ESP-32-POE-ISO: Problems using POE

IoTGSA · January 31, 2022, 03:57:15 PM

Hello,

We have several ESP32-POE-ISO rev. G1 that are totally bricked. After some research on the devices, we have found what we think could be the issue, but we do not know how to solve it or how it happens in the first place.

The boards are experiencing 3 states:
1. Programmed and working without problems. Can be reprogrammed. (We will ignore this ones)
2. Programmed and working, but cannot be reprogrammed.
3. Unprogrammable, and not working; totally bricked.

All the boards are planned to be powered through POE.

I will now explain the issues on number 2 and 3:

About number 2, accessing via

Code Select

idf.py monitor -p COMX -B 115200,
we have the following:

If we power them through POE, we got the following tension between 3.3 and GND:

However, it is not always fixed, as the tension fluctuates between 1.4 and 3.4.

As for the number 3, we get the famous "waiting for packet header" when we try to access through serial.

If we measure between 3.3 and GND, we get the following:

In this devices, we measured in U7 between LX and GND, and found that the voltage was 0.11 Volts.

We suspect that the POE is causing the U7 to break or missbehave. However, we are not really sure that is the real issue, or there is something else, and what is more important, how could we fix it.

Many thanks in advance.

LubOlimex · January 31, 2022, 04:53:23 PM

Board #2 is probably poor quality usb cable, bad solder joints or low power from USB connector as reported here: https://github.com/nkolban/esp32-snippets/issues/168

Board #3 has bigger problems than simply not starting - its PCB antenna seems to be melted.

IoTGSA · January 31, 2022, 08:09:17 PM

Boards number 1, 2 and 3 are all powered through PoE, and programmed using the same USB wire.

With time, board number 1 switches to the state of number 2, and then goes to state of number 3. Sometimes It goes directly from 1 to 3.

As for the antenna in number 3, It is just some epoxy used to fix the device to a box. The ESP32 in state 3 happens either with or without epoxy.

LubOlimex · February 01, 2022, 10:10:17 AM

Do you have any other power supply attached? Can you describe the powering around the time the fault occurs?

IoTGSA · February 01, 2022, 10:34:13 AM

Quote from: LubOlimex on February 01, 2022, 10:10:17 AMDo you have any other power supply attached? Can you describe the powering around the time the fault occurs?

The powering is only through POE, no other system is used.

We have 2 different types of PoE switches:
- First switch
- Second switch

We have checked both of them with a PoE checker and got a good result.

As for the faults, they always happen when you disconnect and connect the RJ45. Currently 10-15% of our devices are in state 3 due to this problem (around 40 boards so far).

IoTGSA · February 01, 2022, 11:50:36 AM

In the boards with state 3, the U7 is giving 0,07 V at the output, it seems like it is not properly converting the voltage from 5 to 3.3.

IoTGSA · February 01, 2022, 01:10:12 PM

Also, we are getting continuity between 3.3V and GND:
-

IoTGSA · February 02, 2022, 02:33:45 PM

Any clue on what could be happening @LubOlimex ?

LubOlimex · November 22, 2022, 10:16:08 AM

It should be something specific to your hardware setup or usage of the boards. We have kept our eyes on similar complaints but we got none so far. Hard to say how exactly this short-circuit between 3.3V and GND happens. It is hard to determine where exactly this appeared, is it 0 ohm short or more than 0 ohm?