Typical bad block counts

[Josh Boyer]

On Wed, 2007-02-21 at 10:15 -0500, Dave Kelly wrote:
> On a 1Gbit single level nand (ST-micro), I have 10 chips all with less
> than 6 blocks that are bad originally, a few with none. 
> 
> The chips all bit flip excessively in a few blocks.
> 
> I am suspicious that somebody erased the original marks. The few blocks
> with 0 for the bad block mark all are stuck on 0 for the entire block,
> so could never have been erased anyway.
> 
> What is a reasonable number for bad blocks on this size chip as it 
> arrives from the mfg?

The datasheet for those chips says up to 20 blocks can be bad.  That
includes both Bad blocks at MFG time, and bad blocks that can develop
later on.

josh