Review, Teardown: Verbatim 8Gb Store’n’Go PinStripe USB 2.0 Drive (49062)

Yet another low-cost USB flash drive gets a review, mainly because I just happened to end up receiving it. Unlike the others though, this one comes from Verbatim which is a brand which has come to be respected due to their reputation built around floppy disks and optical discs. The 8Gb drive costs around AU$5.98, so it’s not the cheapest around, but hopefully it’s of good quality and performance. Lets see how it fares.

Unboxing

2016120311129943 2016120311129944

The drive is packaged inside a retail hanging card, with a shiny glossy finish to the card. The drive itself shows inside a clear plastic bubble, and has a striped plastic design to match its fancy PinStripe name. The rear has some basic information in a number of languages, but basically it’s a USB 2.0 drive with a claimed data retention of 10 years. The unit is Made in Taiwan with a product code of 49062, and password protection software is available, however it is only a trial. I don’t recommend using such software anyway, but other drives do include it as standard, so this seems somewhat curious.

2016120311139945

According to the logo at the front, less plastic is used in the pacakging and it definitely seems to be the case with the plastic bubble only being on one side and minimally larger than the drive itself. However, I somehow doubt that this selling point really resonates with customers.

2016120311149949

The drive itself has the capacity engraved on the connector, and a slide-body cap which means nothing to lose. The end cap has a provision for a lanyard attachment.

2016120311139947

A serial number is engraved at the top side of the drive.

2016120311139948

Unlike some other low-cost drives, this unit has a “real” USB connector with gold plated pins and a raised profile, so should have a fairly good connection to host ports as compared with connections made from PCB traces.

instd hwids

The unit identifies as a Verbatim STORE N GO device, with a VID of 18A5 and a PID of 0302.

vsng-asformattedThe drive is pre-formatted with the FAT32 filesystem without any pre-loaded bloat, and the capacity total is 7,742,701,568 bytes. This is technically less than even 8-weasel-gigabytes which would have been 8,000,000,000 bytes, although it seems that this is a fairly common trend amongst lower cost devices due to the need to reserve space for more effective wear levelling, bad block replacement, and to compensate for higher bad-blocks from manufacture especially when using lower quality flash.

As compared with some other cheap 8Gb flash drives I’ve used, the Sandisk Cruzer Facet had 7,998,799,872 bytes which was much closer to the expected capacity. The Comsol 8Gb USB 2.0 Flash Stick had around 7,808,122,880 bytes available, so it seems the Verbatim drive is on the low end when it comes to the available user capacity. This may have benefits in terms of lifetime, performance and reliability however, and it is clearly stated on the packaging that up to 10% of the capacity may be used for formatting and other functions.

The user capacity is hence 96.78% of the weasel 8Gb capacity or 90.14% of the binary 8GiB capacity. This means that roughly 10% of the chip is overprovisioning/management/cache/bad-blocks/etc.

Teardown

2016120315219950

I started my explore by pulling the front cover off of the body of the drive, only to realize that was not the way.

2016120315239951

Instead, if you just separate the end cap from the device, the whole PCB slides out. Interestingly, the PCB itself has a BGA-mounted flash device, where pads are also provided for the more traditional SMD legged devices. There is also an internal red activity indicator LED which does not shine through the case and is otherwise being wasted.

2016120315239952

A close look under the label shows the flash device appears to be a Toshiba THGBM8G6D1KBAIL from China. This part number doesn’t turn up much, but it might well be an eMMC device (maybe even used raw, as in the Comsol stick). The soldering does look a little funny, as some solder balls seem to be visible at the edges – might these be used as crude test points?

2016120315249953

The underside reveals very little, as an all integrated controller requiring no external crystal is used. The controller is a Phison PS2251-68-5, a common controller amongst “lower cost” flash drives.

Performance Test

HDTune Pro

03-december-2016_16-11

A quick read test shows the drive is capable of reaching 25.1MB/s average read speed. The read speed of most drives are capable of reaching USB 2.0 bus limits as reading is hardly difficult and controllers have gotten more sophisticated. This is a middle-of-the-road performance compared to the Comsol which performs worse (~13MB/s), and the Sandisk Cruzer Facet which performs better (~32MB/s).

CrystalDiskMark

vsng-cdm

The sequential read speeds are fairly consistent with the results obtained with HDTune and superior to the other two cheap 8Gb USB drives. The write speed is just above 10MB/s, making this perform similarly to a Class 10 SD card which just meets the bar. In terms of sequential write, the performance blows the other two drives out of the water.

However, just as with many high speed SD cards, the small block performance takes a negative impact. The 512kB write falls to 1.05MB/s, which is faster than the Comsol at 0.3MB/s but much slower than the Sandisk Cruzer Facet at 4.192MB/s. By the time we get to 4kB writes, the drive achieves just 9kB/s whereas the Sandisk Cruzer Facet is still trundling along at 1.3MB/s.

It seems that the optimization for higher sequential write performance means significant penalties for small block writes probably due to the larger erase block of the flash necessitating longer erase/program times. For saving very small documents, the Sandisk Cruzer Facet might actually perform better, however, for large media files, the Verbatim would win. In most read-cases, the Verbatim would win, except for sequential where the Sandisk Cruzer Facet still has a lead.

ATTO

vsng-atto

Full performance is only really reached by 128kB block accesses. For the Sandisk Cruzer Facet, this was achieved at 64kB, although it did have an anomalous write performance curve which suggested a decline in write performance at higher block sizes.

H2testW

vsng-h2testw

The drive successfully passed the test, with the write speed just shy of 10MB/s and the read speed spot-on at 25MB/s.

Conclusion

The Verbatim Store’n’Go is probably targeted at value conscious consumers with its very basic no-software inclusions, USB 2.0 interface and simple packaging. It tries to be a little “fancy” with the product naming for what is essentially a ribbed plastic case. It does however, sport a proper USB connector and a full PCB-based design rather than the “all-in-package” that some low cost drives are using.

Its sequential read performance at all block sizes was good and in-line with that of “full size” USB 2.0 drives, although its write performance seemed to be tuned for larger block accesses (e.g. sequential). The Sandisk Cruzer by comparison, had better sequential read, and small-block writes and at a lower price. At such a low price point, I don’t think performance is high on the priority list.

Internally, it utilizes a Phison controller as is quite common, paired with Toshiba flash. The longevity is not known, however, the overprovisioning is rather large compared with its competitors, resulting in less user available space. Verbatim does offer a limited lifetime warranty on the drive though.

About lui_gough

I'm a bit of a nut for electronics, computing, photography, radio, satellite and other technical hobbies. Click for more about me!
This entry was posted in Computing, Flash Memory and tagged , , , , . Bookmark the permalink.

7 Responses to Review, Teardown: Verbatim 8Gb Store’n’Go PinStripe USB 2.0 Drive (49062)

  1. ginbot86 says:

    That’s certainly an interesting way to implement a USB 3.0 flash drive.

    The balls that are visible are for eMMC packages in the 12mm*16mm, 12mm*18mm, and 14mm*18mm sizes; they have no electronic purpose and only serve as extra reinforcement since all the BGA balls are centered in the device package. The eMMC used on that board is 11.5mm*13mm, and there are devices as small as 11mm*10mm!

    Looking at the eMMC’s date code, it probably was an eMMC 4.5 or 5.0 device, and if the controller on it was/is functional, it likely would have achieved well over 50 MB/s sequential read and ~3 MB/s 4k write, if it weren’t interfaced as raw NAND like it is here.

    I do wonder if the controller’s still working on that device. It’d be interesting putting it into an eMMC test socket and inserting it into a standard SD/MMC reader, but chances are the firmware’s already long gone. It’s almost sad seeing an eMMC used in this manner…

    • lui_gough says:

      Agreed – I would have expected much better performance since it would be the sort of memory you’d run a tablet or phone on. I suppose we do get what we pay for ultimately, as the Phison chip is a common controller amongst lower-end drives. Maybe these were yield defect chips with a controller issue? I wouldn’t have a clue really since I don’t have the gear to test for it, but from my experience with the Comsol that used microSD-style packaged cards as raw NAND, it wouldn’t function as a microSD card due to lack of internal formatting for its controller.

      – Gough

      • ginbot86 says:

        It could potentially be dud controllers – there was a huge kerfuffle over some older Samsung eMMCs with a firmware bug where a secure erase/trim would corrupt the NAND flash wear leveling table, and if the area affected was near critical OS/bootloader structures, the device would be “hard bricked”.

        Here’s a link to a CyanogenMod page discussing that issue: https://wiki.cyanogenmod.org/w/EMMC_Bugs

        So far, I’ve been looking at ways to use eMMC Flash as fast USB/SATA storage, but I’ve been finding extremely little in terms of controllers that support 8-bit MMC at HS200/HS400 speeds. Dell had some SSDs that were simply 32GB eMMCs with an OZ788WR2 SD/MMC-to-SATA II bridge on an M.2 card; I have three of them – one sacrificed for more in-depth reverse engineering, and two working ones – I intend to upgrade one of them to a Samsung 128GB eMMC 5.1 module when (if?) it arrives in the mail.

        (and it looks like I said “USB 3.0” in my previous comment – whoops, it’s always after one clicks “Post” that mistakes are found!)

        • lui_gough says:

          Thanks for that reply – I do remember the brick bug, but it was too late by the time my S3 got hard bricked out of warranty. That being said, adapting eMMC to anything with high performance sounds like a challenge, but I suppose a cheap USB 3.0 drive might be possible just by using 4-bit MMC mode and a USB 3.0 SD-card reader style chip as the interface. Maybe that might entail paying some royalties on the chip in regards to SDA, CFA etc. in the case of a multi-standard chip, but I suppose it’s another “route”.

          There was one silly SSD on the market which batched together a tonne of eMMC chips … for not much good really …
          http://www.computerworld.com/article/2990446/data-storage-solutions/chinese-ssd-maker-eyes-us-market-for-8tb-drive-intro.html

          I suppose that will all abate as UFS based flash becomes more common in the smartphone space.

          – Gough

          • ginbot86 says:

            I bought a board that uses one of those “10-channel” controllers. It’s pretty crummy – each channel can only do 25 MB/s, it’s limited to SATA II speeds, and it lacks TRIM/secure erase. I’m inquiring with an eBay seller to get a bunch of 64 GB eMMC chips to see whether I can turn it into a somewhat-usable SSD. Could be quite an adventure!

            I’ve had fairly good luck using eMMC on standard SD readers (but the socket was not cheap – $115 USD!). The most I’ve gotten out of any chip is 45-50 MB/s sequential read, which makes sense since we’re limited to 52 MHz 4-bit DDR at 3.3V signal levels; the HS200/HS400 modes are limited to 1.8V signal levels.

            UFS is definitely looking pretty promising, and a few phones out there have already adopted it as internal storage. Too bad there doesn’t seem to be any UFS-to-SATA solution out there yet.

  2. malseni says:

    WOWOWOWOOWOW it’s the first time you forgot to blackout the s/n. let’s edit it quick!

    • lui_gough says:

      WOWOWOW what a very insightful and useful comment. Not!

      Some things I have to due to the nature of supply. Other things I don’t. It’s just how things go.

      – Gough

Error: Comment is Missing!