Interesting supurius non-zero data on brand new WD10EARS



i had received my WD10EARS yestarday, and perfromed some initial diagnostic.

In SMART everything is OK, all statistic essentially zero (1 Hours on Power, 4 load cycles, 2 retract, temperature 38, rest zero). Performed Conveyance test, completed in few minutes without error.

Then i was courious, and becuase I know WD hard drive comes with zeros on the whole surface  of the disk if performed check, to if there indeed are only zeros. I have not written any single byte to this disk, even no partitions was created or MBR.

Well, i used hexdump (which detects repetitive blocks) command on Linux, and dd to recheck.

booted ubuntu from live cd, and told:

sudo hexdump /dev/sda

This will output two lines: first 0000000 00 00 …  …

and second one: * (which means that hexdump detected next 16 bytes which is exactly the same as previous, so again 0x00 bytes.

After few hours on about 51 GB position i started to new lines, and of course they contained non-zeros, after few megabytes, again i saw only zeros. Then at about  59 GB i again saw about 1 MB of non-zero data. Then after 10 hours, again at position 703 GB i saw another 1-2 megabytes of data.

So, 99.99% od data looks ok, as zeros (0x00)

This data looked like noise. It consisted mostly of ‘0xFF’ (all bits set), but i also saw another character often, for example “@” (0x40), or 0x04,  0x01, 0x10, 0x08, 0x80, and few other rererly. It also happened in one block that it contained mostly 0x00, but with sparse occurrences of “@” and some other characters. Then intermixed long block of mostly 0xFF and in it some noise with other characters in it also.

So, it is probably nothing really important, but this pattern and that it only occurs on about 0.1% of surface  made me even more curious what is going on.

I will save this blocks for further analysis. Then rewrite whole disks with zeros, and then reread it again to check if there is only zero.

Any body want to help in analysis, or  have or is planing to buy new WD drive? Please perform “hexdump” check without


There is also very high freqency of dublet “C#” (0x42 0x23) in some noisy block at 51 GB.

Will dump this data and share with you.