link to page 9 link to page 10 link to page 9 link to page 10 ADXRS652 THEORY OF OPERATION 0.1 The ADXRS652 operates on the principle of a resonator gyro. ) s Two polysilicon sensing structures each contain a dither frame m r that is electrostatically driven to resonance, producing the neces- 0.01Hz sary velocity element to produce a Coriolis force during angular °/sec/ rate. At two of the outer extremes of each frame, orthogonal to ( Y0.001IT the dither motion, are movable fingers that are placed between S N fixed pickoff fingers to form a capacitive pickoff structure that DE L A0.0001 senses Coriolis motion. The resulting signal is fed to a series of R T C gain and demodulation stages that produce the electrical rate E P signal output. The dual-sensor design rejects external g-forces and S0.00001E IS vibration. Fabricating the sensor with the signal conditioning NO electronics preserves signal integrity in noisy environments. 0.000001101001k10k100k 21 -0 The electrostatic resonator requires 18 V to 20 V for operation. 20 FREQUENCY (Hz) 88 0 Because only 5 V are typically available in most applications, a charge pump is included on chip. If an external 18 V to 20 V Figure 22. Noise Spectral Density with Additional 250 Hz Filter supply is available, the two capacitors on CP1 to CP4 can be TEMPERATURE OUTPUT AND CALIBRATION omitted, and this supply can be connected to CP5 (Pin 6D, It is common practice to temperature-calibrate gyros to improve Pin 7D). CP5 should not be grounded when power is applied to their overall accuracy. The ADXRS652 has a temperature propor- the ADXRS652. No damage occurs, but under certain conditions, tional voltage output that provides input to such a calibration the charge pump may fail to start up after the ground is removed method. The temperature sensor structure is shown in Figure 23. without first removing power from the ADXRS652. The temperature output is characteristically nonlinear, and any SETTING BANDWIDTH load resistance connected to the TEMP output results in decreasing the TEMP output and its temperature coefficient. Therefore, External Capacitor COUT is used in combination with the on- buffering the output is recommended. chip ROUT resistor to create a low-pass filter to limit the bandwidth of the ADXRS652 rate response. The −3 dB frequency set by The voltage at TEMP (3F, 3G) is nominally 2.5 V at 25°C, and ROUT and COUT is VRATIO = 5 V. The temperature coefficient is ~9 mV/°C at 25°C. Although the TEMP output is highly repeatable, it has only f = / 1 2 × π × R × C OUT ( OUT OUT ) modest absolute accuracy. and can be well controlled because ROUT has been trimmed V during manufacturing to be 180 kΩ ± 1%. Any external resistor RATIOVTEMP applied between the RATEOUT pin (1B, 2A) and SUMJ pin 22 -0 R 20 (1C, 2C) results in FIXEDRTEMP 88 0 Figure 23. Temperature Sensor Structure R = 180 kΩ × R / 180 kΩ + R OUT ( EXT ) ( EXT ) MODIFYING THE ADXRS652 SCALE TO MATCH In general, an additional filter (in either hardware or software) THE ADXRS620 is added to attenuate high frequency noise arising from demodu- The ADXRS652 scale factor can be modified to match the lation spikes at the 14 kHz resonant frequency of the gyro. The 6 mV/°/sec scale factor of the ADXRS620 by adding a single noise spikes at 14 kHz can be clearly seen in the power spectral 1.07 MΩ resistor between the RATEOUT and SUMJ. No other density curve, shown in Figure 21. Normally, this additional performance characteristics are affected by adding this resistor. filter corner frequency is set to greater than five times the required bandwidth to preserve good phase response. CALIBRATED PERFORMANCE Figure 22 shows the effect of adding a 250 Hz filter to the Using a three-point calibration technique, it is possible to output of an ADXRS652 set to 40 Hz bandwidth (as shown calibrate the ADXRS652 null and sensitivity drift to an overall in Figure 21). High frequency demodulation artifacts are accuracy of nearly 200°/hour. An overall accuracy of 40°/hour attenuated by approximately 18 dB. or better is possible using more points. Limiting the bandwidth of the device reduces the flat-band noise during the calibration process, improving the measurement accuracy at each calibration point. Rev. A | Page 9 of 12 Document Outline FEATURES APPLICATIONS GENERAL DESCRIPTION FUNCTIONAL BLOCK DIAGRAM TABLE OF CONTENTS REVISION HISTORY SPECIFICATIONS ABSOLUTE MAXIMUM RATINGS RATE SENSITIVE AXIS ESD CAUTION PIN CONFIGURATION AND FUNCTION DESCRIPTIONS TYPICAL PERFORMANCE CHARACTERISTICS THEORY OF OPERATION SETTING BANDWIDTH TEMPERATURE OUTPUT AND CALIBRATION MODIFYING THE ADXRS652 SCALE TO MATCH THE ADXRS620 CALIBRATED PERFORMANCE ADXRS652 AND SUPPLY RATIOMETRICITY NULL ADJUSTMENT SELF-TEST FUNCTION CONTINUOUS SELF-TEST OUTLINE DIMENSIONS ORDERING GUIDE