WD WD3200JB-00KFA0 Caviar SE (3.5″ 320GB 2006)

In the same year as when I purchased the 300GB version of the Caviar SE, the 320GB version was released and I decided that it was probably something I could do with. After all, the problem with my NAS was a lack of ports – so more storage per port was very welcome. Unfortunately, this was perhaps my least favoured Caviar SE drive as it did cause minor data loss as it developed a weak sector – but the Seagate equivalent also did the same, so this was really pushing longitudinal recording to its limits.

The Drive

While the drive belongs to the same Caviar SE family as the 300GB variant, this one came with a new label on the “same old” lid and tub design. This was because Western Digital was to soon transition to a new case design, leaving the all-black sleek look behind. I suspect the 320GB version was available with the older label as well, for those with particularly deep pockets who may have bought it half a year earlier.

Otherwise, from the outside, it is unremarkable and looks just like the 300GB version. So how did they manage to squeeze an extra 20GB into the box, and does it go any faster?

Performance Testing

CrystalDiskInfo

This drive has a reasonable amount of hours from its server role, before it was pulled due to its weak sector issue. It has firmware 08.05J08 and has the promised 8MiB of cache, and interestingly, shares the same suffix as the 300GB version.

HD Tune

The transfer rate graphs from the drive seem to suggest it is healthy – averaging a throughput of 57.2MB/s which is a healthy and welcome increase over the previous average of 51.1MB/s. I suppose prospective purchasers of hard drives at the time might not have been aware of this – you may have paid on capacity but you also got performance increases (sometimes).

The IOPS performance is relatively smooth and improved slightly, with read caching seemingly working better than the write caching in the extra testing.

CrystalDiskMark

This is reflected in the CDM results.

ATTO

ATTO seems to show a bias towards write performance at smaller blocks, reaching full performance by 16kB while reads need 32kB or so. This is probably not the most desireable outcome for a desktop user where small-block access may be more important to making a system feel “snappy”, but for NAS/media storage, this is actually quite fine.