Hi Anil, I am not pretty sure what is the partitioning problem you are facing but I encourage you to watch RFPro 5mins tutorial playlist on my channel. RFPro provides a more modern way of doing partitioning between layout and circuit components and not only that I can automatically created back-annotate schematic for EM-Circuit cosimulation for you: ua-cam.com/play/PL9OnCetH8TYpxUnG6_9jKZIi1PsJtXXcN.html
Hi, I am trying to simulate a PCB to check for 0Hz magnetic field generation. Does it is plausible to do it with FEM. I already did all the preparations and i can run momentum simulations but it gets stuck when i try to do FEM simulations. In the logs status message it gets stuck in the part where it sais automatic inititial target mesh size. At first it shown a really large number 10^6. But after seeing you video i get it to a similar number as you get. Could you give me some insight pls. By the way. The frequency i am interested in is 0hz to 10 hz. But i swip from 1 Ghz to 0 as if intry to use a slower sweep it gets errors.
Sir, I'm working on 1:4 powerdivider and combiner, results with method of moments were satisfied s11 at desired frequency was -25dB, will the practical results will be matching can I go head with fabrication. 2. Sir if i want to do FEM for same design how to add isolation resistance 100 ohm in layout. Where as on method of moments we added in schematic.
Its difficult to comment when I have not seen the layout you simulated and how well you simulated but in general with MoM you can expect measured results within +/-5% considering the tolerances and FEM should be within +/- 1 or 2%. Process to add 100Ohm resistor is same between Momentum and FEM, jus the emmodel will change behind the scene. You could either use emcosimulation method like what you seem to be using now or directly add 100Ohm in layout as I have shown in the Power Divider design video. Best of luck...👍
Helle thank you for the video it"'s verry useful. I have 2 questions which is the difference between HFSS and FEM in ADS. and the second question is which is the difference betwenn momentum RF et momentum microwave and FEM
Hi Farah, Some of the points are provided here for your help. For more info you can refer to ADS documentation or contact tech support team: 1. Its difficult to answer HFSS vs ADS FEM in detail but just be aware that HFSS is also based on FEM for frequency domain analysis so there is lot of similarity how Mesh is created, refined and problem is solved. 2. Momentum is a Method of Moments based planar EM solver whereas FEM is a full 3D solver that can handle any arbitrary 3D geometry for simulation 3. Momentum RF is based on quasi-static technique of EM simulation whereas Momentum-Microwave is a full wave solver. ADS documentation provides detailed description of Momentum and FEM, it will be worthwhile if you go through the documentation.
Hi Anurag, thank you a lot for your great tutorials, can you please help me, I want to simulate passif structures (Coils, MIM capacitors and transformers) on Silicon substrate. I don't Know what is the best simulator to use it: Momentum RF, momentum Microwave or FEM ? I want to study the variation of Q factor inductance, capacitance, mutuel inductance etc as a function of frequency. Really your channel is SuperGreat Thank you.
For the structure you mentioned Momentum would be sufficient as you really don't need a full 3D simulator like FEM. Whether to use Mom RF or MW, is your choice but mostly MoM-RF is sufficient as those structure are electrically small component and RF mode works very well till half-wavelength. Kindly refer to ADS documentation for more details.
your videos were a great help. I followed all the steps in the fem simulation; I get a "Simulation Error" because the command line is too long I need some help
Hello Mr. Bhargava. Thank you for great tutorials. I have a question regarding the port scaling factors under the options tab in FEM EMSetup. In ADS2016 by default the factors are 100. Isn't this erroneous? This makes mostly the whole surface area of the simulation box as port input/output area. I think those factors also should be scaled about 3-5 instead of a 100 times factor. Would you agree?
Hi, I wouldn't say it is erroneous....its a default setting and designer can change it. 3-5 scale factor wouldn't also work for all cases. Kindly see below text from ADS FEM documentation: The port surface scaling options control how large the port surfaces are that the FEM simulator will generate for calibrated (TML) ports on the bounding box of the layout. You set the following options: Port LATERAL Scale: The port surface will be extended laterally at both sides Port LATERAL scale times the width of the conductor connected to the positive pin of the port in case of an infinite ground plane. In case of a finite ground plane the port surface will be extended Port LATERAL scale times the widht of the smallest bounding box containing both strips. Port VERTICAL Scale: The port surface will be extended vertically Port VERTICAL scale times the vertical distance between the positive and negative pin of the calibrated port. In case of a finite ground plane the port surface is extended both above and below the ground plane. In case of an infinite ground plane only at one side, the side where the positive pin of the port is located. By setting these options to a very large value (say 1.0e6 or larger) you can ensure that the port surface covers the entire boundary surface. This will give the most consistent value for port de-embedding and is recommended when the Reference Offset of the port is non-zero.
@@BhargavaAnurag Thank you for the reply sir. I kindly would like to ask one more question related to the matter: As the ADS documentation indicated setting the port scales very high will ensure port boundaries to cover whole simulation box surface so that, for my case, it is the most accurate. But what happens if multiple ports share the same boundary surface? Wouldn't setting the port scaling factors that high somehow distort the results as we feed the circuit from multiple/different points but practically from the same excitation area?
@@sirdancealot2645 Yes, if multiple ports share the same boundary then it won't be nice to use 100 as scale factor and you can think to set Port Lateral Scale to something like 0.1 etc depending on the design. Again designers need to set it depending on what structure they are trying to simulate... This topic is too detailed to be discussed on UA-cam, I would recommend contacting the Keysight Tech Support team when in doubt. They have detailed training available on this topic...Hope this helps....
Hello sir, I have simulated the same thing what you have done in the video but I got error like 'command is too long's I couldn't able to solve it.. Can u tell me where will be the mistake
I am not aware of the error which says "command is too long..." Sometimes you get a warning if path to the simulation directory is too long but there also it gives you option to simulate it on temp directory and once you say yes then simulation continues. If you are seeing something different then contact ADS tech support at: eesof-europe_support@keysight.com
Hello sir,How do we resolve the partitioning problem....
Hi Anil,
I am not pretty sure what is the partitioning problem you are facing but I encourage you to watch RFPro 5mins tutorial playlist on my channel. RFPro provides a more modern way of doing partitioning between layout and circuit components and not only that I can automatically created back-annotate schematic for EM-Circuit cosimulation for you: ua-cam.com/play/PL9OnCetH8TYpxUnG6_9jKZIi1PsJtXXcN.html
Hi, I am trying to simulate a PCB to check for 0Hz magnetic field generation. Does it is plausible to do it with FEM.
I already did all the preparations and i can run momentum simulations but it gets stuck when i try to do FEM simulations. In the logs status message it gets stuck in the part where it sais automatic inititial target mesh size.
At first it shown a really large number 10^6. But after seeing you video i get it to a similar number as you get.
Could you give me some insight pls.
By the way. The frequency i am interested in is 0hz to 10 hz. But i swip from 1 Ghz to 0 as if intry to use a slower sweep it gets errors.
None of the softwares with FEM solvers are good for simulation around 0 Hz.
My simulation is showing positive S21 and s11 and looks nothing like yours from previous videos. I’m not sure what went wronf
Difficult to say without looking at the data. Contact Keysight support by writing to support@keysight.com for help
Sir, I'm working on 1:4 powerdivider and combiner, results with method of moments were satisfied s11 at desired frequency was -25dB, will the practical results will be matching can I go head with fabrication.
2. Sir if i want to do FEM for same design how to add isolation resistance 100 ohm in layout. Where as on method of moments we added in schematic.
Its difficult to comment when I have not seen the layout you simulated and how well you simulated but in general with MoM you can expect measured results within +/-5% considering the tolerances and FEM should be within +/- 1 or 2%.
Process to add 100Ohm resistor is same between Momentum and FEM, jus the emmodel will change behind the scene. You could either use emcosimulation method like what you seem to be using now or directly add 100Ohm in layout as I have shown in the Power Divider design video.
Best of luck...👍
Helle thank you for the video it"'s verry useful. I have 2 questions which is the difference between HFSS and FEM in ADS.
and the second question is which is the difference betwenn momentum RF et momentum microwave and FEM
Hi Farah,
Some of the points are provided here for your help. For more info you can refer to ADS documentation or contact tech support team:
1. Its difficult to answer HFSS vs ADS FEM in detail but just be aware that HFSS is also based on FEM for frequency domain analysis so there is lot of similarity how Mesh is created, refined and problem is solved.
2. Momentum is a Method of Moments based planar EM solver whereas FEM is a full 3D solver that can handle any arbitrary 3D geometry for simulation
3. Momentum RF is based on quasi-static technique of EM simulation whereas Momentum-Microwave is a full wave solver.
ADS documentation provides detailed description of Momentum and FEM, it will be worthwhile if you go through the documentation.
Hi Anurag, thank you a lot for your great tutorials, can you please help me, I want to simulate passif structures (Coils, MIM capacitors and transformers) on Silicon substrate. I don't Know what is the best simulator to use it: Momentum RF, momentum Microwave or FEM ?
I want to study the variation of Q factor inductance, capacitance, mutuel inductance etc as a function of frequency.
Really your channel is SuperGreat Thank you.
For the structure you mentioned Momentum would be sufficient as you really don't need a full 3D simulator like FEM. Whether to use Mom RF or MW, is your choice but mostly MoM-RF is sufficient as those structure are electrically small component and RF mode works very well till half-wavelength. Kindly refer to ADS documentation for more details.
your videos were a great help.
I followed all the steps in the fem simulation;
I get a "Simulation Error" because the command line is too long
I need some help
Not sure why that could come, would recommend you to contact tech support team by writing to "support@keysight.com" for help.
Hello Mr. Bhargava. Thank you for great tutorials. I have a question regarding
the port scaling factors under the options tab in FEM EMSetup. In ADS2016 by default the factors are 100. Isn't this erroneous? This makes mostly the whole surface area of the simulation box as port input/output area. I think those factors also should be scaled about 3-5 instead of a 100 times factor. Would you agree?
Hi, I wouldn't say it is erroneous....its a default setting and designer can change it. 3-5 scale factor wouldn't also work for all cases. Kindly see below text from ADS FEM documentation:
The port surface scaling options control how large the port surfaces are that the FEM simulator will generate for calibrated (TML) ports on the bounding box of the layout. You set the following options:
Port LATERAL Scale: The port surface will be extended laterally at both sides Port LATERAL scale times the width of the conductor connected to the positive pin of the port in case of an infinite ground plane. In case of a finite ground plane the port surface will be extended Port LATERAL scale times the widht of the smallest bounding box containing both strips.
Port VERTICAL Scale: The port surface will be extended vertically Port VERTICAL scale times the vertical distance between the positive and negative pin of the calibrated port. In case of a finite ground plane the port surface is extended both above and below the ground plane. In case of an infinite ground plane only at one side, the side where the positive pin of the port is located.
By setting these options to a very large value (say 1.0e6 or larger) you can ensure that the port surface covers the entire boundary surface. This will give the most consistent value for port de-embedding and is recommended when the Reference Offset of the port is non-zero.
@@BhargavaAnurag Thank you for the reply sir. I kindly would like to ask one more question related to the matter: As the ADS documentation indicated setting the port scales very high will ensure port boundaries to cover whole simulation box surface so that, for my case, it is the most accurate. But what happens if multiple ports share the same boundary surface? Wouldn't setting the port scaling factors that high somehow distort the results as we feed the circuit from multiple/different points but practically from the same excitation area?
@@sirdancealot2645 Yes, if multiple ports share the same boundary then it won't be nice to use 100 as scale factor and you can think to set Port Lateral Scale to something like 0.1 etc depending on the design. Again designers need to set it depending on what structure they are trying to simulate...
This topic is too detailed to be discussed on UA-cam, I would recommend contacting the Keysight Tech Support team when in doubt. They have detailed training available on this topic...Hope this helps....
@@BhargavaAnurag Thank you again. Also keep up the grest videos, I really find them useful as a newly grad engineer!
Hello sir, I have simulated the same thing what you have done in the video but I got error like 'command is too long's I couldn't able to solve it.. Can u tell me where will be the mistake
I am not aware of the error which says "command is too long..." Sometimes you get a warning if path to the simulation directory is too long but there also it gives you option to simulate it on temp directory and once you say yes then simulation continues.
If you are seeing something different then contact ADS tech support at: eesof-europe_support@keysight.com
awesome thanks