SLAA843A August   2018  – March 2019 MSP430FR2512 , MSP430FR2512 , MSP430FR2522 , MSP430FR2522 , MSP430FR2532 , MSP430FR2532 , MSP430FR2533 , MSP430FR2533 , MSP430FR2632 , MSP430FR2632 , MSP430FR2633 , MSP430FR2633

 

  1.   Sensitivity, SNR, and design margin in capacitive touch applications
    1.     Trademarks
    2. 1 Overview
      1. 1.1 Design Objectives
        1. 1.1.1 Reliability
        2. 1.1.2 Robustness
      2. 1.2 The Designer's Dilemma
    3. 2 Recommended Actions for Developers
      1. 2.1 Run SNR and Design Margin Tests
    4. 3 Terminology
      1. 3.1 Signal (S)
      2. 3.2 Noise (N)
      3. 3.3 Threshold (Sensitivity) (Th)
      4. 3.4 Design Margin
        1. 3.4.1 False Detection Margin (Min)
        2. 3.4.2 Detection Margin (Mout)
      5. 3.5 Signal-to-Noise Ratio (SNR)
      6. 3.6 Advice
    5. 4 CapTIvate Device Performance
      1. 4.1 Minimum Recommended Values
      2. 4.2 CapTIvate Device SNR
    6. 5 Interpreting the Results
      1. 5.1 Interpreting the Advice
      2. 5.2 Check Other Results
    7. 6 Application of Terms
      1. 6.1 Count and Percent Change Analysis With 7.5-mm Overlay, Advice = POOR
      2. 6.2 Count and Percent Change Analysis With 1.5-mm Overlay, Advice = GOOD
      3. 6.3 Count and Percent Change Analysis (1.5-mm Overlay vs 7.5-mm Overlay)
      4. 6.4 Effect of Post-Processing and Sampling Rate
    8. 7 Summary
  2.   Revision History

3.5 Signal-to-Noise Ratio (SNR)

Signal-to-noise ratio, or SNR, is defined as the ratio of signal 'S' with respect to noise 'N'. SNR values greater than 1 imply that the signal magnitude is larger than the noise magnitude. SNR values less than 1 imply that the noise magnitude is larger than the signal magnitude. This parameter is useful for comparing the magnitude of the signal (measured percent change in capacitance due to a touch) to the noise floor (measured percent change in capacitance due to factors other than a touch). Equation 6 shows the SNR equation.

Equation 6. SNR = S N

There are a few differences between this use of SNR and traditional use of SNR.

First, SNR is traditionally a ratio of average powers (average signal power with respect to average noise power). This is not useful for capacitive sensing, because capacitive sensing involves looking at long-term step response signals rather than continuous time changing signals. Noise may indeed be thought of as a time changing signal with a frequency component, but it is more difficult to think of a sustained touch signal this way.

Second, SNR is traditionally presented in decibel (dB) format on a logarithmic scale to allow for comparison of values across a very wide dynamic range. Because capacitive touch signal values are typically within an order of magnitude or two of noise levels, there is little to no advantage to working on a logarithmic scale, and it can often make SNR values more difficult to compare than if a standard scale were used. Therefore, this document use simple ratios for SNR and not decibel values. However, if decibel equivalents are desired, the correct computation is 20 times the base 10 logarithm of 'S' over 'N' (see Equation 7). A multiple of 20 is used, because 'S' and 'N' represent magnitudes and not powers.

Equation 7. SNR =20 log10 S N

When measuring SNR with CapTIvate Design Center, you may find that it can be difficult to obtain consistent values when running the test multiple times. This is usually because, in many CapTIvate setups, the noise value may only be one or two counts as it is quantized in low-resolution measurements. Thus, if one SNR measurement has 1 count of noise and the next measurement has 2 counts of noise, the SNR would reduce by 50%, even though the reliability and robustness of the sensor may not have drastically changed. This is one limitation of SNR as an analysis tool in low-resolution capacitive sensing applications. Due to variance that can be seen in SNR, other metrics like design margin should also be considered.