How to design active circuit in HFss

[LTCC]

Hi,all
I am doing simulation of LNA using HFSS based on LTCC substrate.

The LNA is a MMIC,and input and output ports are matched, so s11 s22 are below -10db. I already have s parameters of LNA,and I want to use it in LTCC substrate.

Does anyone has experience to design this kind of simulation ? What should I do next? I need some advice.

 

[Azulykit]

Ansoft Designer is appropriate for active device analysis. Passive elements of the design can be simulated with HFSS and linked to the Designer project. I do not see how you could design a LNA with HFSS alone.

The advice is consider Ansoft Designer rather than HFSS.

 

[LTCC]

Thank you for Azulykit.

Not just using HFSS.

First step ,I did circuit simulation in ADS.I made some matches in the input and output ports of LNA .

Second,I did full wave simulation in hfss. I simulated the pcb of the entire circuit except the LNA,the model has 4 lumped port , 2 for input and output ,other 2 for interface of LNA.

Third,export the s parameters into ADS.
When the simulation in HFSS is completed, export s parameters into ads
then cosimulate the s para of LNA and the s para obtained by hfss.

Last ,I compare the results in first and third step

My english is not good ,hope you understand.


Sherry

 

[maxwellian]

If you are using ADS, then why don't you use Momentum for your LTCC EM simulation? I would think it would be more efficient, and it would probably have less error in the ports where you want to connect your LNA.

--Max

 

[fishyxl]

HFSS is for passive components

 

[LTCC]

If you are using ADS, then why don't you use Momentum for your LTCC EM simulation? I would think it would be more efficient, and it would probably have less error in the ports where you want to connect your LNA.

--Max

Yes, ads is more efficient.But there are something that ads can'ts handle.
Something like cavities and conductor recessed especially in the surface layer.

I'm still not sure the way I have simulated is correct or not. I don't know how
the LTCC manufacturers or companies simulate their active circuits.

 

[maxwellian]

fishyxl: Yes, and Momentum is also for passive circuitry only. That is why LTCC is simulating the passive parts in the EM simulator, and then importing the EM-based model into ADS to cascade that with the model for the LNA (which is non-linear and active).

LTCC: Thanks for the explanation. Yes, I agree that full 3D is better if you have a recessed cavity. If you're just running lines up to the edge of the cavity and then wirebonding over to the LNA, then you can still probably get by with a planar EM model for your PCB structures, unless the cavity has some serious interactions with nearby PCB transmission lines. It probably doesn't affect the capacitance of the open transmission line to which you might be wirebonding.

Momentum (and other planar EM simulators) can handle metal thickness, and conductor recess into your dielectric layer, so you get correct spacings between broadside transmission lines or capacitor plates.

If you are using lumped internal ports in HFSS for your internal LNA connections, then I would be concerned about the discontiniuties that HFSS adds to the model resulting from these lumped ports; they aren't very good to use at higher frequencies. I don't know about your frequency range, though, so maybe you are still in the range where the error is small enough for your purposes. I saw that Sonnet has created a new type of internal port that is completely de-embedded with other closely spaced internal ports; this might be useful to you as well. I guess they can handle surface mount devices really well, which are used a lot on LTCC and PCB.

Good luck,

--Max

 

[Azulykit]

Hi Sherry,

After reading your more detailed post it looks like you are on the right track.

Regards,

Azulykit

 

[LTCC]

Thanks Max and Azulykit
I have sloved my problem.

Regard,

Sherry

 

[Mohd Najim]

Thank you for Azulykit.

Not just using HFSS.

First step ,I did circuit simulation in ADS.I made some matches in the input and output ports of LNA .

Second,I did full wave simulation in hfss. I simulated the pcb of the entire circuit except the LNA,the model has 4 lumped port , 2 for input and output ,other 2 for interface of LNA.

Third,export the s parameters into ADS.
When the simulation in HFSS is completed, export s parameters into ads
then cosimulate the s para of LNA and the s para obtained by hfss.

Last ,I compare the results in first and third step

My english is not good ,hope you understand.


Sherry

Ansoft Designer is appropriate for active device analysis. Passive elements of the design can be simulated with HFSS and linked to the Designer project. I do not see how you could design a LNA with HFSS alone.

The advice is consider Ansoft Designer rather than HFSS.

Hi Dear, I am Najim, doing PhD from IIT Roorkee, India.
I am new to this field and I want to design Active FSS on HFSS.
can you plz tell me how can I make a active fss on HFSS?
or how to insert a PIN diode in the fss?

Thx

- - - Updated - - -

Hi Dear, I am Najim, doing PhD from IIT Roorkee, India.
I am new to this field and I want to design Active FSS on HFSS.
can you plz tell me how can I make a active fss on HFSS?
or how to insert a PIN diode in the fss?

 

[Azulykit]

I note that you preface your question by saying that you are interested in using HFSS. Usually a debate about which software is "best" results. I am mot looking to go there.

Ansys is able to simulate passive and active components by using a combination of HFSS and Designer. Generally Designer is appropriate for the active side and HFSS is better with passives. Simulating diodes in a FSS sounds like it will tax both with the interaction. I would pose the question to the technical support folks HFSS/Designer?Ansoft/Ansys to see what they would recommend.

Good Luck.