to jian - very urgent
I have sent one mail to you with some geo files. I have simulated the file named check.geo. got some results. also i fabricated that. but the measured results are not good. even the return loss. i also checked the stability of simulated results. but in practical i am not getting.
also i added another few files. I kindly request you yo check all the files. plz consider the meshing, cells per wavelength, highest frequency and other parameters such that the simulation results should not deviate from measured results.
after your reply only i can proceed further. deadline to submit my report is very near. for further correspondance plz use my mail id.
with sincere thanks,
kbmani
Hi, kbmanick:
There are a few things you may want to pay attention to:
1. Your port dimension is quite large. It is easy to be affected by parasitics and uncertainty.
2. The CPW center conductor and the grounds are not completely balanced for your structure. The results will be affected very much by the details of how you feed or measure your antenna. The uncertainly involved is similar to you are measuring a dipole antenna with a coaxial cable with the inner conductor connected to one arm and the outer conductor to another arm without balun. The results will be signfiicantly affected by the detail configuraiton at the port or even by the shape of your cable because there will be unbalanced current going onto the outer conductor of your cable and they will also contribute to the radiation. Your measured input impedance is basicaly not fixed value.
Regards.
================ ==================
|| ||
|| || ||
|| || || \
\
| \ outer
innner
dear sir,
i could not get clear detail. is that 3 mm size is too large? can u plz correct the structure and send it to me? if say me the sample i can change accordigly. i have to complete the work as early as possible.
with thanks,
kbmani
Hi, Kbmanick:
The mainly problem is not the size. The size of 3 mm at 6 GHz is a little bit big. However, it might still be ok. At 6 GHz, the wavelength in free space is 50 mm. Considering the effect of the dielectrics, 3 mm is about 10% of the wavelength. It is not small and not very big. The major problem is that your antenna is a CPW fed antenna. The center conductor and the ground metal are not balanced as it is. Adding extra length in the feed line may make it balanced due to the fact that the higher order modes are not propagating modes. As it is, the measured impedance will be strongly dependent upon the detail of the feed proximity. In some sense, if y ou are using a coaxial cable to measure this antenna, you will see moving your cable will make a big difference in your measured impedance. The reason is that there will be current going onto the outer conductor of the coaxial cable due to the unbalance between the center conductor and the grounding conductors for this antenna. If you still don't understand it, I may try make a draft to you.
Best regards,
dear jian,
i also thought the same reason. now i have modified little bit. but cif u send me the draft i can still get some idea. i am really thank u very much for ur prompt reply and timely help.
with sincere thanks,
kbmani
Added after 3 minutes:
dear jian,
how much extra length i should add? in the previous file - final check - i added some extra length in the files named best1 and best2 etc. is it ok?
with thanks,
kbmani
Hi, Kbmanick:
Here is a picture illustrating it. The center conductor part is a little bit short. It does not make the structure balanced. If you connect a some TLN to measure it (assuming coaxial), you will have current following on the inner conductor and on the inner surface of the outer conductor. When the strucutre is not balanced, there will be signficant current flowing on the outer surface of the outer conductor. This current will be part of the radiator. Depending upon the configuration of the connector is, the current flowing on the outer surface may be different. In this way, you may not be able to get consistent input impedance. Just change the connector slightly and you may see big difference in the Zin. Also, move the cable a little bit also may affect it much.
The unbalance between the center conductor and the gnd lines is basically the higher order modes. Adding additional length will let the higher order modes die down and you will have the main CPW mode left. Then, you should get more consistent results. CPW has another propagation mode (-1, 0, +1). However, your structure is symmetrical and the mode can't be excited.
How long should you add it? You may need to do some test on it. You may find out what Zc the TLN is. Then, you try to normalize the s-parameters ot the Zc of the TLN. When you see the S(1,1) normalized to Zc does not change much with the length, it is an indication the higher order modes have died down. Regards.
I tried to add different length and it seems to me the |S11| normalized ot 61.7 ohms (zc) does not stablized. This is an indication your port's size is too big compared to wavelength. You can't get consistent Zin for such a large dimension port. As the structure is, it is a good radiator. However, you do need to define your feed properly in order to get consistent results. You do need to make sure the + and - terminal of the feed to be close enough and the size of the port to be small enough compared to wavelength. Right now, + and - is close. However, the size of the + terminal is too big. Regards.
thanks jian. you mean that port dimension is too large>? i mean 3mm?
Hi, Kbmanick:
Including the gnd, it is 7 mm wide. Your antenna is resonating at about 6 GHz. The wavelength in free space is 50 mm. The wavelength in the dielectrics is even smaller. The port size is well beyond 15% of the wavelength. You should consider reducing it. Regards.
dear jian,
the port size is 7mm.ok. from calculation i am not getting 15% above the wavelegth.? plz tell me with formula. can reduce the port size to 5 mm?
plz sedn me reply. by today i have to conculde. sirry for the continuous disturbance.
with thanks thanks,
kbmani
Hi, Kbmanick:
I don't have your file here and I don't know the Er of the substrate.
At 6 GHz, the wavelength in air is about 50 mm. If Er = 4, the wavelength in the subsrate is 25 mm. Then, 7 mm out of 25 mm is more than 25% already. If we consider the effective Er or Ereff of your structure is 2. Then, the waveguide wavelength is about 50 / 1.4 = 32 mm. 7 mm out of 32 mm iw more than 20%. Anyway, the dimension is big. You do need to reduce it. I can not tell how much you want to reduce it. You may want to try to see whether the Zc normalized S11 is consistent when the TLN is long enough. If it is not, it is an indication the port dimension is too big. Regards.
dear jian,
I made a very big mistake. thanks alot. now i corrected everything. without u i can not get this much idea. thank u for ur timely help and support
with very sincere thanks from heart,
kbmani