CPW fed patch antenna

[rabiazainab]

Hi,

I am trying to design and simulate CPW fed patch antenna in HFSS. Can any one here help me because I am confused in one thing. How can I know the exact length and width of the substrate (not height) because every time i change the length and width of substrate and ground plane my results change drastically. BTW I am using Rogers 4003c as a substrate.

Also one more question. Is there a way in HFSS to find out the resonance frequency of the antenna. I want to know at which frequency my built antenna is resonating.

any help would be much appreciated. Thanks in Advance

Rabia!!

 

[Sherry1]

is it the bandwidth that changes upon changing length and width?

Resonance frequency is where your return loss is minimum and radiation efficiency is good.

You must be following some formula for patch antenna structure that tells relationship of freq and some dimension of antenna.

 

[WimRFP]

If your ground plane acts as a ground plane, influence of ground plane size variations should be well below the influence of patch size variations.

When your ground plane is not much larger then the patch size, there is dependency between antenna parameters and ground plane size, so it is normal that you experience this. If you want less dependency, you may increase the ground plane size.

Just to make sure: what happens when you reduce the meshing size to (for example) 70% (so reduce it with 30%)?

Using (grounded) CPW: is this a requirement as microstrip works very well also?

 

[rabiazainab]

yes i am using some formulas for the length and width of the patch antenna but i do not know any formula for the length and width of the substrate. Is there any formulas for that.

And when i change the length and width of the substrate, yes the bandwidth changes plus my antenna starts resonating at a frequency other than the operating frequency which is 60 Ghz in my case.

thanks.

 

[rabiazainab]

If your ground plane acts as a ground plane, influence of ground plane size variations should be well below the influence of patch size variations.

When your ground plane is not much larger then the patch size, there is dependency between antenna parameters and ground plane size, so it is normal that you experience this. If you want less dependency, you may increase the ground plane size.

Just to make sure: what happens when you reduce the meshing size to (for example) 70% (so reduce it with 30%)?

Using (grounded) CPW: is this a requirement as microstrip works very well also?

yes unfortunately its a requirement.

I think i figured out the problem. actually i was also changing the length of the CPW feed line which was causing the issue.

Now the issue is that I am not really sure of the length of the CPW feed line which will match with my antenna. I know a tool from where i can get the length and width of microstrip line but using these values as CPW feed line I am not getting the right results. Can you tell me How can i find out the length and width of my CPW feed line

Thank you so much for the reply

 

[reidintransit]

length of feed line shouldn't matter, width can be found with one of many transmission line calculators like TXLine from AWR (free and works well).

 

[WimRFP]

When the characteristic impedance of the CPW equals the reference impedance for your smith chart (the source in the simulator), the length does change S11, but not |S11| (it only rotates around the origin of the Smith Chart).

If Zcpw is not equal to Zsource, then |S11}, VSWR, return loss, etc will change. Very likely your simulator has an add-on utility to design the CPW. When the ground plane is close to the CPW, Z0 reduces significantly.

 

[rabiazainab]

When the characteristic impedance of the CPW equals the reference impedance for your smith chart (the source in the simulator), the length does change S11, but not |S11| (it only rotates around the origin of the Smith Chart).

If Zcpw is not equal to Zsource, then |S11}, VSWR, return loss, etc will change. Very likely your simulator has an add-on utility to design the CPW. When the ground plane is close to the CPW, Z0 reduces significantly.

thank you for the help.

actually when I change only the length of CPW feed line (keeping the patch dimensions same), the |s11| of my model changes along with the shift. it means Zcpw and the Zsourse of my model is not equal. right ??

Can you tell me how can I make these two equal. I am using wave port excitation with full port impedance of 50ohm . Does that mean Zcpw should be 50 Ohms as well ??

 

[WimRFP]

Do you have ground under your CPW? If so, when you change the length of the CPW, you also should change the length of a finite groundplane. Changing the size of a relatively small ground plane, will change the antenna's input S11.

There can be some change in |S11| even when Zcpw = Zsource. This happens when a common mode current exists on the CPW. In other words the CPW becomes part of the radiating structure of the antenna.

Can you post your structure (inclusive (finite) ground plane if present)?

Regarding how to make them equal, try a search on CPW impedance. You can check the CWP impedance by simulating a CPW section terminated on both sides with a port. Do a frequency sweep where the line length varies at least with 0.25 lambda. S11 (or S22) should be close to the origin for the full sweep range. If not, Zcpw <> Zsource, or there is something not well in de port definition. When looking to S12 or S21 you can also calculate the loss of your CPW structure.

 

[rabiazainab]

Do you have ground under your CPW? If so, when you change the length of the CPW, you also should change the length of a finite groundplane. Changing the size of a relatively small ground plane, will change the antenna's input S11.

There can be some change in |S11| even when Zcpw = Zsource. This happens when a common mode current exists on the CPW. In other words the CPW becomes part of the radiating structure of the antenna.

Can you post your structure (inclusive (finite) ground plane if present)?

Regarding how to make them equal, try a search on CPW impedance. You can check the CWP impedance by simulating a CPW section terminated on both sides with a port. Do a frequency sweep where the line length varies at least with 0.25 lambda. S11 (or S22) should be close to the origin for the full sweep range. If not, Zcpw <> Zsource, or there is something not well in de port definition. When looking to S12 or S21 you can also calculate the loss of your CPW structure.

Hi,

Sorry for late reply but I am try hard to get right results. I have attached my model. there are two important things

1. the CPW feeding line should be greater than 7.5 mm
2. Antenna should resonate at 60 Ghz

I have got the results but they are not very good specially radiation pattern is not accurate. Can you plz take a look at my design and let me know that how can I make radiation pattern better and gain more. Also better return loss.
Also have I assigned right boundaries, excitations and radiations in the design

Also I have one more question that does the thickness of the copper also affect the resonant frequency or not?

Thank you

 

[reidintransit]

1. Uncheck the renorm all modes box on your port and look at setup->matrix and see that port impedance is 77 ohms not the 50 you are targeting. For 50 ohms from TXline program linewidth of feedline should be 0.29mm keeping all other parameters same.
2. Remove the PerfE boundaries from copper objects.
3. Thickness of metal will have an effect, more capacitance to ground conductors.
4. Why do you have a finite conductivity boundary on the bottom of the airbox way below the substrate ground? Probably delete this unless your board is floating over a metal object in which case you would use an infinite ground plane I'd think.
5. In general, assign boundaries to faces/planes not 3D objects
6. You may find it helpful to use variables to set linewidth, gap, patch_width, patch_length, then changes are quicker.
7. Consider making the top ground plane smaller than your bottom ground plane and see effects on gain and resonant frequency.

 

[rabiazainab]

1. Uncheck the renorm all modes box on your port and look at setup->matrix and see that port impedance is 77 ohms not the 50 you are targeting. For 50 ohms from TXline program linewidth of feedline should be 0.29mm keeping all other parameters same.
2. Remove the PerfE boundaries from copper objects.
3. Thickness of metal will have an effect, more capacitance to ground conductors.
4. Why do you have a finite conductivity boundary on the bottom of the airbox way below the substrate ground? Probably delete this unless your board is floating over a metal object in which case you would use an infinite ground plane I'd think.
5. In general, assign boundaries to faces/planes not 3D objects
6. You may find it helpful to use variables to set linewidth, gap, patch_width, patch_length, then changes are quicker.
7. Consider making the top ground plane smaller than your bottom ground plane and see effects on gain and resonant frequency.

thanks for the reply

1. what happens if you check the renorm all modes box. Actually I have to fabricate this antenna and then i will feed it with the infinity probe which will probably have 50ohm characteristic impedance. So only this reason I have checked this box so that the full port impedance is 50 ohms. I really do not understand this concept of port and txline impedance match. Also because of this infinity probe i have in my lab, I am bound to keep the linewidth of CPW feedline 80 microns (0.08mm). what shld i do in this case

2. I have removed the PerfE boundaries. Thanks
4. presence and absence of finite conductivity did not effect my results so i did not bother removing it. Now i will remove it.

Thank you for giving the idea of using variables. I have really had great trouble in changing the values.

 

[reidintransit]

1. I assumed you wanted 50 ohm feed line to your antenna. If so you can change most of the line per my suggestion and then use a step to the probe dimensions.

 

[cjrathi]

You need to play around with length, width of CPW line. Changing length and width of substrate should not matter much if substrate is > that patch dimension.

There is some option in HFSS, wherein you can vary any parameters (ie you can define range of h,w,t, etc) and start simulation. After simulation completes, you can see all results and choose the best one and analyze it further. It really helps. That's what I had done.