Battery voltage monitor and AD conversion for XIAO_ESP32C

[msfujino]

The battery pad of XIAO_ESP32C3 is not connected to any port, so the battery voltage cannot be read and there is a risk of over-discharging the battery.
(XIAO_BLE has a port connected to the pad and can read the voltage.)
The battery voltage was divided by 1/2 with 200k and connected to the A0 port so that the voltage could be monitored.

About the AD conversion of the ESP32C3
The datasheet says nominally 2500mV full scale AD conversion, but there is a large variation from chip to chip, actually ±10%. My chip was 2700mV full scale.
Fortunately, the calibrated correction value for each chip is written in the fuse area, and by using the function alalogReadMilliVolts(), I can read the corrected voltage value without doing anything special.
The result of AD conversion and the voltage measured by the multimeter agree well with each other with an error of about 5 mV, which is not a problem in practical use.
In addition, during communication in particular, spike-like errors occurred, which had to be averaged out 16 times to remove them.

Reference:
“Analog to Digital Converter - ESP32-C3 - — ESP-IDF Programming Guide v4.3-beta3 documentation”
“\Arduino15\packages\esp32\hardware\esp32\2.0.5\cores\esp32\esp32-hal-adc.h or .c”

void setup() {
  Serial.begin(115200);
  pinMode(A0, INPUT);         // ADC
}

void loop() {
  uint32_t Vbatt = 0;
  for(int i = 0; i < 16; i++) {
    Vbatt = Vbatt + analogReadMilliVolts(A0); // ADC with correction   
  }
  float Vbattf = 2 * Vbatt / 16 / 1000.0;     // attenuation ratio 1/2, mV --> V
  Serial.println(Vbattf, 3);
  delay(1000);
}

[braddo]

I wonder if there’s any way to use the matrix mapping in the ESP to actually attach a port/pin to the battery? It’s not so hard to add a voltage divider, but it seems almost impossible to understand why this board has a charger but can’t read the battery charge!

[PJ_Glasso]

Yea, Your on the bus with the rest of us SMH…
GL :-p

[msfujino]

As far as the schematic shows, the battery terminal is not connected to any port.

[PJ_Glasso]

Hence the Voltage divider youtube video :-), either intentional?(without explanation) or a HUGE, GIGANTIC fail by whomever at Seeed Engineering.
SMH ;-p

[3tcubed]

Just what I was looking for, this sounds great, any idea how much power it would draw, hence reducing battery life? Any way to prevent/decrease loss. I’ve seen similar diagrams 45k and 100k resistors (different multipliers) which would be more efficient, decreasing power draw. Any chance the -C6 version corrects this and uses a pin to monitor directly?

[ernstkl]

Using two 100k resistors will result in a current of ca I = U / R = 3.7 V / (2 x 100 kOhm) = 18.5 micro A.
So I would rather use 1 MOhm resistors.

[Zhairgling]

Good morning,
Having only 1M resistance on hand, I would like to test what you suggest. Should I change the sample code?
Sorry I’m not very strong in electronics, I’m more of a developer.
THANKS

[ernstkl]

The code needs to stay the same. As long both of the two resistors are the same, the voltage will be divided in half.

[Zhairgling]

Thank you, I have the impression that the measurement is not precise on the other hand, I do not have the same value at all as that measured with a multimeter.

[ernstkl]

One source of inaccuracy is the tolerance of the resistors. A typical tolerance would be 5%. Say in the worst case one resistor is off from its nominal value by +5%, and the other is off by -5%.

Then the voltage result you get will be off by 5% as well:

V_pin = V_true * R_2 / (R_1 + R_2)

R_1 = R_nominal + 5% = R_nominal * 1.05
R_2 = R_nominal - 5% = R_nominal * 0.95

V_pin = V_true * 0.95 / 2

V_result = 2 * V_pin = 0.95 * V_true

[Zhairgling]

Thank you for all those informations.
However, I don’t know how but I feel like I burned my board. it must surely be the 3V3 regulator, the card is no longer recognized when it is plugged in via usb and the 3V3 pin is at 0V.

[esat]

Thanks for the technique. I am using this and it works wonderfully. I have a concern though:
My max measurement is around 4.2V on full charge, which is as expected. However, my lowest measurement could go all the way down to 2.8V (or may be further), which I believe may damage lithium batteries.

Doesn’t the xiao esp32c ldo have a cut off value that is safe for lithium batteries? If not, how could I enforce this in HW?

[msfujino]

The easiest way would be to use a battery with a protection circuit. Both over-discharge and over-charge should be protected.

[Robert_Ferencek]

Anyone had any luck or solution reading battery voltage via grove shield? https://www.seeedstudio.com/Grove-Shield-for-Seeeduino-XIAO-p-4621.html

[msfujino]

I have examined the schematic and it does not seem to include a function to read the battery voltage. The battery terminal voltage must be resistively divided and connected to the XIAO port for AD conversion.
(The battery pad on the back of the XIAO is not connected to this Grove-Shield.)