How to utilize s-param data supplied by manufacturer for device simulation correctly?

[afz23]

I am designing an amplifier at 12ghz using infineon part cfy25.
I used it's s-param for design,but did not get the desired performance after fabrication.

I am having concern the way ,the manufacturer has extracted these s-param for this device,on inquiry he has informed that,he has extracted this data after mounting device on a PCB and choosing reference planes at device lead edges.
When I simulated this amplifier,I considered package edges as reference.
Please give suggestion,for using this s-param data correctly in simulation.As additional device lead lengths of 1.5mm on each side,will spoil the performance of amplifier,if not accounted properly.

 

[volker@muehlhaus]

Yes, 1.5mm length can make a big difference at 12GHz. Why don't you go back to simulation and check the effect on your design?

Also, if you have used measured S2P data, these already include some ground connection and if your's is different, differences must be expected.

 

[afz23]

Please see this figure,to see the measured reference and simulated(desired) reference plane


Do you think,leads need to be de-embeded for simulation purpose to match the practical condition of mounting the device.

 

[volker@muehlhaus]

As said: yes, this is a problem. And as said, you should move the ref plane in your simulation to match the measurement ref plane. If the leads are included in the S2P data, do not include them again in your schematic or EM simulation. They can be part of the layout footprint, but should not be part of simulation.