60 GHz phased array, need help!!!!
Patch antenna is used as the radiator whose dimension was decided by the very classic equations for pactch antenna.
Rogers 5880 of 0.127mm is the substrate.
When I placed four radiators with half free-space wavelength, and feed each with signals of same amplitude and 0, 45, 90 and 135 degree in phase of each from left to right, I could not see a pattern in th desirable direction and the side lobe is very high.
When I trid to adjust the separation between each radiator, the problem is still there.
But when I did it at low frquenies like 2.45 GHz, it worked pretty good.
So can any body help me with this?
I drove me carzy 'cause it is such a simple thing but I really could not find where the problem is.
Can anyone kindly reply and advice? I really wanna know what's happening on earth with my simple model.
Theoretical design on a simulator, or actual testing?
Things change pretty dramatically going from 2.4 Ghz to 60 GHz! There are many parasitic reactances and aspect ratio problems that are huge at 60 GHz, but insignificant at 2.4 Ghz.
You will either need a EMAG field simulator program, or if you are doing this empirically, go from 2 GHz to 10 Ghz and test. Then go from 10 Ghz to 20 GHz and retest/reoptimize. Then maybe go from 20 to 60.
As far as I know, theoretically, the array can scan certain agles with proper feeding in phase, regardless of working frequency.
But I built a very simple linear patch model (attached) in HFSS, the simulation result prove in vain. I am wondering what's the main factor for this?
I mean it works pretty well at low freq, but higher what I got is totally mess.
did you read my post? "are many parasitic reactances and aspect ratio problems "
Thank you biff44 for your post. Since I am a rookie ni this area, can you explian "parasitic reactance" and "aspect ratio"?
'Cause when I check Z11 or Z21 and some other self impedance and mutual impedance, they seem not being affected by the existence of adjacent radiators.
As for my understrading, I'd like to believe somthing is happening in the far-field region. Because nearly all the impedance whether self or mutual seem not be affected.
Look at the attached sketch. At 2.4 Ghz, the 50 ohm lead-in microstrip line is so small compared to the patch. Stray inductances and capacitances where the microstrip meets the patch resonator are almost negligible. Also, you can play around with the attachment point to get a better 50 ohm impedance match.
At 60 Ghz, because the patch size has to be so small to resonate at 60 Ghz, the attachment point of microstrip to patch is crucial. Those small inductances and stray capacitances probably influence 50% of the entire design. Also, becasue of the small size of the patch, it is harder to do any sort of impedance matching structure.
Hi biff44,
Thank you very much for your sketch, it is quite interesting.
Since I am using probe to feed the patch, and the parasitic reactance see to affect only the matching conditions of the patches.
For my understanding, as long as each patch has a smooth raidaiton pattern, whether resonant or not, they can form a tilted beam in the far-field region with proper phase distribution in feeding.
But in my design, I can not see this with 0.5 lambda patch spacing and 45 degree phase lag when feeding.