| There have been several requests on net.audio for my Magnavox modification
article, so I am sending it out with many updates of new information.
Any comments, technical or stylistic before I send it?
Thanks,
Mark
Modifying Magnavox/Mission/Meridian/Revox CD players, rev. 1
Would you believe a $170 Magnavox CD player can sound as good as $2,000
worth of turntable, tonearm, and cartridge? Without modifications most
quality is lost in an analog audio section on par with a $50 preamp.
These simple modifications to the audio circuitry can be made by you or
a professional. This posting is for those who hear player differences,
those who wonder what is inside, and audio hackers!
I explain design flaws in CD players that until recently corrected, caused
great scorn for the CD among audiophiles. This is my report on
experimentation and analysis on correcting problems in a CD player's
analog audio circuitry. Lastly, some personal thoughts and details on
future products are given.
The modification techniques used here are not new. They have been
applied on amps and preamps for years. What is new however, is the
application to CD players and identifying sonicly important changes
in circuits with digital to analog converters. These techniques are
probably also applicable to LV disk players and "hi-fi" VCRs.
I. Why modify a Magnavox CD player?
Don't be put off by the Magnavox name or the low price tag. Magnavox,
Sylvania, and Norelco are US trade names for Philips, codeveloper of the
CD with Sony. Philips based CD players use quadruple oversampling, two
digital to analog converters, and both digital and analog anti-aliasing
filters. The result is less phase shift than other techniques.
Most players exhibit no correlation between price and sound quality.
All you get for extra money is more useless features you try once. The
marketing of CD players slightly bewilders this software engineer.
Nearly all the top CD players use the Philips technology. Brands based on
Philips hardware include: PS, Cambridge Audio, Meridian, Mission, Discrete
Technology, McIntosh, Revox, Marantz, and Magnavox. This is an elite group
where the best player(Cambridge) costs $2500, yet a Magnavox can be had for
well under $200. Several of these players are an insulting rip-off when
considering the minor differences for which they charge hundreds of dollars.
Magnavox uses the right technology, but its execution is wrought with
cost compromises necessary in a mass market consumer product where
sound quality is generally lost on the masses. Compromises took a big
toll on the audio section with essentially transistor radio quality parts.
Simply substituting high quality parts in the audio section will make
dramatic improvements. It will solve most of the sonic problems that The
Absolute Sound and others have identified. While I can't prove this
conclusively to each reader, I will try to by identifying sonic improvements
that occur with specific changes. I will also recommend other changes that
I have not tried, but others found valuble.
II. The weaknesses of Magnavox players
The sound is compressed - lacking low level dynamics. Imaging is shallow.
Transients are distorted and there is a general glare and edge. Problems
show up on the tough test of horns, violins, and piano. These problems
can all be traced to weaknesses in the audio section and are not because
"digital" has sapped the soul from the music.
The problems are: mediocre opamps, a noisy, poorly decoupled power supply
with bandwith and current limited regulation, the use of ceramic and
electrolytic capacitors in the signal path, and a very poor audio cable.
Magnavox 1040s and 2041s use numerous, compact Surface Mounted Components
(SMCs) of dubious sonic quality. Magnavox 2040s do not.
Internal noise from clocks and digital signals is a problem, but requires
moving the audio circuitry into a separate shielded enclosure. There also
seems to be a DAC related problem of supersonic spurae centered near fractions
of the sampling frequency. Additional filtering would alleviate this.
III. What changes to the Philips do the audiophile players make?
The Mission 7000R appears to be just a Magnavox 1040 with added remote
control and a black epoxy filter module. That's it. The filter module
attenuates the supersonic spurae and restores phase linearity altered by
the anti-aliasing filter. There must be a whole $10 worth of parts in it.
The Meridian MCD player incorporates their own analogue circuit board with
improved power supply regulation on a Magnavox 1000 chasis. They have
received criticism in Stereophile and HFN&RR for poor DC blocking capacitors.
The Meridian Pro has taken the same chassis and put the digital to analogue
circuitry in a separate shielded enclosure underneath. This cut noise by
30db. Noise was also diminished by using a separate power supply instead
of sharing the supply which powers the laser tracking mechanism. The
DAC's output feed a better opamp for better current to voltage conversion
which is then amplified by a discrete, DC coupled servo amplifier. A more
stable clock has improved high-end detail. Different DACs are used (?).
IV. Here are the modifications I performed.
First, I ordered the Magnavox FD1040 service manual and CD service training
manual from NAP (North American Philips). I called them in Tennesee at
(615) 475-3801 and asked for the publications department. The service
manual costs $12.50 and the training manual is $4.00. The training
manual contains unnecessary but interesting theory of operation. They take
credit cards or will bill amounts under $15.00. Delivery is prompt.
Second, I bought a #T-10 Torx driver at Sears. The player is full of torx
screws. They are used so people who don't know what they are doing won't
electrocute themselves. Lucky for me, hot wiring is out of the way!
The first change I made was to bypass the 22uf dc filtering electrolytic
caps and used a pair of 6.8uf WIMA metalized polycarbonate caps just
before the output cables. Musical Concepts replaces the 22uf with a
10uf metalized polypropelene, and the Mod Squad uses a series of bypassed
polypropelenes adding up to over 20uf. Alternatively, you could replace
these capacitors (2608 and 2573) with wire and use large films in an
external chasis. Room is tight inside for large film capacitors.
Big difference in sound. You don't have to be a "golden ears" to hear
this one. There is now much more detail. Dynamic range increased
a little upwards and lots downwards. A lot more depth and ambience.
Bass through the midrange was tighter and didn't break up on crescendos.
However, the low-end improvements left the high end weaknesses glaring.
My next job was smoothing out the high end. This required some observation.
The power supply feeding the opamps only has two 22uf electrolytics
after the linear regulator. I replaced one pair with 220uf electrolytic
caps and shunted with 2.7uf polycarbonate films. I also added the same
films before the regulator. The opamps themselves are poorly decoupled.
On the +/- 12v supplies there is a 100 ohm resistor and 22nf ceramic
capacitor to help cut noise from the servo section. That is not enough
capacitance and one of the capacitors is even missing! I added 22uf
electrolytics by each of the supplies. The sound got a little smoother
and the bass tightened up some.
I made a few other substitutions, mainly replacing a couple of the
surface mount resistors and capacitors with metal film resistors and
mica capacitors. These did not have the dramatic improvements that
previous changes had. John Hillig at Musical Concepts sent me his
opamps to try (numbers removed) and they were a definite improvement
with tighter bass and better imaging.
The high end has remained a problem for me because it is still coarse,
especially on transients. Crescendos and piano highlight it. After
discussions with the modifiers at Musical Concepts and The Mod Squad,
I'm convinced that the cause is the supplied audio cable I still use.
I wanted to keep the modification cost at a minimum, and a pair of
Tiffiny sockets or good interconnect cost more than the combined cost
of components I used for the modification. Alas, it does matter.
I stopped making mods. Crescendos don't make me cringe as badly
and additional mods are physically harder with diminishing returns.
There are quite a few surface mounted capacitors and resistors. They
make small Passport radar detectors possible, but sound lousy in audio
circuits. Space is fairly tight on the 1040 and they were necessary
for the designers to use. For example, there are 10 caps to filter
the 10 active bit outputs on the DACs. The 1040 used two mylars and
eight SMC ceramics. The 2040 uses five mylars and five ceramics.
One area I have not investigated is lowering the noise by grounding
improvements. I have heard of someone that does this and has succeeded
in reducing problems affecting tube preamps. I'm still investigating.
In short, in order of sonic importance, the changes are as steps,
take a Magnavox 2040 and first replace the DC blocking capacitors.
Second, I suggest replacing the audio interconnect. Third, improve
the power supply decoupling on the DACs and opamps. Fourth, find some
op amps better than the stock NE5532's. I can't help here because I
don't know what good, available ones are. Fifth, replace the row of
10 filtering caps on the DACs with polypropelenes and upgrade/match
all the components that form the current references on pins 15, 16,
and 17 of the DACs. Lastly, upgrade the remaining passive components,
which is nearly impossible on a 1040 or 2041 with all of its SMCs.
The first two changes are the easiest and bring the most improvement.
Most audiophile could live with just them (until better players appear).
The changes also limit your time/money investment. Its a good place to
stop, listen, and decide whether to continue. Additional changes are riskier.
V. Caveats
These players are hard to work on. The traces are very thin and pull up
easily. I've got several wirewrap patches because of it. Use both a solder
sucker and Solderwick and be careful. Beware of component labeling errors
in the service manual. One of the DC blocking capacitors is misnumbered on
the PCB print for the 1040. The laser and some chips are static sensitive.
Putting audio sockets into the rear heat-sink panel is quite a chore.
Philips uses a really tough alloy. Players made after September 1985
do not have this problem as sockets are now standard. Better sockets
and better interconnect to the circuit board could be used on those.
Modifications may violate the 1 year warranty. NAP's minimum service
charge is $65 and they charge steeply for parts.
If you don't think I've been explicit enough describing which components
I changed, have the changes done professionally or read the the articles
mentioned under "Suggested Reading."
VI. Professional Modification.
Let a professional do the modifications. They know what they are doing
and have done many players. They use the finest sounding components
available. The prices may seem a little high for some components and
their installation, but you are also getting the time these places have
spent listening to different capacitors choosing the best.
The two professional modification houses for CD players are:
Musical Concepts The Mod Squad
#1 Patterson Plaza 542 Coast Highway 101
Florissant, MO 63031 Leucadia, CA 92024
314-831-1822 619-436-7666
Both places do basicly the types of changes I have outlined. Both use
better opamps and capacitors and improve the power supply. Musical
Concepts uses 10uf metalized polypropelene DC blocking capacitors in
all their modifications, whereas Mod Squad can't fit their 20+uf
capacitors in a Magnavox 1040 and uses a good electrolytic bypassed with
a polypropelene. The Mod Squad's approach may have a deeper low-end.
Musical Concepts replaces the nasty polystyerene capacitors in the anti-
aliasing filter with quality, custom polypropelenes. This is available
in their supermod and not available from the Mod Squad.
The Mod Squad and Musical Concepts both modify Mission and Meridian CD
players with great benefit. These players are better than stock Magnavoxen
but need improvement to compare with a modified unit. The Mod Squad
will modify the MCD Pro and Kyocera units, and Musical Concepts will
modify a Revox and is considering doing Sony's.
Musical Concepts prices are $180 for the basic modifications and $280 for
the supermod. The Mod Squad's prices range from $200 for a standard
Magnavox 1040 mod, to $500 for a Meridian MCD with the analog circuits
housed separately and fed with a separate low impedence regulated power
supply. There are several options in between and accessories like
the CD damper, VPI magic brick and Tiptoes. Contact them for details.
VII. Is it worth it?
Yes, if your stereo system and you ears are up to it. If you are looking
for a CD player now but have the ears of Julian Hirsh, buy a Magnavox
player now and have the option of later modification. As your listening
and system improve, you don't have to get a new player and Magnavoxen are
cheap. This won't be your last CD player. CD players are getting better
and cheaper. I consider them almost disposable because of limited life,
high repair cost, low replacement cost and improving technology.
A modified Magnavox is one of the best five sounding CD players available
at any price, and also the least expensive. However, modified anythings
do not command the resale value they are worth. I think all CD players
will have poor resale value though, because they are getting better and
cheaper so fast. You may want to limit your modification investment because
of this and the limited product life. Philip's new 16-bit players
might sound worse than current players - there is no telling until we
get to hear them. Another option may be to buy one of the modified players
comming out from PS or others. There is no telling how they will sound
yet either, but they will cost more than a modified Magnavox and hold value.
VIII. What player to buy?
The Magnavox 2040 is recommended over the 1040 because it has better
components inside and there is more room for modifications. Players
made after September 1985 (FD2040SL02) have jacks in the back instead
of the captive cable. This saves difficult drilling work. Discount
price is $179 in the US. Britians pay $350 for it!
The new Magnavox 2041 player is available, but the service manuals are
not. It is made of plastic and has surface mounted components. On
the positive side, it sounds better out of the box due to better power
supply decoupling, better opamps, better capacitors in the anti-aliasing
filters, FETs that replace the relays in the signal path, and the
elimination of the very nasty sounding audio interconnect. There is
also cavernous room inside for a better audio board and power supply.
IX. Do transports and number of lasers matter?
No. Bits are bits. All players on the market read them competantly.
While the soft error rate is high (1 in 10**4 bytes), hard errors
requiring interpolation (guessing) are typically 1 in 10**9 bytes.
On the other hand, people say that a CD damper really works.
A CD damper disk on top that Mod Squad likes or a second CD that HFN&RR
uses while playing perhaps improves the sound by lowering internal
electrical noise caused by the laser focusing servo. Laser tracking
arm servo noise can be reduced by buying Philips disks! Their CDs
are laser centered before the hole is punched. Accuracy is within a
few microns! Other CDs have the hole molded in and can be off center
by a milimeter - as bad as LPs!
X. What's down the road?
Philip's 4x oversampling 16 bit chip set is due out sometime in 1986.
Little is said about these new chips from Signetics. They may have more
low-level details and high frequency "air" than the current 14 bit ICs.
Mission's contribution is their filter module. It restores phase linearity
to the high frequencies that was altered by the anti-aliasing filter.
Hopefully an implementation with better parts than Mission's will appear.
Sony's best advance was their one clock digital section. The use of
many good, strong clocks causes intermodulations or "beat frequencies"
between them that are picked up in the analog section. Hear them on
your FM tuner. A single clock does not have beat frequencies.
Onkyo showed a CD player in Japan that had separate chassis for analog
and digital sections connected by fiber optic cable! This allows total
electrical isolation between circuits for minimum noise figures.
XI. Personal Thoughts.
I used to be ardently anti-digital. That is until I heard about and
tried upgrading CD players. Changing the DC blocking capacitors alone
convinced me that there was much unrealized potential in my player.
It's really frustrating to see the simplest last 2% of electronics destroy
all the magic of reading .6 micron divots. Its just like putting $2500
worth of turntable, tonearm and cartridge through a $50 preamp. The
sound quality is lost in the cheap electronics. CD players that put
a few extra dollars into the analog section sell for many hundreds
of dollars more retail. Most top of the line players are a rip-off.
I have lost interest in my CD player. The reason is CD prices. LPs
usually cost half as much and I have yet to see a CD cut-out bin. In
the last few months I have bought 25 direct to disk recordings from
cut-out bins for an average price of $3.00 each. Stores are dumping
audiophile records to make room for CDs and import pressings are nearly
on parity with domestic pressings. I will take a chance on a new or
used LP. I won't on a $13 CD. Few CD stores allow auditioning CDs.
Another reason I have lost interest in CDs is that since the original
posting in September, I replaced my Kenwood turntable with a Linn Sondek,
replaced my interconnects with Straightwire Flexconnect, and upgraded
the capacitors in my preamp to teflons. Now there is not enough sonic
difference for me to justify paying twice as much for CDs as LPs.
XII. Suggested Reading
Picking Capacitors, Walter G. Jung and Richard Marsh; Audio Magazine,
February and March, 1980.
Modifying the Marantz 7C or St. Pooge and The DRIAAGON, Walter Jung
and Chuck Hollander; The Audio Amateur, 1/1981.
POOGE-2 A Mod Symphony for your Hafler DH200 or other Power Amplifiers,
Walter Jung and Richard Marsh; The Audio Amateur, 4/1981.
Omitted Factors in Audio Design, John Curl; Audio, September 1979.
Stereophile, Vol 8, #7 due out soon will supposedly have an article by
George Graves on how to modify a Magnavox. The Dec. 1984 issue of
High-Fi News and Record Review had "CD sweetening" suggestions.
XIII. Evaluation tools used during CD player modifications
CDs: DMP's Flim and the BB's - Tricycle, Telarc's Chopin, CBS's
Dimeola, Delucia, McLaughlin - Friday Night in San Francisco. I also
used the Telarc LP and the Japanese Sony/CBS LP of McLaughlin.
Equipment: Kenwood KD500/Grace 707/Grado Signature 8M, Musical Concepts
modified Hafler DH110 and DH500, modified M&K speakers, Signet TK33
electret headphones, Kenwood interconnects, Monster cable, Tiptoes.
DISCLAIMER: The only relationship I have with Musical Concepts and The Mod
Squad is that of satisfied, paying customer. I found this adventure bene-
ficial and instructive. I wanted to share it with net.audio before CD
player modification information is available in audiophile journals.
Mark Kaepplein (617) 480-6652 Digital Equipment Corporation
Kaepplein%amber.dec@decwrl 150 Locke Drive LMO4/H4 Box 1015
[email protected] Marlboro, Massachusetts 01752-9115
...!decwrl!dec-rhea!dec-amber!kaepplein or [email protected]
|
|
Newsgroups: net.audio
Path: decwrl!dec-rhea!dec-amber!kaepplein
Subject: Modifying Magnavox CD players / Why most CD players sound awful
Posted: 9 Jan 86 23:40:02 GMT
Organization: Digital Equipment Corporation
Modifying Magnavox/Mission/Meridian/Revox Compact Disc players, rev. 1
Would you believe a $170 Magnavox CD player can sound as good as $1,500
worth of turntable, tonearm, and cartridge? It can, but most sonic
quality gets lost in an analog audio section on par with a $50 preamp.
You or a professional can modify the audio section to reveal CD's
potential. This posting is for those who can or want to hear player
differences, those who wonder what is inside, and audio hackers!
I explain design flaws in CD players that until recently corrected,
caused great scorn for the CD among audiophiles. This is my report on
experimentation and analysis correcting problems in a CD player's
analog audio circuitry, demonstrating why players sound different.
Also, some personal thoughts and details on future products are given.
The modification techniques used here are not new. They have been
applied to amps and preamps for years. What is new however, is the
application to CD players, and identifying sonically important changes
in circuits with digital to analog converters (DACs). These techniques
are also applicable to laser video disk players and "hi-fi" VCRs.
I. Why modify a Magnavox CD player?
Don't be put off by the Magnavox name or the low price tag. Magnavox,
Sylvania, and Norelco are US trade names for Philips, codeveloper of the
CD with Sony. Philips based CD players employ quadruple oversampling,
two DACs, and both digital and analog anti-imaging filters. The result
is less phase shift than other techniques.
Nearly all the top CD players use the Philips technology. Brands based on
Philips hardware include: PS, Cambridge Audio, Meridian, Mission, Discrete
Technology, McIntosh, Revox, Marantz, and Magnavox. This is an elite group
where the best player(Cambridge) costs $2500, yet a Magnavox can be had for
well under $200. Several of these players are an insult, considering the
minor changes for which they charge hundreds of dollars.
Magnavox uses the right technology, but its execution is wrought with
cost compromises necessary in all mass market consumer products. Sound
quality is generally lost on the masses. Compromises ravage the audio
section with essentially transistor-radio quality parts.
Simple substitution of high quality parts in the audio section will make
dramatic improvements. It will solve most of the sonic problems that The
Absolute Sound and others have identified. While I can't prove this
conclusively to every reader, I will identify the sonic improvements
occurring with specific changes. I will also list other changes that I have
not tried, but that others have found valuable.
II. The weaknesses of Magnavox players
The sound is compressed -- lacking low level dynamics. Imaging is shallow.
Transients are distorted and there is a general glare and edge. Problems
show up on the tough test of horns, violins, and piano. These problems
can all be traced to weaknesses in the audio section and are not because
"digital" has sapped the soul from the music.
The problems are: mediocre opamps, a noisy, poorly decoupled power supply
with bandwith and current limited regulation, use of ceramic and electro-
lytic capacitors in the signal path, and very poor audio cables. Magnavox
1040s and 2041s include numerous, compact Surface Mounted Components (SMCs)
of dubious sonic quality. Magnavox 2040s do not.
Internal noise from clocks and digital signals is a problem. Solution
requires separate grounding of the audio circuitry and preferably,
evacuation into a separate shielded enclosure. The DACs produce noise too,
concentrated at fractional frequencies of the 176.4 khz sampling rate.
Additional filtering would alleviate this.
III. What changes to the basic Philips do the audiophile players make?
The Mission 7000R appears to be just a Magnavox 1040 with added remote
control and a black epoxy filter module. That's it. The filter module
attenuates the supersonic spuriae and restores phase linearity altered by
the anti-imaging filter. There must be at least $5 of parts in it.
The Meridian MCD player incorporates their own analogue circuit board with
improved power supply regulation on a Magnavox 1000 chassis. They have
received criticism in Stereophile and the British publication Hi-Fi News
and Record Review (HFN&RR) for poor DC blocking capacitors.
The Meridian Pro has taken the same chassis and put the DACs and analogue
circuitry in a separate shielded enclosure underneath. This cut noise by
30db. Noise was also diminished by adding a separate power supply instead
of sharing the supply which powers the laser tracking mechanism. The
DAC's output feeds a better opamp for better current-to-voltage conversion
which is then amplified by a discrete DC coupled servo amplifier. A more
stable clock improves treble detail. There is a phase reversal switch.
IV. Here are the modifications I performed.
First, I ordered the Magnavox FD1040 service manual and CD service training
manual from North American Philips. I called them in Tennessee at
(615) 475-3801 and asked for the publications department. The service
manual costs $12.50 and the training manual is $4.00. The latter
contains an unnecessary but interesting theory of operation. They take
credit cards or will bill amounts under $15.00. Delivery is prompt.
Second, I bought a #T-10 Torx driver at Sears. The player is crammed with
torx screws. They are used so people who don't know what they are doing
won't electrocute themselves. Lucky for us, hot wiring is out of the way!
The first change I made was to bypass the 22uf dc filtering electrolytic
caps. I used a pair of 6.8uf WIMA metalized polycarbonate caps just
before the output cables. Musical Concepts replaces the 22uf with a
10uf metalized polypropelene, and the Mod Squad uses a series of bypassed
polypropelenes adding up to over 20uf. Alternatively, one could replace
these capacitors (2608 and 2573) with wire and use large films in an
external chassis (room is tight inside).
Big difference in sound. You don't have put on golden ears to hear
this one. There is now much more detail. Dynamic range is increased
a little upwards and a lot downwards. Lots more depth and ambience.
Bass through the midrange was tighter and didn't break up on crescendos.
However, the low-end improvements left the treble weaknesses glaring.
My next job was smoothing that out. This required some inspection.
The power supply feeding the opamps has only two 22uf electrolytics
after the linear regulator. I replaced one pair with 220uf electrolytics
and bypassed them with 2.7uf polycarbonate films. I also added the same
films before the regulator. The opamps too are poorly decoupled.
On the +/- 12v supplies there is a 100 ohm resistor and 22nf ceramic
capacitor to help cut noise from the servo section. Not enough, and
one of the capacitors is even missing! I added 22uf electrolytics to each
supply. The sound got somewhat smoother and the bass tightened up some.
I made several other substitutions, mainly replacing a couple of the
surface mount resistors and capacitors with metal film resistors and
mica capacitors. These did not have the dramatic improvements that
previous changes produced. John Hillig at Musical Concepts sent me his
opamps to try (numbers removed) and they were a definite improvement
with tighter bass and better imaging.
The treble remains a problem for me because it is still quite coarse,
especially on transients. Sudden crescendos and piano expose it. After
discussions with the modifiers at Musical Concepts and The Mod Squad,
I'm convinced its caused by the supplied audio cable I still use.
I had resolved to keep the modification cost low, and a pair of good
sockets or interconnect cost more than the combined cost of components
I used for the modification. Alas, scrimping doesn't pay..
I ceased making mods. Crescendos no longer make me cringe, and additional
mods are physically harder with diminishing returns. There are too many
surface mounted capacitors and resistors. They make miniature radar
detectors possible, but sound awful in audio circuits. Space is tight on
the 1040 so the designers used surface mount technology. For example,
there are 10 caps to filter the 10 active bit outputs on the DACs. The
1040 used two mylars and eight SMT ceramics. The 2040 uses five mylars
and five ceramics.
In summary, in order of sonic importance, the changes come in steps.
First take a Magnavox 2040 and replace the DC blocking capacitors.
Second, I suggest replacing the audio interconnect. Third, improve
the power supply decoupling on the DACs and opamps. Fourth, find some
op amps better than the stock NE5532's. I can't help here because I
don't know what the good, available ones are. Fifth, replace the row of
10 deglitching caps on the DACs with polypropelenes and upgrade/match
all the components that form the current references on pins 15, 16,
and 17 of the DACs. Lastly, upgrade the remaining passive components,
which is nearly impossible on a 1040 or 2041 with all of its SMCs.
The first two changes are the easiest and bring the greatest improvement.
Most audiophiles could live with just them until better players appear.
These two changes also limit your time&money investment. Its a good place
to stop, listen, and decide whether to continue. Additional changes involve
more risk of breaking your machine. I strongly discourage them on a 1040.
V. Caveats
These players are hard to work on. The PC traces are very thin and pull up
easily. I now have several wirewrap patches because of it. Use a solder
sucker, Solderwick and extreme caution. Beware of component labeling errors
in the service manual. One of the DC blocking capacitors is misnumbered on
the PCB print for the 1040. Use anti-static grounding.
Putting RCA audio sockets into the rear heat-sink panel is a chore.
Philips uses a really tough alloy. Players made after September 1985 do
not have this problem as sockets are now standard. Better sockets and
better interconnect to the circuit board could be used on those.
Modifications may violate the 1 year warranty. NAP's minimum service
charge is $65 and they charge steeply for parts.
If you think I was too vague describing my changes, its deliberate.
Unless you know what you're doing, have the changes done professionally
or read the the articles mentioned under "Suggested Reading."
VI. Professional Modification.
Let a professional do the modifications. They know what they are doing
and have done many players. They use premium components. The prices may
seem a little high for some components and their installation, but you are
also getting the time these places have spent listening to different
capacitors to choose the best.
The two professional modification houses for CD players are:
Musical Concepts The Mod Squad
#1 Patterson Plaza 542 Coast Highway 101
Florissant, MO 63031 Leucadia, CA 92024
314-831-1822 619-436-7666
Both places basically perform the types of changes I have outlined. Both
use better opamps and capacitors and improve the power supply. Musical
Concepts use 10uf metalized polypropelene DC blocking capacitors in
all modifications, whereas Mod Squad can't fit their 20+uf capacitors into
a Magnavox 1040, so use a good electrolytic bypassed with a polypropelene.
Musical Concepts replaces the nasty polystyrene capacitors in the anti-
imaging filter with quality, custom polypropelenes. This is available
in their "supermod" and not available from the Mod Squad.
The Mod Squad and Musical Concepts both modify Mission and Meridian CD
players with great benefit. These players are better than stock Magnavoxen
but need improvement to compare with a modified unit. The Mod Squad
will further modify the MCD Pro and Kyocera units, and Musical Concepts
will modify the Revox and is considering doing Sonys.
Musical Concepts prices are $180 for the basic modifications and $280 for
the "supermod". They also sell a complete modified unit. The Mod
Squad's prices range from $200 for a standard Magnavox 1040 mod, to $500
for a Meridian MCD with the analog circuits housed separately and fed with
a separate low impedance regulated power supply. There are several options
inbetween and accessories like the CD damper, VPI magic brick and Tiptoes.
[ I just heard that Acoustic Electronics, maker of the AIR amplifier,
modifies Magnavox players for $400. P.O. Box 13, Highlands, NJ 07732.]
VII. Is it worth it?
Yes, if your system, your ears, and your wallet are up to it. If you are
looking for a CD player now but identify with Julian Hirsh, buy a Magnavox
now and have the option of later modification. As your listening acuity
and system improve, you won't have to get a new player and Magnavoxen are
cheap anyway. This won't be your last CD player. They are getting better
and cheaper. I consider them almost disposable because of limited life,
high repair cost, low replacement cost and improving technology.
A modified Magnavox is one of the best sounding CD players available
at any price, and also the least expensive. However, modified anythings
do not command the appropriate resale value. I think all CD players
will have poor resale value though, because they are getting better and
cheaper so fast. You may want to limit your modification investment because
of this and the limited product life. Philip's new 16-bit players
might sound worse than current players - there is no telling until we
get to hear them. Another option may be to buy one of the modified players
coming out from PS or others. There is no telling how they will sound
yet either, but they will cost more than a modified Magnavox and hold value.
VIII. What player to buy?
The Magnavox 2040 is recommended over the 1040 because it has better
components inside and there is more room for modifications. Players
made after September 1985 (FD2040SL02) have jacks on the rear instead
of captive cables. This saves difficult drilling work. Discount price
is $169-179 in the US. Britons pay #225 for it!
The new Magnavox 2041 player is available, but the service manuals are
not. It is made of plastic and is crammed with surface mounted components,
making modification very difficult. On the positive side, it sounds
better out of the box due to better power supply decoupling, better opamps,
better capacitors in the anti-imaging filters, and no nasty audio cables.
Unfortunately, electrolytic DC blocking capacitors are still used. There
is room inside for better power supply regulation and discrete amplifiers.
IX. Do transports and redundancy of lasers matter?
Maybe. Bits are bits. All players on the market read them competently.
While the soft error rate is high (1 in 10**4 bytes), hard errors
requiring interpolation (guessing) are typically 1 in 10**9 bytes.
On the other hand, people say that a CD damper really works.
A CD damper disk on top that Mod Squad likes or a second CD that HFN&RR
uses while playing perhaps improves the sound by lowering internal
electrical noise caused by the laser focusing servo. Laser tracking
arm servo noise can be reduced by buying Philips disks! Their CDs
are laser centered before the hole is punched. Accuracy is within a
few microns! Other CDs have the hole molded in and can be off center
by a millimeter - as bad as LPs! CD manufacturing quality will undoubtedly
continue to decline as manufacturers cut costs and maximize profits.
Fortunately player error correction continues to improve and can handle
bubbles, holes, eccentricities, and warps. Another CD perfection dispelled.
X. What's down the road?
Philip's 4x oversampling 16 bit chip set is due out sometime in 1986.
Little is said about these new chips from Signetics. The DACs will
be a full 16 bit. There will be subcode and digital output and improved
error correction. Judging from preliminary published specifications
and circuits they will still use electrolytics in the signal path,
so modifications will be possible and necessary for good sound.
Mission's contribution is their filter module. It restores phase linearity
to the high frequencies that was altered by the anti-imaging filter.
An implementation with better parts than Mission's will undoubtedly appear.
Another boon would be selectable phase restoration circuits to compensate
for the 9 or 13 pole anti-aliasing filters used in studio recorders.
Expecting the studio to compensate is almost as far fetched as them
converting to non-linear quantization and oversampling, but let's hope.
Sony's primary advance was their one clock digital section. The use of
many strong clocks causes intermodulations or "beat frequencies" that are
picked up in the analog section. A single clock solves this problem.
Onkyo showed a CD player in Japan that had separate chassis for analog
and digital sections connected by fiber optic cable! This allows total
electrical isolation between circuits for minimum noise figures.
XI. Personal Thoughts.
I was once ardently anti-digital. Then I heard about and tried upgrading
CD players. Changing the DC blocking capacitors alone convinced me that
there was vast unrealized potential in my player.
It's amusing to see the simplest last 2% of electronics destroy all the
magic of reading .6 micron divots, just like putting $1500 of turntable,
tonearm and cartridge through a $50 preamp. All sonic quality is sacrificed
to the cheap electronics. CD players that put a few extra dollars into the
analog section sell for many hundreds of dollars more retail. Yet, this is
better than no correlation between price and performance. Most top of the
line players are overpriced, but Magnavox, PS, and Meridian provide value.
Lately I have lost interest in my CD player. The reason is CD prices.
LPs usually cost half as much and I have yet to see a CD cut-out bin.
In the last few months I have bought 25 direct to disk recordings from
cut-out bins for an average price of $3.00 each. Stores are dumping
audiophile records to make room for CDs and import pressings are nearly
on parity with domestic pressings. I will take a chance on a new or
used LP. I won't on a $13 CD. Few stores permit audition of CDs.
Another reason I lost interest in CDs is that since the original
posting in September, I replaced my Kenwood turntable with a Linn Sondek,
replaced my interconnects with Straightwire Flexconnect, and upgraded
the capacitors in my preamp to teflons. Now there is not enough sonic
difference for me to justify paying twice as much for CDs as LPs.
XII. Suggested Reading
Picking Capacitors, Walter G. Jung and Richard Marsh; Audio Magazine,
February and March, 1980.
A Passive Role?, Martin Colloms; HiFi News & Record Review, 10/1985.
Capacity to Change, Martin Colloms; HiFi News & Record Review, 12/85.
Modifying the Marantz 7C or St. Pooge and The DRIAAGON, Walter Jung
and Chuck Hollander; The Audio Amateur, 1/1981.
POOGE-2 A Mod Symphony for your Hafler DH200 or other Power Amplifiers,
Walter Jung and Richard Marsh; The Audio Amateur, 4/1981.
Omitted Factors in Audio Design, John Curl; Audio, September 1979.
Stereophile, Vol 8, #7 due out soon will supposedly have an article by
George Graves on how to modify a Magnavox. The Dec. 1984 issue of
High-Fi News and Record Review had "CD sweetening" suggestions.
XIII. Evaluation tools used during CD player modifications
CDs: DMP's Flim and the BB's - Tricycle, Telarc's Chopin, CBS's
Dimeola, Delucia, McLaughlin - Friday Night in San Francisco. I also
used the Telarc LP and the Japanese Sony/CBS LP of McLaughlin.
Equipment: Kenwood KD500/Grace 707/Grado Signature 8M, Musical Concepts
modified Hafler DH110 and DH500, modified M&K speakers, Signet TK33
electret headphones, Kenwood interconnects, Monster cable, Tiptoes.
XIV. Late breaking information on the PS Audio CD1 Player
Yet Another Modified Magnavox (YAMM) is now available. It comes pre-
modified from a largish American manufacturer of audiophile equipment.
Costing $800, it is more expensive than sending a unit to a modifier,
but will hold resale value. PS may be a good alternative for many.
Most noteworthy about the PS modifications are two new problems they
solved: power supply and ground isolation, and TIM distortion in the
anti-imaging filter and buffer stage. PS did a lot of work on the power
supply, beefing all of them up, even the +5v logic supply. They also
isolated the grounding to prevent noise from reaching the audio section.
PS realized that high frequencies, high negative feedback, and moderate
bandwith opamps is a formula for slew rate limited distortion generation.
PS solved this by first passively filtering the output of the DACs, which
is composed of 176.4khz square wave components. Later it is actively
filtered in a discrete amplifier circuit. Last, it goes through a discrete
class A buffer instead of the unity gain opamp buffer. Measured wide-
band THD was reduced from a stock .21% to .05%.
The PS has been hot in west cost audio. Strong reviews are coming out
from Stereophile and IAR. Based on the modifications they do, it is
probably better than any player available directly from a manufacturer
with the exception of the $2500 Cambridge. Comparison with Mod Squad
and Musical Concepts machines would require extensive listening. They
solve many of the same problems the PS does, but also additional ones.
DISCLAIMER: The only relationship I have with Musical Concepts and The Mod
Squad is that of satisfied, paying customer. I found this adventure bene-
ficial and instructive. I wanted to share it with net.audio before CD
player modification information is available in audiophile journals.
Mark Kaepplein (617) 480-6652 Digital Equipment Corporation
Kaepplein%amber.dec@decwrl 150 Locke Drive LMO4/H4 Box 1015
[email protected] Marlboro, Massachusetts 01752-9115
...!decwrl!dec-rhea!dec-amber!kaepplein or [email protected]
|