Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Data Storage Hardware

SanDisk Extreme Pro Failures Result From Design and Manufacturing Flaws, Says Data Recovery Firm (tomshardware.com) 38

Anton Shilov reports via Tom's Hardware: A new report from a data recovery company now points the finger at design and manufacturing flaws as the underlying issue with the recent flood of SanDisk Extreme Pro failures that eventually spurred a class action lawsuit. It became clear in May that some of Western Digital's SanDisk Extreme Pro 4TB SSDs suffered from sudden data loss; at this point, the company promised a firmware update to owners of the 4TB models. However, the 2TB and 3TB models also suffer from the same issue, and Western Digital did not promise any firmware updates for these drives.

Markus Hafele, Managing Director of Attingo, a data recovery company, told FutureZone that the problem lies in hardware, not firmware, which could explain the lack of corrective firmware updates for those models and SanDisk's continued silence about the source of the issues. Attingo, which has been in the data recovery business for over 25 years, normally sees these failed SanDisk Extreme Pro SSDs at least once a week. The problem appears to be rather complex. According to HÃfele, the components used in these SSDs are too big for the circuit board, causing weak connections (i.e., high impendence and high temperatures) and making them prone to breaking. He also says that the soldering material used to attach these components is prone to forming bubbles and breaking easily.

It remains unknown whether the cause is cheap solder, the componentry, or both contribute to the issues observed. However, newer revisions of these SanDisk Extreme Pro SSDs seem to have been modified with extra epoxy resin to secure the oversized components. This suggests that Western Digital might know about the hardware problems. Nevertheless, these newer models are still failing, thus sending data recovery service customers to firms like Attingo. According to the head of Attingo, the issue seems to be affecting multiple product lineups, including both SanDisk Extreme Portable SSD as well as the SanDisk Extreme Pro Portable SSD.

This discussion has been archived. No new comments can be posted.

SanDisk Extreme Pro Failures Result From Design and Manufacturing Flaws, Says Data Recovery Firm

Comments Filter:
  • Ah, the modern era: where trusting your hard drive is akin to trusting a cat with a goldfish. Remember when 'drive failure' meant you forgot where you parked? Now, it's all about SanDisk playing Jenga with our data. I'm just waiting for the day they announce SSDs with 'optional data retention' as a premium feature. And let's not forget, in a world where cloud storage is king, local drives are like that one friend who still thinks flip phones are peak technology. So, here's to hoping our trusty SSDs don't de
    • by myowntrueself ( 607117 ) on Monday November 13, 2023 @06:52PM (#64003559)

      Ah, the modern era: where trusting your hard drive is akin to trusting a cat with a goldfish. Remember when 'drive failure' meant you forgot where you parked? Now, it's all about SanDisk playing Jenga with our data. I'm just waiting for the day they announce SSDs with 'optional data retention' as a premium feature. And let's not forget, in a world where cloud storage is king, local drives are like that one friend who still thinks flip phones are peak technology. So, here's to hoping our trusty SSDs don't decide to 'Thanos snap' our files out of existence. Because, let's face it, the only thing worse than a faulty SSD is realizing your backup plan was just wishful thinking.

      Trusting one *single* *piece* of storage media with your data... that was never a good idea.

      • by Potor ( 658520 )

        Trusting one *single* *piece* of storage media with your data... that was never a good idea.

        Redundancy and the cloud is truly the answer. I cannot remember the last time I've lost anything critical in the last 30 years (I used to do nightly FTPs to my university's servers), although I have had my share of lost drives.

        • Trusting one *single* *piece* of storage media with your data... that was never a good idea.

          Redundancy and the cloud is truly the answer. I cannot remember the last time I've lost anything critical in the last 30 years (I used to do nightly FTPs to my university's servers), although I have had my share of lost drives.

          That said, just finding the data you wanted becomes the hard part... especially if you use something like Teams or Slack...

        • by networkzombie ( 921324 ) on Monday November 13, 2023 @10:09PM (#64003881)
          “The cloud" is never the answer to anything. A specific company's "storage solution" maybe an answer, but “the cloud" is bullshit.
          • by Potor ( 658520 )

            “The cloud" is never the answer to anything. A specific company's "storage solution" maybe an answer, but “the cloud" is bullshit.

            Which is why I said the cloud and redundancy, including hard copies (I write books, and my biggest nightmare is losing all soft copies - I print fanatically).

            • You did not understand my statement. “The cloud" is never the answer to anything. “The cloud" can steal, share, leak, and sell your data. Never tell anyone to use “The cloud". Recommend a specific company's storage solution. “The cloud" is a bullshit term even if you combine it with redundancy, hard copies, or magic beans.
              • "The cloud" is never the answer to anything. "The cloud" can steal, share, leak, and sell your data.

                So can a hacker, with your on-premise data. Or a disgruntled employee. Or your IT-illiterate grandma.
                Point is, every tool could be used well or in a shitty way.
                Note: I am not advocating the cloud over on-premise, or the other way around. Personally, I use both: local+cloud backups of important data, parity drives and/or RAID for the rest.

        • Redundancy and the cloud is truly the answer.

          So the only two choices I have is "I don't have access to my data because it's gone" and "everyone has access to my data because I trusted the cloud"?

          Tough one.

        • Can confirm. I had a RAID 1 set up at home with spinning disks that stored all my families home videos and photos. Around 18 years worth of media starting from the day we got married, to when we had kids, to the present day. I copied them from my phone to my RAID array religiously and every so often, monitored the health of the drives.

          When we got fiber, I uploaded the entire lot to my Google drive. It was about 2TB worth and promptly forgot about it. The drive app on my desktop kept on uploading stu
    • by Anonymous Coward
      You don't have to buy SanDisk stuff you know.

      Just buy something else with a better track record.
    • There's nothing modern about this. Catastrophic dataloss has always existed on all harddrives. And most companies have released products which were such stinky turds that they almost ended the business, and in some cases did.

      No your SSD is not less reliable than your HDD. ... Well unless you have a SanDisk Extreme Pro. But then again is that less reliable than a 40GB IBM Deskstar?

  • Stick the SSD in a toaster oven for a few min to reflow the solder? My guess is that's what this data recovery guy is doing.
  • by Powercntrl ( 458442 ) on Monday November 13, 2023 @07:49PM (#64003665) Homepage

    These drives are just an OEM version of WD Blue NVMe SSDs mated to a USB to NVMe adapter board. Since the bare NVMe drives don't seem to experience the same sort of failures when installed inside a PC, it certainly stands to reason that they could be breaking because the sandisk extreme's external enclosure is quite flimsy. The plastic "door" which sits directly over the NVMe PCB is thinner than a credit card (probably for improved heat transfer, since plastic is a terrible conductor of heat).

  • "According to HÃffele, the components used in these SSDs are too big for the circuit board, causing weak connections (i.e., high impendence and high temperatures) and making them prone to breaking." " oMg wUhT dO wE dO wItH aLl oF tHeSe lEfT oVeR cIrCuIt bOArDz!".. I know! Let's just install components on.them that they were never designed for instead of cutting our losses with the old stock and designing proper circuit boards for thr new components. :-/
  • I learned the hard way to never trust a hard drive to store anything important. I trust an SSD even less because, chances are, it's not going to warn you that it's about to fail.

  • by hirschma ( 187820 ) on Monday November 13, 2023 @09:18PM (#64003791)

    I used to run a PCBA [sfxpcb.com] factory. Here's what I'm seeing:

    * The capacitor that's elevated at an angle is called a "tombstone". At least, think it's a cap - "CS" isn't a standard reference designator.

    * Tombstones happen when uneven solder paste is applied, or improper heating is applied to the board during manufacturing, or the actual PCB design is crap, or the components are placed by a crap pick and place machine. So, one side of the component is pulled up by the surface tension of molten solder, as the other side hasn't melted yet.

    * Wrong-sized components could exacerbate the situation, but I'm guessing that they're doing this to actually fix the situation (larger sizes may be less likely to exhibit this).

    * There should be an inspection step in the manufacturing to catch defects like tombstones, before any "potting" is applied (like the epoxy shown). Such boards should be discarded or repaired.

    * The potting happenedafter the heating step (called reflow, which melts the solder paste). That is entirely useless, as all they did was apply epoxy to defective electronics.

    Our factory used to get a lot of bad designs, so we often just applied epoxy under certain components to prevent tombstones. Has the effect of gluing them down. It's called "staking".

    * There should be an inspection step at the end, which again, should catch defects like tombstones. Such boards should be discarded. They can't be economically fixed after potting.

    So, the upshot is that this manufacturing was outsourced to a factory with sloppy manufacturing, non-approved component substitutions, and zero quality control. OR the PCB design is terrible, and was not designed to prevent thermal gradients during manufacturing, and the factory had terrible habits on top.

    This type of shoddiness is pretty shocking for a brand name like SanDisk, or Western Digital.

    • Just to add - if "CS" means current sense (resistor), it's a wonder that the drive ever worked at all. Thing probably fried itself.

    • by AmiMoJo ( 196126 )

      Agreed. The tombstoning should have been caught by computer vision inspection, or manual inspection.

      The capacitor being lifted like that could have been detected during testing, either using a probe or possibly even in the controller IC itself if it can measure voltage somehow (ADC or comparator). It might even be possible to measure the surge current into the USB connector at the start of heavy loads.

      My guess would be that they reduced the amount of testing they were doing after failure rates stabilized at

    • This type of shoddiness is pretty shocking for a brand name like SanDisk, or Western Digital.

      It would be shocking for a brand name like SanDisk if you didn't know they were now owned by Western Digital. It's not shocking for WD though, they haven't been known for high quality for decades... just for being better than Seizegate. Yeah, I go back to the days of whacking ST225s with a screwdriver to cure stiction. WD used to make the best MFM controllers, that's the last time I can remember being impressed by them.

  • Well, it is WD. Should not be surprised.

  • This sounds exactly like Ford and others trying to rush out electric vehicles too late and doing a crap job of the engineering in multiple ways. It's not like hard drive companies didn't know SSDs were coming. They had easily 10+ years to get out ahead of this and that did they do? Acquire SanDisk (or partner with them, I don't remember) and throw together this defective, rushed, cheap, poorly engineered piece of garbage. What else did you expect from execs that react like dinosaurs instead of the hyper fas
  • I have no clue what this 'recovery guy' is on about, but as someone that manufactures PCBs and does the SMT assembly work, you could put a component triple the size of it's intended form factor down and as long as the solder connection is good it just doesn't break free. I've got oversized inductors on pads barely the size of a pencil eraser and those go into off-road vehicle lighting. Guess what isn't breaking and coming back to my doorstep, and hasn't for almost 4 years?

    Someone at Sandisk failed to do a p

    • by Khyber ( 864651 )

      Oh, looking at the images, it's even funnier.

      One is a tombstoned component because of another component under it on the pad. Another is tombstoned because the other pad it should attach to has no solder, visually. Another picture you see an excess of solder and poor reflow, the pads are the correct size (they match up to the traces at package edges) but might not have been properly-exposed during mask removal, hence why the excess solder shows.

      All basic QC fails, and not at all what the data recovery guy, w

To stay youthful, stay useful.

Working...