Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
20 years ago

Nios II Flash Problems

We have a Nios II board design that has been populated with two types of CFI Flash chips. Toshiba and the ExcelSemi equivalent? 16Mbit CFI Flash devices. Checking the data sheets all pins are the same and timing is about the same on the two chips. The boards with an ExcelSemi chip run flash memory tests all day without errors using the Altera CFI flash drivers and the Altera memory test example code. On multiple boards, the Toshiba Flash chips always return an error code -5 when we try to write or erase them and no data seems to change. It seems like we always read the same "random" data back from all of the Toshiba Flash chips. Any ideas or suggestions? We tried slowing down the flash bus cycles and nothing changed.

2 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    20 years ago

    From your posts it's not clear whether all Toshiba's don't work or only some of them. In case they all don't work: have you tried stepping through the initialization of the CFI flash drivers? I have a small incompatibility with my chips although it still works. When I stepped through the source I noticed that the assumption of AMD or Intel chips is being made and even then not all AMD chips behave the way the driver expects. Try setting a break point in in alt_sys_init where it calls the flash driver init here:

    ALTERA_AVALON_CFI_FLASH_INIT( FLASH, flash );

    then step through that code and check whether the chip behaves as the source code expects.

    Andrew
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    20 years ago

    They all do not work. Thanks will give it a try. I assumed the CFI Flash standard chips would all be the same in software, but I guess not!