Phison \ Toshiba TF17G2JAJA PS2251-03-Q

[HaQue]

Hi all, Just did a long and detailed post which dissapeared into the ether.. if it shows up and doubles up, sorry!

I have a TDK Gold USB3 16Gb flash drive I am trying to recover.

The chip is labelled TF17G2JAJA, and after googling, this page http://www.usbdev.ru/databases/fidmemory/ says it is a PHISON rebranded Toshiba chip.

The ID is 98 3A A8 92, and corresponds to these 3 on that page:

Phison TF17G2JAJA Toshiba TLC-8K 98 3A A8 92 76 D7
Phison TF58G2JAJA Toshiba TH58TEG8T2JBA4C TLC-8K 98 3A A8 92 76 D7
Phison TJ79G2JAJA Toshiba TLC-8K 98 3A A8 92 76 D7

I can dump it with SC NR using the config for FZBWTHS58TVG8T2JTA20 but I have to change the banks from 2 to 1. No Idea if I should be doing that but the ID is the same, though the above page has a 6 byte ID.

Anyone have any ideas for this chip? Apologies for spamming all the forums, but the customer wants the Flash Drive back and this is a last ditch effort for my own benifit.

[arvika]

Yes, use config from NR and change banks from 2 to 1. Other parameters do not change.

[HaQue]

Many thanks Arvika, I was able to dump the chip, though the BB file was 2.26Gb or so. I am guessing that is not good.

PS2251-V 1151_1_1143 Xor 0561_86 fits best.

After Cut Bad Bytes, mix found:

Block Pair
Block 0x204000
Page 0x2000

after this Mix not found. Played around with a few things based on sililar cases with no result, and ended up saving image and using GDB.

I have a partial recovery, a recovery of sorts by saving the image and using getdataback on it. It detected it as FAT12 filesystem, but recovered files ignoring FAT.
271 seemingly good Word documents, 65 corrupted Word docs, 12 good .jpg

I have no idea how much was on the drive.

Customer is about to pull pin, so looks like thats it for this case.

cheers

[Corsari]

Hello I have this chip
TF58G2JAJA.

PIN 1 is marked on the "wrong" corner and unfortunately this chip has a lot of broken pins due to a drop/fall

So we had to do a surgery operation to sold it on adapter.

Adapter (which is not the best Ace product) + weird pin position on the chip, makes it hard to guess which is it correct PIN 1 (the read chip ID attempts failed)

Having the string TF58G2JAJA oriented correctly (readable) the pin is in the low left corner of the chip instead than on top left corner (as expected)

So... which is it the correct PIN 1

The other two corners (top left and low right) have two identical circular depressions.

Thank you for any hint

Cor

Attachments

The deep small pin (hard to see) is in lower left

[dick]

@Corsari

Is it possible the chip was fitted upside down? The written identification would be on the wrong side and not be viewable in your reader. That way pin one would be in the expected position. I have seen this configuration in the past.

[Corsari]

WOW!
Interesting theory...
The pendrive was almost disrupted and the chip got's completely detached from it!!!

But you may believe it or not, no one single pin was ripped off and remain on the body of the circuit.

All the pins, completely bend, were still attached to the chip... but bended sooo much.

So you do mean that the chip marking TF58etcetc is burned at bottom of the chip instead than on top as usual? Or what?

Pins were bended but their original shape has been figured out (looks like correctly)

Thank you for your support and appreciated replies

[fzabkar]

Pin 1 should be obvious from the circuit.

[dick]

So you do mean that the chip marking TF58etcetc is burned at bottom of the chip instead than on top as usual? Or what?

Yes that's right.

There are a couple of different scenarios here:

1) The chip has a blank top face without markings or the markings were removed.

2) The chip was mounted directly on top of a second chip. In a kind of piggy back style? Not sure how the connections would work but it might mean a chip is missing?

Sorry this is not my area of expertise, I'm just mentioning these possibilities in case it helps you to find a solution.

[Corsari]

@ fazbakar! Nice to read you! It was some time now. The customer carried here only part of the circuit and chip was completely detached from it.

mmmmh...
Now that you write this...
Aren't the others two circular depressions typical of the tsop48 chips' bottoms?

Could be. The pendrive was a 32GB and according to what is written in the previous posts of this thread looks like it was a single 32GB chip.

Last edited by Corsari on April 26th, 2015, 13:44, edited 2 times in total.

[arvika]

Is it really hard to take the measurement and other chip as reference and compare?

By the way, this type of chip sometimes is not recognize by Ace reader.

[Corsari]

Hello arvika, I didn't worked personally on this chip, it has been processed by my colleague and today I was at the lab to read it, but it has been soldered on the adapter.

Comparing require to desolder it, it can be done tomorrow.

[Corsari]

Generally speaking, I've written in this thread also because in one of the posts above it is mentioned the same chip with the precisation that the chip is a toshiba rebranded by Phison

So it would be interesting to ear from posters above about any particularity of it.

[fzabkar]

I don't know about NAND flash, but some chips are produced with both a "normal" pinout and a "reverse" pinout.

For example, here is a Toshiba TC551001CF SRAM:
http://aoisonic.com/shop/pdfdata/tc551001cft.pdf

[HaQue]

I have a few chips with the pin 1 mark on the "wrong" side. The chips are still inserted in reader "as normal", not upside down, so pin 1 mark is at pin24 position. example http://flash-extractor.com/forum/viewtopic.php?t=6922

BTW, "JAJA": , I believe first J denotes WL chip and second J denotes DDR chip.

GAJA would be standard data rate, DDR
GAHA would be Standard and SDR,
JAHA would be WL chip and SDR.

I haven't been through my research catalogue but I think this is how it goes.

These chips may need additional VCC supplied to other pins

they should also ID ok, so check each pin carefully

The other round circles(larger) are just marks from the mold of the package, and sometimes a factory mark for batch ID used internally