HDD GURU FORUMS

Hi, I have an OCZ ARC100 480GB to recover. Imaging fine until it hits 33GB then imaging halts. Controller is Indilinx IDX500M10-BC and there are 16 Toshiba TH58TEG8DDKBA8C NAND chips.

Anyone know of any issue already discovered for the symptoms?

Or if chip off is similar to the earlier Indilinx Barefoot chips such as the IDX110M00 and IDX400N00?

I see mention of AES256 support, is this native encryption or user through a utility that may not be implemented?

any help would be great, thanks

Nothing after that point, even in reverse?

Could we see the SMART report?

lcoughey wrote:Nothing after that point, even in reverse?

It hangs windows at that point. I don't have hardware tools for SATA DR.. But good point about reverse, using DMDE, it is at 154GB, no errors cloning in reverse. It detects and is accessible in windows, and I can access the file system fine until I touch certain files or folders. then windows hangs.

Of course the customer is mainly concerned oubt the 4.7GB .pst file. said pst copies merrily until 2.61GB then hangs. Of course, who would expect any different!

My guess is a bad nand chip or part of.

will see how far it gets going backwards... thanks for the idea.

fzabkar wrote:Could we see the SMART report?

Sure, here are 2, one from Crystal and the other the OCZ util:




Thanks guys

If you have use PC3K to clone the drive.

Thanks, but No I don't have it, I don't have any SATA DR tools, only flash.

I know PC3K has barefoot 3 tool and some solutions for various issues using technological mode.

I'm guessing it will reverse clone back to damaged chip and stop, and the pst file will be unrecoverable. not sure if PST files are only part copied if anything can be extracted from them.

Some time ago I helped a user recover data from an SSD with a similar problem. He played with the NAND supply voltage among other things. IIRC SMART reported some problems that pointed us in that direction.

Just out of curiosity I checked the numbers for attribute 0xAB.

User accessible capacity + Available OP Sector Count

= (480.1 GB) + (152 734 032 x 512 bytes)
= 519.96 GiB
= 512GiB + 8GiB
= (16 x 32GiB) + 8GiB

I would have thought that the sum should correspond to the total capacity of the NAND array but it appears that there is an additional 8GiB capacity (0.5 GiB per chip).

fzabkar wrote:Some time ago I helped a user recover data from an SSD with a similar problem. He played with the NAND supply voltage among other things. IIRC SMART reported some problems that pointed us in that direction.

Just out of curiosity I checked the numbers for attribute 0xAB.

User accessible capacity + Available OP Sector Count

= (480.1 GB) + (152 734 032 x 512 bytes)
= 519.96 GiB
= 512GiB + 8GiB
= (16 x 32GiB) + 8GiB

I would have thought that the sum should correspond to the total capacity of the NAND array but it appears that there is an additional 8GiB capacity (0.5 GiB per chip).

yes this drive is overprovisioned.

having no luck any other way, I de-soldered the first BGA132 NAND and read it. Well 1/2 of it, I guess. connected to D0_0 - D7_0 and CE_0_0, RB0_0, CE1_0, RB1,0 for 2x 8GB banks. I am supposing when I wire the D0_1 - D7_1 I will get the other 2 banks.

problem is I cant see any data / SA structure in the dump. Read it DDR, and looks right (no double bytes)

I know the first barefoot had join by byte, second didn't. This one could be AES as well but shouldn't I still see page structure?

The point I'm making is that the following relationship should hold true, AISI:

User accessible capacity + Available OP Sector Count = total capacity of NAND array

However, the numbers suggest that the total capacity of the NAND array is 512GiB + 8GiB rather than 512GiB. This in turn suggests that each chip has a capacity of 32.5GiB. Therefore the SMART attributes don't make sense.