Flash Drive Research project

[HaQue]

Hi,
See here for details of a project I have started, It may interest you.

cheers,
HaQue

http://malthus.zapto.org/viewtopic.php?f=59&t=605&p=1578

[HaQue]

these are hard to solder

Attachments

[craig6928]

thanks for the photos i dont think they too hard to solder
we had worse ones then that

good project for sure

[fzabkar]

If you're doing a lot of these, maybe you could rig up some kind of "bed of nails" setup.

[Vulcan]

@HaQue,

Wow, that's quite a collection of flash drives

The first challenge with some drives, seems to be - how to get the PCB out of the case? On cheapo drives, that process is easy, of course, but I've seen some metal-encased drives in adverts where there wasn't an obvious "way to get in".

Anyway, thanks for sharing, and I look forward to any updates as you go through that pile of drives

[HaQue]

Thanks fzabkar, I was talking about a bed of nails setup actually earlier today, trying to figure out something like that. Still not an easy thing to do, there is 40 pins needed in a row for this one and crowding them would be an issue and how to hold them. I am thinking it through..

I think my wire is still too thick at .25mm

Thanks Vulcan, I have seen some of the metal ones. One of them, a small monolith, I bought 2 just to see how it comes apart. I cut one open - and killed the monolith - but really they just pushed it in the metal rectangular tube case with double-sided tape. Some monoliths also have a small plastic L shaped piece, and you can lift the monolith slightly and slide it over the L . Sometimes they put double-sided tape on these as well. There are some metal ones by HP and others that have a collar holding the drive guts in, and glued with some crazy strong glue. I think heat would work well on these.

[Vulcan]

I think my wire is still too thick at .25mm

Similar thinner wire is available - I can try to find a link if you're interested.

I cut one open - and killed the monolith

That would be my worry, working "blind" with a tube-like case.

but really they just pushed it in the metal rectangular tube case with double-sided tape.

Yuck.

There are some metal ones by HP and others that have a collar holding the drive guts in, and glued with some crazy strong glue. I think heat would work well on these.

Double yuck. Perhaps a solvent might help to loosen the glue too, but would then be a fire hazard if heat was applied with any of the solvent still "in the cracks".

Thanks for explaining - they're definitely not designed for repair!

[arvika]

One of them, a small monolith, I bought 2 just to see how it comes apart. I cut one open - and killed the monolith - but really they just pushed it in the metal rectangular tube case with double-sided tape. Some monoliths also have a small plastic L shaped piece, and you can lift the monolith slightly and slide it over the L . Sometimes they put double-sided tape on these as well. There are some metal ones by HP and others that have a collar holding the drive guts in, and glued with some crazy strong glue. I think heat would work well on these.

If you don't try open Iron Key you don't know nothing about disassembly flash from case Look at the picture:

First challenge - metal case (enclosure), 2mm thickness
Second challenge - something like glue/plastic layer, 2mm thickness too, very tough

I remove nand with 100% success without damage it, read it, but flash is crypted

[digitalferret]

Maybe make friends with a local dentist or airport security guard and get some of those X-Rays layed down

[fzabkar]

Instead of a "bed of nails" setup, could you rig up a pre-soldered setup by sacrificing an old card and then pressing it onto your patient with Z-Axis Electrically Conductive Tape?

http://www3.3m.com/catalog/us/en001/arc ... utput_html

[HaQue]

For anyone interested in my Flash Drive stuff, I have been busy formalising the info to publish on the web.

after spending way too many hours on html tables, I came to realise they were just too cumbersome and awful to use.

I spent a bit of time at the wonderful w3schools http://www.w3schools.com and decided on XML to keep the data and be able to access it and add/subtract easily when needed. Some benefits are very easy to use and learn, able to update web page contents without having to refresh the whole page, filtering lists of data etc. I will be able to use JavaScript or some other technology to easily retrieve whatever data is necessary using the one master database file, without building a complicated client-server database.

The free editor from Microsoft is really all you need, and is pretty good http://www.microsoft.com/en-us/download/details.aspx?id=7973

basically:

- Each single device I will call a flash device "case" with a unique ID number to keep it in the same vein as normal DR.

xml1.jpg (23.33 KiB) Viewed 25874 times

Each case will have an id attribute and certain elements:

xml2.jpg (32.57 KiB) Viewed 25874 times

some elements have child elements such as pictures:

xml3.jpg (49.85 KiB) Viewed 25874 times

Information:

xml4.jpg (37.1 KiB) Viewed 25874 times

child elements of information is broken up by the sources of info,
The general section is visual and vendor supplied info:

xml5.jpg (49.37 KiB) Viewed 25874 times

I am breaking down the data to this level so that it is easy to parse in many ways:

xml6.jpg (22.54 KiB) Viewed 25874 times

other "Information" children are from the excellent tool by Nirsoft, USBDeview http://www.nirsoft.net/utils/usb_devices_view.html:

xml7.jpg (54.1 KiB) Viewed 25874 times

and the built into windows, Microsoft's MSinfo32.exe, specifically the "components-->storage-->disks-->(case flash device)" section :

xml8.jpg (32.25 KiB) Viewed 25874 times

Then the last 2 elements are the data recovery results and any comments about the case I want to add:

xml9.jpg (27.56 KiB) Viewed 25874 times

Here is the diagram of the complete element of a case:

status:
- separated 185 devices, 1 device into each bay of 5-bay containers and marked with case numbers.
- recorded USBDeview and msinfo32 from 74 devices. Each device is taking about 2 - 3 minutes now, so should get through these quick.
- photographed 200 devices for the initial picture only
- have another 30 or so almost ready for cataloguing
- have created a test set of files and have md5 hashes to check integrity after recovery, and some documentation going on the test set of files.
- have the XML format pretty close to being happy with it.
- have been playing with making the data into a pretty and functional web page/set of pages
- have a domain etc already this will go on.
- have even more enthusiasm and am now looking for more devices.

I welcome any suggestions and critisisms.

If anyone wants to hire or loan me a PC3K.... will definitely be very happy..

Thanks

[HaQue]

If you missed what this was about, the initial details I posted are here: http://malthus.zapto.org/viewtopic.php?f=59&t=605&p=1578
forgot to add that and editing time for the post ran out

there are some elements that seem duplicated by different tools. I thought it valid to see the differences in what the tool is badged as, what the system sees it as in each software tool.

[fzabkar]

Thanks very much.

[HaQue]

Update
- Loads of documentation done
- Formatted and copied test data to around 10 drives.
- Un-Soldered 3 NANDs so far, read 2 of them with 2 different tools, to get a feel for exactly what data I will end up with, and how exactly to code the webpages that will display the results.
other one I need info on power issues.

The coding.. well after a 6 hour session trying different tutorials and testing a LOT of code trying to get the XML to display, I found some obscure post saying "This might not work on IE if on the local system, but upload it to a real web server and it works fine..."

AAAARGGGH!!! if you heard a load scream from southern Australia yesterday, now you know what it was!

More to come soon.

Edit - and I bought a few more drives..

[sophia]

Thanks! for the photos of flash drive, anyway i am looking forward to your updates which might be good conversation regarding the Flash drive.

[fzabkar]

I'm looking forward to them, too. However, I can't see the HDD Oracle coming back online anytime soon, if ever.

[HaQue]

I'm looking forward to them, too. However, I can't see the HDD Oracle coming back online anytime soon, if ever.

Certainly seems that way. Not to worry, I will post here still, and a proper blob/site with all the research. I am not affiliated with The Oracle in any way.

I have been working on things - a lot.

I have changed the way I am doing certain documenting of each device, because of things discovered about the way certain software reads the drive. Also added a few things to document, bought about 40 more drives, taken better pictures and done some coding.

I had been working on some way to add a socket type arrangement to each device to allow adding or removing the NAND easier, but this seems like it isn't workable and soldering each time might be just easier.

Also been looking at tools to purchase and some controller research.

[HaQue]

A Lesson on Over-Engineering

Each device I am testing, I wanted to take 5 separate dumps of the NAND. 1 for each different tool to test how well it recovers, and 2 other dumps of certain byte patterns to help reverse engineer/learn about the structure the controller uses to store the data. I wasn't looking forward to soldering on the NAND 5 times for each device!

I spent about 10 hours in AutoTrax PCB software trying to design a NAND shaped PCB with 48 pads in the center, to make a thin pcb, and some kind of pin arrangement to press down on them. I got a pretty good design (tough working out all the fine traces and pads, gaps etc).

Well Turns out I was thinking too big. Instead of getting hundreds of little PCBs made up, and soldering them on, and rigging up some dodgy pin contraption, I solved it quite cheaply, and IMHO a lot easier.

From my Favourite Electronics store that has free shipping, Ebay, I bought some cables and connectors. I believe the connectors are a lot easier to solder to a Flash drive than a NAND that has been removed and soldered back on a few times. The cost is minimal as well.

Simply solder a 24 pin, 0.5mm pitch connector on each side of the PCB in place of the NAND, a Flash drive with 1 NAND chip costs only $1.18.

For the socket, 2 cables, 2 connectors and 1 TSOP48 Programming socket is around $20. Most drives I have are 1 or 2 chips, so for now I have many 1 chip drives to gather data on, and can make another later easy enough.

The socket is pretty easy to work with.

Just add a NAND chip and the drive is complete again

Then just add connectors to all the drives. Didn't take me long to do these 3. I can now test at will and take the chip out to read, no soldering required.

A test shows the drive works fine. I tested all 3, with 3 separate dumps of the NAND each and no errors.

r008.jpg (40.38 KiB) Viewed 24676 times

Hopefully I can now make some quicker progress on the 330 or so drives, and the many new ones Im always buying.

In summary, a very simple solution, nothing fancy but works very well.

[labtech]

That's pretty cool, I like it.

Been thinking about something similar in concept, but with regard to the controller. Wondering if there could be a way of masking the main on board controller (presumably one that has failed) with one that is connected to it from a outer socket. Or very similar in concept to the masked in ROM in HDD that could be bypassed by soldering an external ROM. Specifically, targeting SSDs, since many times there is no feasible solution due to encryption and much desoldering work.

[einstein9]

Excellent work HaQue
in fact more than Perfect, i just saw it now