Troubleshooting¶

Symptoms you may see while bringing up a TLE92466ED, their causes and the fix for each, cheapest check first.

Start with the error a call returns and, once Init() has passed, with GetAllFaults(): the error tells bus problems (no reply, bad replies) apart from chip state (wrong mode, a fault the chip reports).

Bring-up¶

Symptom	Cause	Fix
`Init()` returns `DeviceNotResponding`	MISO idle: no VIO, CSN not reaching the chip, MISO not connected	Measure VIO at pin 8; scope CSN and SO during `Init()`
`Init()` returns `FaultDetected`	Every reply is a critical-fault frame	VDD (4.5 V to 5.5 V at pin 29) first, then the clock source
`Init()` returns `WrongDeviceID`	`ICVID` is not `0xC1xx`: bits shifted by a wrong SPI mode, or another device on the same CSN	SPI mode 1; one CSN per device
`Init()` returns `ConfigurationError`	`OTP_VIRGIN`: the part's calibration OTP is not programmed	Use a production part
`CRCError` or `SPIFrameError` now and then	Frames not exactly 32 bits, CSN high time under 600 ns, clock too fast for the wiring	One 4-byte transfer per CSN; lower SCK to 1 MHz; shorten the wires
Wrong values on a shared bus	Another device's frame between the command and its reply	Hold the bus lock for the whole `TransferMulti()` (Porting)

Channels¶

Symptom	Cause	Fix
`EnableChannel()` returns `WrongMode`	Still in Config Mode	`EnterMissionMode()` first
`ConfigureChannel()` or `SetChannelMode()` returns `WrongMode`	In Mission Mode	`EnterConfigMode()`, configure, then go back
Channel enabled, no current	EN low (all channels off), setpoint 0, or the chip dropped to Config Mode	`Enable()`; `IsMissionMode()`; `GetAllFaults()` for supply or watchdog faults
Channel drops out after a while	SPI watchdog expired: the chip returns to Config Mode	Reload it faster (Diagnostics) or leave it off
Current far from the setpoint	No freewheeling diode at FWDx, setpoint above the specified 1.5 A, `parallel_mode` missing on a pair	Hardware, Current control
`SetCurrentSetpoint()` returns `InvalidParameter`	Above 2000 mA (4000 mA in parallel)	Lower it; pass `parallel_mode = true` for a pair
A parallel pair drives half the current	The slave was configured instead of the master	Configure and drive CH0, CH1 or CH4 only

Feedback¶

Symptom	Cause	Fix
`GetAverageCurrent()` returns `FeedbackNotReady`	`DITHER_CLK_DIV` still 0, so the measurement period is 0	`ConfigureDither()` or `ConfigureDitherClock()` on that channel
Feedback frozen at an old value	`FB_FRZ` left set by an interrupted snapshot	`ReleaseFeedbackFreeze()`
`SampleChannelCurrents()` sets `rejected_chains`	Replies arrived out of order on the bus	Bus integrity: shared-bus lock, wiring, SCK speed
`ReadChannelFeedback()` returns `TimeoutError`	No feedback update within the timeout	Longer timeout, or a faster dither (shorter measurement period)

Faults¶

Symptom	Cause	Fix
FAULTN low before `Init()`	RESN low, or latched power-on flags	`Init()` releases RESN and clears the power-on flags; then read `GetAllFaults()`
A fault returns right after `ClearFaults()`	The cause is still there	Read `GetAllFaults()`; the flag names the cause
`open_load` on a fitted coil	Open-load threshold too high for the current, or the load is wired to the wrong LOAD pin	Lower `open_load_threshold`; check the pin table
`overcurrent` on a channel	A short, or a load below the coil resistance it was sized for	Measure the load; the protection trips between 2.05 A and 6 A
`REG_ECC_ERR` comes back after every clear	Known on some parts; `Init()` reports it and continues	No action if the channels work

Debugging tips¶

Route Log() in your port to a console: the driver logs every mode change, each failed attempt and the decoded critical-fault byte.
PrintAllFaults() prints every fault flag; IsFault(true) does it when FAULTN is low.
ReadRegister(CentralReg::ICVID) is the cheapest end-to-end read; ReadAllSupplyVoltages() the cheapest 22-bit one.