AWR1843BOOST: High-Gain Antenna

Hello Hasan,

     One of the concern with the high gain antenna is beam tilt across the frequency. If the significant beam tilt happens then you may miss the target. Small beam tilt is acceptable but need to analyze with the actual link budget to meet the desired detection target goals. 

You may want to review the antenna gain and beam tilt across desired operating frequency band. It may not be big issue for narrow frequency bands, for example 76-77GHz. If you are planning to use broadband like 76-81GHz then you may have to look closely at this parameter. 

Another aspect is you may have to match the antenna with the transmission lines. Antenna simulations should include transmission lines, any radiations & coupling from the transmission line will affect the antenna performance. At the edge of the Package is the 50Ohm reference plane for that transmission lines need to be connected and need to ensure good matching  i.e. S11 < -10dB for all the antenna ports.

Package to PCB transition guideline is provided in the below app-note in section 2.1.2.

https://www.ti.com/lit/an/spracg5/spracg5.pdf

Along with the antenna, whole MIMO array simulation with the board routing need to be performed, to see any coupling effects and via, grounding effects that could alter the antenna performance. 

Thanks and regards,

CHETHAN KUMAR Y.B.

Hello Hasan,

     Thank you for sharing the simulation results.

     Typically high gain antennas are used in longer distance and would be used for narrower frequency bandwidths, This is traded-off with the range resolution to meet the max distance, IF bandwidth, memory, MIPs etc computational goals. One would be using narrow RF frequency bandwidth typically 500MHz to 1GHz for longer distance. Hence, it should not be a big problem in your application.  Beam tilt is a typical problem in the series fed patch array PCB based Antenna designs for high gain antenna. You may want to re-run your link budget analysis by keeping desired frequency of operation, if the roll of of beam due to tilt is not acceptable, then you may want to reduce the gain and increase the beam with at the cost of reduced Tx/Rx gain.

Another option is to consider beam steering capability offered by the Tx phase shifters. Device offers 6 Deg beam steering capability. This also warrants simulations to confirm grating and sidelobes are suppressed sufficiently. 

Series fed patch antenna for higher gains also increases higher losses and over all decreased efficiency of the antenna. You may have to pay attention to this point as well, Lower substrate loss (like R03003 is one of lowest loss material even compared to Ro3003G2) will help in lowering the overall antenna loss. There are other consequences like peel strength, manufacturing constraints, bonding with other layers of substrate materials, drill, high temperature requirements etc also have to be carefully need to be weighted for the application requirements.  

In your design, you have good match between 76 to 79GHz. That is good. And your best matching frequency is 77.7GHz, Due to limited etching tolerance from the PCB manufacture this frequency move up or down depending upon under etching or over etching of the traces and Antenna geometry. You may want to run simulation depending upon the min-max over etching and under etching tolerance limits of PCB manufacturer. If the frequency deviation is under your design limits then you are ok. Typically design is done such that Antenna/Transmission lines meets the design goals of PCB tolerance limits. 

It may be wise to run the test after the fabrication to find out what would be optimum frequency at which best matching is obtained and design your chirps at those frequencies. 

Thanks and regards,

CHETHAN KUMAR Y.B.

Hasan,

    Thank you for sharing the antenna structure. This structure is not uncommon and published in literature in many places. Below are some example  published in IEEE.  

IEEE Xplore Full-Text PDF:

IEEE Xplore Full-Text PDF:

Conventional lamda/2 spaced virtual array would limit going into higher angular resolution, one has to explore other higher angular topology such as sparse array /MUSIC algorithms with some trade-offs. In sparse array topology one need to explore grating lobes/spectral leakage emanating due to gaps in the virtual array as compared to linear arrays. This would be limiting the dynamic range of angle estimation. Are you planning to use other Tx2 and Tx3 also to maximize the number of virtual antenna pairs?

For further improving angular resolution, one could explore multichip cascaded solution in AWR2243 kind of product, But it would be more complex than single chip solution.   

Yes, It's not clear from the picture for the feedlines, But care must be taken to minimize the transition losses and well matched to antenna ports.

R03003G2 is relatively newer product introduced by Rogers in recent times, hence there may not be issue for the samples for proto developments, Where as R03003 has been existed in the market from many decades running high volume for most of the mmWave Automotive sensors. It may be challenging to acquire sample proto types, But one of the lowest losses material in the industry today. 

Thanks and regards,

CHETHAN KUMAR Y.B.

Hello Hasan,

     This appears to be simulation done for the Rx antenna at 77GHz for Azimuth and Elevation cuts. You may wan to look at all the Antenna pairs. 

If you are operating chirps from 77-78GHz or 77 -79GHz then you may want to review the whole frequency band for the antenna radiation pattern at say 500Mhz interval, and review antenna gain at the desired field of view and review does it meets the link budget requirement or not.

Thanks and regards,

CHETHAN KUMAR Y.B.

Hasan,

     Thank you for sharing feedback from the Rogers, Yes there are manufacturability challenges for the RO3003, Hence with the some performance trade-offs RO3003G2 is introduced.

As I mentioned before which need to be dealt with, for avoiding the thinning process, typically thin 0.5Oz base copper is used and minimalistic plating process need to be adhered to avoid thinning process. In the current design two types of vias are used L1-L2 and PTH, these requires plating process. If there are different types of vias used this will warrant a additional plating process and increases the thickness of the top cu, then thinning process need to be employed. This will create uneven copper surface this is undesired for the transmission line and Antenna arrays, If it's not done properly especially for the large panels and also it degrades the surface roughness. Some PCB manufacturer use scrubbing process to thin down this is even worse, which causes even higher surface roughness causing more insertion loss. All these aspects need to be carefully to be discussed with PCB manufacturer. Hence one need to go with the PCB manufacture who has handled and perfected the art of manufacturing process for these material otherwise there would be considerable learning process to handle this material.

Thank you for sharing the Genetic Algorithm used for getting Higher angular resolution.. This seem to unconventional but looks like promising. We don't have experience to comment on this. Our best wishes for your success on this project. 

Thanks and regards,

CHETHAN KUMAR Y.B.

Hello Hasan,

   Thanks for your confirmation. And once again best wishes for your success on this project. 

Thanks and best regards,

CHETHAN KUMAR Y.B.