ABOUT SPICE2: HOW SHOULD THE SUBCKT CIRCUITS BE CONNECTED? And what are the shorthand for RE and CO in the circuit?

[wan1235/*-1-]

* RLC admittance subckt

.SUBCKT bbs_RL_main 1 2 3 

X1 1 3  bbs_RL_id_1_1
X2 2 3  bbs_RL_id_2_2
X3 2 1  bbs_RL_id_2_1

.ENDS bbs_RL_main
 **************************************************


.SUBCKT bbs_RL_id_1_1 1 2

Rgs     1     2      5.00000000000000e+01

Rre1     1     10001      1.04846705084131e-01
Lre1     10001     2      4.53866569122965e-09

Rco2      1     12      -1.22951940170065e+02
Lco2      12     22      8.56793345142387e-09
Cco2      22     2      6.38685779263559e-14
RGco2     22     2      1.07754867125530e+03

Rre3     1     10003      -2.72922003692118e+20
Lre3     10003     2      -1.18537624402029e+13

Rco4      1     14      7.51523839890950e+23
Lco4      14     24      5.53833940683497e+12
Cco4      24     2      9.03847957116950e-36
RGco4     24     2      -8.17225035349824e+23

.ENDS bbs_RL_id_1_1
**************************************************


.SUBCKT bbs_RL_id_2_2 1 2

Rgs     1     2      5.00000000000000e+01

Rre3     1     10003      1.03666977501187e-01
Lre3     10003     2      4.50254544363957e-09

Rco4      1     14      -1.78464545052696e+02
Lco4      14     24      9.11849751545625e-09
Cco4      24     2      5.49654477833988e-14
RGco4     24     2      9.14975239401263e+02

.ENDS bbs_RL_id_2_2
**************************************************


.SUBCKT bbs_RL_id_2_1 1 2

Rre3     1     10003      2.72922003692118e+20
Lre3     10003     2      1.18537624402029e+13

Rco4      1     14      -7.51523839890950e+23
Lco4      14     24      -5.53833940683497e+12
Cco4      24     2      -9.03847957116950e-36
RGco4     24     2      8.17225035349824e+23

.ENDS bbs_RL_id_2_1
**************************************************

 

[FvM]

Your owing us some basic information, what's the purpose of the circuit, what does it model?

Regarding your question, the subcircuits are already connected in bbs_rl_main which exposes three nodes. If you ask about their intended connection, depends on the purpose, see above.

 

[wan1235/*-1-]

Your owing us some basic information, what's the purpose of the circuit, what does it model?

Regarding your question, the subcircuits are already connected in bbs_rl_main which exposes three nodes. If you ask about their intended connection, depends on the purpose, see above.

There is no actual model of this circuit, this is the SP2 file that I simulated with ADS with known S-parameters, in fact, I am trying to build an actual circuit model from this file.
What I want to know is whether the number of nodes in the subcircuit is too large will have an effect, and whether the alphanumeric numbers after R, L, and C in the subcircuit have special meanings, such as whether Rre refers to the resistor itself or the actual part of the resistor, etc.
I haven't seen anything like Rco or RGco before, so I'm not sure what that means.

 

[FvM]

O.k., it's apparently an extracted circuit generated by circuit extraction tool. Nodes 1, 2 and 3 are given by s2p definition. But what is the purpose of the original system? Where do the S-Parameters come from? Which ADS model generator has been used?

The extracted SPICE circuit has physically implausible negative resistor values. Is an active circuit, e.g. an amplifier?

For SPICE, component identifiers are arbitrary alphanumeric strings. They have no specific meaning.

 

[volker@muehlhaus]

and whether the alphanumeric numbers after R, L, and C in the subcircuit have special meanings, such as whether Rre refers to the resistor itself or the actual part of the resistor, etc.

Yes, the first character tells you if this is R,L,C

As I already described in your previous thread on the same topic, the equivalent circuit should be built on physical understanding of the device, i.e. its series and shunt pacth characteristics. The broadband SPICE model is a mathematical fit and does not provide an intuitive topology.

The question is why you want/need an intuitive physical model at all? The circuit from bbextract will work in SPICE type simulators, if you cannot use S-parameters directly.

 

[FvM]

There's a reason why forum rules strictly require to post similar and particularly follow-up questions in the same thread. Please respect it in future.

Thanks Volker for hinting to https://www.edaboard.com/threads/ho...rom-s-parameter-touchstone-file-in-ads.409245
I realize that bbs stands for broadband spice model generator. I believe everything has been already said.

 

[wan1235/*-1-]

Yes, the first character tells you if this is R,L,C

As I already described in your previous thread on the same topic, the equivalent circuit should be built on physical understanding of the device, i.e. its series and shunt pacth characteristics. The broadband SPICE model is a mathematical fit and does not provide an intuitive topology.

The question is why you want/need an intuitive physical model at all? The circuit from bbextract will work in SPICE type simulators, if you cannot use S-parameters directly.

I'm sorry I've been asking similar questions lately, but my goal is to make a physically implementable model based on the known S-parameters and the initial SPICE model generated with ADS, and make it in reality, which was given to me by my teacher.
I also hope to stop at the SPICE model to meet the requirements. The last post made me confirm that this SPICE model can't directly generate the corresponding theoretical circuit model, so I ended up trying to create a physical model directly from this SP2 file.
I need to generate an exact achievable .cir file to be able to continue researching, and I'm sorry if it bothered

--- Updated Dec 14, 2023 ---

Yes, the first character tells you if this is R,L,C

As I already described in your previous thread on the same topic, the equivalent circuit should be built on physical understanding of the device, i.e. its series and shunt pacth characteristics. The broadband SPICE model is a mathematical fit and does not provide an intuitive topology.

The question is why you want/need an intuitive physical model at all? The circuit from bbextract will work in SPICE type simulators, if you cannot use S-parameters directly.

Or if SPICE models can't be used, is there another way to create a physically implementable model with known S-parameters?

 

[BigBoss]

Copy and paste this netlist in a text file then import it in ADS.

 

[volker@muehlhaus]

Or if SPICE models can't be used, is there another way to create a physically implementable model with known S-parameters?

I still don't understand what you mean by "physically implementable model"? From what I saw in your previous thread, you created the S-parameters from Momentum, so that layout is your physical implementation. The corresponding equivalent circuit of an RF layout can be very very very complex, if you want to implement it as RLC.

The method shown by BigBoss makes no sense to me, because you seem to start in ADS from layout, use Momentum to create S-Parameters and these data blocks can be cascaded in ADS directly, with no need for equivalent circuit models.


So I still don't understand what your teacher expects you to do. Can you show the simulated layout here?

 

[BigBoss]

I think he's looking for a equivalent RLC circuit for a Coil, right ??
Like that in Sonnet.

 

[wan1235/*-1-]

I still don't understand what you mean by "physically implementable model"? From what I saw in your previous thread, you created the S-parameters from Momentum, so that layout is your physical implementation. The corresponding equivalent circuit of an RF layout can be very very very complex, if you want to implement it as RLC.

The method shown by BigBoss makes no sense to me, because you seem to start in ADS from layout, use Momentum to create S-Parameters and these data blocks can be cascaded in ADS directly, with no need for equivalent circuit models.


So I still don't understand what your teacher expects you to do. Can you show the simulated layout here?

Yes! MYes! My initial S-parameter is the best result calculated by matlab, and there is no actual circuit, I hope to get the corresponding RLC circuit from this S-parameter! y initial S-parameter is the best result calculated by matlab, and there is no actual circuit, I hope to get the corresponding RLC circuit from this S-parameter! Just like bigboss shows, but I don't quite understand how this RLC diagram is obtained?Just like bigboss shows, but I don't quite understand how this RLC graph came to be.

--- Updated Dec 14, 2023 ---

I think he's looking for a equivalent RLC circuit for a Coil, right ??
Like that in Sonnet.
View attachment 186912

I used to import lib files but only got a black box like a Symbol

 

[volker@muehlhaus]

My initial S-parameter is the best result calculated by matlab

But what is the device/circuit then? What did you optimize in Matlab?

 

[BigBoss]

I can import this netlist very easily. ADS will create a block for each subcircuit ( here there are 3 subcircuits ) then combine all into one as shown blow.
bbs_rl_main

 

[wan1235/*-1-]

But what is the device/circuit then? What did you optimize in Matlab?

What I did was to cancel the stealth of the antenna, and the goal was to extract the best S-parameter curve for the impedance at the end of the antenna, which is my simulation process in matlab.

--- Updated Dec 14, 2023 ---

I can import this netlist very easily. ADS will create a block for each subcircuit ( here there are 3 subcircuits ) then combine all into one as shown blow.
bbs_rl_main
View attachment 186916
View attachment 186917

Yes, the process is very easy to implement, so how do you get the RLC circuit you mentioned above

 

[volker@muehlhaus]

I can import this netlist very easily. ADS will create a block for each subcircuit ( here there are 3 subcircuits ) then combine all into one as shown blow.

Sure, but does this help him? ADS can use the S-parameters directly (without the broadband fit), and importing the SPICE model into ADS doesn't give him the desired "meaningful" physical RLC model.

impedance at the end of the antenna

Ok, we are getting closer. Not sure why your supervisor wants you to extract an RLC model for antenna input impedance, nobody seems to use (or need) that. What might be useful is complex impedance vs. frequency.

That said, you can look at the model above and remove the "extreme" components that do almostnothing, like very small series impedance or very lange shunt impedance. Examples are paths with Rco4=7E23 Ohm and Rre3 with -2E20 Ohm.

Also, the model will be less complex if you narrow down the frequency range.

 

[wan1235/*-1-]

Sure, but does this help him? ADS can use the S-parameters directly (without the broadband fit), and importing the SPICE model into ADS doesn't give him the desired "meaningful" physical RLC model.


Ok, we are getting closer. Not sure why your supervisor wants you to extract an RLC model for antenna input impedance, nobody seems to use (or need) that. What might be useful is complex impedance vs. frequency.

That said, you can look at the model above and remove the "extreme" components that do almostnothing, like very small series impedance or very lange shunt impedance. Examples are paths with Rco4=7E23 Ohm and Rre3 with -2E20 Ohm.

Also, the model will be less complex if you narrow down the frequency range.

Thanks again, I'll overlay these three subcircuits in node order to see his S-parameters. At the same time, thank you for your simplified circuit method. After that, I'll upload the graph when the simulation results are out.

 

[wan1235/*-1-]

Sure, but does this help him? ADS can use the S-parameters directly (without the broadband fit), and importing the SPICE model into ADS doesn't give him the desired "meaningful" physical RLC model.


Ok, we are getting closer. Not sure why your supervisor wants you to extract an RLC model for antenna input impedance, nobody seems to use (or need) that. What might be useful is complex impedance vs. frequency.

That said, you can look at the model above and remove the "extreme" components that do almostnothing, like very small series impedance or very lange shunt impedance. Examples are paths with Rco4=7E23 Ohm and Rre3 with -2E20 Ohm.

Also, the model will be less complex if you narrow down the frequency range.

I popped up the latest results, the red ones are the simulation results of the RLCG circuit, the blue ones are the SPICE model simulation results, and the purple ones are the best simulation results of the theoretical model.

I'd also like to ask how to export this RLCG schematic as a .cir file in ADS？