Computer analysis of dual band dish feed (G0MJW, PA3FYM, M0EYT). Part 2 –S Band

  • The same methodology was used to calculate feed parameter for S-band. Time domain solver approach was applied to calculate feed radiation pattern and impedance match. I –solver (MLFMM) was applied to evaluate the whole antenna assembly. 3D radiation patterns and their appropriate 2D cuts are shown in the attachment. The main lobe of the pattern exhibits some asymmetry and squint angle from Z axis. It can be evidently observed in polar plot. Axial ratio (AR) diagrams for angels Phi = 0, 45 and 90 deg. are shown in the following pictures, respectively. The established industry standard for circular polarization antennas is less than 3dB. However, I was curious about the changes of AR after reflection of the energy from parabolic reflector. The result is shown in the next diagram, from which it is obvious that reflector has not changed this parameter. Just a small deterioration occurred. Despite the authors report of very good impedance match, I was not able to get better value than 16 dB at 2.42 GHz. I got only one dip, just a little bit higher above the working frequency. I used very fine mesh setting with the mesh refinement around the edges. I have used this software setup in many of the projects, resulting in a very good agreement between the calculated and measured values. It is possible, that I used different N-style connector for modelling than the authors. See the attachment. For the impedance match adjustment, authors recommend bend patch to – from feed reflector. This technique may cause another deformation of the radiation pattern. However, it is better to sacrifice a few percent of the antenna efficiency than loose power amplifier or transventer. I guess, that better option is to insert some sheet of low loss dielectric material between the patch and its reflector.

    Remarks: The established industry standard of the impedance match (or returned loss RL) for a UHF fixed narrow band Tx antennas is around 20 dB (for example. GSM antennas have VSWR warning threshold 1.4, i.e.RL= 15.5dB). I measured a “professional“ product – helical feed for 2.4 GHz for EsHail 2 with VSWR =3, RL= 5.8dB recently).

    The phase center of the S band section lays 11.3 mm in front of the feed reflector plane. For X band it is 23.6 mm. Since, I consider more important to achieve better efficiency for X- band than for S-band, I calculated the feed optimally positioned for the X band. This introduces axial defocusing error of about 0.1λ. The efficiency calculation is involving axial defocusing error, X - polarization losses, diffraction losses, illumination losses, phase error losses, spillover losses, and aperture distribution losses. The efficiency curve is not pure concave shaped, it has also a convex part. I intentionally did not smooth the efficiency curve in the MATLAB for better detection of this behavior. The feed on S-band operates on the edge of its farfield region. For close reflectors, the field emanating from the feed has no point source properties. resultnig in phase error losses.More can be find in my article at: This is not this desing‘s fault but a general problem of small reflector antennas.


    Engineers of Mitshubishi electronics did an excellent job. Sensitivity of the transponder despite twice lowering its gain is amazing. It opens a gate for hamradio experiments on UHF/SHF bands. There are reports of sucessful work even with 8 dBi Tx antennas. RF antenna designer can try and verify their design easily using an SDR technology or web-site receivers. Possible deficiencies in the design can be easily compensated by increasing the transmitter power. This simplicity thus enables utilization of the transponder not only for common amateur radio applications but also for education or emergency purposes. Behalf of QARS I wish you to enjoy the EsHail 2 satellite!

  • Interesting - I am not sure if your conclusion is the antenna is good or bad.

    Remember I do not have access to the fine mesh optimisation of CST and had to do this in the frequency domain under rather limited mesh size, so getting even close is not a bad result. I am not an expert in antenna design. This was developed from scratch over the course of a week over Christmas, with no previous experience of CST or any other simulator. We did a little tweaking later on but it is pretty much the first attempt. I did not expect so many people to copy it.

    Some observations on your comments:

    An efficiency of just over 50%, with all the losses included is not a bad result. Especially considering this is a dual band antenna with a thumping great hole in the middle.

    I think we only claimed 18 dB return loss so not that far out. I did not have the capability to include the connector in the simulation, the mesh limits again...

    The axial ratio, 3,6 dB is not so bad given the compromise of a single feed point and all the limitations in my simulation. Perhaps even a bit better than expected. If it is properly tuned, it might improve. I noted this parameter varies strongly with frequency. Do you see the loop if you switch to the frequency domain?

    Squint - yes, I saw that too but it wasn't too serious, well maybe with a very large dish. A penalty of the feed arrangement. I was more worried about overspill loss at the time.

    I aimed to get the phase centres close but was not too concerned about S-band for the reasons you set out. It is 4 times less critical and 2.4 GHz power is not so hard to come by. All this was done last year, well before the satellite opened to use so we had no idea it was so sensitive.

    The question is now, can you do better with access to the full software? Are there better dimensions we can aim for?


  • Adding some words of my own to Mike words. First of all I appreciate A75GR 's efforts to simulate the feed. However, we do not live in a simulated world. This implies compromises. Like Mike said, during Christmas last year we had to take some decisions concerning 'what is good enough'? The result is a feed which performs well enough, even better than (most criticists) expected (!)

    One simulates 3 dB AR, somebody else simulates 0.1 dB AR ...

    Concerning the double dip, this was simulated and also measured in practice.

    The trick is to get Z=1 (50 + j0) in the centre of the 'balloon' and that took some (trail and error) efforts to find the right feedpoint (several attempts/prototypes have been constructed).

    Not preaching to my own parish, the result is an ultra simple solution to dual feed a single dish with relative good efficiency, easy to construct with simple metal working tools.

    Two sheets of brass, a piece of copper pipe (with some nylon) compared to a solid piece of aluminum with four ports, (phase) couplers etc ...

    So, I also challenge people to come with a better solution given the simplicity and ease of the design: two plates, a piece of copper pipe (and maybe some nylon (or other dielectric)).

  • PA3FYM Remco I appreciate your comments... your work was a very good one for most of us.. The lens issue is a minor one but can improve the hole system. I took a file to get down the four edges of the feed a very little bit to put it into a in Germany so called HT110 plumb pipe endcap. this you can buy for 0,8 Euro in a brigstore.. that makes everthing waterproof between the plates in heavy rain and there is no impact reguarding the signal strenght of the transponder at all from that material..if you use the gray one and not the orange one.

    This design is a milestone for the whole community and a great work from the team to share expirience and knowledge to most of us.. Many thanks again..

  • To Mike,

    I tried to modify your design, just adding colar ring on reflector. See attachment. I achieve impedance match to be on the working frequency and below 20 dB, substantially suppressed side lobes, significantly improve axial ratio and shift phase center almost on the same position as it is on X band. However, feed still suffer with squint of main lobe. I think, that it is due to asymmetric position of the feeding point. I am attaching some samples from antenna book. I did not study an influance of colar ring on X-band radiation pattern and I did not calculated efficiency in dish. It is job for one day. So it is the way how you can improve your design. Play with reflector size, size of colar ring (It can be also with conical shape. I saw similar design at University) and feeding point.. I am leaving Eu for Duha now. No more time for antennas... Maybe for some band activity.

    73 & BR


  • Thanks Rasto

    That's really interesting. I will see if I can get this to work within the limits of the student edition. I did try a similar approach earlier but it was not successful as it tended to be more suited to a lower F/D. However, this was with a circular patch at 6mm spacing and before I realised how to get CP so maybe. I will look again. I looked at that book myself and concluded the same and also that the small squint probably didn't matter too much.

    It is more complex to make, but not excessively so and also maybe a retro-fit - someone (not me) needs to compare and contrast perhaps. If it turns out to be a worthwhile improvement, we can call it the "Patch Of Next Year" or PONY for short and sell it for £25. (only Brits of a certain age will get this joke)

    If you can tell me the dimensions, it would help.

    Enjoy Doha.


  • Hello Rasto,

    we also built some of these feeds and have reduced the reflector disc to a diameter of

    100mm in order to fit it into a 100m tube to protect if against rain etc.

    Do you think you 16mm high collar would still make sense or are the edges of the radiating element possibly to close to the collar ? Do you have any concerns to make
    the collar only 0.5mm thick instead of 1mm ?

    I like you improved design especially as you managed to get the same focal point for
    13cm and 3cm.

    Many thanks in advance for your feedback.

    By the way I measured also only one dip and think I understand why. If the crossing point of the alpha-shaped S11 is at 50 Ohms you get 2 dips. If the tip of the alpha shaped S11, i.e. the center frequency is at 50 Ohms you get one dip.

    I wish you a safe trip to Dohar. I hope to have another QSO with you soon from there.

    Kind regards

    Matthias DD1US

  • Hello Mathias,

    the goal to properly feed patch, is to excite proper dominant mode. Usually it is TM11. In circular polarization patch, two orhogonal dominant modes must be excited. Changing feeding point position, also higher modes could be excited (TM21, TM31 ...), resulting in radiation pattern deformation or unwanted side and backlobe radiation. The proper feeding point, could have impedance different than 50 Ohms. An impedance match must be subsequently performed. To find the feeding point for the best VSWR for 50 Ohms system, means various trade off for radiation pattern changes.

    I think, that it is not good idea to decrease reflector size. You increase backward radiation this way and increase main lobe width. You can use collar ring also for your smaller reflector, but you can expect higher impedance changes, since collar ring is closer and more affects the patch. The collar hight can be up to lambda/4. It can be 1 to 2 mm thick. It is not important value.

    73, Rasto

  • Matthias great, thank you. I'll order soon. But I have no experience how to glue Teflon and PP together.

    Perhaps anyone has a recommendation of an adhesive that is useable in this outdoor application.

    73, Mike

  • I got around to trying this and it seems good in the student edition too. That's all subject to the errors of a coarse mesh. The ring moved the resonance down a little and the VSWR is better, 1.2:1 and the axial ratio much better if the patch is size is scaled down from 65.5 to 65 mm but I am sure there are optimisations here. Axial ratio of 3dB is OK but not great. I tried 64.5 mm - the axial ratio is excellent, 1 dB, but the VSWR rises to 1.3:1. All of this was with the feedpoint at 29mm radius and 16 degrees, repeating with 18 degrees (the published design) gave a better VSWR and axial ratio but a narrower bandwidth.

    One thing we never included in the publication was the radiation pattern! There are some examples below for the curious.

    I will see if I can optimise, but we are going to need Rasto's help to verify the real performance according to the full CST.