I could not reproduce the results of an UWB antenna in one paper, could anyone please help me?

[joshuacp]

Dear friends,

The reference paper [1] reports an antenna that could operate in the frequency range of 0.1 - 30 GHz while the size is 60 mm x 60 mm x 1.6 mm.

I found the results of the reference paper [1] were very exciting and I tried to follow the paper by simulating the antenna in HFSS. All of the antenna size modeled in HFSS are the same as that in [1]. But I could not reproduce the results in [1]. The operating frequency range in [1] is 0.1 - 30 GHz, but my simulation shows that the lower edge frequency for |S11| < -10 dB is about 1.7 GHz and the antenna I modeled could not operate below 1.7 GHz.

The attachments contain the reference paper [1], my HFSS model, and my simulation results.

My simulation tool is HFSS (Ansys Electronics Desktop 2022R1).
The simulation tool used in the reference paper [1] is also HFSS.

Could anyone please help me check my simulation model?

Any suggestions are appreciated.

Thank you very much in advance!

1) Das S, Mitra D, Chaudhuri SRB. Staircase Fractal Loaded Microstrip Patch Antenna For Super Wide Band Operation. Progress In Electromagnetics Research
C. 2019;95:183–194. Available from: http://www.jpier.org/PIERC/pier.php?paper=19070105.

 

[volker@muehlhaus]

Looking at the results that the authors of [1] show for simple patch antennas in Figure 3, I doubt that this research work was done properly.

 

[D.A.(Tony)Stewart]

“0.1 GHz to 30 GHz with a ratio impedance bandwidth of 300:1 for S11≤ -10 dB”. This means its voltage could be proportion to frequency for a constant E-field without remarkable nulls.

How far off were your results?

Does it cite all the necessary 1D, 2D and 3D variables that affect the results?

What is the resistance ratio of a 1cm length of wire? (Terminated by 50 ohms)


It may be remarkable to not have any resonances more than 10 dB return loss, The 3D characteristics can differ from simulation if not done in an anechoic chamber.

Does this paper cite all the design and test assumptions? How was it validated?

 

[FvM]

Curiously the paper reports S11 and normalized directional patterns but not S21 or radiation efficiency. In principle it's not impossible to achieve high bandwidth and good S11 with an electrical small antenna - at cost of efficiency by introducing large artificial losses. But it's not possible with the shown simple patch geometry.

I suggest to read Balanis Antenna Theory, paragraph 11.5 "Fundamental Limits of Electrically Small Antennas" to understand why the reported results are simply impossible.

 

[volker@muehlhaus]

I agree with FvM, this paper is worth nothing, and I wonder about the quality of peer review for this journal.

The "great" S11 results for all types of antennas seem to be a result of excessice substrate loss in their FR4 material, resulting in low radiation efficiency. We already see the negative impact of FR4 on gain & efficiency for patch antennas at 2 GHz, and they use it up to 30 GHz...

Also look at the photo of their hardware and how they soldered the ground connection ... these guys seem to have very limited experience designing for this frequency range.

 

[joshuacp]

Looking at the results that the authors of [1] show for simple patch antennas in Figure 3, I doubt that this research work was done properly.

Yes, I also doubt that this research work was done properly from the designs in Figure 3 of the reference paper [1].

The reference patch antenna in Figure 3 of [1] is a normal planar rectangular-shape monopole antenna. K. P. Ray gave the empirical formula for initial design of the normal planar monopole antenna in [2]. The reference paper [1] also used the similar empirical formula for initial design, but the difference from the formula used in [2] to that used in [1] is the coefficient, k and εeff. I think the formula used in [1] is wrong. As εeff used in [1], actually, is for the microstrips or microstrip antenna, not for the planar monopole antenna. There is a ground plane under the microstrip antenna while there is no ground plane under the planar monopole antenna. The electric and magnetic field distributions for the two kinds are not the same. So, εeff used in [1] could not be used in the formula in [1]. The reference paper [2] was published in 2008 and the formula used in [2] has also validated by lots of papers ([2] has been cited by about 334 papers).

For the sizes given in the reference paper [1], using the formula in reference paper [1], the calculated low edge frequency is 0.76 GHz, while using the formula in reference paper [2], the calculated low edge frequency is 1.34 GHz.

For the wrong formula used in [1], [1] shows that the HFSS simulation confirms that the low edge frequency of the reference patch antenna or the 0th iteration antenna in the Figure 3 of [1] is 0.75 GHz which is almost the same as the calculated result using the wrong formula in [1]. Maybe, this shows the authors of [1] start to make cheat from this step.

Reference
1) Das S, Mitra D, Chaudhuri SRB. Staircase Fractal Loaded Microstrip Patch Antenna For Super Wide Band Operation. Progress In Electromagnetics Research C. 2019;95:183–194. Available from: http://www.jpier.org/PIERC/pier.php?paper=19070105.

2) K. P. Ray. Design Aspects of Printed Monopole Antennas for Ultra-Wide Band Applications. International Journal of Antennas and Propagation 2008. Available from: https://doi.org/10.1155/2008/713858.

--- Updated Apr 4, 2024 ---

“0.1 GHz to 30 GHz with a ratio impedance bandwidth of 300:1 for S11≤ -10 dB”. This means its voltage could be proportion to frequency for a constant E-field without remarkable nulls.

How far off were your results?

Does it cite all the necessary 1D, 2D and 3D variables that affect the results?

What is the resistance ratio of a 1cm length of wire? (Terminated by 50 ohms)


It may be remarkable to not have any resonances more than 10 dB return loss, The 3D characteristics can differ from simulation if not done in an anechoic chamber.

Does this paper cite all the design and test assumptions? How was it validated?

For the exact size and substrate material in [1], the low edge frequency in [1] is 0.1 GHz, while my HFSS simulation shows the low edge frequency is 1.7 GHz.

There is no parametric analysis in [1], but there are step by step design procedures, detailed size, and substrate material information about the antenna in [1]. Normally, if there are no mistakes in my HFSS model, I should get the similar results as in [1].

--- Updated Apr 4, 2024 ---

Curiously the paper reports S11 and normalized directional patterns but not S21 or radiation efficiency. In principle it's not impossible to achieve high bandwidth and good S11 with an electrical small antenna - at cost of efficiency by introducing large artificial losses. But it's not possible with the shown simple patch geometry.

I suggest to read Balanis Antenna Theory, paragraph 11.5 "Fundamental Limits of Electrically Small Antennas" to understand why the reported results are simply impossible.

Thank you very much for your information about "Fundamental Limits of Electrically Small Antennas" in Balanis's classical antenna text. I got Chu's paper long long ago, but I have not spent time to read the paper. This time, to check the results in [1], I must spend time to grasp the fundamental limits of ESA.

Although, the reference paper [1] had not presented the radiation efficiency results, but presented the gain information in Figure 12. The gain results in Figure 12 of [1] seems a little abnormal from other UWB antenna. Normally, the gain in the upper frequency range, e.g., in the range of 25 - 30 GHz, should lower than that in the middle frequency range, such as, 5 - 15 GHz. But the gain vs. frequency in Figure 12 of [1] is monotonic.