All trademarks are the property of their respective owners.
TI’s mmWave radar sensors include an internal processor to stabilize the radar front end performance across temperature and process by running calibration routines. The processor also enables the sensor’s functional safety by periodically determining RF/analog performance parameters and detecting functional failures by running monitoring routines. The processor is programmed by TI and is dedicated for RF calibration and functional safety monitoring.
This document describes the various calibration mechanisms available in TI’s mmWave radar sensors and their configurability.
Figure 1-1 illustrates the radar front-end architecture in a TI mmWave radar device. The performance parameters of the RX LNA, IF amplifiers, TX PA, X4 (frequency multiplier), LO distribution buffers, and the clock sources shown all vary with process and temperature.
The purpose of calibrations is illustrated in Figure 1-2 using RX gain and TX power as examples. The gain of the RX LNA and the TX PA vary from device to device due to manufacturing process variations and also across temperature. The purpose of calibration is to ensure the RX gain and output power are maintained as configured by the user despite variations in process and temperature. To achieve this, the internal processor adjusts the mmWave circuit configurations at initialization (to mitigate effects of process variation) and periodically at runtime (to mitigate effects of temperature drifts). Figure 1-2 illustrates how calibration can be used to maintain the RX Gain and TX Power close to the configured settings across temperature drifts. These charts are illustrative and may not reflect actual device performance. Even with these calibrations done across temperature there would be some gain variations between devices, which must be considered in the user application.
These are representative plots on TI's first generation radar devices. Some of the calibrations (for example, the gain and power calibrations) are implemented as adjustments of circuit configurations based on measurement of RF/analog parameters. Other calibrations are implemented as adjustments based on process/temperature look up tables.
To enable functional safety, such as in automotive applications, the monitoring mechanisms in the device can be configured to periodically provide the host processor with RF/analog health and diagnostic information. These mechanisms enable determination of RF/analog performance parameters and detection of failures arising from transistor and interconnect faults in the field. The diagnostic information they provide can also be helpful during development and optimization of designs integrating TI mmWave radar devices.
The calibration and monitoring mechanisms in TI’s mmWave devices are implemented using a combination of hardware and firmware. Some of the hardware infrastructure blocks enabling these are illustrated here.
Several TX, RX RF and IFA parameter measurements are enabled by the mmWave power detectors coupled to the TX PA outputs and RX LNA inputs, and the TX-RX RF and RX IF loopback structures in the device, illustrated in Figure 2-1.
For example, the Tx output power calibration is enabled by measuring the internal Tx power using the power detector at the Tx power amplifier output port. The voltage level of the power detector is read using internal general purpose ADCs. These ADCs are also used for measuring other internal voltage levels, such as PLL control voltages, during VCO and APLL calibrations.
Some calibrations, such as RX IF filter calibrations, use the internal IF loopback structure. The loopback signal at different IF frequencies is fed, the IF frequency response analyzed, and appropriate resistor and capacitor bank adjustments are made to realize desired cutoff frequencies. Other calibrations, such as Rx gain calibration, use internal RF loop structures to feed a known amplitude of signal level from the TX chain to the Rx chain. The Rx gain is analyzed by processing the ADC data amplitude and accordingly the Rx chain biases are set to calibrate the gain.
In some other calibrations, fixed look up tables (LUTs, derived from nominal design simulations) are evaluated in the firmware based on measured temperature and analog bias settings adjusted.
In AWR294x, RX Gain Calibrations use the RF Loopback structure to feed known amplitude of signal level from synthesizer to the Rx chain. There is no phase shifter loopback available. The loopback structures in AWR294x are shown inFigure 2-2.