Michael B. Martin [Michael B. Martin is a retired electroacoustics engineer who spent approximately half of his 40-year career designing consumer and professional recorders (analog and digital) and the other half in the field of magnetic recording tape. The last three years of his career were devoted to optical recording.]
Prev: Recordable CD: Promises & Problems (part 1) (Jun. 1992)
Magneto-optical recording, discussed last month, is hardly the only record able CD system being investigated: Phase-change and dye-polymer technologies are also in the running. Phase-change materials are those which, when heated, change from amorphous, non crystalline molecular arrangements to organized crystalline structures or vice versa. In thin layers of metallic phase-change materials, such as alloys of antimony and selenium, the change from the amorphous to the crystalline phase causes a change in reflectivity.
This change is triggered when the layer is heated to about 170° C (338° F). It is irreversible and can therefore be applied to WORM media. If a disc coated with a thin layer of antimony-selenium is heated by a modulated recording laser, data can be recorded as spots of differing reflectivity. When a recorded disc is read by a playback laser, the variation in reflected light permits the data to be recovered.
Other materials have the property of reversible phase change, making them suit able for erasable media. Complex combinations of rare earths and elements, including layers of gallium antimonide with indium antimonide and tellurium alloyed with elements such as germanium and indium, have been developed. In their neutral state, these compounds are usually in the stable crystal phase, but when heated to a temperature a little above their melting point, the heated spot solidifies in the amorphous state. Be cause the crystalline phase is more stable than the amorphous, the coating will tend to return to this form; thus, erasure can be achieved by heating the recording layer to a temperature just below the melting point, which triggers a return to the crystalline phase and erases the data. Erasable phase-change media can be used as WORM or archive systems by increasing the power of the recording laser to the point where a hole is melted in the coating, thereby creating a permanent record.
The difference in reflectivity between a crystalline and an amorphous surface is typically 30% or less. The specification for a commercially produced CD calls for a reflectance between 70% and 90%. The overall reflectance of a phase-change surface does not reach 70%; therefore, reliable play back of a phase-change recording will require CD players to be more sensitive to low levels of reflectance and to the relatively small changes in this reflectance generated by data. This does not create any major technical difficulty, but it is probable that many players now in the field will not be capable of satisfactory performance; there fore, phase-change recording can only realistically be considered a forward-compatible system. Another difficulty with phase change appears to be with volume production at an economical cost: Years of work by several technically powerful companies have not yet yielded a completely satisfactory process.
Both private and public demonstrations of prototype systems have been made, but there has been no announcement of a system at consumer prices.
The materials used for the active coatings of phase-change discs are somewhat similar chemically to those of MO discs; therefore, it is not surprising that the corrosion problems can be similar. As with MO, the maximum information density that can be recorded is a function of the wavelength of the recording laser, and it is diffraction-limited.
The recording systems known as dye polymer rely on the use of optical dyes as absorption filters. All dye-polymer media use the principle of energy absorption by an optical dye loaded into a layer of normally clear polymer that has the same refractive index as the plastic or glass substrate of a disc.
When a laser beam passes through a plastic layer containing a dye that absorbs at the wavelength of the laser, the layer is heated very rapidly (hundreds of degrees per micro second). A dye-polymer medium can be de signed to achieve one of several possible results. The heat generated can be used to burn a hole in the dyed layer, bleach the dye, locally vaporize the surface of the substrate (thus forming a pit in the substrate surface), or cause rapid local expansion of the dyed layer to create a bump or pit.
The majority of dyes in existence were developed for applications such as textile coloring and photography; few are suitable for the dye-polymer field. Suitable dyes can be and have been synthesized in small quantities; some dyes designed for other applications have been used for experimental studies but have less than desirable efficiencies. It is a drawback of all dye-polymer systems that even if all the media used the same dye and the annual production were to be in the tens of millions of discs, the dye requirement would still be insufficient to justify even a small manufacturing unit for this special dye. The established dye manufacturers are conditioned to producing vast quantities for traditional applications such as textiles and photographic products. They do not become enthusiastic about producing annual quantities in the 1-ton region; they only begin to get interested at 1,000 tons! As a result, dyes for optical recording will be produced with laboratory-scale equipment at very high cost. Fortunately, because of the minuscule amount of dye used in a disc coating, the impact on the final cost of the media will be very small.
Systems that burn holes in the active coating are known as ablative systems. Obviously, an ablative disc is a WORM medium, but it has the advantage that, by coating the polymer with a reflective layer that is burned off with the polymer, the disc can support a CD-compatible recording. The master discs for a commercial CD are made by an ablative system, i.e., a photo-resist coating on a glass substrate is burned off using an argon-gas laser. The beam produced by an argon laser is blue in color and has a wavelength of 488 nm, which permits the production of precise pits at a very high density.
Unfortunately, practical laser diodes now available to de signers of consumer recorders emit at much longer, near-infrared wavelengths (840 or 780 nm) instead of blue. As a result, the typical holes prod aced in a dye-polymer layer are larger and less well formed than the pits in a commercial CD.
Large holes force a lower information density than can be achieved by CD manufacturers, resulting in an overall play time of perhaps 60 minutes for an ablative disc, much less than the 78 to 80 minutes possible for a CD.
The development of laser diodes that emit a shorter wavelength of light will definitely improve this situation because they will permit proportion ally smaller holes to be burned into the coating of the disc.
A dye-polymer approach to making an erasable and compatible CDR disc is the multi-layered layer-distortion or bump-forming system. A laboratory sample was demonstrated early in 1988, but there have been no public demonstrations since then. As with all dye-polymer approaches, the bump-forming system depends on optical dyes absorbing energy from a laser beam, with the difference that there are two layers that absorb at wavelengths different from each other. Two possible configurations for bump-forming media are shown in Fig. 5. The bumps produced by the upper structure (the system demonstrated in 1988) will project away from the substrate. This allows the recording, play, and erase lasers all to operate from the substrate side. The bumps produced by the lower structure will project towards the substrate, and the record and erase lasers will operate from the coated side, opposite the play laser on the substrate side. With the first type, the recorded media will be fully compatible with all CD players that have three-beam tracking. A two-laser recorder will be possible because one of the lasers can be dual-purpose, used in a high-power mode for erase or recording and at much reduced power for playback. With the second type, the recorded media will be compatible with CD players that have either single-beam push-pull or three-beam tracking, but the recorder will be more complicated. Also, the recorder will require three lasers-one each for recording, playing, and erasing.
The principle of operation is exactly the same for both structures. The dye in the expansion layer absorbs energy from the re cording laser and appears transparent to both the play and erase laser beams. The dye in the retention layer soaks up energy from the erase laser and is transparent to the other two beams. At normal temperatures, the retention layer is rigid compared to the expansion layer, which is relatively elastic.
When the recording laser is energized, the expansion layer is heated to very high temperatures in picoseconds and expands locally with explosive force. The retention layer is thus heated by conduction to its softening point and deformed into a bump by the swollen expansion layer. The reflective layer is in turn deformed by the retention layer. When the recording laser is de-energized, the retention layer becomes rigid before the expansion layer is fully contracted and the bump becomes permanent, with the expansion layer in a state of tension. Erasure is achieved by heating the retention layer to its softening temperature with the erase laser, which allows the tension in the expansion layer to relax. The relaxing expansion layer pulls the bump flat in all three layers, thus erasing the information.
The principal attraction of the bump-forming technology is that presently it is the only one which has a chance of being fully compatible with commercial CDs. Unfortunately, several difficult problems must be solved be fore bump-forming discs can be considered fully developed and out of the laboratory.
Some of these problems are:
The number of times the disc can be erased is limited by fatigue in simple reflective-layer materials or the formation of small artifacts in the polymer layers. Both of these phenomena worsen the signal-to-noise ratio and decrease the response to the shorter information pits.
Fig. 5--Cross-section of two types of bump-forming media. Note how each dictates
a different laser layout in the recorder.
Unusual eutectic alloys of metals with low melting points are needed to overcome the reflective layer fatigue problem and will also require the development of new coating techniques.
Disc-controlled tracking during the recording process is difficult, because the refractive indices of the polymer layers and polycarbonate are the same.
Existing dyes are far from ideal, and the close proximity of the wavelengths of avail able diodes (780 and 840 nm) makes the development of special dyes very difficult.
The introduction of 640-nm laser diodes will simplify the development of dyes whose absorption spectra do not overlap; they will also improve the recorded information density. (See Fig. 6.)
Fig. 6--Dye-polymer absorption and laser wavelengths; see text.
Interrelationships of Hardware and Software
As a product for home recording, record able CD will have unique relationships with commercially produced software. Unlike any other home recording medium, either audio or video, when CDR is introduced into the consumer market a huge library of program material will already exist on a non-erasable medium--the commercial CD. CDR will have a structure different from that of the CD, unlike magnetic tape systems where both the blank media and the prerecorded product have identical structures and are recorded by essentially the same process. With the cassette and DAT systems, the prerecorded library came after the blank media system was launched; DCC will fit into this pattern. In all probability, CDR will also have quite a different appearance from the well-known shiny CD. For example, a dye-polymer disc will have a blue or cyan color because its polymer layers absorb red light. The relatively high cost of CDR discs, coupled with the difficulty of high-speed duplication, is likely to make CDR an uneconomical system for high-volume commercial applications, including piracy.
As with other memory systems, CDR must be developed with an eye towards both the price of the media and the price of the recorder. Many blank tapes or discs are purchased for each recorder. Thus, a small increment in the cost of the media to the consumer can result in a much higher total operating cost over the life of a recorder than will a larger addition to the recorder's initial cost. An important consideration is whether to have tracking during the recording pro cess controlled by the disc itself, by a head-positioning mechanism in the recorder, or by a combination of the two. In addition to cost considerations concerning the tracking method, there are significant technical problems to be solved with any technique. Currently, the general opinion is that recorder-controlled tracking is not cost-effective.
The reasons for the strong bias towards disc-controlled tracking are to be found in the physical dimensions of CD tracks and the data recorded on them. (These dimensions were given last month in Table I.) At a pitch of 1.6 there are over 15,000 tracks per inch, and the information pits are comparable in size to smoke particles. The accuracy required of any mechanism that must work to these dimensions is almost beyond comprehension. It is believed that the pits in a CD are the smallest artifacts produced in a controlled manner by any manufacturing process. The lathe used for mastering a CD is built like the proverbial battleship, the motion of the head carrying the argon laser is totally computer-controlled, and the position of the sled carrying the head is controlled by an elaborate feedback system which constantly refers to the pits just cut by the laser.
The sled is mounted to the lathe on air bearings in order to eliminate any irregular motion generated by mechanical imperfections in the supporting unit. In addition, the lathe is mounted on a massive steel plate or granite block which is isolated from outside vibration by complex shock absorbers. Within the limitations of a consumer-oriented recorder, isolation of this order is impossible to achieve.
Consumer CD players track satisfactorily by virtue of the fact that they follow the information recorded on the disc. Therefore, it is logical to believe that recorders can be de signed to track to the necessary tolerances if tracking information is buried in the blank discs. The favored means of providing guidance for the recording head is to mold a spiral groove into the disc's transparent substrate on the side to be coated with the recording material.
The molding die used to produce a grooved substrate is made in a manner similar to the way a CD mold is fabricated, and the master groove is cut on a CD mastering lathe. The dimensions of the groove are chosen so that interference takes place at the wavelength of the recording laser. Similarly, the tracking servo is designed to position the recording head in the area of minimum interference, in other words, between the grooves. When the recording is played back, there is no interference from the grooves because the playback head tracks on the recorded information and does not see the grooves in the substrate. A problem arises with this particular system if the refractive index of the active coating is the same as that of the substrate and the coating flows into the tracking channels. If this happens, the grooves become invisible to the recorder. A solution to this difficulty is to add a primer coating which fills the grooves and makes them visible to the recorder's laser. In dye-polymer systems, the primer may have an adverse impact on the overall performance of the media.
The recorder's servo system must be tolerant to the errors created by disc run-out created by the interface between the mounting chuck and the disc's center hole, flatness errors in the disc itself, and molding errors in the groove. These errors can add up to eccentricities that are equal to several times the pitch of the groove and take place at least once per disc revolution. Some disc-con trolled tracking systems use a more complicated guide than a simple groove, such as having recorder-control information embedded in the pilot channel.
All tracking controls embedded in the disc substrate increase the manufacturing cost of what is already an expensive piece of plastic. The substrate blank for a CDR must be free from optical and mechanical flaws which can cause errors, such as dropouts, during the recording process. The tight specifications for substrates result in a higher manufacturing cost than that of a finished CD.
When the costs for coating materials and processing are added, the result is a high basic cost. This generates a retail price for a blank CDR greater than that of a commercial CD that already includes music.
The high costs associated with the blank discs cause the developers of the media to pressure the designers of recorders to consider ways of having the recorders control the tracking. No recorder designer has yet developed a cost-effective tracking system.
Also, as discussed above, the technical difficulties are such that unless a brilliantly conceived floating-point servo is designed, which automatically isolates the focused light beam from outside influence, a recorder-controlled tracking system will never appear on a consumer machine.
Conclusion The audio community has been allowed to believe that the principal block to CDR is opposition by the music industry to digital recording equipment for the consumer market. The release of Digital Audio Tape recorders to the stores suggests that the music industry's objection is no longer the overriding problem facing home digital media such as the recordable CD, especially since the Serial Copy Management System now built into consumer-level DAT recorders to prevent multiple copying can also be included in a DAD machine. It is far more likely that the important reasons why consumer CDR has not appeared are that the systems the engineers are offering to the marketeers have too high a level of projected costs and that they lack compatibility with CD. Without CD compatibility, rapid random access to recorded programs is the only advantage CDR has over such less costly alternatives as DAT and DCC. Quick random access may well be important to the professional user, but it may not weigh very heavily in a consumer's decision to buy when DCC is available at a much lower price. If the only compatible CDR avail able is a WORM system, it is questionable whether the ability to play the recordings on standard CD players will outweigh the severe disadvantage of not being able to correct mistakes and erase and reuse the media.
The CD is a remarkable and successful product, one of those rare developments that is a technical success, a commercial success, and a revolutionary product. The CD has caused rapid and permanent change in the audio industry as a whole because the convenience in use and the matchless quality of the reproduced sound proved to be irresistible to the public. Nevertheless, there is a definite possibility that a commercially successful CDR will never exist.
Bibliography
For further information, the following books are recommended. They were of great assistance to me in the preparation of this article and are very valuable for the extensive reference bibliographies given in each.
-M.B.M.
1. Pohlmann, Ken C., Principles of Digital Audio, Second Edition, Howard W. Sams & Co. (Indianapolis, Ind., 1989).
2. Pohlmann, Ken C., The Compact Disc: A Handbook of Theory and Use, A-R Editions (Madison, Wisc., 1989).
3. Watkinson, John, The Art of Digital Audio, Focal Press (London, England and Boston, Mass., 1988).
4. Mee, C. Denis and Eric D. Daniel, editors, Magnetic Recording, Volume Video, Audio, and Instrumentation Recording, Chapter 6, "Magneto-optical Recording" by Dan S. Bloomberg and G. A. Neville Connell, McGraw-Hill (New York, N.Y., 1987).
(adapted from Audio magazine, Jul. 1992 )
Marantz CDR-1 Professional CD Recorder (Equip. Profile, Apr. 1992)W
= = = =