SLAAEO8 October   2024 MSPM0C1103 , MSPM0C1103-Q1 , MSPM0C1104 , MSPM0C1104-Q1 , MSPM0G1105 , MSPM0G1106 , MSPM0G1107 , MSPM0G1505 , MSPM0G1506 , MSPM0G1507 , MSPM0G1519 , MSPM0G3105 , MSPM0G3105-Q1 , MSPM0G3106 , MSPM0G3106-Q1 , MSPM0G3107 , MSPM0G3107-Q1 , MSPM0G3505 , MSPM0G3505-Q1 , MSPM0G3506 , MSPM0G3506-Q1 , MSPM0G3507 , MSPM0G3507-Q1 , MSPM0G3519 , MSPM0L1105

 

  1.   1
  2.   Abstract
  3.   Trademarks
  4. 1ADC Introduction
    1. 1.1 SAR ADC Principle
    2. 1.2 ADC Parameters
      1. 1.2.1 Static Parameters
      2. 1.2.2 Dynamic Parameters
        1. 1.2.2.1 AC Parameters
        2. 1.2.2.2 DC Parameters
  5. 2ADC Noise Analysis
    1. 2.1 ADC Noise Classification
      1. 2.1.1 ADC Noise
      2. 2.1.2 Reference Noise
      3. 2.1.3 Power Supply Noise
      4. 2.1.4 ADC Input Noise
      5. 2.1.5 Clock Jitter
    2. 2.2 How to Reduce Noise
      1. 2.2.1 Reducing Input Noise Through RC Filtering
      2. 2.2.2 Layout Suggestions
      3. 2.2.3 Improving Signal-to-Noise Ratio
      4. 2.2.4 Choose a Suitable Reference Voltage Source
      5. 2.2.5 Software Methods for Reducing Noise
  6. 3ADC Oversampling
    1. 3.1 Sampling Rate
    2. 3.2 Extraction
    3. 3.3 Application Conditions
  7. 4ADC Application Based on MSPM0
    1. 4.1 ADC Configuration of MSPM0
    2. 4.2 ADC DC Test Based on MSPM0G3507 ADC EVM Board
      1. 4.2.1 Software/Hardware Configuration
        1. 4.2.1.1 Hardware
        2. 4.2.1.2 Software
      2. 4.2.2 Test Result
      3. 4.2.3 Result Analysis and Conclusion

2.2.5 Software Methods for Reducing Noise

The most direct way to reduce signal noise by software is to increase the sampling frequency for oversampling, collect more samples than needed, and reduce noise in the signal by taking the average, thereby improving effective resolution and signal-to-noise ratio. The mean process also helps to eliminate the DNL error of the ADC transfer function. For the code lost in the ADC output due to the large DNL error, taking the average can make the code appear again, so oversampling can be used to effectively improve the dynamic range of the ADC. There are several points to note about oversampling:

  • If an accuracy higher than 1LSB is required, hardware averaging cannot take the average of the actual sample size. For example, when oversampling and collecting 16 data points, you cannot directly average the 16 points on hardware, which results in a result of 12 bits resolution and a maximum quantization error of ½ LSB. It is possible to average every 4 out of 16 sampled data to obtain a 14-bit quantization result. In this case, it is then converted into a 12-bit floating-point data by software. The maximum quantization error in this case is 1/8 LSB;
  • Appropriate noise can improve the noise reduction effect of hardware averaging. If the input signal with very low noise (noise peak to peak value less than 1LSB), due to the ADC resolution is only 12 bits in hardware, the output result remains constant regardless of the number of oversampled samples, and higher resolution cannot be achieved by oversampling. Therefore, an appropriate amount of noise exceeding 1LSB can improve the effectiveness of averaging;
  • Usually, the more noise there is, the more samples are needed to obtain high precision through oversampling, which results in lower effective sampling frequency for actual input signal.