ESP doesn't boot when it is powered via the 5V input

[GeorgeFlorian1]

I am continuing an issue that I have opened here: https://github.com/espressif/esp-idf/issues/7127

Development Kit: ESP32 Wemos
Module or chip used: ESP-WEOOM-32
IDF version: v4.2.1
Build System: [Make|CMake|idf.py]
Compiler version: xtensa-esp32-elf-gcc (crosstool-NG esp-2020r3) 8.4.0
Operating System: Linux
Using an IDE?: VS Code 1.56
Power Supply: External 5V

I am trying to implement this project: https://github.com/DroneBridge/ESP32 and I can't connect to the ESP's AP mode.
I can see the AP (DroneBridge ESP32) in the WiFi list but it doesn't connect to it. I've gotten two different error messages:

Couldn't authenticate connection
Couldn't get IP address

The project basically puts the ESP32 in AP Mode and makes it a bridge between the user and the drone.
The ESP is powered by a drones battery at 5V.
If I power it via the USB,it works.
I have measured and the power input is 5V. I have also tried the other 5V input on the ESP board but to no avail.

esp wemos.jpg (129.78 KiB) Viewed 23404 times

Ignore the RX and TX wires on the ESP32 WEMOS as they are not connected to anything.

I have switched to a new ESP32: the DevKit V4 Wroover-B . The red LED is on, but there is no AP listed. If I hit the EN button on the board the AP shows up but I can't connect to it, giving me the same error messages as the previous ESP.

second_esp.jpg (125.09 KiB) Viewed 23404 times

This one also works only when connected to USB.

I've got a multi-meter and it measures 5.05V on the 5V pin and 3.3V on the 3V3 pin.

[ESP_Sprite]

On the boot issue: Does it boot when you momentarily connect EN to GND?

[WiFive]

A drone battery should have enough current to deal with esp32 current spikes in WiFi mode but isn't it unusual to have a 5v battery voltage? Shouldn't it be a multiple of (3.7-4.2)? Is it really the main battery output?

[GeorgeFlorian1]

On the boot issue: Does it boot when you momentarily connect EN to GND?

Yes, it does. And after it boots I can also connect to EPS's AP; so it works as intended, but only after briefly connecting EN to GND.

A drone battery should have enough current to deal with esp32 current spikes in WiFi mode but isn't it unusual to have a 5v battery voltage? Shouldn't it be a multiple of (3.7-4.2)? Is it really the main battery output?

I've got a PDB with 12V and 5V outputs.

[ESP_Sprite]

Likely solution, in that case: wire an 10uF capacitor between your EN and GND pins. The power supply is likely coming up slower than the duration of the reset signal as generated by the on-pcb RC network, leading to issues.

[GeorgeFlorian1]

Likely solution, in that case: wire an 10uF capacitor between your EN and GND pins. The power supply is likely coming up slower than the duration of the reset signal as generated by the on-pcb RC network, leading to issues.

I've even tried powerin the ESP via the 3V3 pin using a Power Step-Down from 5V to 3V3 and it still doesn' boot properly. It ONLY boots after I press the EN button.
All I have left is trying your solution and adding an 10uF capacitor between the EN and GND pins.

[GeorgeFlorian1]

Likely solution, in that case: wire an 10uF capacitor between your EN and GND pins. The power supply is likely coming up slower than the duration of the reset signal as generated by the on-pcb RC network, leading to issues.

What type of capacitor do I need to use ?

[ESP_Sprite]

What type of capacitor do I need to use ?

Any type should work.

[GeorgeFlorian1]

What type of capacitor do I need to use ?

Any type should work.

I've soldered a 10uF capacitor on the ESP32: the DevKit V4 Wroover-B between EN and GND and the ESP finally boots.
I have tried powering the ESP using 5V and then 3V3. In both cases the AP is visible but I can't connect to it.

I still get the same two error messages:

Couldn't authenticate connection
Couldn't get IP address

Only after I press the EN button can I connect to the AP.

[GeorgeFlorian1]

Likely solution, in that case: wire an 10uF capacitor between your EN and GND pins. The power supply is likely coming up slower than the duration of the reset signal as generated by the on-pcb RC network, leading to issues.

I did that. The ESP boots and the AP is visible, but I can't connect to ESP's AP.