Ganglion documentation, conversion from ADC sample to uV

Comments

• pushtheworld

  January 2018

  The value is suppsoed to be in microvolts
  
• Billh

  January 2018 edited January 2018

  Ah, found it elsewhere. In the file

  
  

  https://github.com/OpenBCI/OpenBCI_GUI/blob/master/OpenBCI_GUI/DataLogging.pde

  
  

  it says:

  
  

    // Digital max and mins

    private String bdf_digital_minimum_ADC_24bit = "-8388608"; // -1 * 2^23

    private String bdf_digital_maximum_ADC_24bit = "8388607"; // 2^23 - 1

    private String bdf_digital_minimum_ADC_12bit = "-2048"; // -1 * 2^11

    private String bdf_digital_maximum_ADC_12bit = "2047"; // 2^11 - 1

  
  

    // Physcial max and mins

    private String bdf_physical_minimum_ADC_24bit = "-187500"; // 4.5 / 24 / (2^23) * 1000000 *  (2^23)

    private String bdf_physical_maximum_ADC_24bit = "187500"; // 4.5 / 24 / (2^23) * 1000000 * -1 * (2^23)

    private String bdf_physical_minimum_ADC_Accel = "-4";

    private String bdf_physical_maximum_ADC_Accel = "4";

  
  

    private String bdf_physical_minimum_ADC_24bit_ganglion = "-15686";

    private String bdf_physical_maximum_ADC_24bit_ganglion = "15686";

  
  

  So microvolts per digital unit is 15686 / 8388607 = 0.001869917138805

  
  
• mark.o.s

  April 2018

  Hello,

  
  

  I have come across this same problem.

  Is the equation below from the MCP3912 data sheet correct?

  
  

  
  

  The typical equation for conversion from ADC value to Volt is more like:

  Vout = ADC_value * (Vref / 2^(n-1))

  
  

  I can not seem to get correct data using the formula specified in the learning section of OpenBCI "Ganglion Interpreting the data"

  http://docs.openbci.com/Hardware/08-Ganglion_Data_Format#ganglion-data-format-interpreting-the-eeg-data
  

  
  

  Or @BillH is the correct scale factor ADC_value * (15686 / 8388607)? Can you explain where this value came from?
• Billh

  April 2018

  Mark,

  
  

  15686 is the given physical voltage range from 0 of the values produced by the hardware, and for a 24-bit integer 8388608 is the digital range from 0. The data the computer gets is in the digital, 24-bit range but in general we are interested in the scalp voltages which are in the physical range. The two are directly linearly related via that conversion ratio, so the graphical displays of the two will be relatively equivalent (frequencies will look the same) but for the units on the axis of the graphs.. The actual non-railed EEG signal should be much less that the extremes of such a range.

  
  

  The ADC value is calcualated from the other values in the equation you placed above.