No transient model for microstrip lines #681
-
|
I tried to simulate following design and do not get any output. The pulse generated is not visible on the out line and as well the cpld line is quiet... this does not look right. |
Beta Was this translation helpful? Give feedback.
Replies: 17 comments 2 replies
-
|
It is not a bug. Qucsator never had transient model for MS lines. The MS lines are considered as zero resistnace for all analysis except SP. |
Beta Was this translation helpful? Give feedback.
-
|
The Qucsator from Qucs-0.0.20 gives the same result. |
Beta Was this translation helpful? Give feedback.
-
|
Sorry I mixed QUCS-S and QucsStudio... there is see: |
Beta Was this translation helpful? Give feedback.
-
|
QucsStudio is closed source project not related to Qucs-S and may have incompatible extensions. |
Beta Was this translation helpful? Give feedback.
-
|
It looks like a missing feature... since if it is possible in QucsStudio it should be available as well for Qucs-S. |
Beta Was this translation helpful? Give feedback.
-
|
Adding transient models for MS lines is technically possible if someone will provide equations. |
Beta Was this translation helpful? Give feedback.
-
|
... here we need an RF expert. Anyone in the community able and willing to share? |
Beta Was this translation helpful? Give feedback.
-
|
I have converted this issue to discussion. Probably someone could look at this and point a solution. |
Beta Was this translation helpful? Give feedback.
-
|
I asked the RF expert in my company and he told me to point you to the Qucs-Studio documentation chapter 6.8 for the algorithm. |
Beta Was this translation helpful? Give feedback.
-
|
You can use the ngspice TLine or Spice model of the needed device to make most transmission line elements like a coupler. See attached examples. Do you have the BFP193W amplifier schematic used in this example?
|
Beta Was this translation helpful? Give feedback.
-
|
Hello Tom, Yes, I already used the TLINE for simulating microstrips in combination with the IBIS models using ngspice... this is now possible and would be a good extension for Qucs-S as well... since KiCAD does the translation of the IBIS model to spice models, such an extension looks feasible and would be useful for CPU/FPGAs, memory and MGT designs running at several GHz. Question to you: Would the proposed implementation of MS in transition simulation be something you could use as well? I'm wondering, if I'm the only one preferring such a solution where I'm able to add the PCB dimensions and simulate the effects of design compromises. Cheers |
Beta Was this translation helpful? Give feedback.
-
|
I looked at that KiCAD example when I was troubleshooting issues with using TLine in Qucs-S. Unfortunately there seems to be few IBIS models available and the analysis only accounted for input/output "parameters" of digital gates. As far as PCB design, trying to accurately replicate PCB traces and vias on a multi-layer PCB in "Qucs" doesn't seem feasible. You would be hard pressed to make an accurate schematic or find IBIS models for the desired devices. Don't you really need a 3D EM solver that would take in the actual PCB layout? |
Beta Was this translation helpful? Give feedback.
-
|
Hello Tom, The IBIS models are in deed for digital electronics and MGTs as used for example for FPGAs (https://www.xilinx.com/support/download/index.html/content/xilinx/en/downloadNav/device-models/ibis-models/virtex-series-fpgas.html) or DDR SDRAMs (https://www.micron.com/search-results?searchRequest=IBIS) or many other digital components used up to few GHz. With many companies delivering exactly this IBIS model for the signal integrity checks... not more and not less. The main purpose if the IBIS models is stated as follows: Hence, the industry does not need to publish the functionality of a component like in SPICE models but solely the IO buffer structure. This IBIS model combined with a model of the traces are used to simulate PCB designs and should potentially allow us to develop stable working PCBs. An EM solver like ADS or openEMS are as well possible for making "precise" models of the traces, but the simulation effort is very high and computing power as well. Here the more simple ngspice and other simulators come into play. Cheers |
Beta Was this translation helpful? Give feedback.
-
|
Seems like a massive amount of work for a feature not many would use. What
actually is the goal?
Having extensively used QucsStudio, its transient analysis capability like
Qucs is abysmal. Also the inability to use most device Spice models limits
what devices are used. What circuits/devices do you want to analyze?
…On Mon, Apr 29, 2024 at 11:07 AM goranmarinkovic ***@***.***> wrote:
I asked the RF expert in my company and he told me to point you to the
Qucs-Studio documentation chapter 6.8 for the algorithm.
—
Reply to this email directly, view it on GitHub
<#681 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AYQTCTHMN2ZBIJ3MWRH2OZLY7ZO4LAVCNFSM6AAAAABGA5OWZKVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TENRUGE2DK>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
|
The goal is still to be able to simulate the coupling between adjacent tracks on a PCB. A regular PCB designer / electronics engineer can not use ADS or openEMS but he/she is able to see the effects of the coupling and do counter measures like separation, guard wires, serial termination at the source or the proper termination at the drain of the signal. QucsS could be the tool bridging the functional simulation between SPICE and EM solvers. Cheers |
Beta Was this translation helpful? Give feedback.
-
|
You're assuming the models used in Qucs are accurate AND more importantly
you could make a schematic that accurately models PCB traces/coupled lines
and multiple types of vias meandering on a multi-layer PCB.
…On Tue, Apr 30, 2024 at 11:22 AM goranmarinkovic ***@***.***> wrote:
@tomhajjar <https://github.com/tomhajjar>
The goal is still to be able to simulate the coupling between adjacent
tracks on a PCB. A regular PCB designer / electronics engineer can not use
ADS or openEMS but he/she is able to see the effects of the coupling and do
counter measures like separation, guard wires, serial termination at the
source or the proper termination at the drain of the signal. QucsS could be
the tool bridging the functional simulation between SPICE and EM solvers.
I guess there are more PCB designer / electronics engineers world wide in
need for such an analysis than RF engineers doing the precise design of a
"simple" digital PCB. However, the current approach to build a PCB and then
measure is not up to the art of engineering.
Cheers
Goran
—
Reply to this email directly, view it on GitHub
<#681 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AYQTCTAFVFXUN26WC453NJDY76ZMLAVCNFSM6AAAAABGA5OWZKVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TENZWHA2TO>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
|
@tomhajjar |
Beta Was this translation helpful? Give feedback.
-
|
I have created ra3xdh/qucsator_rf#7 The adding MS lines transient models for QucsatorRF is planned for the near future. Help is highly appreciated. I can provide a necessary guidance. |
Beta Was this translation helpful? Give feedback.
-
|
I have started with implementation of the TRAN model for microstrip lines etc. The help from RF experts is appreciated. See ra3xdh/qucsator_rf#16 |
Beta Was this translation helpful? Give feedback.
It is not a bug. Qucsator never had transient model for MS lines. The MS lines are considered as zero resistnace for all analysis except SP.