Troubleshooting¶

Symptoms you may see while bringing up an AT25040B, their causes and the fix for each.

Start with LastRdsr() after a failed Probe(): it tells a dead bus (0xFF), a looped-back bus (0x05) and a working chip apart without a logic analyser.

Symptom	Cause and fix
`Probe()` returns `Absent`, `LastRdsr()` is `0xFF`	Nothing drives MISO: wiring, CS on the wrong pin, no VCC. Fix: Check SO (pin 2) to MISO, CS (pin 1), VCC (pin 8)
Same, right after an MCU reset	The chip is finishing a write cycle the old firmware started: all status bits read 1. Fix: Call `WaitReady()` before `Probe()`
`Probe()` returns `Absent`, `LastRdsr()` is `0x05`	MISO echoes MOSI one byte late. Fix: Look for a short between SI and SO, or a missing device on that CS
`Probe()` returns `Absent`, the two reads differ	Bus contention or noise. Fix: Another device drives MISO while this CS is low; lower the clock
Calls return `Busy`	RDSR shows a write cycle in progress: one started outside the driver (an MCU reset during a write, a WRITE or WRSR sent through your port). Fix: `WaitReady()`, then retry
`Read()` and the writes return `Busy`, `Probe()` returns `Absent` with `0xFF`	Nothing drives MISO, which then reads RDY = 1. Fix: Check SO (pin 2), CS (pin 1), VCC
`SetBlockProtection()` returns `Ok`, `GetBlockProtection()` shows the old level	WP is low: the chip ignores WRSR. Fix: WP (pin 3) high
Every call returns `Timeout`	Your port's `transfer()` returned `false`. Fix: Bus not initialised, or the bus lock timed out
`WritePage()` returns `Ok`, data unchanged	WP low or the address is block-protected: the chip ignores the write. Fix: WP (pin 3) high; `ReadStatus()` BP bits
`WritePage()` or `WaitReady()` returns `Timeout`, other calls work	RDY still 1 after 10 ms: MISO reads `0xFF` (floating), or the chip is damaged. Fix: Check SO (pin 2) and VCC; `LastRdsr()` shows the last status byte
`WritePage()` returns `Arg`	More than 8 bytes, or the write crosses an 8-byte page. Fix: Use `Write()`, which splits at page boundaries (API guide)
`Read()` returns `Arg`	Null buffer, zero length, or `addr + len` past `0x200`. Fix: Keep the range inside the array
Data at `0x100` and up appears at `0x000`	An AT25010B or AT25020B: no A8. Fix: Check the part marking; keep to its size
Reads fine, writes fail at high clock	Clock above the limit for your VCC. Fix: 5 MHz at 1.8 V, 10 MHz at 2.5 V, 20 MHz at 4.5 V
Random failures, worse with long wires	HOLD floating. Fix: Tie HOLD (pin 7) high