#122 Wrong S-Paramter simulation

[Stefan]

I want to design an oscillator at 7GHz with the BFP405. Because the current model was in the library I used the non-linear npn model with four ports. The fourth port is for the substrate. I inserted all available parameter and added the parasitic elements of the chip and housing as lumped element. Uce=3 V Ib=116µA F=1 ... 10GHz
The S-parameter simulation shows for S11 and S22 a unstable curve along positive imaginary part of the Smith-diagram.

I did the simulation with AWR Microwave Office and Qucs-Studio and got similar results to the available S-parameter of the transistor.
I simulated only the transistor, which should be the chip, and got sensible results. I checked the transistor data and the value of all components.

[Stefan]

Sorry. I forgot the other pictures

[magneto98]

Could you please also post the qucs schematic file so that the error can be reproduced.

[Stefan]

Yep. Here is a 7-ZIP file including the Touchstone file.

[Guilherme]

Is there a reason to use DC blocks as parasitic capacitances on the BFP405 subcircuit?
I replaced them by normal capacitors and the result approach the measurement. See attached figure. Can you please check it again?

[Stefan]

First, thank you that you looked at that problem. I didn't expect any response after 9 month.

Second. I didn't used the DC block capacitance on purpose. And I can't see the different if I look at the circuit. I will test it.

[Stefan]

I changed the capacitances and I got the same result.

[Guilherme]

I am trying to address pending bugs to move on into a new release. Sorry it took so long.

It seems to work fine o my side.
Hum, can you please extract the attached packaged project and give it a try?

[Stefan]

Hi Guilherme,

The simulation looks consistent. Problem solved!

Thank you for your support.

[Guilherme]

Gook to known. Closing.