Issue when init lwip_init using LAN8742 and STM32H723 custom board RMII

DavidWahlberg · ‎2024-05-14

I am struggling to get ethernet to work on my custom PCB H723. I have succeeded develop an ethernet LWIP application on H723 NUCLEO board though.

I have more or less copied the schematic of the NUCLEO board, however when running the LWIP_init function it ends in the while(1) error handler loop. I am also using the LAN8742 IC with RMII, 25MHz external osci.

When i power my board the LED2 is always lighting up. When I connect a ethernet cable LED1 will light and blink for activity.

Anyone have experienced similar? Or have any thoughts or go-to's? I am out of ideas.

Friendly regards,

David

Muhammed Güler · ‎2024-05-14

Are differential paths with 50 ohm impedance in PCB design? Are the configuration resistors on the LEDs and other pins at the correct value? Are they connected to the correct line?
It would be good to see at least the ethernet-related part of your circuit to find a solution.
Have you checked your solders? A poorly soldered part may be the cause of your problems.
You can start debugging by measuring the PHY oscillator. Most likely, you will not see any oscillation on the lines that should be 25 or 50 Mhz.
I had an issue with LED polarity on the first board I made. I later realized that my connections required connecting one LED Active High and the other LED Active Low.
I have experienced similar problems in mass production, usually due to badly soldered parts.
When I looked at the parts that looked like they were soldered from above, I realized that they were not soldered when I looked at them from different angles.

View solution in original post

Muhammed Güler · ‎2024-05-14

Are differential paths with 50 ohm impedance in PCB design? Are the configuration resistors on the LEDs and other pins at the correct value? Are they connected to the correct line?
It would be good to see at least the ethernet-related part of your circuit to find a solution.
Have you checked your solders? A poorly soldered part may be the cause of your problems.
You can start debugging by measuring the PHY oscillator. Most likely, you will not see any oscillation on the lines that should be 25 or 50 Mhz.
I had an issue with LED polarity on the first board I made. I later realized that my connections required connecting one LED Active High and the other LED Active Low.
I have experienced similar problems in mass production, usually due to badly soldered parts.
When I looked at the parts that looked like they were soldered from above, I realized that they were not soldered when I looked at them from different angles.