This series of articles is aimed at non-technical readers who nevertheless would like to learn a little of the background to Compact Disc, how CD players work, how to choose and use a CD player, and what exciting new developments the CD medium still has in store. The articles have been prompted by many letters from GRAMOPHONE readers. These make it quite clear that, whilst most of the first converts to the CD format were already knowledgeable collectors of LP vinyl records or prerecorded musicassettes, who had decided to add this third 'music carrier' to their existing audio systems, there were many complete beginners attracted by the new CD medium for whom the whole hi-ti mystique was like learning a new language.

Part 4 - How CD players work

The first thing that has to be said about a Compact Disc player is that it is not just a player, like a gramophone turntable or a tape deck. It is also a high-capacity storage medium, a high-speed optical data readout system and a powerful microprocessor. In fact, much less than half its circuitry is devoted to the audio signals. As a result, a CD player is a complex precision instrument mechanically, optically and electronically, despite the apparent simplicity of its operation and external appearance. Although the electronics are complicated, modern technology has reduced everything to just a handful of LSIs (large-scale integrated circuits) and so, once these LSIs have been designed and manufactured, final electronics assembly of a CD player is easy, uncluttered and relatively inexpensive. Traps still exist for the unwary, however, and poor layout or lack of attention to detail in the power supply or the few analogue components can materially affect the subjective performance.

The mechanical and optical design are less capable of simplification, yet here too important advances have continued to be made throughout the half-dozen years of CD's existence. For example, new aspherical lenses are being introduced which greatly simplify the optical readout system. In this way each new generation of players can, in the best examples, be found to possess significant improvements in stability and accuracy.

Mechanical questions
Some idea of the mechanical complexity inside a CD player may be gained by studying Fig. l which shows a typical mechanism (the Nakamichi OMS-7/II) in outline. Like the vast majority of CD players, this is a front-loader and the Eject/Load key can be seen at the front of the diagram, just in front of the disc loading motor. Pressing this key energizes the motor and causes the disc-drawer to glide in or out. When the drawer is firmly closed, in the loading sequence, the disc is secured to the drive spindle by means of a magnetic chuck or other device, the drive motor is switched on and also the laser light source. As a safety provision, the laser cannot be switched on when the drawer is open, though the low-powered source does not constitute a hazard of any seriousness, and is in any case focused only a few millimetres away from the objective lens. Similar reasoning applies to top-loading players such as the mini portables and some professional players, where closing the lid activates the loading sequence.

The tracking motor then carries the laser source and the other components in the optical readout system to the lead-in area on the disc (near the disc centre) so that the play sequence can be got ready. For this to happen, no fewer than three servo control mechanisms are needed: to monitor and carry out continuous fine adjustment of running speed, beam-tracking of the 'groove' and beam focus. These separate functions will now be discussed in turn, and can be seen in schematic form in Fig. 2.

The output of the optical readout system goes to a decoding system whose primary function is to carry out all necessary unscrambling of the digital audio signals and perform digital-analogue conversion to pro duce Left and Right analogue stereo outputs. At the same time, the decoding system feeds the control and display circuitry, and also a servo system for the three functions mentioned above, drive motor speed, tracking (in the Y direction) and focusing (in the Z direction).

Speed control
I have suggested that nothing inside a CD player is quite as simple as it appears, and this is certainly true of the running speed. As described earlier in this series, the CD digital pits track is recorded at a constant linear speed (nominally 1.3 metres per second, though any velocity in the range from 1.2 to I.4m/s is permitted on any given disc). The rotational speed of the motor is therefore continually slowing down, from about 500rpm at the centre down to about 200rpm at the outer edge. The drive motor servo has therefore to monitor some constant feature of the bit-stream (in fact the sampling frequency is chosen as being the most important parameter) and adjust the speed to keep this in synchronism with a sampling frequency crystal oscillator built into the player. During the start-up sequence, the correct speed has to be determined in a two-stage rough search and fine adjust process.

During all subsequent operations, such as normal play, track skip and fast search, the speed servo must again make rapid adjustments to lock the motor to the speed appropriate to each point across the disc. This is why, if a new track is selected for example, the sound will be muted briefly until the new speed has been esta blished.

Tracking control

The optical readout system is built on to a carrying arm which tracks it across the disc, just like a gramophone pickup. Again, as in the case of gramophone pickups, this will often be a pivoted arm (see the Philips example illustrated as Fig. 5 in Part I). Many other players use linear tracking with the carriage moving on a lead-screw along a straight radial line. Three kinds of tracking motion are needed: fast motion during track skipping; a fully controlled slow motion during normal play holding the laser beam dead central over the track being scanned; and a choice of quicker speeds during fast search.

Normal track following calls for very precise servo control. The track pitch is a mere, requiring the laser spot to follow the centre of the track within 0.1pm, whereas manufacturing tolerances for the player chuck and the disc hole may introduce eccentric 'swinging' of the tracks by as much as 300pm, and of course outside vibrations will present additional tracking problems. At least two track-following methods are in wide use, which we can call three-spot and one-spot.

In the three-spot method, used by Sony and others, the optical system is as outlined in Fig. 3 in which a diffraction grating splits the laser beam into one main spot plus two additional offset side spots E and F as indicated in Fig. 4. Separate sidespot detectors will produce unequal amplitude signals when there is a tracking error and instruct the tracking servo to re-centre the main spot.

In the one-spot method, used for example by Philips, there is no diffraction grating or beam-splitting. A single sensor divided into two halves is able to detect any unbalance of the reflected light beam due to off-centre tracking and drive the servo control accordingly. In some cases the tracking servo is modulated at a fixed frequency to help the sensor to identify the correct position. This extra complication may appear to be a disadvantage of the one-spot method. On the other hand, the main beam is reduced in strength in the three-spot method which may also introduce a problem; so both methods have their ad herents.

It will now have become clear why linear tracking is an essential feature of three-spot machines, to keep the three spots in a fixed relationship (tangential) to the recorded tracks. For economy, a rack and pinion drive can be used for coarse movements and lateral motion of the lens in the laser beam path for rapid fine adjustments. The 2-axis device shown in Fig. 3 is essentially a dual moving-magnet drive unit (not unlike that in a loudspeaker) and able to move vertically as well as laterally, as instructed by the focus servo which we discuss next.

Focus control
For proper CD reproduction it is essential that the laser beam remains accurately focused to within 0-51.1m to produce a fixed spot size at the metallized recorded surface. Any loss of focus due to record warps (which can be as much as 1mm at the outside edge) or irregularity in the thickness or optical properties of the clear polycarbonate base will increase crosstalk and introduce errors. Again we find several different focus control methods in use.

In the system illustrated in Figs. 3 and 4 the main spot detector is divided into four sectors and the cylindrical lens has the effect of sending equal light energy to all four sectors only when focus is correct. If the information layer on the disc is too close or too far away, an elliptical spot arrives at the detector which informs the focus servo that the main lens must be moved along the optical axis accordingly.

In the one-spot system, a knifeedge method is used to determine focus. A split sensor is placed at the focal point of the incident and reflected beams so that again any offfocus condition produces left or right sensor bias depending on whether the information layer is too close or too distant.

P and Q subcodes
As we have seen in earlier instalments in this series, the stream of binary pits in the recorded track contains a whole variety of data besides the all-important audio signals. There are 75 so-called frames per second and each frame contains information needed for audio, synchronization, error detection and correction (or concealment) and eight subcode bits designated P to W. Of these only the P and Q subcodes are used so far on audio CDs. The remaining six subcodes comprise the 'user field' and are available for future graphics, data and video developments.

The P subcode carries flag data to control the optical readout system, indicating the start of each track (with Pause), lead-in and lead-out signals and (by a repeated pattern of ones) that the next track is about to start. This repeated 'Start Flag' helps simple players to skip quickly to a new track. The Q subcode looks after error correction and identifies emphasis on/off, copy inhibit on/off and two/four track operation. (So far no four-track discs have appeared.) It also contains track numbers, index numbers if any, track elapsed time and total time measured from the beginning of the disc. Even more important, the Q subcode in the lead-in area contains the TOC (Table of Contents) information which comprises the number of tracks (up to 99), their durations and starting times. This TOC information is repeated throughout the lead-in area which the laser scans at the beginning of the disc.

When the disc is loaded or the start key is pressed, the microprocessor goes through an interrogation routine and acts upon the data according to a flow chart of the kind illustrated in Fig. 5. At first the requirement is to check and bring into action the various servo mechanisms. When focus, speed and tracking are correct the TOC data is stored and the laser can jump forward to the beginning of Track 1 or any other track selected by the user, and sound can then be unmuted. The error flow loops in the chart show why a chattering sound is sometimes heard when a disc is faulty (or upside down) or some other fault condition has arisen such as dust on the lens.

The TOC information as to durations, flag/address points, track numbers, etc. once safely stored in the machine's memory is available for the display and control functions. Thus the machine can access any address position in quick time; it can skip or repeat tracks or selections of tracks and display (recall) the programmed sequence and even add up its total playing time; and it can immediately recognize an incorrect command—such as "go to Track 10" if only 8 tracks exist on the disc.

Instead of attempting to describe all the CD player operating possibilities, it will be enough to take just one typical example. Suppose that the machine has been commanded to start playing Track 7. The overall playing time at which Track 7 starts can be determined by referring to the Table of Contents: the physical distance which the pickup must be moved to reach this location cannot easily be computed directly as, say, a fixed number of tracks to be crossed, since the constant linear velocity means that the tracks near the centre are equivalent to shorter time intervals than tracks near the edge. However, the track spacing (pitch) is constant and so a geometrical calculation by the microprocessor is used which does indeed determine how many tracks must be crossed. The normal track-following servo is then disabled and high-speed drive is applied to the carrying arm to swing the pick-up rapidly across the tracks. As a result, the tracking error signal peaks up and down at each crossing and these peaks are counted until the required address point is reached. In practice, a slight underestimate is deliberately made, or the pickup is slowed down, to enable normal track-following to be resumed and the P flag for Track 7 identified. At the end of the P flag (for which a numerical 'count-down' is displayed on some machines) the muting is removed and Track 7 begins to play. Similar reasoning will explain all the other wonderful facilities which CD has brought to our home music systems.