Over the last 15 years the storage capacities of hard disk drives has risen by a factor of over 1000. Part of that increase is due to the number of platters and heads inside a drive. Yet a bigger part has come through greater tolerances and better manufacturing techniques. We’ve moved from 5.25 inch drives to 2.5 inch drives today, and lately we’ve even reduced the number of platter and heads by increasing the areal density. [The areal density is the amount of data per square inch of media].
Yet the way forward to greater capacities and higher performance drives is changing the way that data is written to the surface of the disk. Current technology relies on Longitudinal Recording (LR) of data. This means that the positive and negative charges of each piece of data are written parallel to the disk surface. More importantly, they are also in the same plane, as opposed to being multi-layered like some optical technologies.
The amount of data that we can store on a disk surface, in this way, is approaching its limit. As we seek to reduce the size of the charged particles we increase the risk of distortion. As a result increased areal density, the holy grail of disk manufacturing in recent years, is becoming ever hard to meet.
Increased areal density has been the key to reducing the costs of drives, by removing the number of platters and heads. It has also allowed the disk format to reduce ensuring that high capacity hard disks have moved from the desktop and server market into the notebook, personal video recorder (pvr), mp3 player and other consumer technologies.
As we keep pushing the areal density of storage higher, we need to find ways to overcome a serious problem - the superparamagnetic limit. As we squeeze more and more data onto the surface of the media, the energy used to maintain the data has to be reduced to prevent corruption. Eventually the energy used to maintain each bit of data becomes so small that it is indistinguishable from the background energy created by the drive. At this point the entire recording field collapses in on itself.
One way to get round the problem is to increase the coercivity (the magnetic field required to write data) of the disk. Material scientists believe that we are fast approaching the limit of what we can achieve in both the disk head and the magnetic media.
To regain these advances in areal density disk manufacturers have been investing heavily in a new technology called Perpendicular Recording (PR). Where LR works in a single plane parallel to the disk surface, PR works at right angles to the disk surface piling the bits on top of each other. The concept is not new. In architecture we have long built large skyscrapers to house more people. Disk technology is finally catching up with that.
How PR works
So how does Perpendicular Recording work? In LR recording, data is laid down by the heads as they pass over the medium. Magnetic particles are magnetised and set out as bits of data. The heads float over the surface of the media emanating magnetic fields from the head. The fields do not just address the media directly under the heads, they also address media in very close contact. Think of a magnet passing across a tray or iron filings. It affects more than just those it passes over.
In PR, the head is much smaller and addresses not only the existing surface of the media but also a new layer known as the SUL (Soft UnderLayer). By using just a single pole, the magnetic field addresses just the media directly under the head which, by comparison to a LR head is also much smaller. Rather than a magnet over a tray of iron filings you now have a pair of tweezers passing over the iron filings.
(Part 2 of this feature discusses how close we are to seeing real products and which suppliers might deliver them.)