pimoroni W5500 Ethernet Co-Processor Breakout Board

AI context

Dense context for pasting into a model prompt — addresses, pinout, voltages, tolerances and the vendor's reference circuit.

BOARD: Adafruit product 6348, the Wiz5500 Ethernet Co-Processor Breakout Board - a WIZnet W5500 hardwired TCP/IP Ethernet controller in a 48-pin LQFP (7 x 7 mm, 0.5 mm pitch) with 32 KB of on-chip TX/RX buffer memory, an embedded 10BaseT/100BaseTX PHY and an SPI host interface. The board adds an RJ45 magjack (PulseJack RB1-125BAG1A) with link and activity LEDs, a 25 MHz crystal, an AP2112K-3.3 regulator, a 74HC4050 hex buffer for level-shifting the host's SPI down to 3.3 V and a 74AHC1G125 buffer on MISO, plus three mode solder jumpers. Pimoroni resells it as SKU ADA6348 at 15.50 GBP as measured; Adafruit is the maker.
ADDRESSES: none in the I2C sense - this is an SPI peripheral, not an addressed device. Its network identity (MAC, IP) is whatever DHCP or the host library assigns; nothing about it is fixed on the board.
PINOUT: 9 pads on JP1, and the order is not the usual one - the SPI group is clock first. Pad 1 = VIN (power in, 3-5 V), pad 2 = 3.3V (regulator output, up to 100 mA), pad 3 = GND, pad 4 = SCK, pad 5 = MISO (fab net MISO_MCU), pad 6 = MOSI, pad 7 = CS, pad 8 = IRQ, pad 9 = RST. Labels are Adafruit's published pin names; the fab file's nets agree, calling the MISO return MISO_MCU because it comes back through the buffer.
SUPPLY: VIN takes 3 V to 5 V and U2 (AP2112K-3.3) makes the 3.3 V rail the W5500 runs on. The W5500 itself is a 3.3 V part: supply 2.97 V to 3.63 V, absolute maximum -0.5 V to +4.6 V on VDD and -0.5 V to +6 V on any input, which is why its inputs tolerate 5 V logic. Operating temperature -40 to +85 degrees C.
LOGIC LEVEL: 3 to 5 V on the host side. Everything the host drives (SCK, MOSI, CS, RST) passes through U1 first, and MISO returns through IC1, so a 5 V Arduino and a 3.3 V Feather both work without extra parts. The W5500's own VIH is 2.0 V minimum and VIL 0.8 V maximum.
REFERENCE CIRCUIT: as built, from the fab file. Crystal Y1 is 25 MHz with C1 and C2 of 15 pF and R2 1 MOhm across it. R1 12.4 kOhm sets the PHY bias on EXRES1 (pin 10) to analog ground. The magnetics side has 49.9 Ohm terminators R4 and R9 on TXP/TXN and R10 and R11 on RXP/RXN, C3 and C4 of 6.8 nF on RX+, RX- to the jack's receive centre taps, C5 10 nF on the RXP/RXN termination node, and R12 10 Ohm feeding the jack's centre tap (X1 pin 3) with C6 10 nF as its bypass. Decoupling is C7 to C10 and C13 and C15 at 0.1 uF plus C11, C12, C14 and C16 at 10 uF, with ferrites FB1 and FB2 separating the analog supply and analog ground from the digital ones. Pull-ups: R3 10 kOhm on IRQ, R8 10 kOhm on RST and R15 10 kOhm on CS, all to the relevant rail; R5, R6 and R7 are 10 kOhm pull-ups on PMODE0, PMODE1 and PMODE2 with SJ1, SJ2 and SJ3 in series to let each be tied high.
TRAPS: RST must be held low for at least 500 us to reset the part. IRQ is active low. MISO does not come straight from the chip - it comes back through IC1, a 74AHC1G125 whose output enable is the level-shifted CS, so a host that drives CS from a pin that is not the one the board's shifter sees will read garbage. The three mode jumpers are soldered on the back and are labelled M2, M1 and M0 on the silk while the fab file binds them to PMODE2, PMODE1 and PMODE0; all three open is the default 10BT half-duplex mode with auto-negotiation disabled, so leaving them alone leaves the link at 10 Mbit until a jumper is set. Adafruit says the mode table is also in the W5500 datasheet on page 10. There is no PoE on this board; a splitter is needed. With the yellow LED off there is no link speed and with the green LED off there is no connection - it is a diagnostic, not decoration.