Hardware / Storage Devices
A trio of researchers from the National Physical Laboratory in the UK have proposed an entirely new read head mechanism that has the potential to overcome some of the looming barriers towards increased data storage density. First, there is bit size and bit spacing. Bits are stored as tiny ferromagnetic crystals, and it is their orientation that signals a one or a zero. These bits can only be made so small and placed so close together before they start to interact strongly enough to flip each other.
Another problem is that the read and write heads must be small enough so that the orientation of a single bit dominates the signal from the head. This means that the heads must be about the size of the spaces between the bits. Then there are problems of heat and power dissipation. Read heads have a small but continuous current flowing through them, constantly draining your laptop battery and heating the read/write head. Reducing the size of the read head means that it will get hotter, which causes problems of its own. Excess heat could cause the read head to flip the bit that it is reading.
Fundamental to these problems is the way that the read head works. The magnetic field induces a small change in the resistance of the read head, which is then detected by measuring the voltage drop across it. The structure that supports this is relatively large, and the prospects for reducing it further are not too good.
The read head structure can be simplified considerably, however, by using a combination of magnetostrictive and electrostrictive material (a magnetostrictive material changes its size in a magnetic field and an electrostrictive material changes its size in an electric field). When layered, these combine to make a material that is termed magneto-electric, which essentially means that a magnetic field is used to generate a current. When the electrostrictive material is sandwiched between two magnetostrictive materials, the magentostrictive material tightens or loosens its grip every time the read head passes over a bit. Squeezing the electrostrictive material generates a detectable current.
The sensitivity per unit area is about the same as for more traditional giant magentoresistive and tunneling magnetoresistive read heads, but the layering requirements are much less strict. The researchers propose a device that is just 7 layers rather than the 15 in today's hard drives. Furthermore, a current is only generated when the read head passes over a bit and the resistance is very low, so the power dissipated in the read head should be less. Beyond that, it doesn't require that pesky constant current source, so it should consume less power and require fewer components.
The biggest advantage is probably in the simpler read head construction. Current read heads need a biasing magnetic (a magnet that generates a relatively strong background magnetic field against which the bits on the disk operate) attached to them, and the authors claim that putting this magnet on the head takes about 100 production steps. Even if no other benefit is derived from the technology, that is probably enough to make it worthwhile to manufacturers.
Finally, a note of caution. This is a simulation of a read head, not the real thing. No experiments have been done, and although I think the idea will work, I have one small concern. Because the read head is now generating a fluctuating current, it is generating a magnetic field that is of comparable strength to that of the bits themselves. The generated magnetic field will oppose the direction of whatever bit is currently being sensed. I want to be sure that the read process is not going to be flipping any of my bits before you can sell me that hard drive.
If there are no major technical gotchas, then this technology should provide a factor of two increase in storage density in its first generation. Beyond that, it's much harder to prognosticate on, but I think it is safe to say that the doubling in storage density every year is set to continue for a while longer.
Journal of Physics D, Applied Physics, 2007, DOI: 10.1088/0022-3727/40/17/003
If you're an experienced IT professional and would like to share your expertise with other Setup32.com readers, why not submit a new article to our Hardware area?
Pass your MCSE certification exams with EasyCert - the most realistic exam simulations available today! Click the links bellow for free MCSE demos.