|Home | Audio Magazine | Stereo Review magazine | Good Sound | Troubleshooting
Departments | Features | ADs | Equipment | Music/Recordings | History
by WALTER G. JUNG [Walter G. Jung is a consultant and writer based in Forest Hill, Maryland.]
The new series of Philips/Magnavox CD players have quickly earned respect for their generally superior audio performance at affordable prices. (It was Philips, the parent company, that originated the CD system with Sony). The players are manufactured in Belgium and are sold under the Philips name in Europe and the Magnavox, Sylvania, and Philco names in the U.S. Steady improvements over the last several years have resulted in better and better sonics, while prices have dropped (or held) and features have been added.
The innovative Philips playback processing system is generally considered to be technically very advanced.
Basic to this system are two key steps.
With the implementation of these techniques, the Magnavox players produce a cleaner, more accurate, and more noise-free CD output signal.
Until recently, all Magnavox players were based on a 14-bit-resolution digital-to-analog (D/A) converter, the TDA1540, in each of the two stereo playback channels. Previous Magnavox players, from the early 1000 series up to the 2040 series, used this same D/A for the audio signal process, as did the FD1041, FD2041, and FD1051 models (which were available up until January of this year). More recently, Magnavox has introduced the long-promised 16-bit models, the CDB650 (reviewed in Audio, March 1987), the related CDB560 and CDB460, and the newest player in this family, the CDB465, which was introduced in May (more below). All of these players use the brand-new TDA1541 D/A, a single IC with two channels of 16-bit resolution. Related models are also available under the Sylvania and Philco brand names.
A 16-bit D/A extends the theoretical resolution and dynamic range of the audio signal from 14-bit's capacity of 16,384 (214) levels and 84 dB to 16-bit's 65,536 (216) levels-four times as many-and 96 dB, an obvious improvement. In addition, the new players use an associated set of digital chips which offer other improvements, such as enhanced error-correction algorithms and better digital filtering (see sidebar).
The CDB650 is the flagship of the current line of CD players from Magnavox, laden with all manner of bells and whistles, both in terms of operational features and with regard to sonic goodies. The CDB560, CDB460, and CDB465 are shorter on the operating bells and whistles, but about as long on the sonic ones. What I want to discuss in this article are the technical features which distinguish these players and the sonic impressions they make in a high-resolution system using some high-quality CDs. I also wish to suggest some easily performed parts changes which will, I believe, improve the sonic performance.
The CDB650, CDB560, CDB460, and CDB465 have similarities as well as differences. The differences are, for the most part, in their operating features and their prices. Interestingly, there are more things alike than different, insofar as the audio signal path is concerned. Thus, the CDB560, CDB460, and CDB465 can deliver about the same sound as the CDB650 but at a lower price.
THE NEW CHIP SET. The Essence of the ICs
By PRASANNA SHAW [Prasanna Shah is an applications engineer at Signetics Corp., Sunnyvale, Cal.]
THE NEW CHIP In the new high-performance CD players such as the Philips/Magnavox CDB650, CDB560, CDB460, and CDB465, a second generation of dig ital audio circuits is used. This new generation has a higher level of integration, with more sophisticated algorithms for the decoding. In addition, the number of circuits has been reduced; this helps to keep manufacturing costs low, making the players more affordable and competitive.
The new CD players use SAA7210 and SAA7220 ICs in the digital de coder section and TDA1541 and NE5532N or LM833N ICs in the analog section. Figure B1 shows a block diagram of the new decoder system, with an indication of the (dotted) first-generation circuits which were re placed; only an indication can be given because the new circuits have much superior performance over the first generation. This new decoder system architecture utilizes an Inter-IC Sound (12S) bus standard for de sign flexibility. The 12S bus is a three-wire serial bus for data exchange between the integrated circuits, and it enables simplified digital audio data exchange between Compact Disc, Digital Audio Tape, digital sound processors, etc.
The SAA7210 is the new decoder circuit which completely replaces the first-generation SAA7010, which handled control and display signals, and the SAA7020, which handled servo control signals. The new SAA7210 also does much of the work of the older SAA7000 and provides many enhancements. It has a demodulator with adaptive data slicer, a fully integrated phase-locked-loop demodulator for bit-clock generation, a sub-code data processor for control and display information, and an advanced motor-speed control to stabilize input data rate information and further reduce wow and flutter. It also has an improved error corrector with an adaptive strategy, an interpolation and sample-and-hold circuit, and an improved error-correction strategy to handle up to a 15-frame error burst out of a 32-frame block (as opposed to seven frames with the SAA7020).
The adaptive error-correction approach discriminates between different types of errors found on a Compact Disc; this enables more corrections to be made, with greater reliability. The error-corrected and interpolated data are then transferred to the SAA7220 via the I2S bus.
The SAA7220 is the digital filter circuit with enhanced concealment of uncorrectable errors. It has two identical phase-linear digital Finite Im pulse Response (FIR) filters with 120 12-bit filter coefficients (as opposed to 96 in the SAA7030). This filter also does the four-times oversampling and offers better noise shaping than the older SAA7030. It performs linear interpolation of up to eight consecutive erroneous samples, thus leaving virtually no trace of errors. Also, the SAA7220 has a soft audio mute when starting, stopping, or pausing and 12-dB attenuation of audio output while scanning. It also provides a digital-format audio output for inter face with digital sound processors that become available in the future.
The older SAA7020 could correct up to seven frames of data out of a 32-frame blocs (each frame having 256 bits) and could generate an error flag identifying unreliable data. To minimize the effect of erroneous data, a second IC, the SAA7000, replaces them with interpolated data. Where the data samples have very large numbers of error bursts, and thus do not lend themselves to signal interpolation, the SAA7000 generates a mute signal. This eliminates the audible annoyance which would result from erroneous data getting through the error-correction circuits, obviously an important feature for less-than-pristine discs. The SAA7220 replaces these functions, with the improvements noted.
The other major new IC is the TDA1541, the dual 16-bit D/A converter on a single chip. Like its tandem TDA1540 D/A predecessors, this converter also uses the dynamic-element-matching technique for high accuracy and stability. With true 16-bit D/A conversion and four-times oversampling and noise shaping from the SAA7220, an effective 18-bit resolution can be approached, pro viding greater than 100 dB of S/N.
Since the dual D/A converters are on the same chip, there is no delay between the stereo channels; there is-J also good matching of full-scale and zero-scale outputs on both channels.
The (sin x)/x response of the hold function on the D/A converters and the Bessel filter's response were taken into consideration when the filter’s coefficients for the SAA7220 were scaled, so that the overall digital-to-analog conversion function would provide a flat response in the audio band. The final low-pass filter is again a third-order Bessel type, using either the NE5532N or LM833N.
All these new and improved enhancements in the decoder electronics should enable the high-end Magnavox CD players (CDB650, CDB560, CDB460, and CDB465) to reproduce marvelous sounds and bring them closer to reality.
1. Matull, J., "ICs for Compact Disc Decoders," Electronic Components and Applications, Vol. 4, No. 3. May 1982.
2. Nijhof, J.. "An Integrated Approach to CD Players, Part 2: The Decoding Electronics," Electronic Components and Applications, Vol. 5, No. 4, 1984 (Philips TP172).
3. Signetics Linear Data and Applications Manual, 1985.
4. Schott, W., "Philips Oversampling System for Compact Disc Decoding," Audio, April 1984.
The CDB650's outstanding new feature is "FTS," which stands for favorite Track Selection. This feature, under the control of its own dedicated, nonvolatile memory, allows one to store a unique track-sequence code for each CD. Thereafter, when a particular CD is loaded, the CDB650 will play its tracks as you have stored them via FTS. Up to 785 different tracks can be stored with this system. It even allows specific start/stop times within a track to be recalled. I can easily see this being used as a tool for instant replays during audiophile "CD shootouts"; it is a replay/evaluation tool which totally eliminates the search-and-stop game.
The CDB650 also has a full-featured remote from which just about all operations can be controlled. It's about as complete a remote as you can find these days, save for the fact that it has no volume control. The related CDB560 has a more limited remote capability, and the CDB460 has none;
neither of these two models has the FTS feature. The CDB465's claim to fame is that it has the basic features of the CDB460, plus FTS. The CDB650 also has a digital output in the form of a single rear-panel RCA jack. This output is transformer isolated and is intended to be used for future CD-ROM applications or external digital processors.
While the CDB650, CDB560, CDB460, and CDB465 all have similar audio paths, there are some distinguishing additions to the CDB650. It sports two sets of gold-plated audio output jacks. The first set has a filter like that of all Magnavox players made thus far, a three-pole Bessel-response low-pass filter using (in this case) two Signetics NE5532N or National LM833N dual op-amps. The second set of jacks adds an additional filter in the form of a pair of op-amps in a five-
pole, low-pass configuration. This player also comes with a pair of stereo cables, as is typical, but these have gold-plated ends. The CDB560, CDB460, and CDB465 have output circuitry like that used for the CDB650's first set of outputs, and more conventional nickel-plated output jacks.
As far as internal construction is concerned, all the newer Magnavox players use electronic parts and techniques that are of relatively high quality for what are reasonably priced consumer units. Parts quality has gotten generally better with succeeding generations of Magnavox units. For example, the now-obsolete 2040 player used a number of nonlinear, high di electric-constant ceramic capacitors, while the newer units employ chip ceramics of higher quality for similar functions. The power supplies of the CDB650, CDB560, CDB460, and CDB465 are ± 15V for the op-amps (as opposed to ± 12 V in previous units). All supplies are regulated and use the common 78M00 and 79M00 series of three-pin TO-220 IC chips.
Additional capacitive bypassing is also employed across the output(s) of these regulators. I note one oddity, though, which is the fact that the players' de signers have chosen to decouple the ± 15 V supplies to the output op-amps with relatively large resistances in each supply line, even though these sup plies are regulated. This effectively ne gates the virtues of regulation for these stages.
In the active filter stages which comprise the common output section of the players, 2% polystyrene capacitors are used with 1% MF-25 style metal-film resistors. The CDB650's additional filter stage also uses these higher quality capacitors and resistors. This filter circuit is housed on a completely separate board set off to the right inside the CDB650; in the others, the corresponding space is simply empty.
These filter circuits are d.c.-coupled within, and a single 220-µF/10-V electrolytic capacitor couples each audio channel to the outside world. The polarizing bias on these capacitors is 4 V (unlike the FD1041 and FD2041, which operate a similar capacitor with no bias).
The p.c. board construction in Magnavox players has lately used a mixed combination of component technologies. While the above-mentioned RC components are the traditional variety (with leads) that mounts on the top of a p.c. board, the FD1041 (and later) players began to also use surface-mounted chip components located on the underside of the board. This includes many 1206-size ceramic capacitors and thick-film resistors, as well as transistors. This is one way that the parts density on the boards can be increased. Perhaps more important, this design lends itself to automated assembly, thus ultimately producing a lower priced unit.
All the latest units use this type of construction also, on both sides of the main p.c. board (it mounts the decoder, microprocessor, digital filter, D/A converter, audio filters, and power-supply circuits). Most of the upper plane of this board is a ground plane, in copper for high conductivity. Additional CDB650 p.c. boards include the headphone driver, the control/display, the FTS board, the aforementioned additional filter, and a servo board.
This servo board is actually integral to the player mechanism, whose performance is, of course, highly critical to the overall success of the unit. The CDB650 uses a plastic mechanism, as did the FD1041/FD2041 series. It would seem that a plastic mechanism would not be adequately reliable, yet I personally have not had any problems with it, as used in three different players (an FD1041, an FD2041, and a CDB650). By contrast, I did have some problems with two early model 2040 servo mechanisms, so to me the newer unit seems to have proven itself.
Nevertheless, good lessons can sometimes be "unlearned." This same mechanism, as mounted in the CDB650, has a hangup: Its position is recessed slightly into the player chassis. As a result, you cannot simply run the tray out and conveniently drop a CD into place as you could in the FD1041/FD2041. One must tilt the CD down in the rear, to slip it under the overhang. Boo on this one, Magnavox! For anyone interested in the more technical aspects of the Magnavox players, I can strongly recommend the purchase of the appropriate service manual. These are available for each series (even for now-discontinued units) from the publications division of NAP Consumer Electronics, Box 555, Jefferson City, Tenn. 37760. You can also call to order by credit card: (615) 475-3801. The prices are very reason able; for instance, the CDB650 manual costs only $6.50. These manuals are truly excellent, with clear markings of parts, values, tolerances, and their locations via an X-Y map. "Three-dimensional" color views of the two board sides can help find the impossibly tiny chip components. (If you are at all interested in learning how to work on the chip components on these p.c. cards, you'd better think twice. Not only are there the usual standard disclaimers about voiding the warranty, but special tools are needed to do any sort of work with the surface-mounted chip components. It is all too easy to take these parts for granted because everything works so well and because the players do not cost an arm and a leg. How ever, don't lose sight of the fact that there still is a powerful assemblage of very high-tech stuff in each of them!) Before I discuss perceived sound, I should note that in my experience just about all CD players are helped sonically by a long warmup and stabilization period. I know some audiophiles who keep their systems on continuously so as to avoid a waiting period each time they listen to CDs.
I should also preface the comments below by saying that they apply to what I feel is a system of higher-than-average resolution. Systems of lesser resolving power will likely not display the same level of detail as I describe.
In any event, I will try to qualify my comments to maintain the perspective.
Several CDs which I have found useful to differentiate specific sonic effects are: Dire Straits' Brothers in Arms (Warner Bros. 25264-2, DDD); Mahler's Symphony No. 5 (Chicago Sym phony Orchestra, Georg Solti, London 414321-1, AAD); Hoist's The Planets (London Philharmonic Orchestra, Georg Solti, London 414567-2, AAD); Stravinsky's "Le Sacre du Printemps" (Montreal Symphony Orchestra, Charles Dutoit, London 414202-2, DDD); Mozart's "Eine Kleine Nachtmusik" (Academy of Ancient Music, Christopher Hogwood, L'Oiseau-Lyre 411720-2, DDD), and Janácek's "Taras Bulba" (Vienna Philharmonic Orchestra, Charles Mackerras, London 410138-2, DDD).
Right up front, I'm willing to say that listening to the CDB650 is a treat com pared to many other CD players. There is little of the harshness or stridency so often mentioned as a characteristic of "CD sound." Most discs sounded quite smooth on the CDB650, particularly those which put fewer demands on the playback system. (Many high-quality analog-mastered ones seem to do this, while high-energy, digitally mastered CDs can cause nonlinear players loads of grief, resulting in that "CD sound.") With the very best CDs, the CDB650 can sound quite exciting.
The CDB650 did extremely well on all the recordings, with good detail and a fairly broad but not very deep sound stage. On recordings which have a lot of depth and a good bit of hall ambience (such as the Hoist and the Stravinsky), the CDB650 reproduced this depth and ambience well. However, it does seem to have a greater ability to reproduce a sonic image in width than it does in depth. Two other Magnavox players (a modified FD1041, with improved amplifiers and power supplies plus direct-coupled, servo-controlled outputs; and a modified CDB650 with similar amplifier/power changes, plus film coupling capacitors) extracted a greater sense of depth and space as well as smoother and finer detail on the Hoist and Stravinsky, compared to the stock CDB650, but the differences were not major. Those two recordings also give a good sense of the acoustic space or "air" around the instruments, which adds greatly to the overall natural liveness. Again, the stock CDB650 did fairly well in this regard, but not as well as the modified units. However, it outpointed a stock FD2041 in detail, imaging, and bass.
When listening to other types of recordings, such as those multi-miked for various special effects, the importance of the ambience factor becomes moot. On Brothers in Arms, there is virtually no depth, and that is what was heard on all players. The CDB650 and the two modified units did well on this one, while the FD2041 sounded a mite "zingy," with loss of detail. As resolution is lost, things blur towards a more homogeneous sound. On the other hand, a very high-resolution CD play back actually will "take apart" the mix-down process of many pop recordings. (You may not like all of what you hear, either.) The point is that if a player can be made which sounds excel lent on the very best naturally miked recordings, it will likely give faithful re production on other material as well.
I did not spend a great deal of time listening to the CDB650's "additionally filtered" outputs for comparison, as the other players had no comparable function. It is valid, of course, to compare it to the CDB650's more conventional outputs, and this was done. The differences between the normal outputs and the additionally filtered ones left me somewhat cool. I could not perceive that the extra filter did anything to actually enhance the sound (at least in my system). The additionally filtered out put was, in fact, a couple of dB down at the top of the audible range; the thinking here, presumably, is that the lower levels of ultrasonic energy will ease overload and/or high-frequency stress downstream. Regardless of the idea behind these outputs, their use was a step backward in sound quality when used in my system. Through them, the sound thickened up and lost the sense of space present at the nor mal outputs. Overall, I'd rate this extra filter as a nice idea in principle, but one that fell short in its execution.
The CDB650 sounds good just as it comes, but can it be easily made to sound better? The answer is generally yes, but with some very important caveats. First, don't attempt to modify your CDB650, CDB560, CDB460, or CDB465 without fully understanding what you stand to gain and lose. If you try it and have trouble, you are out on a limb, as you will have voided your player's warranty by opening it up and working on it. Even if you do decide to take this risk, you should be familiar with modern solid-state construction, use only a low-wattage miniature soldering iron, and be prepared to back out if any confusion develops. It is strongly recommended that you at tempt these modifications only after obtaining the service manual for your player, and be sure to check out all steps thoroughly before you begin the first one.
above: Installing these ICs, resistors, and capacitors in the CDB650 are most-for-least modifications that should take no more than 45 minutes.
On the plus side, the modifications described below are quite low in cost (less than $20), relatively painless if you are skilled with electronics, and don't really expose the player to any hazards. The modifications are simple one-for-one substitutions of parts, plus a rebiasing of the op-amps or better linearity. A list of parts and sources is at the end of this article.
Begin by unplugging the CDB650 and taking the cover off. For this you'll need a Torx screwdriver, or a small-tip conventional unit which "just fits," to remove the back and side screws. The two NE5532N op-amps on the main board are seen near the 8-pin connector in the photo above. They are shown pictorially on page 6-3 of the service manual, as part numbers 6306 and 6307, and schematically on page 6-5.
With the board still in the player, they may be seen in the rear right quadrant of the p.c. board, with part numbers NE5532N or LM833N. Note the orientation of the notch; it is to the left, as noted in the service manual.
With this in mind, now remove the board. Doing so is a mite tricky, as you will need to disconnect the seven cable plugs, unscrew the four inner and one rear screws, and work the board up and out toward the front (some bending of the rear panel may be necessary). Very carefully de-solder the eight pins of the two above-mentioned ICs, taking care to use a solder de-wicking braid. The ICs should drop out easily. Install two new NE5535N devices in their place. Double-check the pin orientation from the top side to make sure it is still correct, and then solder in the ICs.
While still in this area of the board, very carefully solder two 3.92-kilohm, metal-film resistors on the back side of the board, on each of the two op-amp IC pin patterns at 6306 and 6307. Connect one resistor from pin 4 to pin 1, and the other from pin 4 to pin 7. Note that this can be most easily seen using the reverse-side board view (shown on page 6-4 of the service manual). Take care that the resistor leads touch only the tracks noted; use insulating sleeving on the leads as appropriate. Make sure that no solder splashes are left around these two ICs, and dress the resistors slightly away from the board surface to finish them up.
Next, note the locations of the output capacitors, numbers 2366 and 2367, which are 5/6-inch-diameter electrolytic units back near the output jacks (again, see the manual pages noted above). Carefully de-solder the 220-µF/10-V units supplied, and install a pair of 330-uF/25-V low-ESR (equivalent series resistance) electrolytics in their place, using either of the two types noted in the parts list. Be very careful that the polarity of the newly installed units matches that of the old, with the "minus" stripe toward the rear panel. The "minus" polarity stripe is also indicated in the manual, by the dark capacitor plate. This finishes up the modifications. Carefully check your work and re-install the board, taking care to reconnect all cables as they were originally.
While there are many other things which might be done to the player, these relatively modest changes should provide a reward of more smooth, open, and detailed sound, with a minimal outlay of time and money plus a low risk factor in execution.
The "most-for-least" modifications I have described should take only 45 minutes or less.
The adventurous reader can likely expand on my suggestions, perhaps substituting polypropylene or polyester film capacitors for the high-performance electrolytics (a large-value film capacitor will be needed, however).
For those so inclined, the references below discuss other changes that are possible. For those not inclined to tinker, the four latest Magnavox players will still be a great source of enjoyment, no doubt! I'll admit that I'm enthusiastic about the CDB650, and about CD reproduction in general. However, I also can understand the type of sound the CD critics have been rejecting. Admittedly, we have not had very many truly top notch players with the transparency, imaging, and naturalness of the very best analog playbacks--certainly not at medium prices. But I'm also enough of an optimist to feel that this is changing, with quality players and discs.
The CDB650 and related models might impact this situation by moving CD playback a step closer to the best analog reproduction. They might even make some new believers. But even without the new believers, these players will make a lot of people happy, since they are plenty good "out of the box." My hat is off to the Philips organization for bringing these products to market.
Op-Amps: Two Signetics NE5535N dual high-slew-rate ICs.
Resistors: Four 3.92-kilohm, 1% metal-film, MF-25 types.
Capacitors: 330 µF/25 V. Use two Panasonic "Z" series (stock #P6635) or two Panasonic "HF" series (stock #P6714).
The parts listed above may be purchased from Digi-Key Corp., P.O. Box 677, Thief River Falls, Minn. 56701, (800) 344-4539. Audiophile-quality parts (ICs, film capacitors, metal-film resistors, etc.) are also available from Old Colony Sound Lab, Box 243, Peterborough, N.H. 03458, (603) 924-6371.
1. Jung, W. G., M. L. Stephens, and C. C. Todd, "An Overview of SID and TIM," Audio, June, July, and August 1979.
2. Jung,. W. G. and R. N. Marsh, "Picking Capacitors," Audio, February and March 1980.
3. Jung, W. G. and R. N. Marsh, "POOLE-2, A Mod Symphony for Your Hafler DH-200 or Other Power Amplifiers," The Audio Amateur, Issue 4, 1981.
4. Otala, M., "Feedback Generated Phase Non-Linearity in Audio Amplifiers," London AES Convention, March 1980, Preprint No. 1576.
5. Matull, J., "ICs for Compact Audio Disc Decoders," Signetics Linear Data and Applications Manual, Vol. 2, 1985.
6. Van de Plassche, R. J., "Dynamic Element Matching for High-Accuracy Monolithic D/A Converters," IEEE Journal of Solid State Circuits, Vol. SC-11, No. 6, December 1976.
7. Van de Plassche, R. J. and D. Goedhart, "A Monolithic 14-Bit D/A Converter," IEEE Journal of Solid State Circuits, Vol. SC-14, No. 3, June 1979.
8. Jung, W. G., "Op Amp Meets CD," The Audio Amateur, Issue 3, 1986.
9. Childress, H., "Modifying Yamaha's CD2 Player," The Audio Amateur, Issue 3, 1986.
10. Jung, W. G., Audio IC Op Amp Applications, 3rd Ed., Howard W. Sams, Indianapolis, Ind., 1986.
11. Schouwenaars, Dijkmans, Kup, and Van Tuijl, "A Monolithic Dual 16-Bit D/A Converter," IEEE Journal of Solid State Circuits, Vol. SC-21, No. 3, June 1986.
(adapted from Audio magazine, June 1987)
= = = =
Prev. | Next