how to calculate the feed point of microstrip antenna

can anybody plz help me in calculating the feed point for prob fed of microstrip antenna
Epsr=Rt/Duroid 5880
h=1.575mmL=38.4mm w=38.4mm
it is a square patch
plz plz calculate the feeding point for me
i will be very thank ful to you

Hello.
I would like to remind you that in feeding a patch antenna, the location of the feed point determines the patch's input impedance at the resonance frequency. If feeding with standard 50Ohm SMA connectors, the input impedance of the patch should be adjusted to approx. 50Ohm for optimum matching.
Now keep in mind that you don't want the feed location to be neither at the edge of the patch nor at the center. The reason is that (using the cavity model for analyzing the patch and assuming the dominant mode) the E field is zero at the center of the patch (so this would give you zero input impedance) and maximum at the edges (implying very high input impedance). So you generally want to be somewhere in between the edge and the middle to achieve 50Ohm.
The rule of thumb is approx 1/3 of the distance between the two edges.
Any book for microstrip antennas should have the formulas you need to find the location of the feed point, to obtain precisely 50Ohm input impedance. My calculation gave me that you need to place your feed at about 12.88 mm from the edge.
Notice that 12.88* 3 = 38.66 which validates the rule of thumb??
I strongly encourage you to read the analysis of the patch antenna. You would feel much more comfortable with your calculations!!
Take care.

Thank you so much for your explaination on feed point. but tell me one thing that why 12.88 is not giving me the good return loss
how can i get the good return loss?
i set the 12.88mm feed point on both possitive x axis and on negative also but i m not getting the good return loss. why?
does return loss depends also on grnd plain?

actually i m working on a very simple research paper. i just have to verify its result. this paper is based on enhancement of bandwidth and gain.

regards
sameer

Hi again,

So I guess you're saying that your simulated |S11| is not very good? What do you mean not "good return loss"? A typical patch antenna has very good return loss right at the resonance frequency. This means an |S11| < -20dB.

The bandwidth though is small. By bandwidth I mean the frequencies for which |S11| < -10dB. Typically the bandwidth is about 6%. Bandwidth is usually define as (fmax-fmin)/fcenter

with fmax and fmin being the frequencies where |S11| crosses -10dB and fcenter is the center frequency (where you have almost perfect matching i.e. |S11|<-20dB).

If you want to make the discussion easier show me a graph of your |S11| vs frequency and I can help out to fix your matching. It might be for example that the probe you made to feed the patch is not 50Ohms and that's why the patch looks missmatched. But I'd like to see the plot to be able to help more.

The ground plane does affect things a lot. But as long as you've made it larger than the patch (about 2-3 times larger) you should be fine.

If you want attach me the paper you're reading. I'd be interested to see what they're doing.

Best,
Yiannis

hello again

i have attached the paper and i set the grnd plane as 80mm x 80 mm.
the return loss is reaching upto -10dbi only

regards
sameer

i have attached the ie3d file (.geo file) of that antenna which i m designing. plz have a look and guide me.i will be very thankful to you
regards
sameer

Thank you.
Unfortunately, I've never used ie3d for simulations so I can't really open the file and process it :-(
But don't worry, we'll work with what we have

So from what I understand, are you trying to reproduce the green curve on figure 4 of the attached paper?
if the answer is yes I'd like you to validate a couple of things for me:

1) look at your |S11| vs frequency curve. You say that the dip barely reaches -10dB. This is an absolute indication that your matching is not good. So I want you to check your port impedance and make sure you've set it to 50Ohms. Then, if this is set right I want to to play with the feed location a little bit. Try moving it slightly (like 0.5mm step) either towards the center or towards the edge. So run an other 2 simulations with feed location 12.88-0.5mm and 12.88+0.5mm.
Just make a couple of simulations. The reason I'm saying this is that the formula for calculating the feed location is only approximate. You always need to fine tune your design to achieve perfect results. If you see that |S11| gets better for one of these cases, keep moving the feed either faster or slower until you find t he perfect location.

2)Now, you mentioned earlier that
"i set the 12.88mm feed point on both possitive x axis and on negative also but i m not getting the good return loss. why?"

Keep in mind that the feed location must always be on the line connecting the centers of your 2 radiating edges. Since your patch is square it doesn't matter. For example, to make myself clear: Say you have a square patch on the xy plane. Say the patch dimensions are 1mm x 1mm. The patch center is at (0,0). In that case you would need to place your feed on either the x or y axis and at approx. 0.3mm from the edge. It doesn't matter if it is the right or left edge because of symmetry.
So make sure your feed lies on that line, in your design. If you place it off that central line matching would be a little different!
If your patch was not square, say it was 0.8mm x 1mm things would be a little different but lets not get into that since it is not the case!

Finally how did you model your coax cable??

Maybe this is a problem.
Let me know

Yiannis

Hello Yiannis how are you?
Yiannis i have added .5mm to 12.88mm but did not get good s11 parameter even it was not touching the -10db. on 12.88 it was just reaching to -6 db but you know feed p0int at -8mm was giving the resturn loss (s11) parameter to -11db.
keep on guide me plz

regards
sameer

Hello,

How about the other things we talked about?
Which figure of that paper you attached are you trying to validate?
How do you model your coax cable?
Is your port impedance 50Ohms?
S11 = (Za - Z0)/(Za + Z0) where Za is the patch impedance and Z0 is your port impedance. By moving the feed location you are trying to adjust the patch impedance Za to be 50Ohms. But make sure that you've set Z0 to be 50Ohms also, otherwise you'll never get a good |S11|.
What do you think about those questions above?

Yiannis

Yiannis said:

Hello,

How about the other things we talked about?
Which figure of that paper you attached are you trying to validate?
How do you model your coax cable?
Is your port impedance 50Ohms?
S11 = (Za - Z0)/(Za + Z0) where Za is the patch impedance and Z0 is your port impedance. By moving the feed location you are trying to adjust the patch impedance Za to be 50Ohms. But make sure that you've set Z0 to be 50Ohms also, otherwise you'll never get a good |S11|.
What do you think about those questions above?

Yiannis

Click to expand...

Hello Yiannis
well, i m trying to get the better result of that attached paper, once i get that paper verified. i want to do further work just to get better result than that paper. so ye i want validate that figure 4
now abt port impedance, here in ie3d its already fixed to 50 ohm and whenever we get good return loss more than -20dbi it gives input resistaance at 50 ohm or more than 50 ohm like bcoz i have designed another antenna its basically high gain antenna and at -26.54 db its giving s11 at 52 ohm input resistance. and i checked also that by default 50 ohm is fixed.

But, Yiannis i m not getting ur that question that " how did i model coaxial cable"
actually in ie3d we just have the option of " probe feed to patch" we just click it and enters the value like x= -12.88 y =0 and no of segment like 100 segments thats it radius of the coaxial,ie3d calculate itsself.

regards

sameer

Hello again Sammer,

I was asking about the coaxial vable modeling, because sometimes you have to design the actual cable as part of the simulation. Obviously, IE3D offers a built in excitation, so you don't have to worry about that. Now I know that your feeding is not a problem.

I am afraid I can't help further with you problem. To be able to help further I need to see your patch model and your |S11| plot vs frequency.

As a last piece of advice I'd like you to view the following link, if you haven't already, and make sure you follow the steps right.

Modeling of a probe-fed patch antenna with finite ground plane - Mentor Graphics


If you follow all the steps right, and still you are not getting the result you expect, keep tuning the location of the feed. That is probably what's causing your problems.

I hope you get it resolved soon.
Good luck and take care.

Yiannis

Yiannis said:

Hello again Sammer,

I was asking about the coaxial vable modeling, because sometimes you have to design the actual cable as part of the simulation. Obviously, IE3D offers a built in excitation, so you don't have to worry about that. Now I know that your feeding is not a problem.

I am afraid I can't help further with you problem. To be able to help further I need to see your patch model and your |S11| plot vs frequency.

As a last piece of advice I'd like you to view the following link, if you haven't already, and make sure you follow the steps right.

Modeling of a probe-fed patch antenna with finite ground plane - Mentor Graphics


If you follow all the steps right, and still you are not getting the result you expect, keep tuning the location of the feed. That is probably what's causing your problems.

I hope you get it resolved soon.
Good luck and take care.

Yiannis

Click to expand...

thnx Yiannis thnx for ur guide m really thnkful to u

take care u too

sameerccs11 said:

thnx Yiannis thnx for ur guide m really thnkful to u

take care u too

Click to expand...

Hello Yiannis how are you dear?

i need your help again plz reply me

regards Yiannis