Lexar 32GB USB2 Flash drive overheated, now "Dead"

[inflex]

Client bought to me a USB2 Lexar 32GB flash drive, plug it in, the controller gets warm, nothing much else happens.

No visible shorts, photographed it in macro in 8 different ways and nothing profoundly obvious. Only telltale signs were that the case was deformed and the 2R2 resistor is discoloured.

I requested the client purchase an exact replica of the Lexar 32GB, unfortunately they couldn't find one *sigh*. Anyhow, anyone want to throw their experience/options my way?

I've recovered USB sticks in the past, but they've always been physical / mechanical faults requiring some fancy soldering and some part replacements - but this is the first "physically fine, operationally dead" stick.

Thanks in advance.

Attachments

side 2

side 1

[arvika]

If controller is hot, probably is broken. The best idea is read dumps from nands and assembly the image.

[inflex]

If controller is hot, probably is broken. The best idea is read dumps from nands and assembly the image.

Thanks. Would there be any sense in alternatively replacing the controller chip?

[HaQue]

Probably not. The controller is written with firmware. The manufacturers write configure the firmware, and then write it to the device. Especially in the case of SM controllers I would not like the chances. The new controller has a high chance of writing to the NAND as soon as you power it up. There goes customer DATA.

The controller getting hot could be a number of things.

I am in Oz as well, I can read the NAND for you. Can PM me if you like.

I possibly have one of these as well

BTW, a few tips:

get a light and light up the PCB. Then get something fairly small like a match or hairpin and prop up one side to angle it slightly. This will shine the light across the chip markings a bit. play around with it till you can light up the chip markings. It is critical to know both the controller number and NAND markings - all of the numbers. SM3257EN Q AA is different to SM3257EN L AA is different to SM3257EN T AA , and recovery is different for each

cant read either numbers. If you cant get good photos, look under magnifying glass and record them in your post.

Important if you are going to recover data:
don't do any formatting, trying to read in many computers, don't do "tweezer tricks" you see on the net, don't flow solder across "fuses" as seen in blog posts.

The NAND in a lot of cases holds the data okay even when everything else is fubar. for good DR, need to keep that NAND safe.

now edit after your next post:)

There is express post...

Do you have DR software? Even if you can read it, you need DR software to put back the mix of data that the controller did. Normally there are multiple chips with multiple banks and the data is sliced up like a puzzle, and also XORed . You need to be able to put it back together which is EXTREMELY difficult by hand.

Last edited by HaQue on August 27th, 2013, 6:44, edited 2 times in total.

[inflex]

hi HaQue, thanks for the clarification, so that's even if it was another Lexar 32GB V10 drive as per this one?

Client has a time window of until Friday, so that basically throws everything out of the realm of possible.

I've ordered myself a TSOP48 test&burn socket... should have had one months ago since I get phones in here that are dead and need the same sort of thing.

[HaQue]

edited above post

Also Kmart have a large supply of Lexar and I think they just moved to the V10 series, so might be a possible source if you want to try the controller replacement. I wouldn't recommend it, though

Last edited by HaQue on August 27th, 2013, 7:04, edited 1 time in total.

[inflex]

get a light and light up the PCB. Then get something fairly small like a match or hairpin and prop up one side to angle it slightly. This will shine the light across the chip markings a bit. play around with it till you can light up the chip markings. It is critical to know both the controller number and NAND markings - all of the numbers. SM3257EN Q AA is different to SM3257EN L AA is different to SM3257EN T AA , and recovery is different for each

New photos attached.

Important if you are going to recover data:
don't do any formatting, trying to read in many computers, don't do "tweezer tricks" you see on the net, don't flow solder across "fuses" as seen in blog posts.

By personal policy I'm very much a CYA/Do-no-harm sort of person, takes longer a lot of the time, but at least the original data / device stays intact for another person to use if required.

Do you have DR software? Even if you can read it, you need DR software to put back the mix of data that the controller did. Normally there are multiple chips with multiple banks and the data is sliced up like a puzzle, and also XORed . You need to be able to put it back together which is EXTREMELY difficult by hand.

Not likely for NAND Flash recovery. Haven't found anything [yet] for use with non-Windows systems ( Linux etc ). I was going to actually consider it a bit of a large project to write something to do so, though obviously that would not be done by Friday

Attachments

Memory

controller

[HaQue]

Nice photos, well done

I wouldn't think you would write something as almost every single device is different and needs to be researched. I had the same idea till I knew what was involved.

The DR software is something like Flash Extractor from Soft Center or ACE PC-3000 SSD Edition
http://flash-extractor.com/
http://www.pc-3000flash.com

No Rstudio, Winhex or GDB is gonna help

BTW if you try a controller swap, it is critical that the donor drive has the same PCB and exactly the same Memory chip. The firmware is tailored for the exact memory chip. Telling you this because the SM3255EN Q AA is put with a large variety of NANDS.

[inflex]

Nice photos, well done

I wouldn't think you would write something as almost every single device is different and needs to be researched. I had the same idea till I knew what was involved.

Ag damnit... don't you just love it when that sort of spanner gets thrown in your works. Was feeling on a high too because I've been doing some great work recently with improving GNU ddrescue. Still, could always be a new passion in life *cough*. ( What irks me is that almost all the stuff is Windows-centric, I much prefer doing DR in a Unix style environment, be it BSD*/Linux... okay maybe not Solaris)

BTW if you try a controller swap, it is critical that the donor drive has the same PCB and exactly the same Memory chip. The firmware is tailored for the exact memory chip. Telling you this because the SM3255EN Q AA is put with a large variety of NANDS.

Many thanks.

[fzabkar]

Is that a blister on component D1?

What is the resistance between the 2R2 resistor and ground?

[Touchclarity]

Is that a blister on component D1?

I think that's the power / activity LED Franc.....

[fzabkar]

Is that a blister on component D1?

I think that's the power / activity LED Franc.....

Doh! <smacks forehead>

[inflex]

Is that a blister on component D1?

What is the resistance between the 2R2 resistor and ground?

2R2 = 2.35R ( but that's within tolerance of the meter, since it's not a dedicated low ohms meter )

To ground is a little more complex as I imagine there's a few conductions going on, but it ranges from 530R ~ 4MR

[HaQue]

Whats the plan with this one? If you dont want to send it here, maybe you can find some DR Company in NSW with a SC Tool as this definately supports it. Or an ACE for that matter.

[fzabkar]

Is that a blister on component D1?

What is the resistance between the 2R2 resistor and ground?

2R2 = 2.35R ( but that's within tolerance of the meter, since it's not a dedicated low ohms meter )

To ground is a little more complex as I imagine there's a few conductions going on, but it ranges from 530R ~ 4MR

I don't know much about pen drives (obviously), but I would think that you could check the resistance across the flash supply rail, ie between pins 36 and 37, and 12 and 13.

http://tabreview.ru/content/pdf_docs/Mi ... 8CBAAA.pdf
http://www.szyuda88.com/uploadfile/cfil ... 639319.pdf

AIUI, the flash controller incorporates on-chip linear regulators, one of which provides 3.3V for the flash ICs. Your initial resistance measurements would be for the input side of the regulators rather than the load side.

Edit: Here is the product brief:

http://www.sabreadv.com/wp-content/uplo ... -Brief.pdf

The chip does indeed have a 5V/3.3V and a 3.3V/1.8V regulator. The 1.8V supply would the Vcore for the controller.

[inflex]

Whats the plan with this one? If you dont want to send it here, maybe you can find some DR Company in NSW with a SC Tool as this definately supports it. Or an ACE for that matter.

Seems the client has decided that since they can't get it fixed by Friday the damage is already done. Beyond Friday it's an errelevant item of data.

[eellworth24]

I'm having the same problem, but my drive has only 8GB. The metal USB part feels really hot after it's removed from my USB port.

[jermy]

if the data is needed probably chip off recovery have to done

If controller is hot, probably is broken. The best idea is read dumps from nands and assembly the image.