Need help my design before fabricate

horn antenna broadband in hfss

Hai everybody.. thanks for ur time to read my post... =D

I already design my folded patch antenna using hfss v10. My design antenna dimension is 15mm X 15mm for Ultra Wide Band (UWB) application. My bandwidth is 81.8%. My problem is i need someone who can check my design whether it ready to fabricate or need some change since we have to consider other antenna minimum characteristic before we fabricate it... I hope u can help me to solve this problem.. I will send my HFSS file to u if u want to help me....
thankkss....

mr !mad

Hi Mr.Mad,
I can have a look at your design, send me a pm with the file so I can check it. Looking forward for receiving it.

Regards,

I.

Thank Ivan... I already send pm to u...

Hi Mr!mad. I believe your design is very good. The y-polarization gain is of about 3dB. Not so high, but it depends on the application's requirements. Anyway, one can think of using it into an array for getting higher directivities. Anyway, this is a very intertesting structure, in particular for the impressively large bandwidth. Just a question. You used vacuum as dielectric for coax feed. Are u sure? Did you check in your connector datasheet to see which material is used? You are supposing that the dielectric into the coax is lossless and with a unit relative permittivity... I don't think so.
Anyway, apart from this, I think you may try to improve meshing to see what changes refining the solution accuracy.
Let me know how it goes.

Ivan

Thanks Ivan!

I already take note your comments.. thanks very very much!
Can explain me details about the cross polarizatiann and co polarization for both E-plane and H-plane and how to get that graph using HFSS??
If I don't change anything for this design, are this design will be accepted to fabricate and can we call this as an "antenna"? =D

mr !mad

The co-polarization is the component having the same E-field direction as the current direction. In general (apart from particular applications) it is the desired far field component. The cross-pol is the component of the electric field vector normal to the desired polarization component, which is usually undesired cause it's responsible a reduction of desired component power. Copolarized E-component lies on the E-plane, while the cross-E lies on H-plane.
There are lots of possibilities to plot polarizations, I mean, many representation. U can plot x,y and z component of E, directivity, gain, etc., or theta and phi component as well. Choose the correct value for φ (0 or 90deg) and plot the components. You will understand which is the correct one by looking at current on your antenna.
And yes, you can call it 'antenna'. It is a antenna working off-resonance, but always an antenna. This off-resonance operation gives the broadband behaviour (have a look at the input impedance): the real part of Zin has an approximately flat trend over the bandwidth and the imaginary part is close to zero.

Before manufacturing, you'd better simulate with an improved mesh refinement. I think you have to see the exact characteristic of the coaxial connector (the material) and the influences on the mismatch between coax feed and antenna. I.E., if the coax is filled with teflon (epsR=2.1) the reflection coefficient will sligthtly change since the reference impedance (port impedance) will change.
Anyway, S11 shouldn't change dramatically in your case.

Bye

I.

Ivan...

Thanks for the very detail explanation... Its very useful.. I'm using SMA connector as coaxial connector for this design.. Can you tell me which part I need to change the value, isn't the permitivity and dielectric loss tangent only?? and what is the value for SMA connector since I can't find datasheet for that... I'm also need you help to explain me about the graph for polarization (how to know the better antenna from that graph) and also the radiation pattern for gain total.. Here I attach all my result graph in picture. thanks a lot Ivan!!! really appreciate it...

By knowing how to improve my design from polarization and radiation pattern, it will give me correct guideline to have better antenna. Before this, my design only focus on Return Loss graph and ignore all other graph. thankss again for willing help me..

mr !mad

Mr mad,
I apologize for the late reply, but I was pretty busy this weekend, anyway I can answer your last questions now. As for the coax, find any 50ohm sma conn. datasheet and get info about pin radius and dielectruc radius. Teflon is commonly used as dielectric. Whole parameters should give you a 50ohm port impedance.
About gain, plot tot gain in both phi=0 and phi=90° plane. In your case you will see a maximum in the normal direction in just one of them. Then, that is your gain(I guess you want your rad pattern to have a maximum along the normal, so called broadside direction). Then, find out which component gives a high contribution to the gain on that plane to understand polarization. Plot Ex and Ey for that value of phi. If one component has a max value for theta=0 significantly higher than the other (in dB), you can guess your antenna is linearly polarized along that direction. If both the components give a significant contribution in amplitude, that is the case of a circular or elliptical polarization (look at the relative phase to understand which one). It's not difficult. I think that's all. You have enough data to analyze results. Anyway, try a simulation with an improved mesh refinement, in addition to the modifications to the coax to a better validation.
Bye.

I.

Ivan..

Sorry for the very late post... I extremely busy...
Can you look at my result that I attach before...
For graph GainL3X, Isn't unacceptable since the phi=0deg is above the 0...
Because someone say that the graph is wrong because gainL3X must always below the 0..
Please give me the comment.. thanks a lot

!mad

If u want gai3lx to be less than 0 on phi=0, I guess, ur antenna must be y-polarized and be directive for phi=0. If it's true, the behavior could be due to a slight cross-pol component, but I seem from the graps ,gain3lx is >0 in the region of beam ouside the maximum of the desired polarization . Not so bad, so...

Bye

Ivan,

Ivan_ said:

Plot Ex and Ey for that value of phi

Click to expand...

how to plot Ex and Ey????

Ivan_ said:

It is a antenna working off-resonance, but always an antenna. This off-resonance operation gives the broadband behaviour (have a look at the input impedance): the real part of Zin has an approximately flat trend over the bandwidth and the imaginary part is close to zero.

Click to expand...

I already check the graph and i understand it... izzit ok to have off-resonance antenna for UWB application???

Ivan_ said:

try a simulation with an improved mesh refinement

Click to expand...

Nothing change... here are what i did...
1. i select at all main part(copper)
2. HFSS - Mesh operation - assign - on selection - length based
3. put value for maximum length to 5mm

can you explain me how to do correct mesh operation... how can this operation improve my design...

thankss...
mr !mad

how to plot Ex and Ey????

Well, just go to results and plot rEx and rEy on the desired plane.

I already check the graph and i understand it... izzit ok to have off-resonance antenna for UWB application???

Really, u are working off the main resonance, the one that gives the highest radiation resistance, so the highest efficiency. Nevertheless, u are always working at a resonance which is usually called second resonance: that is another frequency at which the capacitive and inductive parts of reactive energy equal, so that it behaves as resonator. Just the real part, which is the radiation resistance is lower with respect to the main resonance. This is typical for having a broadband behavior, since a flat trend of Zinput is required. To see the central frequency of ur bandwidth is a resonance too, plot s11 on the Smith Chart: u will see two circles: the midpoint of each corresponds to the resonances. In the case at hand u should see a couple of circles.

About mesh, probably ur simulation already reached converegence, so a mesh improvement shows no evident refiniments in results. That's not bad. Anyway, another way of improving mesh, instead of setting it manually, is to have more adaptive passes (more than the 6 of default), have a lower limit of converegence and make percent refinement per pass higher than the 30% of dafault.

Bye

I.

When I change coax from vacuum to other material which are the bulk conductivity are not equal to zero... this msg appear..

[error] Port refinement, process abc3d : There are no boundaries to define "P1" correctly. This may be due to incorrect material definitions for those objects that make up the port surface. Please verify the material assignments and properties. (1:38 AM Mar 03, 2008)
[error] Simulation completed with execution error on server: Local Machine. (1:38 AM Mar 03, 2008)

can you explain why??

here are the description about the connector that i will be used...
* JACK, SMA PANEL
* Connector type:SMA Panel socket
* Impedance:50R
* Material, contact:Brass or Beryllium Copper
* Plating, contact:Gold
* Termination method:Solder spill
* Mounting typeanel
* Approval Bodies:BS9210 N0006, MIL-C-39012
* Approval category:BS9210, MIL-C-39012
* Material:Stainless steel
* Plating:Gold
* Temp, op. max:155°C
* Temp, op. min:-55°C
* Voltage rating, AC:500V
* Voltage, working:500V
* Centres, fixing:12.2mm
* Coaxial Connector Type:SMA
* Diameter, fixing hole:2.6mm
* Diameter, panel cut-out:6.5mm
* Frequency, operating max:18GHz
* VSWR:1.07
* Voltage rating, DC:500V
* Voltage, proof:1500V

when i change the coax radius from 1.2 to 2.5... the return loss graph change dramatically... from previous senior experience, the coax doesn't change to much for return loss graph.. can you explain it... isn't any standard to drawing coax for HFSS since many tutorial that i see using vacuum as material for coax...

thanks very much!!!! ....

I ask you: which HFSS version are u using?
Radius modify TEM Z impedance of course, let the material the same. But changing material it will compensate the radius variation giving the impedance specified in the datasheet.
Again: post the file and I will have a try to fix it. I can't guarantee an immediate reply cause I'm very busy with work, but I think I will be able to do it by Sunday.

I.

Ivan...

Thank you very much.. I'm using Ansoft HFSS version 10.0.. I will send the file through PM.. thanks again for your helping...

mr !mad

Have a try using this one's radius and insulator:
**broken link removed**

Find insulator:
TFE fluorocarbon per ASTM D1457 (find characteristics in the net)

It s very nice disscussion. sOrry I need to note it.

Ivan..

I'm usign this connector right now..
**broken link removed**


From now, I'm using this value for my design

Coax pin,

radius = 0.5mm
material = copper

Coax,

radius = 2.0mm
material = vaccum

please correct me if any mistake, i think the coax material shouldn't be vacuum as you mention before... can u check for me from the above site and the technical info that i attach here to correct my coax problem..

thankss a lot..

cuongtran,

thanks for ur note...

mr !mad

Mr Mad,
I apologizer cause I couldn't run and fix the sim. I'm very very busy , I've an important deadline to meet, apart from academic work. Are u sure the material of the coax luncher is vacuum? Don't think so. Maybe some kinda foam??? Anyway, another way of being sure is to simulate the coax alone, just to check its characteristic impedance. If 'll be able to get a few minutes to do something about ur model, I'll immediately let u know.

I.