Correct formula for BATT monitoring on expansion board

[roadfox]

Hi

What's the correct formula for Battery monitoring with a LoPy on the pycom expanysion board.
According to the datasheet it's a 1:2 voltage divider on the expansion board. And the ADC is 12bit (0-4095) and somewhere i was reading the the reference voltage is 1.4 volt

so the formula i came up is:
adcVoltage = adcReading * referenceVoltage / adcResolution
battVoltage = adcVoltage * voltageDividerFactor

my code is:

from machine import ADC

class LoPy(object):
    def __init__(self):
        self.adc = ADC()
        self.batt = self.adc.channel(attn=1, pin='P16')

    def get_batt(self):
        adcVoltage = self.batt.value()
        voltage = adcVoltage*3*1400/4095/1000
        # print("Battery V = %.3f" % voltage )
        return voltage

but the result is too high, while i measure 3.31 Volt with a fluke, the reading is 3.91 from the formula above
what could be wrong, my formula, the adc settings?

[robert-hh]

@roadfox The upper range value from dB is:

10^(dB/20)

so 6 dB results in 1.995, which is about 2

[roadfox]

Ho do i calculate the range from dB attenuation?
So far i have the following Table:
attn = 0 = ADC.ATTN_0DB = 0 db gain = range 0-1V
attn = 1 = ADC.ATTN_2_5DB = 2.5 dB = range 0-1.334V
attn = 2 = ADC.ATTN_6DB = 6 dB = range 0-1.995V
attn = 3 = ADC.ATTN_11DB = 11 dB = range 0 - 3.548V

The Voltage divider consist of a 56k and 116k resistor witch gives a factor very close to 3

so i changed my formula to: voltage = adcValue * 3 * 1.334 / 4095
and it's pretty accurate now

many thanks

[robert-hh]

@roadfox The basic ADC range is 1V. You can set an attenuation factor of 2.5 dB, 6 dB or 11 dB (https://docs.pycom.io/chapter/firmwareapi/pycom/machine/ADC.html). 11 dB results in a 3.548V range. The voltage divider on the expansion board consist of a 56k/100k resistor pair, resulting in an attenuation factor of the resistors of 2.79, resulting in a range of 0..2.79V. So you have to use an attenuation factor of the ADC.
Update: I see that you use ATTN=1, which is 2.5 dB, resulting in a ADC range of 0..1.334 V. But the ADC on ESP32 is anyhow not very precise, and the impedances are not known. So you have to determine your own calibration factor.
Update 2: Some discussion about that is here:https://www.esp32.com/viewtopic.php?t=1045
A while ago I made similar measurements with a Lopy. The table below shows results for the 1 V range (attn=0).

Input	ADC
70	0
100	127
200	550
300	974
500	1820
700	2650
900	3510
1000	3933
1044	4095

You see that it does not match the 0-1 rails. The same for attn=1. For attn=3 it is also non-linear above ADC readings of about 2500.