Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics 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.

Help for Microstrip patch antenna design in Multi-layers

Status
Not open for further replies.

sehun1119

Junior Member level 1
Joined
Sep 30, 2004
Messages
17
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
191
Hi, my name is kook and I am new to this forum and this antenna area.

I am to design a microstrip patch antenna in multi-layer stack-up.

Here is how it looks like. I have 6 metals (M1 to M6) and 5 dielective substrates.
Each metal has a thickness of 0.7 mils, and its width and lengh are not determined yet.

M1

er = 3.38, loss tan = 0.0035, thickness=12mils

M2

er = 2.95, loss tan = 0.002, thickness=1mil

M3

er = 3.54, loss tan = 0.0035, thickness=4mils

M4

er = 2.95, loss tan = 0.002, thickness = 1mil

M5

er = 3.38, loss tan = 0.0035, thickness=12mils

M6

To the best of my knowledge, the microstrip patch antenna has a sandwitch shape of 1 dielectric in the middle of between 2 conductors, but I have got multi-layers. So, I would consider M6 as a GND plane, M1 as a patch and the others as one dielectric substrate. To start designing the dimension of the patch, I would have to know the permittivity and the thickness of the substrate. Well, the thickness of that dielectric would be calculated by summing all the metals and substrates between M1 and M6, but I am stuck with finding permittivity.

One thing I am thinking to solve it is to put all the layers with the specs in HFSS, simulate it and find the resonant frequency. Then,

(wave length) / 2 = length of the patch (or signal line)
(wave length) = c' / f = (c / sqrt(effective permittivity)) / (f)

From those two equations and HFSS simulation, I may obtain the effective permittivity. Is this the right method ?

Also, I have a little problem with HFSS. As I said above, I would try the simulation in HFSS. The operating frequency is 2.5 G Hz so that wave length in the air came out to be 0.12 meters ( wave length = c / f). So, I started it with the dimension of (0.06 meter by 0.06 meter - half wave length) for all the metals and substrates with its thickness. Therefore, the ground plane (M6) and the patch (M1) have the same dimension actually. I wonder how to determine the dimension of ground plane and the other metals except the patch. (I think the patch of M1 is good with 0.06 meters by 0.06 meters, right? ) After that, I wonder what type of port (lumped or wave? ) should I use??

Sorry for too many questions. I do not have really any idea how to start it from.

How do you guys design the microstrip patch antenna in multi-layers ?

Any help or advice would be very appreciated.

Thank you all!!!
 

Hi Sehun119 -- A volume mesher like HFSS is good for this problem if it is important to include the effect of finite substrate size. The difficulty using a volume mesher for this problem is what we call the "big-small problem". You have a 1 mil (I assume your mil = 0.001 inch) thick dielectric in several places and conductor dimensions of 6 cm. The volume mesher mesh size must be small enough to accurately model the 1 mil substrate. This very fine mesh must now also mesh the 6 cm patches. Just a bit difficult.

For this reason, you probably should switch to a planar surface meshing tool. Planar tools do not care how thick or thin the substrate is. Thus, the mesh size is set only by your circuit geometry. A much coarser mesh can be used.

I work for Sonnet, and our full version can handle the problem, but Sonnet is a shielded analysis. Antennas are done by moving the shielding sidewall far away and removing the top cover. If you go this path, be sure to read the chapter in our manual on antenna analysis. Sonnet also has an antenna pattern plotting option. This problem is not appropriate for our free SonnetLite tool (too many layers).

There are several unshielded planar tools that do not need special consideration for radiation that Sonnet does. However, while I have seen a lot of benchmarks from these tools for microwave circuits, I have seen very few benchmarks on antennas, so I can not comment on their relative accuracy for antenna analysis (of course, the vendors all describe their own tools as accurate, but I find such statements uninteresting). I assume that they would all work well enough for most needs. I usually recommend Agilent Momentum, but that is mainly because I can discuss both its advantages and disadvantages fairly freely.

Bottom line: Unless you need to include substrate edge effects (or have other full3-D aspects to your geometry), you can try different geometries very quickly (i.e., a few minutes for each trial geometry) with most any planar EM tool...then you can tell us what you find out.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top