W5500 Start-Up Problem in relation with low temperature

Topkam · 2025-08-04T22:16:57+09:00

Hello Forum, I’m working with a custom board featuring a W5500 Ethernet chip and an STM32 microcontroller. The board has already been produced in large quantities and is running in various test applications. However, I’ve received feedback indicating that the board doesn’t start correctly at low temperatures. I’ve tested this issue and discovered that around 80% of the boards fail to start properly when the temperature drops below 0°C, even though the components are within their specified operating temperature range. Startup Sequence: --------------------- 1. Trigger PHY reset via Register PHYCFGR and RST Bit 2. Wait 200ms 3. Trigger SW Reset via MR (Mode Register) and RST Bit 4. Wait 500ms 5. Set SN_MR (Socket 0 Mode Register) to MACRAW Mode 6. Set SN_MR (Socket 0 Mode Register) Bits for MFEN, UCASTB, ND, BCASTB high 7. Read the Version Nr and SN_MR register and check if these registers deliver correct values. 8. If no correct values are delivered within 200ms I restart at point 1. --------------------- Issue Observations: When the error occurs, the microcontroller is functioning correctly, but I observed the following: Using a debugger, I found that the W5500 always returns a value of 0 via SPI, even when reading the VERSIONNR register, which should return 0x04. The oscillator (either TXP Corp 7M25000014 with 2x 18pF capacitors or HCI XC32M4-25.000-F10NLDW with 2x 10pF capacitors) seems to be working fine and delivering a correct signal. I used slightly different resistor and capacitor values in the design between the RJ45 connector and the W5500 compared to the recommended values in the hardware design guide. The RJ45 connector I’m using is the WIZNET RB1-125BAG1A. Once the board starts correctly, it runs without issues for days. When cooling down the hardware (to -8°C) and reconnecting the power, the initialization routine sometimes takes a few minutes to complete. During this time, the SPI interface only returns 0. After a certain period (which varies by hardware and test run), the initialization works correctly, and the device starts up as expected. I assume this is because the problematic area is warming up due to the room temperature. The issue persists whether or not the RJ45 cable is connected. Suspected Causes: I’m unsure if the problem is related to the oscillator or the incorrect resistor and capacitor values used in the design. Additionally, the traces between the RJ45 connector and the W5500 chip are not routed optimally and differ in length. Has anyone else encountered a similar issue? Could the problem be related to the oscillator or clock? Any advice would be greatly appreciated! Thanks in advance!

Ethernet Chips

by Topkam 2025-08-04 22:16

W5500
STM32

RE: W5500 Start-Up Problem in relation with low temperature

by WIZnet AI AI generated 2025-08-04 22:17

RE: RE: W5500 Start-Up Problem in relation with low temperature

by Topkam 2025-10-06 17:07