The DACx300x provides a feature to save
a few registers into the NVM when the FAULT-DUMP bit is triggered or
when the GPIO mapped to fault-dump is triggered (see also Table 7-18). This feature is useful in system-level fault management to
capture the state of the device or system just before a fault is
triggered, and to allow diagnosis after the fault has occurred. The
registers saved when fault-dump is triggered, are:
- CMP-STATUS[7:0]
- DAC-0-DATA[15:8]
- DAC-1-DATA[15:8]
Note: When the
fault-dump cycle is in progress, any change in the data can
corrupt the final outcome. Make sure the comparator and the
DAC codes are stable during the NVM write cycle.
Table 7-3 shows the storage format of the registers in the NVM.
Table 7-3 Fault-Dump NVM
Storage Format
| NVM ROWS |
B31-B24 |
B23-B16 |
B15-B8 |
B7-B0 |
| Row1 |
CMP-STATUS[7:0] |
Don't
care |
| Row2 |
DAC-1-DATA[15:8] |
Don't
care |
DAC-1-DATA[15:8] |
The data captured in the
NVM after the fault dump can be read in a specific sequence:
- Set the
EE-READ-ADDR bit to 0b in the COMMON-CONFIG
register, to select row1 of the NVM.
- Trigger
the read of the selected NVM row by writing 1 to the
READ-ONE-TRIG in the COMMON-TRIGGER register; this
bit autoresets. This action copies that data from
the selected NVM row to SRAM addresses 0x9D (LSB 16
bits from the NVM) and 0x9E (MSB 16 bits from the
NVM).
- To read
the SRAM data:
- Write 0x009D to the SRAM-CONFIG register.
- Read the data from the SRAM-DATA register to get
the LSB 16 bits.
- Write 0x009E to the SRAM-CONFIG register.
- Read the data from the SRAM-DATA register again
to get the MSB bits.
- Set the
EE-READ-ADDR bit to 1b in the COMMON-CONFIG
register, to select row2 of the NVM. Repeat steps 2
and 3.