|
Page 1 of 9   When asked what hard drive should be purchased, most users would answer: The bigger the better. But is this the right choice? In most cases it is, but we don’t need that much space nowadays, so some of us prefer faster drives, some of silent ones and so on. Luckily there is variety of models to choose from. The time of SSDs is coming for certain, but at the same time hard drives are getting bigger and cheaper. In very near future we can predict transitional period where SSD will be used as system drive while HDD will be used as data storage. Since SSD prices are not showing significant drop tendency, we will be using HDDs for some time in the future. So, we decided to make, maybe for the last time, big hard drive roundup.
Benchmark methodology
Some users think that specifications of hard drives are absolutely enough when making decision on which one to purchase. Unfortunately this isn’t the case, so we tested all relevant parameters of hard drives that we could put our hand on.
Temperature
| Real
Capacity |
You have probably noticed that we have additional category in specifications: Real Capacity. This is
consequence of bad manufacturers’ practice to put “alleged” capacity on hard drive labels like 500GB, 750GB and so on. Real
capacity in some cases can be much smaller than that. For example: ST31500341AS is advertised as 1.5TB hard drive, but its real
capacity is 1.36TB. Don’t tell us that 140GB less capacity doesn’t matter to you? There are notebook (and even desktop) hard
drives of such capacity. This is consequence of manufacturers practice to round 1MB as 1000KB and not 1024KB as it should be.
Of course this is written on each hard drive sticker but in very small font.
Manufacturers’ justification of this practice is
that it is much easier to express total capacity with rounded figures, and that it is prettier. Prettier? Ok, than why 500GB hard
drive is not advertised as 465GB? How many times did you miss 35GB? Interesting fact is that same drives for Apple Macintosh
configurations are advertised with real capacity. If you ask us, this is question of ethics, and it is not nice to make money on small
text.
|
This is parameter that is very often neglected by majority of users. Very often too high operational temperature of hard drive can lead to its premature failure. Ok there is always that warranty period in which “broken” hard drive will be replaced by new one, but what about personal data? In most cases this leaves a mark for life on any user. Hard drive should be cooled actively, which is recommended practice, passively or simply choose a drive that doesn’t heat-up too much.
These tests were conducted in a air conditioned room with temperature set at 23 degrees Celsius. Upon installation of each drive we waited for 10 minutes so it can reach its operational temperature. Thermal prober was secured at its side because it is the hottest area of each hard drive. After that we heated up drive with HD Tune benchmark (read and write tests), waited for 5 minutes after completion and wrote down achieved temperature. Temperature was also measured right after final benchmark was finished, and each test was started immediately after previous one finished. Of course you can measure these temperatures using different software solutions but bear in mind that manufacturers in some occasions can place thermal probes inside hard drive at strategic places so it shows better (lower) temperatures. Because of that we used external thermal probe and an universal meter with thermal probe.
Noise
Another important property of every hard drive is noise that is generated. We combined measured noise levels with subjective impression of each drives noise-print. Seagate ST3320613AS was described as noisiest and it got lowest mark. On the other hand Samsung HD502HI mode was marked with highest mark because it has quietest. Highest mark goes to this drive and all others are somewhere between those two when it comes to noise levels.
Benchmarks
| Hard Drive
Temperatures |
| As you can see from results, some hard drives are “hotter” than others. Lower temperatures guarantee
longer trouble free operational time so some kind of cooling should be applied to hard drives: passive or active. All hard drives
on this test with temperatures below 33 degrees Celsius are considered cool. Everything above that is considered warm, and
above 38 degrees Celsius is considered critical on prolonged operational time. Our test lab was air conditioned at 23 degrees
Celsius. Summer is here so if your room is not air conditioned it can result in hard drive failure, but then again maybe nothing
will happen. To minimize risks try to provide your hard drives with some sort of cooling. |
Most users are interested in results that this segment of our test will produce. We used well known software for measuring HDD speeds. HD Tune with its read and write benchmarks, followed by minimal, maximal and average achieved bandwidth. HD Tach is also used since some users prefer its results. ATTO is also included in our “benchmark battery”. Access times and CPU usage were measured with HD Tune. Beside these tests we also conducted “stop-watch test”. This test is used to measure time that is needed to move data from one partition to another. We decided to use worst case scenario. Each drive was partitioned into 3 partitions. First (1) and last (3) were 10GB and second (2) was occupying the rest of capacity. We made a 3.5GB folder that was consisted of:
- 262MB of small MP3 files and pictures of average size between 100KB and 7MB
- 2.04GB of 3 standard DivX moves
- 1.02GB of 8 video files 130MB in average
- 188MB of 8888 small files in 144 subfolders (icons, files from browser cache ranging in size from 2 to 150KB)
Basically this is one nasty folder consisted of very small files, very big files and a lot of folders.
Test folder was copied to slowest partition (3). Why slowest? Well, magnetic head needs to travel the most to get data from this partition. First partition (1) is fastest since data is stacked closest to the center of HD platter. We measured time that is needed to copy test folder from third (3) to first (1) partition. Of course, lower values present better result.
We managed to collect 23 HDDs from four manufactures: Samsung, Seagate, Hitachi and WD.
|
