Site: Fujitsu Ltd.
Fujitsu Laboratories Ltd.
(optical storage presentations)
Peripheral Systems Lab., Atsugi Facilities
10-1 Morinosato-wakamiya, Atsugi
Kanagawa 243-01, Japan
http://www.fujitsu.co.jp/index-e.html

Date Visited: 11 March 1998

WTEC:

Hosts:

BACKGROUND

Fujitsu, Ltd. was established in 1935. It employs 46,795 employees in Japan and 165,056 worldwide. Its capitalization is about two billion dollars, and consolidated sales in 1996 exceeded $35 billion. Main product categories include computer and information processing systems (66%), telecommunication systems (19%), electronic devices (11.4%) and other operations (3.6%).

Fujitsu Laboratories was created in 1962 through the merger of R&D sections previously managed by separate technical divisions. In 1968, Fujitsu Laboratories, Ltd. was spun off as a wholly owned subsidiary of Fujitsu, Ltd. It has facilities in Kawasaki, Akashi, Atsugi, Numazu, Makuhari, and California. It employs 1,500 employees and is capitalized at $50 million. In 1993, Fujitsu Laboratories, Ltd. was reorganized by field rather than location and now conducts R&D into core multimedia technologies in a wide range of fields including telecommunications, information processing, multimedia systems and devices, personal systems, semiconductors, peripherals and terminals, and materials. The research is undertaken at the following seven laboratories:

There are also a number of project groups working across these boundary lines.

The Peripheral Systems Laboratories develops magnetic and magneto-optical disks, input/output devices (e.g., scanners, plasma display panels, data compression), inspection and manufacturing technologies and mechatronics. The Peripheral Systems Lab employs 250 researchers and another 1,000 employees in Thailand for mass manufacturing. One hundred fifty researchers are involved in MO research.

[Peripheral Systems Laboratories]
Fig. C.1. Peripheral Systems Laboratories.

Storage systems are being developed jointly at Fujitsu Storage Products Group and Fujitsu Laboratories, Ltd., Peripheral Systems Laboratories, as described in Table C.1 below.

Table C.1
Organizational Chart

Fujitsu Storage Products Group

Fujitsu Laboratories, Ltd.

  • Storage Technology Labs.
    • File Memory Lab.
    • HDI Eng. Department
    • Development Department
  • Storage Component Division
  • HDD Division
  • Optical Disk System Division
  • Int. Eng. & Mfg. Support Division
  • M Project Group
  • Peripheral Systems Lab.
    • Magnetic Disk Lab.
    • Optical Disk Media Lab.

Fujitsu is concentrating on producing drives; however, it has the capability of producing several thousand media units that are used for testing purposes. The recording media group performs sputtering, R/W testing and other measurements. The substrate group performs mastering, patterned HDD, injection molding and substrate testing.

DISCUSSION

Comparison with Other Storage Products

MO appears to be central to Fujitsu, Ltd.'s plans for multimedia storage systems, and management perceives 3.5" magneto-optical disks as "the personal files of the multimedia era." This disk format is expected to carry information from office to home or onto portable PCs and for information sharing. In 1995, the number of drive units (230 MB/disk) shipped reached about 800,000 units/year, roughly a 3x increase from 1992 (128 MB/disk). Present shipments (640 MB/disk) average around 1.5 million units/year and Fujitsu researchers estimate future shipments to exceed 5 million units/year shortly beyond 2000. Indeed, in Japan MO drives have 52% of drive market share for removable media and 32% of users have both MO and CD-R drives. CD-R only (4.8%), PD (3.2%) and Zip (6.3%) have relatively small market shares.

An interesting item that opened the discussions was the difference between the Japanese market and the U.S. market for removable storage where the roles of the Zip drive and MO are reversed. Discussions centered on cost per drive, backward compatibility, time to market, and product quality. Also resistance to foreign products in both countries by end users was brought up among possible reasons.

Dr. Ogawa then compared the performance of MO with HDDs and noted that the performance gap between these two technologies has practically disappeared. Indeed, under certain conditions MO can be faster than HDDs in writing but HDDs have a slight advantage in read operations. The reduction in the gap is attributed to "light intensity modulation and direct over write" (LIMDOW) technique that uses over write and an ultra cache to speed up data transfer rates up to 5 Mb/s (asynch) and 10 Mb/s (synch). With a rotation speed of 4300 rpm typical MO seek times are on the order of 28 ms. This performance is sufficient for MPEG II coded motion picture.

Indeed, Fujitsu managers see the following advantages in MO:

Next Generation Products

Next generation 3.5" MO products will have a 1.3 GB capacity and will be followed by 3 GB products. The aim of Fujitsu is to produce one medium for a full hierarchy of personal applications. To this end the company is participating in the Advanced Storage MO (ASMO) Alliance for a 6 GB MO disk for MPEG II motion picture storage that uses DVD format. Fujitsu, Olympus, Sharp, Imation, Sanyo and Hitachi-Maxell are among participants in ASMO. Magnetically-induced super-resolution (MSR), which was proposed some time ago, is one of the key features being used in ASMO. It utilizes the fact that the beam spot is hot in the center and cooler at the periphery. A magnetic film domain is formed on top of the MO recording layer, and operates as a magnetic mask depending on temperature differences. When multiple recording marks are located within a beam spot, a single mark alone can be shown through a "window" in the mask. This prevents interference between read signals. Thus ASMO will provide MO disks offering about 5 Gb/in2 surface recording density. The standards were decided on in the summer of 1997 and production was scheduled to begin as early as late 1998.

Aiming at 70 Gb/in2

To boost the surface recording density to over 10 Gb/in2, however, even MSR is unstable. When the MSR mask is made larger, resolution rises, but at the same time the signal/noise (S/N) ratio degrades. The ratio of light reflected from the mask increases, adding to the noise in the read signal. With a 680 nm laser diode, the top end of surface recording density using MSR is thought to be about 7 Gb/in2.

To increase the performance of MO systems Fujitsu counts on progress to be made on three different fronts. On the magnetic front the company is considering the use of magnetically induced super-resolution (MSR), magnetic field modulation, magnetic amplifying magneto-optical system (MAMMOS) from Hitachi-Maxell or domain wall displacement detection (DWDD) from Canon. On the optical side, the use of shorter wavelength laser diodes, larger NA lenses, solid immersion lenses (SIL), and near-field optics are the subject of current studies. Finally, various approaches grouped under multi-value recording are being evaluated. These include multi-edge (SCIPER), multi-layer and parallel recording using multiple wavelengths.

The magnetic amplifying magneto-optical system (MAMMOS) jointly developed by Hitachi Maxell, Ltd and Sanyo Electric Co, Ltd. applies the concepts of MSR in reverse. Instead of masking inside the beam spot, the tiny recording mark in the recording layer is enlarged to read. Because most of the reflected light contains signal, S/N degradation is minimized. With a 680 nm light source it is possible to reach 20 Gb/in2, and in the future blue-purple light source could achieve 70 Gb/in2. With a surface recording density of 20 Gb/in2, the same 120 mm diameter disk as is currently used for digital video disk (DVD) would hold 30 GB of data per side. This storage capacity is sufficient to hold 10 hours of compressed motion video data, with images of standard television broadcast quality. A capacity of 90 GB would be provided at 60 Gb/in2. However, Fujitsu researchers pointed out difficulties associated with production since synchronization with the magnetic field is difficult with the MAMMOS technique. It was suggested that perhaps DWDD technique might be more promising. However, a 30 GB product was suggested as a desirable capacity for video-MO systems for media to be shared between TV and PCs in a post-VCR era. Data transfer rates of 30 Mb/s would be required. A possible roadmap suggested by Dr. Ogawa is provided in Table C.2.

Table C.2
Possible Roadmap
 

Present (1997)

5 Years (2002)

10 Years Later (2007)

Recording Density (Gb/in2)

1.17

20

200

Capacity 3.5 inch (GB)

120 mm *GB)

0.64

2.6

 

(DVD-RAM)

15

36

150

360

Transfer Speed (Mb/s)

20

50~100

100~300

Seek Times (ms)

30

10~20

5~10

Optical Systems

680 nm

NA 0.55

410 nm

Large NA lens

205 nm

SIL

Media

 

Magnetic

Expansion

Multi-value &

Multi-layer

Video MO

Video MO with 36 GB capacity and 30 Mb/s speed can be used both for TV and PC applications. For PC applications it will hold 5 hours HTTV, will be capable of multi-TV channel recording or could be used as a personal video server. It would have direct link to peripherals through a IEEE 1394. On the PC side, for content management the video MO would enable easy video stream editing. It would be capable of fast index search, multi-TV channel, summary scene and fast feed playing.

Speculative Research Directions

It was pointed out that it was the proper time to start research for 0.1-1 Tb/in2 areal density storage products. For example certain applications such as camcorders require 100 Gb/in2 densities and at those densities mechanical positioning is crucial. Therefore piezoelectric and MEMS actuators are important directions to look at. Near field optics (NFO) is also a promising direction, and Fujitsu is working with the University of Tokyo for simulation of the physics. More importantly, Fujitsu is pursuing research for parallel addressing of MO drives using VCSEL arrays (Multi Beam Optical Head Project) jointly with Prof. K. Goto's group at Tokai University.

LAB VISIT

The WTEC team was shown an apparatus that was capable of packaging at high speeds laser diodes and photodetectors with 5 micron and 0.5o alignment accuracy by using robotic vision techniques. This machine would be capable of producing integrated optical heads and also can be applied to fiber optic products.

OTHER DISCUSSION


Published: June 1999; WTEC Hyper-Librarian