Any way to read the voltage on the 3.3V pin?

[dave0003]

On a previous project, I used CircuitPython with a Pico W, and I was able to read the internal voltage without any special hardware setup. For my new project, I'm using MicroPython with an ESP32. Is there anyway to do something similar in MicroPython?

For example, this is how I did it in CircuitPython:

import analogio
from board import VOLTAGE_MONITOR

def get_vsys():
    try:
        global vsys_voltage
        pin = analogio.AnalogIn(VOLTAGE_MONITOR)
        adc_reading  = pin.value
        adc_voltage  = (adc_reading * pin.reference_voltage) / 65535
        vsys_voltage = adc_voltage * 3
        pin.deinit()
    except Exception as e:
        print("get_vsys problem: ", e)

get_vsys()

[GitHubsSilverBullet]

I used CircuitPython with a Pico W, and I was able to read the internal voltage without any special hardware setup.

That's because the PICOs have the ›special hardware‹ already on board.
It's a very simple voltage divider made from two resistors.

[dave0003]

I see, thanks.

[bikeNomad]

It depends on which ESP32 board you have. For instance, some of the XIAO boards as well as the Unexpected Maker boards provide a resistive voltage divider that can be used to measure the battery voltage (vs. the 3.3V supply voltage, which is generally less useful).

[dave0003]

I got the ESP32 board from China.

DevKitC-32 ESP32-WROOM-32 TYPE-C USB CH340C

There's not much more info on it, or branding, that I can see.

The back is blank.

[bikeNomad]

Yes, that's not an Espressif board or module (or even a real FCC ID number).

If it's a reasonably close clone of the Espressif ESP32-DevkitC, there is no internal provision for reading the supply voltage. The ESP32 has relatively poor ADC performance anyway (see https://www.espressif.com/sites/default/files/documentation/esp32_datasheet_en.pdf).

If it's really important to you to read this, you can add a voltage divider in order to read the supply voltage:

3.3V --[ R1 ]---*---[ R2 ]---*
                |            |
             <ADC_IN>    <GPIO_OUT>

Where R1 and R2 are selected to provide a voltage around 2V at the <ADC_IN> pin (using the highest attenuation setting on the ADC). To read the voltage, configure the <GPIO_OUT> pin as an output, drive it LOW (near 0V), and read the voltage at <ADC_IN>. Then change <GPIO_OUT> back to an input (or open-drain output driven HIGH-Z).
Multiply the measured voltage at the ADC pin by the inverse of the voltage divider ratio.
So if R1=10kΩ and R2=20kΩ, you are dividing the 3.3V by 2/3 and you'd expect to read about 2.2V at <ADC_IN>. So multiply your ADC reading in volts by 3/2.

Take into account both the accuracy spec of the ADC (6% or so) and the accuracy of the resistors when using the resultant number.

[dave0003]

Thank you!

[peterhinch]

If the module's a clone (which it certainly appears to be) everything is up for grabs. The ADC accuracy is anyone's guess.

[shariltumin]

This won't help you read your board's internal voltage, but it can help you test the ESP32's ADC and DAC.

Here's a simple script you can try:

from machine import Pin, ADC, DAC
import time

# DAC output on GPIO25 
dac1 = DAC(Pin(25))    # DAC channel 1 on GPIO25, Channel 2 on GPIO26

# ADC input on GPIO36 (32-39)
adc = ADC(Pin(34))
adc.atten(ADC.ATTN_11DB)  # Full 0-3.3V range

w = 0 # 0-255:0-3.3v - use to set DAC
while True:
    w += 1
    if w>255: w = 0
    dac1.write(w)
    v = adc.read()  # 0-4095
    print(f"ADC: {v}, DAC: {w}, VOLT: {(v/4095)*3.3}")
    time.sleep(0.4)

Use a jumper wire to connect pins 25 and 34, then run the script.

You will get something like this:

...
ADC: 3958, DAC: 253, VOLT: 3.189597
ADC: 3984, DAC: 254, VOLT: 3.210549
ADC: 4010, DAC: 255, VOLT: 3.231502
ADC: 0, DAC: 0, VOLT: 0.0
ADC: 0, DAC: 1, VOLT: 0.0
ADC: 16, DAC: 2, VOLT: 0.01289377
ADC: 35, DAC: 3, VOLT: 0.02820513
ADC: 48, DAC: 4, VOLT: 0.03868132
...

[davefes]

Take into account both the accuracy spec of the ADC (6% or so) and the accuracy of the resistors when using the resultant number.

The 6% is chip to chip variation. If you calibrate , ie changing 1675 below and stay in the linear part of the curve then within reasonable temperature limits you will do better than 6%.

  # read remote battery voltage
    while count < 8:
        battery_ADC_reading = adc1_ch4.read()
        result = result + battery_ADC_reading
        time.sleep_ms(10)
        count += 1

    battery = (result * 3) / 1675 / 8 #  8 samples
    battery = round(battery, 2)
    count = 0
    result = 0

[krismc]

I've had a couple of goes at calibrating esp32s/rpi picos using auto generated lookup tables. One thing to be aware of is that the max 4095 count with 11 dB attn rarely occurs at the nominal 3v3 volts, the batch of boards I was working with it was closer to 3v1.