s2p Co Simulation in ADS

[Atif Hussain Shah]

Hi,
I am Simulating LNA S2P file in ADS using Cosimulation method. The results are in accordance with datasheet when i use ideal ground( Global schematic ground) but when i use circuit ground(local ground)the results are not good. What could be the reason?
I want to simulate the effects of local ground which will be the real practical situation.

 

[Hawaslsh]

I'm not 100% familiar with cosimulation in ADS, but I assume it's combining the S-parameter results from the layout structure and the s2p files for the LNA?
Few questions:
1) How do you define the potential of the ground plane in a cosimulation? When i used momentum in ADS, the port definition defined what metal layer is the reference.
2) As a quick sanity check, could add an inductor to ground and see how added vias would affect your grounding. (I dont think this is the case)

3)Perhaps you can define a "more" CPW input by adding (ground or differential) ports at the inputs if ADS can let you? That way you ensure your metal planes are defined and allow return currents to terminate on the top and bottom plane.

I lost my license to ads, I cant test any of this, so take all of this with a mountain of salt.
Sami

 

[Atif Hussain Shah]

Thanks Hawaslsh for Reply
1) How do you define the potential of the ground plane in a cosimulation? When i used momentum in ADS, the port definition defined what metal layer is the reference.
The vias on top conductor are connected to bottom ground plane. I am using a pin on top conductor where vias are placed and connect that pin to my S2P ground.

 

[volker@muehlhaus]

I would expect that your S2P already includes a ground path (as measured) and if you insert that into the EM co-sim, you now have included a ground path twice: in the EM model and (implicitely) in the s2P data. And that might matter because there is quite some length (=inductance) between the device pins and your ground pin in layout.

Besides that, I see no mistake in your model shown above, if your ports for Momentum refer to global ground.

The comment by Hawaslsh on CPW side grounds isn't wrong, but might create some issues when adding these ground pins on the edges (port size). From experience, I would leave them out here because of the thin substrate with many vias connecting both grounds.

BR,
Volker

 

[Atif Hussain Shah]

@volker@muehlhaus Thanks for reply.
I have read the concept of globle and local ground from your comments in keysight forum. From that what i have understood is
That the schematic ground is global ground and Rf ground( connected to Infinite ground plane through vias) is local ground.
Have i understood right? I am confused between local and global ground.

--- Updated Nov 19, 2020 ---

I would expect that your S2P already includes a ground path (as measured) and if you insert that into the EM co-sim, you now have included a ground path twice: in the EM model and (implicitely) in the s2P data. And that might matter because there is quite some length (=inductance) between the device pins and your ground pin in layout.

Besides that, I see no mistake in your model shown above, if your ports for Momentum refer to global ground.

The comment by Hawaslsh on CPW side grounds isn't wrong, but might create some issues when adding these ground pins on the edges (port size). From experience, I would leave them out here because of the thin substrate with many vias connecting both grounds.

BR,
Volker

How to ensure that my ports in momentum refer to global ground?

 

[volker@muehlhaus]

How to ensure that my ports in momentum refer to global ground?

If the minus-terminal in port editor is "Gnd" then the port reference is global ground (infinite ground at bottom of substrate). This is what you want for your model.

 

[Atif Hussain Shah]

If the minus-terminal in port editor is "Gnd" then the port reference is global ground (infinite ground at bottom of substrate). This is what you want for your model.

Thanks @volker@muehlhaus for your reply.
I have another question." In my model i have an LNA connected to attenuator through transmission line. Individually they give desirable results but when i connect them, S11 is unrealistic(greater than >0).
I have attached figure in which port 21 is connected to the transmission line. I have attached port definitions and momentum warning as well.
I have tried to solve that warning by using edge/area pin but warning does not go away?

 

[BigBoss]

What is the Width of the Transmission Line ?

 

[Atif Hussain Shah]

What is the Width of the Transmission Line ?

For microstrip it is 0.54mm and for CPWG it is 0.48mm.

 

[BigBoss]

The warning message is coming from this too Large Edge Port.
If you have put this port for Ground, it's a mistake.You shouldn't..
Since Momentum has a Global Ground, you don't have to place another extra Port for that.Global Ground ( implicit) ties all GND Pins of ALL Ports so all grounds are tied together on a single infinite GND Plane.
Check also "Physical Connection" of the VIAs in 3D mode...( Tools Menu )

 

[Atif Hussain Shah]

View attachment 165796
The warning message is coming from this too Large Edge Port.
If you have put this port for Ground, it's a mistake.You shouldn't..
Since Momentum has a Global Ground, you don't have to place another extra Port for that.Global Ground ( implicit) ties all GND Pins of ALL Ports so all grounds are tied together on a single infinite GND Plane.
Check also "Physical Connection" of the VIAs in 3D mode...( Tools Menu )

Thanks for reply
"If you have put this port for Ground, it's a mistake." No pin is placed at ground. I am attaching picture which will clearify.