This simply means your structure is accepting input to some extent but not radiating it. e.g. in a transmission line the reflection coefficient is good but you don't get radiation. What type of antenna are you simulating by the way?

Keep in your mind that the radiation effieiecny has not anything to do with the reflection becasue it only counts the losses due to dielectric or metallic and if you want to consider the mismatching losses you need to deal with the total efficiency.

You can estimate both of them using these steps as in the attached photos, also I attached to you a file called Eff.aoutvar which contains the radiation efficiency and the total one so you can directlly import them to your design.

Just to be sure from you design you can assume perfect conductor for all metal objects and lossesless material for all dielectric objects, and with this, you should have a radiation efficiency of 100%.

Good matching is achieved when the input impedance of the antenna is equal to that of the transmission line connected to it.

The radiation efficiency depends on the losses in the conductors and dielectric materials in the antenna. If the current is not properly distributed across the antenna then it will be a poor radiator (having low radiation efficiency).

Nitin Kumar I simulate microstrip antenna array 4x4

Khalil H. Sayidmarie Thanks. I know that. I can't get normal results in HFSS, in other programs all is alright, e.g FEKO.

Kindly try driven modal solution type instead of driven terminal. Driven modal is more suited to microstrips.

You may be losing 75% of the power in the feed network and the patches. How long are your feedlines and what attenuation do you expect in them?

Serhii V. Siden

What is the value of the radiation efficiency you get from HFSS at first and is it the same if you use the variables in the file Eff.aoutvar I attached?

Maybe you have small mistakes in estimation it using HFSS, but it can not be so different compared to other simulators

Dear colleagues! There is another problem, even for 1 element.

If I calculate rectangular patch with offset feeding - radiation efficiency is about 0,5. Can somebody help me?

All parameters were calculated using formulas from Balanis "Antenna Theory"

My variables: h = 1 mm, eps =4,4, Patch_width = 17,4 mm, Length = 13,35 mm, x_feed = 4,5 mm (from edge), 50 Ohm MS line = 1,91 mm

Program is attached.

Operating frequency is 5.25 GHz

Are you radiating linear and looking for circular, or radiating linear 45 deg and looking for linear 0 deg, or radiating circular and looking for linear? Any of these would give about 50% efficiency.

Malcolm White No. I'm just want to compute a rectangular patch antenna with linear polarization. In model which I attached, all is matched, but gain and directivity is different (total values).

I can't read hfss files, and have no experience with hfss, so can't advise about the behaviour of hfss, only about antenna predictions/measurements in general.

Your reply to me still leaves the possibility that the results are for a different polarization than you had intended. The antenna would still be matched, but only half the power would get to a linearly polarized far-field at 45 deg or a circularly polarized far-field.

Does hfss have a similar quirk to CST in that all powers are calculated relative to 0.5 W rms input power?

Thank you, I think your answer will be useful to me.. But I'm new with HFSS. I tried export mode from CST, where I have a normal results, but gain in HFSS less than in CST. Also I tried FEKO where difference between gain and directivity is minimum (abs Gain).

Initially, I tried to reproduce example from youtube - radiation efficiency 0,9, but when I use my parameters it doesn't work, though all sizes and excitation set correctly..

I'm confused. ;)

Dear Serhii V. Siden

After a cautious looking at your design, I think I can provide you now with an explanation for what is happing, please read it carefully and ask me if you need any further details.

Firstly, as a general advice before simulating your structure, you can go back to the theoretical solutions in Balanies book to have a rough estimation of the value of eff that you can get from the attached curve using the substrate height and the resonance frequency and you can see that for your design and for your freq you have h/lambda of about 0.02 and with Er of 4.4 so you can see that you will have an approximate Eff of around 85, for sure the exact value will depend on the losses of your material and your actual design.

Secondly, you are using FR4 epoxy which has losses of 0.02 this is so high and will significantly reduce your efficiency. To be sure from that and confirm that you do not have any problems in your simulation, you can simply simulate without the losses of the dielectric/metallic (for the current design you already ignore the metallic losses). This is what I did for the file called New_lossless and you can see that with this you have a radiation efficiency of 100% which means that the drop in your efficiency is coming from the losses of the dielectric.

So as a conclusion, you can start with the updated design I attached, try to use another material with lower losses such as Roger TMM4 or Arlon AD430 which have almost the same permittivity with losses 10 times less than FR4 epoxy

Also, some other small issues regard HFSS simulation

1- The airbox should cover all the object but for your case, the substrate and the ground are outside.

2- Usually for antenna problems in HFSS, it is better to use Driven Modal with a waveport and it needs to big enough to cover all the fringing fields, you can calculate the dimension using this website

http://hfss13.blogspot.com/2012/05/wave-port-sizing-guidelines.html

3- Better the airbox should be far away from all the objects by a distance of lambda/4 or more from all your frequency points.

Kind regards and hope that this discussion will help you

Attached are the design curve from Balanies V3 Ch14 and the updated design after changing it following all previous points which you can continue using it.

Kerlos Atia Abdalmalak Thank you! Your answer is very useful.

Now I know more about HFSS features.

Thank you one more time for your time!

Serhii V. Siden

You are welcome

Good luck

HFSS some times give Raf. Effi. > 1 which is just impossible. Therefore, no results on Rad. Effi for any antenna obtained from HFSS is not acceptable if proper care is not taken. I suggest, Do not use the direct plot option (rad. effi. vs freq.). HFSS has got limitation in that case.

Instead of using that one, please apply Wheeler cap method using HFSS. Because, in practice we use Wheeler cap method (one possible solution). In that case, you will get error free results. Otherwise, your microstrip antenna can show rad. effi 90% if you use direct plotting option (rad. effi. vs freq.). In practice, microstrip antenna gives 70-75%, not 90%.

So choose proper process to get correct results.