Philips - An integrated approach to CD players - 1984

Part 1: The optical pick-up

A much-publicised advantage of the Compact Disc (CD) system is its high level of immunity to dust, fingerprints and scratches on the compact discs. To this list can be added its immunity to (within-specification) disc defects such as pinholes and track eccentricity introduced by the pressing equipment. Less-publicised, but even more important, are the marked differences in the ability of individual players to handle these defects and, in addition, the knocks, vibration and g-forces experienced in cars.

This article and Part 2 will describe some of the novel aspects of our new ICs and optical pick-up that can be used to manufacture players that will play mishandled discs under adverse conditions without producing any detectable sound colouration. Benefits for the player manufacturer include:

• simple designs (few ICs with virtually no external components),
• flexible designs owing to the use of a standardized interIC digital audio bus,
• designs that can be upgraded to accommodate future trends in digital signal processing.

A full-performance player using the new ICs is shown in Fig.1. All player functions are integrated in the new chip set, reducing the cost of a player's electronics significantly. Furthermore, no factory adjustment of players is necessary. Two servo ICs, with automatic gain and offset correction, control the pick-up, and a single decoder chip incorporates the demodulator PLL and VCO, error corrector and a basic concealment function for uncorrectable audio data. For lull-performance players, an additional chip containing a digital oversampling filter and an enhanced concealment function is available. The chip set is completed by a dual 16-bit DAC for maximum fidelity and an integrated analog filter with outputs for an audio amplifier and for a headphone.

CD Optical pick-ups

Before we describe the CDM2 pick-up*, it will be useful to summarize the stringent requirements to be met by a pickup servo-system, and to review the pick-ups in current use.

* a product of Philips Consumer Electronics Division.

The function of the optical pick-up of a CD player is to generate a high-quality r.f. signal (eye pattern) for the player’s demodulator from a disc’s pit/land transitions which contain the audio data, track number, title, etc. To generate an eye pattern from which all the data on the disc can be retrieved, the laser spot should foljow the centre of the 0,6 gm wide track to within about ±0,1 gm, with no interference from adjacent tracks just J,6pm away. Since the track on the disc may be slightly eccentric due to tolerances in disc pressing equipment, and to handle the ffects of vibration, the servo should furthermore be able to accommodate a maximum side-to-side track swing of about 300gm. The radial tracking servo system capable of meeting these, already demanding, requirements should in addition be able to absorb the effects of knocks to the player and external vibration.

The requirements for the focusing servo are equally severe. For adequate read-out, the focusing servo has to keep the laser beam focused on the reflecting layer of the disc to within +0,5 μm, even with a maximum disc warp of 1 mm.

Manufacturers use a variety of techniques (with varying success) to generate the error signals for the tracking and focusing servos.

Tracking

Four basic techniques are available for generating, a radial tracking error signal:

• single-beam push-pull tracking
• single-beam differential phase detection (DPD) or heterodyne tracking
• single-beam h.f. wobble tracking
• three-beam tracking (using two auxiliary beams to generate the error signal).

Focussing

Three basic techniques are available for generating a focus signal:

• astigmatic (cylindrical lens) focusing
• Foucault (knife edge) focusing
• critical angle focusing.

The three most popular tracking-focusing combinations found in CD pick-ups are:

• single-beam push-pull tracking with Foucault focusing (a design of Philips Audio Division)
• three-beam tracking with astigmatic focusing
• DPD with critical angle focusing.

The first two are in common use and are shown in Fig.2. Philips CDM2 uses a version of the first combination with a lowfrequency wobble injected onto the radial error signal for optimum tracking. The pick-up is mounted in a swing arm which describes an arc across the disc during playback. In contrast, a three-beam pick-up is usually mounted in a sled capable of movement along a disc radius. A singlebeam pick-up can also be mounted in a sled-driven system; a 3-beam pick-up on the other hand won't work in the swing arm since the three beams have to be aligned and maintained in a fixed position relative to the track. Table 1 shows the relative merits of CDM2 and a typical three-beam pick-up. Table 2 gives brief data on CDM2

Philips CDM2 pick-up

Mechanics

The pick-up is mounted at the end of a low-inertia balanced arm pivoted at its centre. A coil housed in the arm, see photo, and a permanent magnet attached to the chassis form a linear motor used both to track and to fast-feed the pickup across the disc. When the motor is energised, the pick-up can be directed quickly to any track (<1 s from inner to outer track), and can follow the centre of the track accurately to within 0,1 gim. Large track eccentricities can be followed because there is no vignetting (the whole pick-up moves, not only an objective as in a three-beam pick-up). Since there is only one moving part (the arm pivot), the mechanism is extremely reliable and wear is minimal.

For comparison, it will be useful to say a few additional words about a three-beam pick-up. A three-beam pick-up needs a linear movement. A disc is tracked by moving the pick-up objective along a radius of the disc. Often, a moveable mirror is inserted in the optical path to deflect the laser beam quickly over small distances, to handle track eccentricity for example. Two auxiliary beams arc used to generate a tracking error signal. Alignment of these beams is critical, and the sled carrying the pick-up and the lead screw and guide all have to be precision components. Any misalignment of the auxiliary beams during the life of the player will impair tracking. Fast-feeding is done by moving the complete optical assembly radially across the disc, for example, by means of a lead screw. The thread of the leadscrew is always a compromise between the fine thread required for normal tracking during playback (35 mm per hour to take up the tracking movement) and the coarse thread required for fast feed (35 mm in 1 second, say).

Optics

A feature of the CDM2 pick-up is its simple diffraction limited optics. The laser point source is focused on the information layer of the disc by two lenses: a spherical glass objective with a plastic aspherical skin, and a sperical glass collimator (most CD pick-ups, including our own CDM1, use a three-element objective and a two-element collimator). The thin plastic skin of the objective is attached using a proprietary process (Ref.l). Owing to the glass body, the objective is stable in humid conditions and temperatures up to 85 °C, while the aspherical skin provides the desired aberration-free optical performance

The low-profile optics of CDM2 has been obtained by interchanging the positions of the laser and photodetector used in our previous pick-up (Fig.2(a) and Ref.2). Furthermore, an inexpensive semi-transparent mirror is used instead of a beam-splitter cube. Astigmatism introduced into the reflected beam by the mirror is corrected by a plastic component (the wedge, in Fig.2(a)) which also dissects the beam into the two halves from which the tracking and focusing error signals are generated.

One could ask, why not use the reflected astigmatic beam to generate the error signals directly (i.e. astigmatic focusing in combination with push-pull tracking)? Our experience indicates that the error signals generated in this way are noisier than with Foucault focusing/push-pull tracking.

Virtually no interference in the radial error signal

The CDM2 pick-up derives the tracking error, the focusing error and the audio signals from one spot on the disc. Compared with a three-beam pick-up, this has two advantages:

• virtually no interference from the r.f. signal in the radial error signal
• allows lower mechanical tolerances in the pick-up, giving long-term optimum tracking with no adjustment of the radial error signals needed.

Shock tolerance

The tracking servo-system for CDM2 is optimised to handle disc defects and the effects of vibration. The latter can be reduced further by mounting the CDM2 mechanism in a well-designed floating suspension. This suspension should have the characteristics given in Table 3 with progressive damping for large-amplitude shocks.

Electronics

Some aspects of the decoding electronics are described in Part 2.

References

1. I. BRAAT,J.J.M., 'Aspherics', Philips Technical Review, Vol. 41, No. 10, 1983, p.258.
2. 2. CARASSO,M.G„ PEEK,J.B.H. and SINJOU,J.P.; The Compact Disc Digital Audio System, Philips Technical Review Vol.40, No. 6, 1982, pp. 151 to 155.

Part 2: The decoding electronics

In Part 1, we introduced a new single-beam optical pick-up and servo ICs for Compact Disc (CD) players, the advanced designs of which allow the manufacture ofsmall players that can handle disc defects such as scratches-, pin holes and track eccentricity, as well as the knocks, vibration and g-forces encountered in portable systems. Hand in hand with this advance come exciting developments in the signal-processing circuitry. Developments that, among other things, allow full advantage to be taken of the error-correction capability of the CD system, to produce high-performance players with no detectable sound colouration.

This article describes some of these new developments, in particular the integrated CD decoder. The decoder requires no adjustment and is suitable for all CD players — those with a single-beam pick-up and those with a threebeam pick-up. Besides the audio application, the decoder can be used in many other applications requiring a highperformance high-reliability Reed-Solomon decoder, e.g. CD ROM.

Supporting the decoder IC, several circuits have been developed to facilitate the manufacture of players with:

• fewer ICs and fewer peripheral components
• greater design flexibility
• improved sound reproduction.

Such a player is shown in Fig.l.

Figure 2 shows the new decoding electronics and the six ICs (shaded) it replaces (Ref.l). The SAA7210 decoder IC incorporates the functions of demodulation, error correction and basic interpolation. The SAA7220 digital filter IC incorporates enhanced interpolation circuitry and a phaselinear digital FIR filter. A dual 16-bit DAC TDAI541 (operating at 176,4 kHz) followed by a stereo analog output IC TDA1542 replaces two DACs and two discrete analog filters.

The data format between the SAA7210, the SAA7220 and the TDA1541 is according to the PS (inter-IC signal) specification*, giving the player manufacturer maximum design flexibility. This would allow the SAA7220 to be omitted in Fig.2, for low-cost players, leaving the manufacturer free to design his own low-pass filter.

* Inter-IC signal (FS) comm unication is a communication format for digital audio. The FS bus is a three-line bus comprising: clock, serial data line, and a control line used to select left and right channel words.

Features of the new decorder

Further integration of the demodulator PEE

A feature of the SAA72L0’s demodulator PLL is that it requires virtually no peripheral components. The VCO is a fully-integrated RC oscillator; it requires no peripheral components. The differential filtering circuitry of the PLL uses a self-balancing charge pump design that requires only one filter comprising just three low-tolerance components (Fig.3).

More efficient processing of subcoding data

Besides the audio information recorded on a compact disc, information representing track numbers, playing times, titles and composers is recorded so that tracks can be played in any desired sequence and titles and elapsed playing times etc. can be displayed. This information, termed Q-channel subcoding data, is derived continually in the SAA7210 from the demodulated r.f. signal. A new handshaking protocol between the SAA7210 and control processor reduces the time the control processor spends handling subcoding data. When the processor wants data, a request is sent (via QRA, see Fig.l) to the SAA7210 which, when a full Q-channel frame is ready, acknowledges the request and enables the serial data output QDA. The processor -sends a clock signal (QCL) to shift the data out of the SAA7210.

The first negative-going edge of the clock signal' resets the acknowledge signal thereby releasing the request line. If the processor doesn’t require all the subcoding data, say only the number of a track (contained in the first sixteen bits), it can reset the request line after these bits have been received, thereby disabling the QDA output of the SAA7210 which resumes collecting new subcoding data.

Adaptive error-correction

Until now, no CD player has made full use of the errorcorrecting capability of the CD system’s Cross-Interleaved Reed-Solomon Code (CIRC). This code enables up to four erroneous symbols* (in a 32-symbol block) to be corrected if the CIRC decoder is given prior information about the position of the errors. This type of correction, where the position of an error is known, is called an erasure correction. When the positions of errors are unknown, up to two erroneous symbols can be corrected. A decoder with the maximum error-correction capability (like that of the SAA7210) can make the corrections shown in Table 1.

* 14 bits

For optimum error-correction, only the following corrections of Table 1 are relevant:

• t = 2 look for and correct two errors, positions unknown
• e = 1, t = 1 make one erasure correction, and look for and correct one additional error
• e = 2, t = 1 make two erasure corrections, and look for and correct one additional error
• e = 3 make three erasure corrections
• e = 4 make four erasure corrections

A new approach to error-correction in the SAA7210 known as adaptive error-correction discriminates between the errors found on a compact disc. This discrimination:

• enables more corrections to be made (e.g. longer burst errors)
• makes the corrections more reliable.

As shown in Fig.4, these improvements are achieved by using additional error flags generated by the EFM decoder (using the bit run-length criteria to flag symbols likely to be in error) and by using multi-level error flags generated by the Cl and C2 corrector*.

* Two Reed-Solomon codes Cl and C2 are used to correct erroneous audio samples. Cl being used to correct small errors in adjoining symbols, C2 to correct burst errors

With the error flags, several error-correction strategies are available, the strategy chosen depending on the type and number of error flags which are set by the defects on a disc

To determine the typical defects to be found on a disc, and therefore the strategy for correcting them, special test discs with known data patterns were manufactured on normal production equipment. These discs enabled typical manufacturing defects to be identified. Normal handling defects such as scratches and fingerprints were then introduced on each disc. Extensive testing with these discs demonstrated that the error-correction was significantly improved by using multi-level flags to signal symbols in error, three flags (two bits) being sufficient:

• hard error-flag (most reliable flag)
• medium error-flag
• soft error-flag (least reliable flag).

A no-error flag is indicated by setting both bits to zero.

With these flags, the best correction strategy is selected by a flag processor, see Fig.4. For example, when two symbols in a block of thirty-two are flagged with soft errorflags, it is best to attempt a t = 2 correction (rather than an e=l and a t=l, or an e = 2, t = l correction). When two symbols are flagged with hard flags, it is usually better to attempt an e = 2 correction and to look for and correct another symbol in error with the remaining t = 1 capability. After a correction, the input flags are compared with the new flags produced, by the corrector to update (harden) the flags for the next stage of processing. Flag hardening progressively improves the reliability of all flags, so that very reliable e = 3 and e = 4 corrections can also be made.

Adaptive error-correction (with about 60 routes' to the correction possibilities in Table 1) is significantly better than the present correction which, owing to the use ofsinglelevel flags, has a somewhat limited correction capability.

All the error-correction routes for the CD audio application system are programmed in ROM in the SAA7210. They can be altered for non-audio CD storage applications.

Large FIFO memory for car players and portables

Part of the memory used to de-interleave the error-correction data is used as a FIFO memory to remove variations in the demodulated-data rate due to, amongst other causes, the g-forces on car players and portables. In addition, this FIFO can be used to increase the allowable mechanical tolerances for the disc drive of a player. The SAA7210 is designed to operate with a 16Kx4DRAM which can accommodate up to 64 frames compared with the 4-frame FIFO of first-generation players. This larger FIFO is also used to store the increased number of-error flags used in adaptive error-correction.

8-sample interpolation

In low-cost players, the SAA7210 can be used as a singlechip decoder, requiring only D/A conversion and low-pass filtering. The SAA7210 contains an audio hold function and single-sample interpolation as found in most other decoders on the market. Full-performance players can be made by adding the digital filter circuit SAA7220 which contains an 8-sample interpolator (Fig.5) that can conceal large burst errors. When more than eight consecutive samples are in error, a hold function operates before the interpolation.

Figure 6 shows the combined effect of improved errorcorrection and interpolation.

Low-pass filtering

The low-pass filter of a CD player should attenuate any high frequency components while preserving the audio band intact. Any intermodulation products should be too weak for perception, and possibly even for measurement, see Table 2.

Table 2 Requirements for a low-pass filter for a CD player
pass-band	0 to 20 kHz
pass-band ripple	< +/- 0,02 dB
transition band	20 to 24.3 kHz
stop-band injection	> 60 dB
linear phase response
low thermal noise
no ageing effects
no adjustements

The choice of low-pass filter in a CD player is very important, a choice that explains, in part, the difference in the sound reproduction of different players. Designs ranging from wholly analog to predominantly digital, with either two-times or four-times oversampling, can be found in current players, see panel. We favour predominantly digital filtering with four times oversampling followed by a loworder analog filter. This arrangement gives:

• linear phase response (at a much lower cost than with all-analog designs, and with no component adjustments necessary)
• flat pass-band
• optimum roll-off for excellent impulse response
• low noise
• long-term stability (no component ageing effects)
• smaller filter than discrete analog designs

Digital filtering is probably the best-known feature of our approach to creating true high-fidelity sound. Our new digital filter circuit SAA7220 followed by a low-order analog filter produces no detectable sound colouration.

The digital filter is a (stereo) phase-linear four-times oversampling FIR filter with 120 filter coefficients. The (stereo) analog filter is an active simple third-order Bessel filter constructed with the TDA1542. Connected between the two is a 16-bit dual DAC TDA154I operating at 176,4kHz, see Fig.l. This DAC introduces no delay between the stereo channels - highly desirable for the listener of normal stereo sound, and essential when mixing audio signals, tor example, to generate 4-channei audio or spatial stereo sound.

D/A conversion

The Philips TDA1541 dual DAC, located after the digital filter, uses a method of current division called dynamic element matching for high-accuracy binary-weighted currents with long-term stability. In addition, for the chip manufacturer, dynamic element matching eliminates resistor trimming. Bit switching is performed with a diode-transistor configuration for fast and accurate switching without the need for external deglitching circuitry.

The TDA1541 has been designed to also meet future trends in digital audio. For example, it is generally acknowledged that digital signal processing such as digital tone control and equalization increases the requirements for the dynamic range of the output signal. Eighteen-bit resolution can be achieved using the TDA 1541 with four-times oversampling and noise shaping, because oversampling and noise shaping shift the quantization noise to above the audio band. As a result, a digital tone control using the TDA1541 can produce a frequency boost of about 12 dB while maintaining a 16-bit dynamic range over the whole audio band.

References

1. MATULL.J., "ICs for Compact Disc decoders”, Electronic Components and Applications, Vol.4, No.3, May 1982,pp. 131 to 141.
2. LAGADEC,R. and STOCKHAM,T., “Dispersive models for A/D and D/A conversion systems”, Paper presented at the 75th Convention ofthe Audio Engineering Society, March 1984, Paris.
3. CARASSO,M.G„ PEEK,J.B.H.and SINJOU,J.P, "The Compact Disc Digital Audio System", Philips Technical Review Vol.40, No.6, 1982, pp. 151 to 155.

Acknowledgment

1. The author wishes to acknowledge assistance given while preparing this article. Particular thanks are due to colleagues at the compact Disc laboratory, Philips Consumer Electronics Division and at Philips Research Laboratories, Eindhoven.
2. This article is based on an article that appeared in Japanese in the February 1985 issue of Audio, Video, Electronics and Record, Tokyo, Japan; permission to publish is gratefully acknowledged.

Source: Philips Electronics components & applications. VOL. 6 NO. 4. August 1984