Wiki

Analog Devices Wiki

English
English
简体中文
简体中文
日本語
日本語
Руccкий
Руccкий

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
Next revisionBoth sides next revision
resources:technical-guides:best_engineering_practices_when_using_adxl_accelerometers [12 Feb 2021 20:26] – Pablo del Corroresources:technical-guides:best_engineering_practices_when_using_adxl_accelerometers [29 Mar 2021 19:16] – Pablo del Corro
Line 8: Line 8:
  
 The first and most important advise we would like to offer when using our ADXL portfolio is to **carefully read the product datasheet**, this can save you precious time.  The first and most important advise we would like to offer when using our ADXL portfolio is to **carefully read the product datasheet**, this can save you precious time. 
 +
 +===== Terminology =====
 +
 +=== Full-Scale Range (FSR) === 
 +The FSR is the guaranteed dynamic range at the output of the signal chain. FSR is specified as a minimum value and is guaranteed across all conditions. Acceleration measurement may be possible beyond this minimum value. However, performance characteristics are not guaranteed.
 +
 +=== Nonlinearity === 
 +Device nonlinearity is the maximum deviation of any sensor data point from the least squares linear fit of the acceleration data set at an equivalent input acceleration level. The acceleration data set can encompass any range of applied acceleration, up to the complete FSR. Nonlinearity is defined mathematically as
 +
 +{{ :resources:technical-guides:nonlinearity.png?nolink&250 |}}
 +
 +where:
 +ACCMEAS is the measured acceleration at a defined gn.
 +ACCFIT is the predicted acceleration at a defined gn.
 +gn is the input acceleration level.
 +
 +{{ :resources:technical-guides:nonlinearity2.png?nolink&400 |}}
 +
 +=== Cross Axis Sensitivity ===
 +Cross axis sensitivity is the measured output of the device in response to input stimuli orthogonal to the intended sense axis. It is measured as a percentage of the applied acceleration, as follows:
 +
 +{{ :resources:technical-guides:crossaxissen.png?nolink&250 |}}
 +
 +where:
 +ACCMEAS(gx) is the measured x-axis acceleration.
 +gy is the applied y-axis acceleration.
 +gz is the applied z-axis acceleration.
 +
 +The cross axis sensitivity specification accounts for device level cross axis components only. These components include variations in sensor fabrication and the alignment of the sensor to the orthogonal axes of the package (also known as package alignment error). The cross axis specification does not account for system level sources of misalignment (for example, on the PCB or module).
 +
 +=== Resonant Frequency (fo) ===
 +fo is the natural frequency at which the MEMS element has  a higher gain when subjected to acceleration events. Input acceleration at this resonant frequency causes the sensor to displace by an amount equal to the applied acceleration multiplied by the quality factor (Q).
 +Some parts use different sensor types for the horizontal (x- and y-axes) and the vertical (z-axis) sensing axes. Therefore, the resonant frequency responses of these sensors are not the same.
 +
 +=== Quality Factor ===
 +The quality factor is a scalar factor that governs the increase or decrease in amplitude of an acceleration signal applied at the resonant frequency of a MEMS element.
 +
 +=== Sensitivity ===
 +Sensitivity is the slope of the line of best fit for the acceleration transfer function, as measured across the output FSR. The sensitivity defines the change in output (LSB) per unit change of input (g). The inverse, scale factor, is in units of g/LSB.
 +
 +{{ :resources:technical-guides:sensitivity.png?nolink&400 |}}
 +
 +=== Measurement Resolution ===
 +Measurement resolution specifies the number of data bits in each acceleration data-word. For example, the 14-bit measurement of the ADXL317 has 16,384 bits of resolution. For an FSR of ±16 g (32 g total), this resolution yields a sensitivity of 500 LSB/g and a scale factor of 2.0 mg/LSB.
 +
 +=== Zero g Bias Error ===
 +The zero g bias error (also called offset) is any static error term on the sensor output. Zero g bias error is measured as the deviation from 0 g with no externally applied acceleration (including gravity).
 +To more accurately measure offset, take measurements at orientations of +1 g and –1 g and average the results. Each measurement must be taken over a sufficiently long time window to reduce the influence of external physical stimuli 
 +that may exist in the measurement system.
 +
 +{{ :resources:technical-guides:offset.png?nolink&400 |}} 
 +
 +=== Initial Zero g Output Deviation ===
 +Initial zero g output deviation is the error level at ambient conditions, measured immediately after completion of device manufacture. The initial zero g output deviation value denotes the standard deviation of the measured offset values across a large population of devices.
 +
 +=== Cutoff (−3 dB) Frequency ===
 +For applied ac acceleration, the cutoff (−3 dB) frequency (also referred to as bandwidth) is the frequency at which the input stimulus is attenuated in amplitude by 29.3% (1 − √2/2) at the output of the signal chain. The −3 dB corner is set according 
 +to the low-pass cascaded integrated comb (CIC) filter and the low-pass infinite impulse response (IIR) filter setting as selected by the user. A high-pass filter can also be turned on by the user but is disabled by default. All other signal chain elements have an appreciably high bandwidth and are not significant contributors to the cutoff frequency.
 +
 +=== Noise Density ===
 +Noise density is a measure of the inherent noise and is a combination of all internal noise sources. This density is fixed by the architecture of the device and is independent of bandwidth. See the Filtering: Noise and Latency Considerations section for more information on noise density.
 +Output Noise
 +Output noise is the realized noise in reported measurements. Whereas noise density expresses the inherent noise in the device, output noise is the union of density and bandwidth. Filters with lower bandwidths provide more aggressive filtering and, therefore, greater noise reduction than filters with higher bandwidths.
  
 ===== Power supply considerations =====  ===== Power supply considerations ===== 
Line 21: Line 84:
  
 === Power Cycling ===  === Power Cycling === 
-It is common in ULP applications to power cycle the accelerometer to reduce power consumption. When using this technique, please remember that it is highly recommend to always be started up the accelerometer from ground level (0 V) to ensure proper operation. +It is common in ULP applications to power cycle the accelerometer to reduce power consumption. When using this technique, please remember that it is highly recommend to always start up the accelerometer from ground level (0 V) to ensure proper operation. 
 If this is not possible, care must be taken regarding the following specifications: If this is not possible, care must be taken regarding the following specifications:
  
-  * __**VS supply start-up threshold**__: During start-up or power cycling, the VS and VDDIO supplies must always be started up from below 200 mV. When the device is in operation, any time power is removed or falls below the accelerometer power supply lower range voltage, VS and VDDIO supplies must be discharged below 200 mV. **This specification is mandatory**. +  * __**VS supply start-up threshold**__: During start-up or power cycling, the VS and VDDIO supplies must always be started up from below 100 mV. When the device is in operation, any time power is removed or falls below the accelerometer power supply lower range voltage, VS and VDDIO supplies must be discharged below 100 mV. **This specification is mandatory**. 
-  * __**Hold time**__: VS and VDDIO supplies must be held below 200 mV for at least 100 ms before re-powering the part. +  * __**Hold time**__: VS and VDDIO supplies must be held below 100 mV for at least 200 ms before re-powering the part. 
-  * __**Rise time**__: For the worst case scenario, 200mV Vs starts up and 100ms hold time, VS and VDDIO supplies rise time must be linear and within 45 μs to reach the supply lower range voltage. For example, for the ADXL372, the power supply lower range voltage is 1.6V, thus the voltage supply should raise from 0V to 1.6V within 45μs. +  * __**Rise time**__: For the worst case scenario (100mV Vs starts up and 200ms hold time) VS and VDDIO supplies rise time must be linear and within 250 μs to reach the supply lower range voltage. For example, for the ADXL372, the power supply lower range voltage is 1.6V, thus the voltage supply should raise from 0V to 1.6V within 250μs. 
  
-Hold time and rise time specifications can vary. For example, if hold time is much longer than 100ms, the rise time criteria might be relaxed. The user should carefully select the start-up specifications that suits their application.+{{ :resources:technical-guides:adxl372_revc_por_graph.jpg?nolink&400 |}}
  
-<note>Fully discharging the power supply to ground level allows a much more relaxed rise time (≤ 500µs, from 0V to the power supply lower range voltage) and hold time as low as 10µs.</note>+<note>Fully discharging the power supply to the ground level allows a much more relaxed rise time, ≤600 µs, from 0 V to supply lower range voltage, for a 200 ms hold time.</note>
      
 To enable supply discharge, it is recommended to power the device from a microcontroller general-purpose input/output (GPIO), connect a shutdown discharge switch to the supply, or use a voltage regulator with a shutdown discharge feature.  To enable supply discharge, it is recommended to power the device from a microcontroller general-purpose input/output (GPIO), connect a shutdown discharge switch to the supply, or use a voltage regulator with a shutdown discharge feature. 
resources/technical-guides/best_engineering_practices_when_using_adxl_accelerometers.txt · Last modified: 15 Mar 2022 14:41 by Pablo del Corro

Page Tools