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.

HFSS Airbox size and fabricated antenna measurement results

Status
Not open for further replies.

kae_jolie

Full Member level 4
Joined
Oct 13, 2006
Messages
234
Helped
9
Reputation
18
Reaction score
8
Trophy points
1,298
Activity points
3,158
I modeled a microstrip patch antenna in HFSS with airbox only lambda by 4 over the patch antenna. The airbox sidewalls and base touched the antenna, so none of the substrate edge scattering ripples are seen in the S11 results.

I, then, fabricated the antenna and saw bunch of ripples in the S11 results, though the pattern matched well with the simulation. The ripples, though, affect the S11 bandwidth since they go above -10dB.

I, then, simulated the microstrip patch antenna in HFSS with a bigger airbox that has its sidewalls and base some distance away from the sidewalls and the base of the microstrip antenna. The S11 results showed almost the same ripples seen in the measurement results and both results agreed well.

My question is: is it more accurate to simulate the airbox where the sidewalls are a distance away from the sidewalls and base of the antenna to capture the substrate edge scattering effect? Or, the substrate edge scattering ripples in measurement could have been avoided if measurement was done some other way?

Thanks.
 

It's better to simulate the airbox on all side of the antenna. I've found the best ways to do this are to either use a lumped port, or simulate the coaxial cable (or whatever you use) extending to the edge of the vacuum box, and use a waveport there. You also might want to try increasing the size of the radiation boundary -- nothing to lose there except simulation time.

Good Luck
 

I typically default the extents to 0.015, for both vertical and horizontal padding. This will also reduce simulation time substantially.
 

Could you elaborate on what you mean by "default the extents to 0.015 for both vertical and horizontal padding"?

Also, I must add that MoM simulation with infinite substrate and no airbox does not show the ripples. So, this begs for the question of how accurate are these MoM simulations with infinite substrates that are all over the literature.

I wonder which bandwidth I should consider: the measurement one with the ripples or the ideal-looking simulated one with no ripples?

Can substrate edge scattering ripples be avoided?

Thanks.
 

So, this begs for the question of how accurate are these MoM simulations with infinite substrates that are all over the literature.

Doesn't it indeed.

I wonder which bandwidth I should consider: the measurement one with the ripples or the ideal-looking simulated one with no ripples?

Which ever one is real.

Can substrate edge scattering ripples be avoided?

How far is the patch from the edge of the substrate? In my experience, this only happens if its too close, or if the radiation efficiency of the antenna is relatively low (< 0.5).
 

Could you elaborate on what you mean by "default the extents to 0.015 for both vertical and horizontal padding"?

Also, I must add that MoM simulation with infinite substrate and no airbox does not show the ripples. So, this begs for the question of how accurate are these MoM simulations with infinite substrates that are all over the literature.

I wonder which bandwidth I should consider: the measurement one with the ripples or the ideal-looking simulated one with no ripples?

Can substrate edge scattering ripples be avoided?

Thanks.
Under "HFSS Extents"
airbox.png
 

How far is the patch from the edge of the substrate? In my experience, this only happens if its too close, or if the radiation efficiency of the antenna is relatively low (< 0.5).

It is farther than lambda by 4 for sure, which is what is recommended, I thought.
 

One wavelength on each side is 30 cm on each side for frequency of 1 GHz. This means the substrate would be 60 cm plus length on one side and 60 cm plus width on another side. This would be a very huge microstrip patch antenna.

On another note, the gain on my antenna is very bad below zero, which indicates a low radiation efficiency. What are possible reasons for low efficiency? How can I improve it?
 

One wavelength on each side is 30 cm on each side for frequency of 1 GHz. This means the substrate would be 60 cm plus length on one side and 60 cm plus width on another side. This would be a very huge microstrip patch antenna.

Usually, wavelengths in the dielectric. But ya, anything is going to be big at 1 GHz. The extra space buffer is only if you want best antenna performance; obviously in real life there are usually more important considerations.

On another note, the gain on my antenna is very bad below zero, which indicates a low radiation efficiency. What are possible reasons for low efficiency? How can I improve it?

That makes more sense, and is probably directly related to your return loss ripples. As I indicated above, a low radiation efficiency often causes this behaviour -- power is coupled into the antenna, but instead of radiated is just dissipated in the dielectric. This causes coupling to/diffraction off of the ground plane edges to play an exaggerated role in the operation of the system if they are closeby.

Is it just a vanilla rectangular patch? If so, a drop in radiation efficiency can typically be attributed to the dielectric -- either its too lossy, or too thick (more thick = more coupling to external features).
 

Is it just a vanilla rectangular patch? If so, a drop in radiation efficiency can typically be attributed to the dielectric -- either its too lossy, or too thick (more thick = more coupling to external features).


The dielectric is somewhat lossy with er=4.4 and loss tangent=0.02. Also, the substrate thickness is 16mm which is 0.1 lambda (for min sweep freq of 2GHz) and 0.26 lambda (for max sweep freq of 5GHz).

So, I think you are right in that it is a combination of these two elements that is contributing to the low radiation efficiency.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top