IWR1443: Effects of Antenna Etch tolerance

Hello Kevin,

The RF structure Grounded Coplanar Waveguide, stripline extension to antenna, and the antenna all contribute to the Tx and Rx return loss.  The return loss has both an RF band 77-81Ghz, where there is more than (6) 10db S11.  

We have seen that over and under etching of the antenna can result in a phase difference, and higher losses.  

When you make the PCB, there is an RF stripline structure, for checking the PCB fabrication, so the return loss of this stripline can be measured.

There is a SPRACG5 mmwave Radar sensor RF PCB Design, Manufacturing and Validation Guide.    The guide indicates that tight tolerances are recommended, for a new EVM, we have the PCB fabricator check the bare board, against a dimensioned drawing.   We normally use a 1mil tolerance for this test.  

The Rogers 4835 LoPro, 4mil dielectric, 1/2oz Copper  - to use the  EVM RF traces/antenna, you need to use this RF Laminate material.

If your design has a max range of < 10meters, Field of view for azimuth and Elevation of ~120 degrees, and you want a simpler to fabricate antenna.   You can use a single patch design for the (4) Rx, and (2 or 3) Tx.    Please look at the EVM User Guide swru518c Figures 11, 12 for the existing EVM.  

The vertical patches combine in phase to reduce the elevation field of view.   If you make each one a single patch, there is less geometry for the under and over etching to affect.    There are 6 sets of Rx; outside left shield, Rx1, Rx2, Rx3, Rx4, outside right shield.  The Tx orientation allows for Tx1, Tx3 for MIMO.  Tx2 is for elevation.  All of these could be a single patch, reducing the board size and complexity.

The spracg5 document discusses EM simulation.   There are several papers that discuss antenna geometry, field of view, some discuss that a 15um disturbance can be measured in the testing of the patch antenna.  

suggestion: if you want to first copy the EVM antenna, and RF traces, use the 1mil error criteria

                    if you want to build an EM simulation from the DXF single antenna attached, you could introduce for a single patch other errors.

https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/1023/iwr1443boost_5F00_reva_5F00_dxf.dxf

I have attached a System Estimator example for 15meters, RCS .1, IWr1443, 4Rx, 2Tx , Range Resolution 4.5cm, with the 5dbi Tx and Rx gain.

At the top right the Max range is calculated at 26meters, with an RCS target minimum of .11.

Sensing Estimator Tool "  dev.ti.com/.../   " 

Attached configuration is loaded for System design - "iwr1443_15m_4Rx_2Tx_4p5cm.json"

{
    "platform": "xWR1443",
    "num_rx": 4,
    "num_tx": 2,
    "tx_gain": 5,
    "rx_gain": 5,
    "frequency_range": "77 - 81",
    "maximum_bandwidth": 4000,
    "tx_power": 12,
    "ambient_temperature_degC": 70,
    "maximum_detectable_range": 15,
    "range_resolution": 4.1,
    "maximum_velocity_kmph": 10,
    "velocity_resolution_kmph": 2,
    "measurement_rate": 10,
    "typical_detected_object": ".1",
    "detection_loss": 2,
    "system_loss": 2,
    "implementation_margin": 2,
    "detection_SNR": 14,
    "maximum_radar_cube_size": 256,
    "maximum_RF_bandwidth": 15,
    "maximum_sampling_frequency": 18.75,
    "sensor_maximum_bandwidth": 4000,
    "maximum_allowed_bandwidth": 4000,
    "starting_frequency": 77,
    "maximum_velocity": 2.7777777777777777,
    "velocity_resolution": 0.5555555555555556,
    "idle_time": 7,
    "adc_valid_start_time": 12.2,
    "excess_ramping_time": 1,
    "periodicity": 100,
    "ambient_temperature": 343.15,
    "noise_figure": 16,
    "num_virtual_rx": 8,
    "non_coherent_combining_loss": 3,
    "rcs_value": 1,
    "combined_factor_in_dB": -17,
    "combined_factor_linear": 0.0199526231496888,
    "valid_sweep_bandwidth": 3658.5365853658545,
    "inter_chirp_time": 20.2,
    "aux_comp_coeff_a": 78829268292.68292,
    "aux_comp_coeff_b": -11863014.634146342,
    "chirp_time": 150.48997524752477,
    "ramp_slope_init": 24.310832527869856,
    "ramp_slope_parameter": 504,
    "ramp_slope": 24.33300018310547,
    "aux_comp_T1": 1,
    "maximum_beat_frequency": 2.433300018310547,
    "sampling_frequency": 2.7036666870117188,
    "number_of_samples_per_chirp": 407,
    "total_sweep_bandwidth": 3984.195602416992,
    "idle_time_minimum": 7,
    "ramp_end_time": 163.73630758377425,
    "carrier_frequency": 77.2968638450859,
    "aux_comp_T2": 6.2,
    "adc_valid_start_time_2": 7.2,
    "lambda": 3.881140644997492,
    "max_chirp_repetition_period": 349.3,
    "chirp_repetition_period": 341.5,
    "num_range_fft_bins": 512,
    "min_num_of_chirp_loops": 11,
    "max_range_for_typical_detectable_object": 26.188911143061215,
    "min_rcs_detectable_at_max_range": 0.10762056499385238,
    "num_doppler_fft_bins": 16,
    "active_frame_time": 5.464,
    "range_inter_bin_resolution": 3.2591796874999996,
    "velocity_inter_bin_resolution": 0.3819444444444445,
    "radar_cube_size": 176
}

Hello,

Here is the single element patch.  You can put this into an HFSS simulation to look at the Field of View, estimated S11 return loss, and estimated antenna gain.   Note: in replacing the IWR1443 antenna, please remember to match the Tx to Tx trace lengths within 1 mil, and separately to match the Rx to Rx trace lengths to 1mil.

Still looking for the other report.

Regards,

Joe Quintal

RO4835_1E.zip