Sunday, June 06, 2010

Chinavasion's Gold Bar 16GB USB Review

The honeymoon period:
Shortly after receiving my 16GB Chinavasion Gold Bar USB Stick the other week I went about immediately using and testing it and found it wasn't very reliable. Since Chinavasion refused to publish my review I'll provide the fully damning one here. Incidentally as of the time of writing, 6/6/10, and after my quick search of reviews Chinavasion don't appear to have many (any?) 2 star or less product reviews (I only saw 2x 3 star reviews, the majority are 5 star, with a few 4 stars).

The casing:

The casing, weight and feel is as good as it looks on Chinavasion, however the stick I received has a tiny chip on the top right corner (not very noticeable) and one of the reverse side screws was not gold coloured but regular grey and looked very out of place (the type and size was correct). I have been tempted to rip out another USB stick memory and use it inside this because of the next problem...

The electronics:
The memory stick appears to be a genuine 16GB, that's the good part. The bad part is that it appears to be Grade B lower quality RAM which has data failures. Initially my investigation required verifying the make of the controller chip to ensure it was the correct memory size, somewhat suspiciously the controller and RAM chips inside the USB stick seem to have extremely faded manufacturer and type details printed on them. This made it impossible to determine what the specific makes were.

The fix:
Starting up fake USB stick software and testing took me a better part of a day as there are literally a hundred different "Alcor" Chinese language low level format and controller modifying applications I eventually found one that was able to recognise the stick. The memory is marked as Generic and the type is an AU6981 (or 6986), although this information can be falsified. Initial analysis with an old version of ChipGenius and its old database initially identified it incorrectly as the 98xx series, through trial and error using a newer version and other programs eventually got the right result.
The version of AlcorMP tool you need if you have one of these USB sticks is available in a zip called (or AlcorMP_AU698X_091111.rar).

Running this tool with some modified settings* and checking the "Status" determined on one run I had 3 bad blocks (with default settings), another run told me 4099 bad blocks and corrected the available size down to 12.1GB. Actual hit and miss testing via USB disk partitioning on Linux had already determined I only had 12GB of reliable memory on the stick as anything above would become corrupted on retrieval (it appeared to write but would consistently fail to read or files vanish and a corrupted filesystem warning show if I used the full 16GB capacity).

The result:

So why didn't I send this back? Well because shipping it back to China after I had just waited 3 weeks to get it (cheap postage) meant another round trip of about 6 weeks and there was no guarantee that the replacement would not be even more defective. The relatively cheap cost of £23, or half a night out, was another factor in that it just wasn't worth the hassle. Basically I used the Alcor tool to mark the USB key's bad blocks and to only use the reliable 12.1GB and this seems to have worked.

* Follow this translated guide. Under the Flash Type tab use the following LLF Check: Natural Check, Scan Level: Full Scan4, Scan Mode: Low Level, RW Cycle Time 33ms (as far as I can tell the memory is rated at 25ms however the stick is faulty so better to be safe than sorry - my write speed is still a slow 3.3MB/s).

No comments: