[SOLVED]All GPIO show Digital despite being Analog

[JasonInRVA]

I am unable to get the web UI to show any of my pins as analog. They are all listed as [D]igital, despite actively using them to input analog data.

Here is a simple program that demonstrates the issue I'm seeing:

#include <Arduino.h>
#include <gpio_viewer.h>
#include "secrets.h"

#define POT_PIN 11   // GPIO11 (ADC2_CH0)

GPIOViewer gpioViewer;

void setup() {
  Serial.begin(115200);
  analogReadResolution(12);                      // 0–4095
  analogSetPinAttenuation(POT_PIN, ADC_11db);    // full 0–3.3 V span
  gpioViewer.connectToWifi(WIFI_SSID, WIFI_PASS);
  gpioViewer.setSamplingInterval(125);
  gpioViewer.begin();
}

void loop() {
  int raw = analogRead(POT_PIN);
  float volts = raw * (3.3f / 4095.0f);

  Serial.printf("GPIO%d raw=%d  volts=%.3f\n", POT_PIN, raw, volts);

  delay(1000); // once per second
}

Output:

20:08:25.924 > GPIO11 raw=1470  volts=1.185
20:08:26.924 > GPIO11 raw=72  volts=0.058
20:08:27.924 > GPIO11 raw=100  volts=0.081
20:08:28.924 > GPIO11 raw=105  volts=0.085
20:08:29.924 > GPIO11 raw=346  volts=0.279
20:08:30.924 > GPIO11 raw=538  volts=0.434
20:08:31.924 > GPIO11 raw=539  volts=0.434
20:08:32.924 > GPIO11 raw=539  volts=0.434
20:08:33.924 > GPIO11 raw=613  volts=0.494
20:08:34.924 > GPIO11 raw=1075  volts=0.866
20:08:35.924 > GPIO11 raw=1543  volts=1.243
20:08:36.924 > GPIO11 raw=2325  volts=1.874
20:08:37.924 > GPIO11 raw=3194  volts=2.574
20:08:38.924 > GPIO11 raw=4095  volts=3.300

So you can see the GPIO is correctly being read as analog, but the gpioViewer web UI only ever shows Digital pins.

My board is an ESP32-S3-DevKitC-1-N32R16V.

I searched but did not see any similar issues. Thoughts?

[thelastoutpostworkshop]

I will test that

[thelastoutpostworkshop]

What is the version of the Espressif board you are using and the GPIO library version ?

[JasonInRVA]

Thanks for the quick reply!
Per your request:

• Arduino-ESP32 core (framework-arduinoespressif32): 3.20017.241212
• GPIOViewer library: 1.6.3

Hardware:

• ESP32-S3-DevKitC-1-N32R16V

Full environment details from pio pkg list -e esp32-s3-devkitc-1 below in case it helps:

Resolving esp32-s3-devkitc-1 dependencies...
Platform espressif32 @ 6.12.0 (required: espressif32)
├── framework-arduinoespressif32 @ 3.20017.241212+sha.dcc1105b (required: platformio/framework-arduinoespressif32 @ ~3.20017.0)
├── tool-esptoolpy @ 2.40900.250804 (required: platformio/tool-esptoolpy @ ~2.40900.0)
├── tool-mkfatfs @ 2.0.1 (required: platformio/tool-mkfatfs @ ~2.0.0)
├── tool-mklittlefs @ 1.203.210628 (required: platformio/tool-mklittlefs @ ~1.203.0)
├── tool-mkspiffs @ 2.230.0 (required: platformio/tool-mkspiffs @ ~2.230.0)
├── toolchain-riscv32-esp @ 8.4.0+2021r2-patch5 (required: espressif/toolchain-riscv32-esp @ 8.4.0+2021r2-patch5)
└── toolchain-xtensa-esp32s3 @ 8.4.0+2021r2-patch5 (required: espressif/toolchain-xtensa-esp32s3 @ 8.4.0+2021r2-patch5)

Libraries
├── ESPAsyncWebServer @ 3.6.0+sha.ad3741d (required: git+https://github.com/me-no-dev/ESPAsyncWebServer.git)
│   └── AsyncTCP @ 3.4.7 (required: ESP32Async/AsyncTCP @ ^3.3.2)
└── GPIOViewer @ 1.6.3 (required: thelastoutpostworkshop/GPIOViewer @ ^1.0.7)
│   ├── Async TCP @ 3.1.4 (required: Async TCP)
│   └── ESP Async WebServer @ 3.0.6 (required: ESP Async WebServer)

[thelastoutpostworkshop]

It's like you are using an older version of the GPIOViewer library.
Because the v1.6.3 consider all pins by default as analog.
Check the version of the GPIOViewer library reported by the web application.
I have tested with your setup :

[JasonInRVA]

The GPIOViewer web UI reports:

• GPIOViewer v1.6.3
• Web application v2.1.6
• Arduino core version 2.0.17 (ESP32-S3)

So it looks like I am on the latest release of the GPIOViewer library, not an older version.

Here is the screenshot:

I can successfully get a pin labeled a PWM (output) by using it as such, but still not joy on analog.

[thelastoutpostworkshop]

But you are using the old ESP32 core v2.017, not 3.2.
When using a version of the ESP32 core 2.x, the GPIOViewer is limited because that version of the ESP32 has limited public APIs.
And I do not focus on it any more because Espressif is no longer working on it

[JasonInRVA]

Ok. Got it working.

This has been a very difficult issue to diagnose, with symptoms that seemed logically impossible. My build environment, configured with platform = espressif32@6.12.0, consistently showed in the build logs that it was compiling against the modern Arduino Core v3.x. However, the binary running on the device would always report Arduino Core v2.0.17. This discrepancy survived every possible troubleshooting step, including a full pio system prune, erasing and reinstalling my whole VSC/PIO stack, a complete flash erase of the chip, and creating a minimal test case in a brand new project folder.

The Solution: Bypassing PlatformIO

Once I abandoned the PlatformIO environment to see if the issue was with the hardware or the toolchain itself, I was able to get the demo working perfectly by using the Arduino IDE directly.

This suggests that the hardware is fine and that the root cause lies somewhere within PlatformIO's handling of my specific board.

Hardware Details:

The board is an ESP32-S3-DevKitC-1-N32R16V, which has 32MB of Octal Flash and 16MB of Octal PSRAM. This is a less common variant, which I believe may be key to the problem.

The Working Arduino IDE Configuration:

For anyone else with this specific board, these are the settings that worked perfectly in the Arduino IDE (v2.x) with the latest esp32 board package (v3.x.x):

• Board: ESP32S3 Dev Module Octal (WROOM2)
• Flash Size: 32MB (256Mb)
• Partition Scheme: 32M with LittleFS (or your preferred 32M scheme)
• PSRAM: OPI PSRAM

Conclusion and Hypothesis:

The fact that the Arduino IDE works flawlessly with a more specific board profile strongly suggests that the generic esp32-s3-devkitc-1 board definition in PlatformIO is either buggy or incomplete. It likely lacks the proper configuration to handle the N32R16V variant's 32MB/16MB Octal (OPI) memory, causing the build system to fall back to an old, incompatible configuration that results in the v2.0.17 binary being flashed.

This is likely not a "bug" in the traditional sense, but an oversight for an uncommon hardware variant. Hopefully, this information can help anyone else in my situation.

Thank you again for your help.