8GB USB Drive with AU6987HL-5

[MasterT]

I have an 8GB USB drive with the following controller:

ALCOR
AU6987HL-5
CA8C0M-10
1042T1953

And 2 memory chips with the folowing markings:

60072979
530002460A
M1QJ0856AJ1

They are recognized in PC-3000 flash like:

$ADD7949A - 7442ADD7HY27UBG8T2ATR [SDR] [Page 8640 bytes (16 sectors). Block Size: 4096 sectors (256 pages)] (x8)

I could read both chips OK but cannot find any ECC or XOR.

Stick looks like these but instead of 16gb mine has just 8gb:
http://www.pc-3000flash.com/solbase/sol ... 0&lang=rus
http://www.flash-extractor.com/library/ ... 24_4b__2x1

I also tried solution here but with no success:
http://www.pc-3000flash.com/solbase/sol ... 4&lang=rus

Can anybody help me? I can provide dumps.

Thanks.

[Roman_TS]

Hello MasterT,

Please contact ACELab Technical Support dpt. Our engineers will help you.
http://ts.acelaboratory.com/index.php?/Tickets/Submit

[AJ2008]

first step in solution you posted is join by byte. Try this then search for ECC or XOR, if nothing then swap order of joined bytes - if joined bytes are not in correct order, no result will be found

[MasterT]

I swaped order of joined bytes and it found ecc.

Still i have 354 mb of data with invalid ecc out of 8gb.

If i want to make reread sectors on invalid ecc sectors how do i proceed because now i have an 8gb map after joined bytes?

Does the software still know at this step each sector from which memory chip belong to make reread correct?

Thanks.

[Roman_TS]

It's impossible to do readout in case if you have two physical chips and ECC could be found only after Joining by Bytes.
Next step would be Page transformation (you can use ECC detection parameters for building new page structure) and probably Interleave. Then - image building with Translator AU.

[AJ2008]

Still i have 354 mb of data with invalid ecc out of 8gb.

This should be ok. Now you can follow solutions in solution center, or try to proceed with Roman instructions manually.

[MasterT]

If you have 2 chips say chip A And chip B.

Sectors in chips are like this:

A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 etc ...
B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 etc ...

After joining by bytes it looks like this: A1B1 A2B2 A3B3 A4B4 A5B5 A6B6 A7B7 A8B8 A9B9 A10B10 etc ...

So you do ECC and find that A7B7 is invalid.

You can make reread on diffrent voltages on sector A7 on chip A, on sector B7 on chip B, and compare results after join by bytes and ecc until you find the one with correct ECC.

Of course it cannot be done manually, but it can implemented in software.

[Roman_TS]

Source of the post MasterT

It's impossible. Readout is based on ECC codes, and till the time you will not find the working ECC codes, readout will be also unavailable. Splinting by bytes - is a physical operation but controller of Flash case works with the whole virtual image of data, where two chip results are joined.

If you want to increase the quality of your result, you can read both chips with Autoverification x3 - it will help to reduce the number of errors, and totally fix them after joining with ECC.

[MasterT]

What if you have 2 readers, one for Chip A and one for Chip B, that can read simultaneously, would it than be possible to reread based on ecc?

[Roman_TS]

Source of the post What if you have 2 readers, one for Chip A and one for Chip B, that can read simultaneously, would it than be possible to reread based on ecc?

Theoretical, if we make such extension - it will be possible. But nowadays it useless, because AU cases with ECC after Joining By bytes are very rare, and usually we can fix 99% errors with a common ECC. Your case is outstanding, and for some reason you have more errors than usually. That's why I'm asking about reading with auto-verification - it should help.

[MasterT]

I made another reading with Autoverification x11, result improved but still had 272mb (vs 354mb before). Maybe one or both chips have damaged areas.

Overall a good recovery case. On the first try i had 1.4GB good files, on the second try with autoverification i had 1.8GB good files.

Thank you for all your help.

Have a nice day.

[Amarbir[CDR-Labs]]

Great Post ,
Really Helped in learning stuff ,I request you roman to make more live cases in blog ,Flash stuff has really taken a backseat since few months

[AJ2008]

Source of the post Great Post ,
Really Helped in learning stuff ,I request you roman to make more live cases in blog ,Flash stuff has really taken a backseat since few months

what kind of information are you looking for? I have many case sample and documented cases and flash related guides, if you are looking for something specific maybe I have something you will find useful. What tools do you use for Flash recovery?

PM me if you would prefer not to disclose anything here