Imagers speed holographic-data retrieval

Micromirrors and CMOS imagers combine to write and read digital data.

Mar 1st, 2007
Th 0703vsd Imagers01

Micromirrors and CMOS imagers combine to write and read digital data.

By Andrew Wilson, Editor

The development of holographic-data-storage products has been limited by the lack of low-cost system components, the complexity of holographic multiplexing techniques, and the absence of suitable recording materials. However, the introduction of digital micromirrors to the consumer display market and CMOS-based active-pixel detector arrays for high-speed machine-vision applications is set to change this scenario. While micromirrors can effectively be used as spatial light modulators (SLMs), high-speed CMOS imagers can read digital data contained within the holographic medium.

In traditional optical-memory-storage devices, data are recorded only on the surface of the medium. To increase the data-storage capacity, holographic techniques can be used to record data at multiple levels within the media. While other technologies record serially onto data media, holographic techniques allow millions of bits of data to be written and read in parallel. This enables transfer rates significantly higher than current optical storage devices.

Light from a single laser beam is split into a signal beam that carries data and a reference beam (see figure). At the point where these two beams intersect the recording media, a hologram is formed. To encode data onto the signal beam, a SLM translates digital data into an optical checkerboard pattern of light and dark pixels. The data are arranged in an array or page of about a million bits.


To form a holographic image onto a recording medium, light from a single laser beam is split into a signal beam that carries data and a reference beam. At the point where these two beams intersect the recording media, a hologram is formed. To read the data at high speeds, the reconstructed array is projected onto a CMOS imager.
Click here to enlarge image

At the point of intersection of the reference and signal beam, a chemical reaction occurs in the storage medium causing a hologram to be formed. By varying the reference beam angle, wavelength, or media position, many different holograms can be recorded in the same volume of material. To read the data, a reference beam deflects from the hologram in the storage media that reconstructs the stored information. This hologram is then projected onto a CMOS imager that reads the data in parallel.

LEGACY CHALLENGES

In the past, three major challenges have faced developers of holographic-based storage system. The fabrication, production, and use of expensive SLMs were limited to mainly military applications. Now, the advent of low-cost devices such as the digital light processor (DLP) from Texas Instruments offers an opportunity for memory storage vendors to use DLPs as effective SLM replacements.

Perhaps most important, the major challenge to implementing holographic storage has been the development of a suitable storage medium. Now, InPhase Technologies, a Bell Labs spin-off, is set to capitalize on commercializing the more than seven years of development work Bell Labs has already invested in the technology. According to InPhase, storage densities of 31.5 Gbits/in.² (a density that would yield approximately 45 Gbytes on a 5 1/4-in. disk) have been demonstrated by recording and retrieving more than 3000 digital data pages. Newer “two-chemistry” materials and custom optics can store data at greater than 100 Gbits/in.² densities. Although InPhase is developing a rewritable material and system, others are also being developed by DCE Aprilis, a company recently acquired by Dow Corning.

“In the development of these drives,” says Tim Baeyens, director of marketing and sales at Cypress Semiconductor, “it is very important to read large amounts of data quickly and in parallel from the reconstructed array as it is projected onto a CMOS detector. To do this, Cypress has developed a 1696 × 1710-pixel CMOS detector array, using 8 × 8-µm pixels, that features LVDS outputs and a frame rate of a few hundred frames per second.

“This,” says Baeyens, “allows the imager to transfer data at rates of a few giga-pixels per second from the device.” Better still, because the device features an on-board 8-bit ADC, digital data can be directly output. “This is important in embedded applications,” says Baeyens, “where reduced system costs are important.” Costing less than $1000 in large quantities, this type of imager will initially be targeted toward tape-storage replacement systems and ultimately toward consumer storage drives.

Click here to enlarge image

null

More in Image sensors