Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

[SOLVED] ADS EM simulation

Status
Not open for further replies.

dirac16

Newbie level 5
Joined
Jan 20, 2021
Messages
9
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
72
Hi there;

I am new to the ADS layout tool. I have a simple impedance matching circuit for which I want to plot the input reflection coefficient and check whether it agrees with the EM simulation performed on the generated layout. Below is the single stub matching circuit created using MLINs to match a ZL=100+75j to a 50 Ohm line.

cir.jpg,.jpg

The center frequency is set at 1GHz. As is expected the magnitude of input reflection coefficient should approach ~0 at 1GHz. Running a schematic simulation and plotting S(1,1) the result is simply verified.

str1.jpg,.jpg

Now in order to prepare for Momentum simulation I generated the layout as shown below.

lay.jpg,.jpg

I then went to the EM menu and chose Simulation Setup. In the pop-up window I created a new substrate with the same characteristics as MSUB1. In the Ports setting I set Port1=50 Ohm and Port2=100+75j. In the Options setting I also activated Edge mesh to increase accuracy near the edges. Then I ran S simulation and it ended without any warning or errors. The following diagram for S(1,1) is plotted:

str2.jpg,.jpg

Quite weirdly the center frequency has now shifted to ~1.65GHz. I am not able to physically justify the result because it doesn't make sense at all. I am really stuck at what made this happen. Does anyone know what's going on here?
 

Attachments


BigBoss

Advanced Member level 5
Joined
Nov 17, 2001
Messages
4,943
Helped
1,498
Reputation
2,994
Reaction score
1,339
Trophy points
1,393
Location
Turkey
Activity points
29,808
Assign P2=100+j*75 Ohm doesn't make sense for Momentum.You should simulate your Physical Layout with P1=P2=50 Ohm then you have to simulate your layout in a schematic view by using Co-Simulation option now by assigning P2=100+j*75 Ohm.
GND cannot be used to tie an element to GND in Momentum because it's a symbolic and used for verification between schematic and layout views only.So you Open Stub is not really connected to GND.The error is coming from that point..
 
Last edited:

dirac16

Newbie level 5
Joined
Jan 20, 2021
Messages
9
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
72
Assign P2=100+j*75 Ohm doesn't make sense for Momentum.You should simulate your Physical Layout with P1=P2=50 Ohm then you have to simulate your layout in a schematic view by using Co-Simulation option now by assigning P2=100+j*75 Ohm.
GND cannot be used to tie an element to GND in Momentum because it's a symbolic and used for verification between schematic and layout views only.So you Open Stub is not really connected to GND.The error is coming from that point..
So my shorted stub is acting as Open, now what should I do? How can I make a GND connection in Momentum simulation?
--- Updated ---

Assign P2=100+j*75 Ohm doesn't make sense for Momentum.You should simulate your Physical Layout with P1=P2=50 Ohm then you have to simulate your layout in a schematic view by using Co-Simulation option now by assigning P2=100+j*75 Ohm.
GND cannot be used to tie an element to GND in Momentum because it's a symbolic and used for verification between schematic and layout views only.So you Open Stub is not really connected to GND.The error is coming from that point..
Also I replaced the shorted stub with MLEF and did both schematic and layout simulation for the case Port1=50 Ohms and Port2=50 ohms. The results were different again.

stra.jpg,.jpg

The red curve shows the magnitude of S(1,1) for the simulated schematic, and the blue curve for the layout. What's wrong really
--- Updated ---

Sorry I meant MLSC, I replaced the shorted stub with MLSC
 
Last edited:

volker@muehlhaus

Advanced Member level 5
Joined
Apr 11, 2014
Messages
2,547
Helped
994
Reputation
1,992
Reaction score
978
Trophy points
113
Activity points
15,127
The solution shows by BigBoss is simple, but the actual via effect in not included in EM. You get an ideal connection from schematic, which is not very accurate. The better solution is to include the via in EM, with realistic dimensions.

So my shorted stub is acting as Open, now what should I do? How can I make a GND connection in Momentum simulation?
You need to draw the actual via.

Choose a layer (e.g. layer 'hole') and draw your via in layout. That can be a line for a 'sheet' via, or a more realistic circle that represents a round (drilled) via.
Then, you need to map that layer as a via in your Momentum substrate definition.

Check your model in 3D EM preview, to make sure the via is included in the actual EM model
 

BigBoss

Advanced Member level 5
Joined
Nov 17, 2001
Messages
4,943
Helped
1,498
Reputation
2,994
Reaction score
1,339
Trophy points
1,393
Location
Turkey
Activity points
29,808
Thank you, it eventually worked! Now what should I do if I want to enclose the structure into a box with a certain height? Should I add a metal cover above the substrate? Or what?
It's you who will decide that a cover is necessary or not.Also please take Volkers' warnings into account. Your layout is not %100 realistic, VIAs should be considered too..
 

dirac16

Newbie level 5
Joined
Jan 20, 2021
Messages
9
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
72
The solution shows by BigBoss is simple, but the actual via effect in not included in EM. You get an ideal connection from schematic, which is not very accurate. The better solution is to include the via in EM, with realistic dimensions.



You need to draw the actual via.

Choose a layer (e.g. layer 'hole') and draw your via in layout. That can be a line for a 'sheet' via, or a more realistic circle that represents a round (drilled) via.
Then, you need to map that layer as a via in your Momentum substrate definition.

Check your model in 3D EM preview, to make sure the via is included in the actual EM model
Thank you, great point. I'll take that in mind

Also please take Volkers' warnings into account. Your layout is not %100 realistic, VIAs should be considered too..
Yes, sure.

It's you who will decide that a cover is necessary or not.
Yes I want to include a cover too and consider the effect of magnetic fields within the box. How should I go about that? I have not found the documentation useful, or it's so overwhelmed that it is hard to follow.
 

BigBoss

Advanced Member level 5
Joined
Nov 17, 2001
Messages
4,943
Helped
1,498
Reputation
2,994
Reaction score
1,339
Trophy points
1,393
Location
Turkey
Activity points
29,808
Metallic Closures shouldn't be very close to layout so that as magnetic field won't be affected.You can also simulate this using with HFSS or CST but it will be time consuming and tricky.But as a rule of thumb 10mm or more height seems to me appropriate.
 

volker@muehlhaus

Advanced Member level 5
Joined
Apr 11, 2014
Messages
2,547
Helped
994
Reputation
1,992
Reaction score
978
Trophy points
113
Activity points
15,127
In Momentum stackup you can add a ground top cover, instead of open boundary, with some air thickness between circuit and cover.

For the sides, you can try modelling with vias that go all the way from bottom GND to top GND.
I never tried that, but I think it should work.

Also, there is some information in the ADS help on simulating inside a box.
 

dirac16

Newbie level 5
Joined
Jan 20, 2021
Messages
9
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
72
Alright, about the vias, I h
In Momentum stackup you can add a ground top cover, instead of open boundary, with some air thickness between circuit and cover.

For the sides, you can try modelling with vias that go all the way from bottom GND to top GND.
I never tried that, but I think it should work.

Also, there is some information in the ADS help on simulating inside a box.
Alright. Still I am not sure how to add vias. I have a shorted stub that must be connected to bottom ground plane. Where actually do I need to put a via on the stub?
 

volker@muehlhaus

Advanced Member level 5
Joined
Apr 11, 2014
Messages
2,547
Helped
994
Reputation
1,992
Reaction score
978
Trophy points
113
Activity points
15,127
Still I am not sure how to add vias. I have a shorted stub that must be connected to bottom ground plane. Where actually do I need to put a via on the stub?
It is your layout, so you decide. An ideal ground connection (as you used in schematic) is not realistic. The MSLC circuit model assumes a 2D sheet via, but in PCB we want to use a drilled round via. ADS offers a circuit model for that round via, if you want to include it at schematic simulation already. Otherwise, you can draw it in layout yourself and include that in EM simulation (circle on layer "hole" which is mapped as a via in Momentum stackup).

I have attached a simple workspace example, hope this helps!.
 

Attachments

Last edited:
Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top