How can I realize probe feed for the patch antenna in HFSS 9?

Hi,

How can I realize probe feed for the patch antenna? Do I have to draw the coaxial cable feeding the patch? If so, how long it should be?

Another problem is what type of excitation use in this case? And what boundary conditions I should use for simulating the antenna?

Thanks in advance,
Redi.

hfss probe feeding

I have not used HFSS V9, but in version 8, I would draw the antenna horn as a air with PEC boundary on outside (this is the default). On the open end, I assign radiation boundary. For probe, draw a small cylindar and subtract it from the air. It is sort of a hole in the antenna now. Length should be about quarter wave. To feed probe, extend a small cylindar out of antenna around probe and assign as port.

hfss conducting cap

Hi linkup11,

I was thinking about ralizing microstrip patch antenna in HFSS. I was looking about some examples for this program and I didn't find them. Horn antennas are not in my concern. Maybe you or anyone have an experience with designing microstrip antennas on HFSS?

Greetings,
Redi.

mws completely filled with metal

to get (accurate) radiation pattern, you can use only "gap" port; to obtain both (accurate) input impedance and pattern, you better draw a coaxial line and use wave port on the far end of the coax to simulate your probe feed. Another option is to use a microstrip and wave port.

It seems the appropriate length is dependent on the problem. If the problem is not big, I would try different setting to see the comparison.

cst + completely filled with metal

Hi loucy,

I have tried several options with the coaxial line and without. Of course without coaxial line I had to use Lumped port. With Wave port and coaxial line the results were worse than without it. I have the questions what is the difference between Lumped and Wave ports? Is any of them better, more accurate or it just depends on the problem solved? What are the differences between solution setups and for what each of them is the best?

Greetings,
Redi.

Re: Probe feed in HFSS 9

My guess is that you have not set the wave port correctly.

First we need to clarify the question. I am assuming that the "probe feed" refers to a coaxial probe. I am not sure what is the "fix antenna" in the top post (meaning of fix?).

Take the example of a microstrip patch antenna fed by a coaxial line. To set and simulate the source, we have two options: 1) draw a short vertical conducting pin from the patch, apply lumped gap source on a small gap between the pin and the ground; 2) draw a coaxial line cutting through the ground and extend the center conductor to tough the patch, draw a conducting "cap" and apply the "wave port" source at the end of the coaxial line (i.e. the source is backed by a shorting plate).

Notice the "cap" is important. Some might say that its purpose is to prevent energy from leaking through the end of the coaxial line, but I think it is actually necessary to have the cap to simulate an incoming wave originating from the infinity (imagine that you have an infinite long coaxial line and send a TEM wave toward the patch). Without this cap, you have to formulate the fields differently to acchieve an incoming wave.

Accuracy of the two schemes is quite obvious once we see that the EM fields at the location where the gap source is applied are very complicated because of the discontinuity. In fact, one should question the return loss calculated with the gap source because the return loss is supposed to be defined for a TEM mode (i.e. what you measure with the coaxial feed). I guess people have done enough numerical experiment and actual measurement to support utilization of this gap source. In other words, we wave our hands and say it is "enough accurate" (when the radius of the pin and the vertical length of the gap are small).

On the other hand, you can say the same thing about the coaxial feed scheme (it is not "rigorous"). But you should trust the 2nd scheme better than the 1st.

Re: Probe feed in HFSS 9

Hi loucy,

Maybe you are right that I have not set the wave port correctly. I haven't heard about conducting "cap". What do you mean by this term? Is this some kind of metal part to connect probe to the patch? How should it look like?

Probably my problem with wave port was that I didn't knew exactly the dielectric of coaxial cable, but only the diameters of the inner and outer conductors. Connecting 50 Ohm wave port to coax cable with wrong dielectric made results wrong, because of bad matching, I suppose. Now I know the dielectric in my coax cable and I will check the results.

I used "p/a/t/c/h antenna" but it was automaticaly changed to "fix antenna". Just look at your own post.

Greetings,
Redi.

Re: Probe feed in HFSS 9

Let's say you draw a z-directed coaxial line from z=0 to z=10, the ground plane of the p_atch antenna is at z=10, the p_atch is at z=11. You would have drawn three cylinders along the z-axis: 1) one is with the radius of a the center conductor of the coaxial line, extend from z=0 to z=10, and it will be assigned the material "perfect conductor", 2)one with the radius of the centor conductor, extend from z=10 to z=11, it will be the conducting pin, 3)one with a radius the same as the inner radius of the outer conductor of your coax, extend from z=0 to z=10, and will be assigned a material the same as your dielectric. You would also draw the cap: rectangular or circular cylinder exending from z=-1 (say) to z=0, so it covers the z=0 face of the dielectric cylinder (the 3rd one). The cap will be assigned as perfect conductor.

you would then: step 1) select the dielectric object and assign a PEC boundary to the it; step 2) select the (planar) face of the dielectric object at z=10 and assign a continuity condition (PHC ? don't remember the exact name, I am talking about HFSS 8); step 3) select the face of the dielectric object at z=0 and define your wave port, set a line on this plane to define the voltage, etc.

Re: Probe feed in HFSS 9

Hi loucy,

Thank you, now I know what did you mean by "cap". I have add this and I think it more or less works. But still I have some doubts if I made everything right. So, let me make a checklist:
1. For the antenna problem I should make a radiation boundary condition. It will be a vaccum or air box big enough to have faces quarter wavelength (of solution frequency) far from the structure.
2. The wave port should be outside of the box with the radiation boundary.
3. I have used the "cap", but I don't know if it's needed when the port is outside the box with radiation boundary condition and it's not even meshed.
4. I have put the port to 50 Ohms (renorm option) and deembeded to position of ground plane (deembed option).
5. I should have to change mesh options (length option) to make the mesh for the radiation boundary condition more dence following the rule from manual (2*lambda)^0.5/10. But I couldn't do this, because the needed value from 24 mm for (maximum length) to around 2 mm, which makes calculation almost imposible to solve on my computer in the resonable time. Is this rule very important for S-parameters calculations for antenna or just for far-field patterns?

Is this what I have done was all right? Another question is continuity condition, I could not find this option in manual. Is it really exist?

Greetings,
Redi.

Re: Probe feed in HFSS 9

the wave port should be inside the box, and the cap should stay away from the radiation boundary (i.e. you don't want any object other than the air to be close to the outer boundary). With the wave port outside, I think of your coax is much longer than necessary.

You can use the Perfect H.. as the continuity boundary.

Re: Probe feed in HFSS 9

Is it necessary to use continuity boundary? I think it will be enough to put first and second dielectric just next to each other. On the other hand one should put everywhere these boundaries where two dielectrics are "touching" each other.

BTW, I am checking the same configuration in CST MWS, because it will be counting faster I hope. For now I am getting very strange results, completely inacurate or even impossible like reflection coeficient bigger than one for a passive antenna. I think that I am defining the port wrongly, but this is the subject for the new topic.

Greetings,
Redi.

Re: Probe feed in HFSS 9

Redi said:

Is it necessary to use continuity boundary? I think it will be enough to put first and second dielectric just next to each other. On the other hand one should put everywhere these boundaries where two dielectrics are "touching" each other.

BTW, I am checking the same configuration in CST MWS, because it will be counting faster I hope. For now I am getting very strange results, completely inacurate or even impossible like reflection coeficient bigger than one for a passive antenna. I think that I am defining the port wrongly, but this is the subject for the new topic.

Greetings,
Redi.

Click to expand...

Redi,

if you encounter problem with your structure in MWS, please post it here - as attachment and without results.
Rgz,
eirp

Re: Probe feed in HFSS 9

Redi,

if you encounter problem with your structure in MWS, please post it here - as attachment and without results.
Rgz,
eirp

Click to expand...

Hi eirp,

I have a question about CST MWS. How to make feeding coaxial line and attach a port to it. I created coaxial line with the inner cnductor, dielectric inside and outer conductor. The only posibility to attach the port was to pick the ending faces of dielectric, inner and outer conductor. But the results are crazy. Am I doing this OK?

Boundary conditions are all open(add space) to check the radiation in every direction.

Greets,
Redi.

Re: Probe feed in HFSS 9

You need the "continuity" condition to overwrite the "perfect conductor" you set when selecting the whole dielectric object (a cylinder) and assign perfect E to it. If you are not using this step, I guess you just need to cut a hole through the ground. You can plot the fields to check.

Re: Probe feed in HFSS 9

loucy said:

You need the "continuity" condition to overwrite the "perfect conductor" you set when selecting the whole dielectric object (a cylinder) and assign perfect E to it. If you are not using this step, I guess you just need to cut a hole through the ground. You can plot the fields to check.

Click to expand...

Hi,

Now I understand. I have done this in a different way by selecting just the side face of the dielectric, without selecting the whole object. So now everything seems to be OK. But I am curoius how to use mesh operations to make calculations more accurate, wihthout increasing much the calculation time. And there is still the question about mesh operations (length) on the radiation boundary. Leave it like it is or make i t smaller like they say to (2*lambda)^0.5/10, which makes the calculations extremely long. I am not so sure if the results will be much better.

Greetings,
Redi.

Re: Probe feed in HFSS 9

Redi said:

Redi,

if you encounter problem with your structure in MWS, please post it here - as attachment and without results.
Rgz,
eirp

Click to expand...

Hi eirp,

I have a question about CST MWS. How to make feeding coaxial line and attach a port to it. I created coaxial line with the inner cnductor, dielectric inside and outer conductor. The only posibility to attach the port was to pick the ending faces of dielectric, inner and outer conductor. But the results are crazy. Am I doing this OK?

Boundary conditions are all open(add space) to check the radiation in every direction.

Greets,
Redi.

Click to expand...

Hi, Redi!

I don't know, what you are doing wrong now (from your explanation it seems OK).
Please PM me your MWS zipped project - without results so it will have few kB.
I will check it!!
Regards,
eirp

Re: Probe feed in HFSS 9

Hi,

I think I have found my what was the problem with my probe feed in MWS. When I made my coaxial cable it was obviously circular, but when I put the port the port was square and some parts of it were extending to the background material, which is air. In this situation MWS was saying that my port is completely filled with metal which was not true. When I made square cable (outer conductor was square, bigger than square port dimension) the results were quite ok.. not from outer space I would say. BTW it is quite weird that I had to do that. My sructure now is not like in reality, but results are better. In HF$$ I didn't have these problems. When I made port circular, it was circular.

I would like to add that my structure is on the finite size of the ground plane which is feeded from back by coaxial cable.The reason is that I wanted to check the back radiation. The problems mentioned are not appearing when you use infinite ground plane, because you put the port on the back of the large metal plane and the port edges are not "going out" of it, like in my case with circular cable.

I am a beginner user of CST MWS and could do something wrong, but these are my observations so far.

Greets,
Redi.