ASSEMBLE YOUR MATERIALS:

• Selecting the processor, motherboard, and memory
• Selecting a case
• Selecting a hard disk
• Extra drives, aka "should I bother with a floppy?"
• Video cards
• Audio cards
• Networking
• Monitors, printers, and all that extra stuff

PUT IT ALL TOGETHER:

• Tools and precautions
• Inspection and preparation
• Pre-assemble the motherboard
• Install the motherboard and expansion cards
• Install the drives and power supply
• The MAGIC SMOKE test...

Selecting a Hard Disk

Your hard disk is the central storehouse of everything that makes your rig yours. It holds your operating system, your office software, and probably most of your games, but more importantly it stores your bank records, your personal emails, and all those pictures you downloaded from the internet... But all of that aside, the point is that this decision ranks alongside the choice of processor, memory, and motherboard as the most critical decisions to make about your new rig.

The key factors to consider are:

• Storage capacity
• Interface type
• Number of drives
• Access speed
• Buffer size
• Recommendations

Storage Capacity

First, a word about terminology: The amount of data that a hard disk can hold is the "storage" capacity. Some will erroneously refer to storage capacity as "memory," but when talking about a PC, memory should only ever refer to data processing apparatus, either hardware (RAM) or software (virtual memory). There is also a distinction to be made between a "hard disk" and a "hard disk drive." The "hard disk" refers to the actual disk itself, whereas the "hard disk drive" is the oblong box that holds it along with the electronics required to read from and write to it, in other words, to drive it. Technically, you do not store data on a hard disk drive; you store data on a hard disk located within a hard disk drive. References in this section will make the distinction between the disk and the drive as follows: The drive will be abbreviated "HDD;" whereas the disk will be referred to as the hard disk. The term "hard drive" is a technically inaccurate blending of both, and will not be used in this page although it may be used in other pages on this site.

That being said, what you are selecting is a drive and not just a disk. The amount of stored data is no more important than the means by which your computer accesses it. Total storage capacity should reflect an economical choice of how much empty space you want to have left over once all of your favorite emails, songs, videos, photos, and bank records are accounted for. My personal rule of thumb is that once your rig is complete, after initially installing all of the above, you should still have at least as much empty space as full.

You will need a certain amount of space for your operating system and essential software, but not nearly enough to justify, for example, a 500 GB monster disk all by itself. A full installation of Windows XP and a passel-load of software should occupy something like 30 GB. That leaves all that leftover space for whatever you want.

Base your decision on your usage. If you're a Gamer, you'll want at least enough to hold your operating system and your primary game(s) of interest. A Surfer will need significantly more, since they'll be more likely to download songs, movies, and perhaps other large file types. Since the rig you build will likely not be the first PC you ever owned, you should have a good idea of how big a hard disk would have to be in order to accommodate your usage. If you simply can't decide, then figure that storage costs between 25 and 50 cents per gigabyte and adjust your requirements as needed.

Back to top

Interface Type

There are two primary choices: PATA and SATA.

PATA ("Parallel Advanced Technology Attachment") is the term that has officially supplanted IDE ("Integrated Drive Electronics") and EIDE ("Enhanced" IDE). It refers to an interface in which several data bits are sent at once as a packet. Typical data transfer rates are between 30 and 50 MB/s in practice, although the theoretical maximum speed is 133 MB/s. The discrepancy lies in the manner in which data transfer rates are measured. The actual practical or "sustained" data transfer rates (sDTR) are based on the amount of data that can be transferred in a given time period from the hard disk to the HDD Buffer and thence to the drive controller on the motherboard—also known as the "host controller". The advertised 133 MB/s speed is the speed at which data can be transferred between the host controller and the HDD buffer, without having to access the actual hard disk.

SATA ("Serial ATA") is a newer type of interface, using a completely different cable and optional different power connectors. SATA HDDs operate at higher frequencies, currently 150 MHz or 300 MHz as opposed to the 133 MHz of a PATA HDD. This allows for greatly increased sDTR, typically around 80 MB/s for SATA-150, and about 150 MB/s for SATA-300.

PATA is older technology, but some would argue more reliable. I disagree. I found it no more reliable than the SATA connection I currently use. Plus, SATA supports RAID ("Redundant Array of Independent Disks/Drives"), which allows much greater data security, although additional HDDs are required.

Back to top

Number of Drives

RAID is a standard covering disk arrays, or drive arrays. A disk array is a group of disks acting as a single disk. There are two basic types of RAID arrays, each requiring a minimum of two SATA HDDs—RAID arrays can be set up using PATA HDDs, but only with additional specialized hardware or software. The first type, RAID-0, is also referred to as a "Striped" array. This essentially means that the two HDDs are combined side-by-side like the two pages of a book. Unlike a typical book, however, each line of the page would be written from the left-hand side of the left page clear over to the right-hand side of the right page before carriage-returning to the next line. When writing to or reading from a RAID-0 array, both pages are written or read at once since each HDD is controlled independently. The overall effect is that the sDTR of the array is double that of an individual HDD. In order to maintain compatibility, both disks must be of equal capacity independently, but the total capacity will be the sum of the two disks' capacities. If the disks are different sizes, the total capacity will be twice the smaller of the two capacities. To summarize:

The second type of array, RAID-1, is also referred to as a "Mirrored" array. In this configuration, identical data is written simultaneously to each HDD, and the array is seen by the operating system as a single HDD, with the capacity and sDTR of a single HDD. If the array is constructed from unequal HDDs, the capacity will be that of the smaller disk, and the sDTR will be that of the slower interface, so a 300-GB SATA-150 HDD mirrored with a 150-GB SATA-300 HDD would be seen by the operating system as a single 150-GB SATA-150 HDD. This is true redundancy and provides excellent data protection. If a drive fails, every byte of data is still perfectly intact on the other.

With these two array types alone, one's choices would be limited to increased performance or increased security and never the twain should meet. Gladly, such is not the case. There are more advanced array types that can be constructed with more HDDs, and are typical of corporate servers and data archives, such as RAID-5 and RAID-6 configurations. The downside, of course, is cost. Unless you create a large enough array to get a volume discount, you will pay as much for the second (or third, fourth, or ninth) HDD as you did for the first.

If the cost of a HDD array is too prohibitive, go with a single SATA drive over a PATA. There is little cost difference between the two; the SATA drive will be faster and more reliable in and of itself, and the technology is more recent, so it will take longer to reach obsolescence.

Back to top

Access Speed

This is not the same as sDTR. Access speed refers to how quickly the disk will return a requested datum once it's been accessed. Generally this is a function of several factors, including the size of the buffer, type of interface, and even the rotational speed of the disk. Rather than attempt to calculate the access time from these parameters, most HDDs will have a standard access time listed on their spec sheet. For example, a randomly chosen 300-GB 7,200-rpm SATA-300 HDD listed an access time of 9.3 milliseconds (ms). Another randomly chosen 150-GB 10,000-rpm SATA-150 HDD listed an access time of 4.6 ms.

Practically speaking, this is probably not that important, as the access time is rarely experienced as a real lag. I recently suffered a crash and replaced the failed HDD with a HDD that was virtually identical in every technical specification, except that it showed a faster access time. I've noticed no difference.

Buffer Size

The HDD Buffer consists of a small memory chip that acts in much the same way as the L2 Cache of the CPU; it acts as a facilitator and ferryman for the data flowing back and forth between the hard disk and the host controller on the motherboard. The maximum data transfer rate discussed above is the rate at which data can flow along the drive interface (the SATA interface). A larger buffer will allow for somewhat more efficient data transfer and less data swapping during large transfers. Larger drives almost have to have larger buffers in order to compensate for the inevitable lag when accessing files at either end of the drive. Six or seven years ago when a 20 GB HDD was considered large, a HDD buffer of 2 MB was normal, perhaps even impressive. Today's 300- and 500-GB HDDs have buffers of 8, 16, and 32 MB.

Recommendations

Gamers:

• Low End: Single 200+ GB SATA-150 HDD
• High End: Two 200+ GB SATA-300 HDDs in a RAID-0 array, plus a krunk-looking case

Surfers:

• Low End: Single 300+ GB SATA-150 HDD
• High End: Two 400+ GB SATA-150 HDDs in a RAID-0 or -1 array

Workers:

• Low End: Single 200+ GB SATA-150 HDD
• High Performance: Two 200+ GB SATA-300 HDDs in a RAID-0 array
• High Security: Three 200+ GB SATA-300 HDDs in a RAID-5 array

Back to top

Next Step: Selecting Additional Drives