .s2p file. See attached "S-PAR Sch.JPG".
• I used the built-in s-parameter.sch template from QUCS to complement the existing schematic with the correct simulation and graphs.
• I ran the simulation and got the attached S11 and S22 results at 10.7MHz "S-PAR S11 S22.JPG".
• From the smith chart graph and clicking on the S11 marker, I selected the "power matching option" to create an LC matching to have an input impedance of 330 ohm (output of the a Murata 10.7 ceramic filter). See attached "S-PAR Sch with 330 ohm inp match.JPG" and "S-PAR S11 S22 with 330 ohm inp match.JPG" for the results.
• I can see that now I am getting the required input impedance but the S22 and output impedance has changed. I then tried to put a matching network at the output to match a 1500 ohm (the input of an SA612 mixer) but then the input impedance changed.
• I remember reading earlier that the input and output matching network affect each other, and it may need some kind of iterative approach to reach the required input and output.

1. Are these the correct steps to follow? I do not thing that QUCS has a feature to adjust by itself both input and output matching networks taking into consideration the interaction between them.
2. Is there another free simulation SW that can do that?
3. If none, would I just have to decide which matching network is more important to my use case and live with the consequences?
4. Is there a way to reduce the interaction between the input and output matching networks to almost nil? This will be my best case scenario.

 

[FvM]

Part of your design goals makes no sense. Forget about matching the amplifier output to SA612. It serves no purpose and only causes trouble. Loading the IF filter with 330 ohm is suggested, but doesn't neccessarily need LC matching. If you aren't noise limited (shouldn't be the case at this point of the signal chain), a simple matching resistor can do the trick.

 

[BigBoss]

s-parameters are not used with DC biasing conditions, Useless..
Matching 1500 Ohm to some S22 will not work. The difference is huge.. Matching is also not necessary.
Design your LNA under 50 Ohm loading condition for lowest noise then connect to SA612.That's it..

 

[aht2000]

Part of your design goals makes no sense. Forget about matching the amplifier output to SA612. It serves no purpose and only causes trouble. Loading the IF filter with 330 ohm is suggested, but doesn't necessarily need LC matching. If you aren't noise limited (shouldn't be the case at this point of the signal chain), a simple matching resistor can do the trick.

I thought that impedance matching is always needed but seems that it is not the case, and in fact seeking matching in some cases will cause troubles.

Can you please give few examples for when matching is not mandatory, or may be the other way around, when matching is a must.

For example, I always thought that a high output impedance to a low input one will need matching to avoid signal loss, while never understood why the opposite needs matching. What is wrong with having a low output impedance amplifier for example feeding a high input impedance following stage, specially that everything is on the same PCB, and we are in the VHF range.