How to feed a dipole in h/f/s/s v8?
First I should mention that I'm an almost total beginner at h/f/s/s.
I'm trying to get a directivity pattern and input impedance of a standard very thin lambda/2 dipole in h/f/s/s v8.
I drew two (perfect conductivity) cylinders on the Z-axis with a small gap between them and altogether length is lambda/2 at the frequency of interest.
How do I feed this dipole? Using ports or voltage drop or current?
What do I apply on the two faces of each cylinder facing each other to feed a dipole correctly?
Directivity pattern is well characterized in literature and the input impedance should be around 70 ohms, right?
Please help!
Thanks,
rfmw
Hi
It is very easy,you can draw a very little rectangle between the dipoles(There is a small gap like Helon's integral equation).Then,point the voltage on the rectangle from bottom to top,it is means set a lumped ports on it.Set radiation boundary conditions and begin calculation:)
CAoCao
OK, thanks for your help!
Shoudn't I draw instead of little rectangle in the gap a little-thin cylinder?
And then I define a "lumped port" to those two faces at dipoles and voltage drop? is this what you mean?
I would try this right now, but I'm at work now and the h.f.s.s is at my home computer.
hi there..... im a begginer in hfss too, did you get to feed the dipole correctly?? could you give me some detail on the boundary conditions?, where did you apply the lumped port??.... .
thanks.....
yes, I did just like caocao suggested. My length of half-wave dipole is 200 mm, which coresponds to wavelength 400 mm or 750 MHz. The radiation pattern looks OK (is symetrical and the curve is kinda right), but Directivity 0.98 and Gain around 0 dBi are wrong, at least I think so. Gain should be around 2.1 dBi, right?
When I change frequency to 1.5 GHz where dipole length equals a whole lambda the radiation pattern curve doesnt change at all!!! except that it has larger directivity and gain. At 2 GHz and above the pattern curve finally changes to multi-beam curve.
I'm doing something very wrong and I don't know what to do...
Look at page 8 in this hfss training material to see how to apply lumped port (caocao's suggestion) or wave port.
I will try to simulate the dipole later, maybe your problem is the radiation boundary... it may be too close to the dipole, try a 0.75*lambda radius sphere as the radiation boundary.....
I've been simulationg this half-wave dipole for the whole day now and found the following:
using the port - wave port the directivity (around 1.6) pattern and gain (around 2 dBi) pattern both look OK. But the S11 is way off from 73 Ohms what I was hoping for.
using the port - lumped port the directivity and gain patterns depend on the radiation boundary distance from dipole. I tried several from less than 0.25 lambda to 0.75 lambda and at 0.75 lambda looks somewhat better but absolute values (1.6 for Directivity and 2.1 dBi for Gain) are way off again. But S11 is pretty good, i got values from 70+j30 Ohms to 80+j40Ohms. I dont know why is this inductive component present...
I would be very grateful if someone experienced in HFSS simulates half-wave dipole, lets say 200 mm (100 + 100 mm each wire) long with 1 mm round wires. Gap between wires is small, around 0.5-2mm, lets say 1mm. Simulate Directivity (pattern), Gain (pattern) and Impedance.
Theoretical Directivity should be 1.64 and Gain 2.14 dBi I believe.
@aaron412 I tried your suggestion to replace cylindrical rad. boundary with spherical, but I hadnt found any significant improvement.
regards,
rfmw
hey..... I tried out the dipole...and i got perfect results for radiation pattern and gain (2.2 dBi)..... but my S11 is not so good.... (-3 dB).......I will continue to try and improve my model....
Here's a few of things to consider:
-a dipole will not resonate exactly at the frecuency of interest, because the lambda/2 formula is an approximation, actual impedance depends on the radius of the dipole and the size of the feed gap... you could try varying those 2 parameters..... aditionally a frecuency sweep could help you out in order to find the resonant frecuency for your model.
-hfss has several defenitions for computing the impedance at the port,you might be using the wrong one so this could be causing trouble with the s11 parameter, actually Im going to look in to this myself because I could be having the same problem, I'll let you know if I make any progress with that.
-the radiation boundary must be at least lambda/4 from the antenna at every point, separating the radiation boundary any further would only cause the problem to get bigger and shouldnt produce significcant variations in the results.
I hope this can be somekind of help
sure it helps, thanks a lot for your info!!! :)
I am sticking to THIN half-wave dipole (that is diameter of wires is 1/100 or smaller than wavelength) because there are known theoretical results for thin dipole which I am comparing with simulations.
What is your structure (dipole dimensions) and what port did you use?
also hf$$ 8 or 9?
about the gap between wires, I was varying that and I think that too tight gap (like 0.1mm for 1mm diameter of wire) is not OK for S11 to be around 70 ohms (larger gap like 2mm for 1mm diameter wire gives larger inductive component in S11, so I've set it to 0.5mm and changed my focus on varying radiation boundary from size of cylinder to sphere).
One more thing, most of my simulations give Radiation Efficiency slightly larger than unity, around 1.01 to 1.05, and this is most strange to me. Would this indicate problem with port definition or boundary? I am using perfect_conductivity wires and vacuum rad boundary.
I am anxiously waiting for you further results :)
best regards,
rfmw
Well I found in Kraus's book whats the impedance of thin linear center-fed half-wave dipole and guess what: Z11= 73 + j 42.5 Ohms 8O
Thats almost exactly the impedance I've got with the lumped port :D .
So to obtain the pure resistant impedance, the dipole length should be a few percent shorter to make the reactance zero.
For 3/2 lambda dipole the theoretical impedance is Z11=105.5 + j 45.5 Ohms if you will simulate at this frequency, I know I will...
I think using HFSS or CST microwave studio is like using a truck to carry a small cat. You will wait too much for the simulation to finish, especially for higher frequencies and/or large dimensions.
You may try to use SuperNEC (A demo version is available on the internet). It's much more faster and more suitable to the job.
Concerning the results, your's are correct. The pattern won't change untill the length of the dipole is more than 1.4 times the wave length, only the directivity does.
You may also use a small free program named APV for the purpose of check. 8)
thanks for your input on this matter.
I'll try to find out more about this supernec proggy. Somebody told me that NEC4 or something is pretty good too, perhaps its similar to supernec.
Well about h/f/s/s I found out that it's VERY mem hungry, but it gives pretty good results (at 40000 tetrahedra).
So far I've verified radiation patterns for thin half-wave dipole for lengths of lambda/2 (impedance OK too), lambda, 3/2*lambda (impedance OK too) and 2*lambda with the theoretical results.
hey,Im using hfss for a dipole just as a training exercise.... because it is easy to know if results are accurate my goal is to simulate more complex antennas, but its true if you want a more efficient simulation in terms of calculation time you must try any NEC based program, I personally recommend EZNEC...theres a demo available on the web which is perfect for the dipole.... Ive done extense work with EZNEC, a dipole simulation time will not exceed 10 seconds no matter what computer you have, and results are very accurate