Modifying Sony CDP-777ESA Part1
modding till Sep.13 2004
It was 1991 when Sony introduced CDP-777ESA to the Japanese market. The
CD player stayed at the top of Sony's ES series models then. Suggested
retail price was JPY200,000. It was categorized to the third generation
of Sony's CD players using their Pulse DAC technology (first introduced
in 1989), by the use of CXD2562Q voltage-output DAC LSI. CXD2562Q was next
used as a front-end DAC chip in their later models, such as CDP-777ESJ,
followed by a current-output LSI.
In addition to the Pulse DAC technology, CDP-777ESA claimed its advantages
as follows:
- Descrete MOS-FET single-ended output stage with DC-Servo
- GIC(General Inpedance Converter) type active analog LPF
- Using two CXD2562Q chips, and making balanced output signals directly
from balanced output of DAC chips
- Center locating CD dirive mechanism
You can find more general description on CDP-777ESA in the webpage below
(Japanese only, but with many pictures).
http://cdp101.hp.infoseek.co.jp/cdp777esa/cdp777esa.html
Frankly I am not a guy who loves Sony's Hi-Fi sound very much. It often
sounds without enough resolution or solidness for me, and it sometimes
makes me even think Sony's equipment spoils passion of music. This characteristic
may not be bad for some type of classical music, but particulary bad for
most of POPs or Jazz.
On the other hand, I cannot completely hate the rigid mechanical construction
and quality-feel finish in their products.
----
Spring 2004 in Akihabara, I happened to find a second-handed CDP-777ESA
in excellent condition at JPY50,000, with being claimed recent service
on the pickup mechanism. I finally carried it by hand to my home despite
its 17kg mass. Still I am not sure why I did do that. Some invisible aura
might caught me then.
I set the player up and listened to its sound without much expectation.
Yeah, the sound was nothing different from what I had known about Sony's
or most of Japanese made Hi-Fi equipments. It's very smooth and soft and
of no passion, no resolution or no solidness.
At that moment, I just started modifying my integrated amplifire (Arcam
A85). It was my first trial of modification to any kind of Hi-Fi equipment.
I made up my mind someday I would try on the player after I could finalize
the Arcam. The goal should be to improve the sound, at least, up to its
mechanical construction.
----
During summer of 2004, after I did some minor mods on Arcam A85 (just rectifiers
and electrolytics, leaving tons of TL072 as it were), I modified Arcam
CD72 CD player. That was the player my friend bought as second hand, and
he requested me to modify. Detail modifications are described here
in my site, in Japanese only, but with many pictures. (I will translate
these into English, if many many guys request it)
The original sound of CD72 is good, passable, but nothing special. It was
(of course,) far behind in any aspect to my reference CD player - Wadia16.
I love Wadia16, both in its sound and appearance, but some CDs sometimes
make me feel Wadia sounds a bit softly. So, the primary goal of my modding
CD72 was to get higher resolution, harder than Wadia16 as second, so that
I can buy and enjoy this Arcam even when it fails to satisfy my friends
:-)
 Fig.1 Arcam CD72 Modified |
 Fig.2 Arcam CD72 Original Main PCB |
I think, at least I believe, the modded Arcam achieved above two goals.
Major effective mods on CD72 were as below in order of the amound of (subjective)
improvement.
- decouple DAC(PCM1716E) analog power separately (Vcc1, Vcc2L and Vcc2R)
- separate opamp decoupling electrolytics for L and R
- change opamps (NJM2114D -> AD8065x2)
- change Xtal to precision oscilator module
- change rectifiers to SBDs
- independent opamp power supply (separate transformer, increase electroltics)
After all the above mods, CD72 has still one significant disadvantage against
Wadia - atomosphere. I tried many things to get it, but finally I gave
up. So then, reproducing atmosphere became one of the major goal of my
modding Sony.
----
It's a Sony.
Fig.3 Sony CDP-777ESA (original)
CDP-777ESA has much more rigid mechanical construction than Arcam. It seems
worth trying to get maximum out of the capability of the cnstruction by
modding its circuitry, despite its known difficulties such like many many
opamps or two xtals architecture.
A fact encourages me, however. Sony is one of the originators of CD format.
It is possible to believe, for me, they must have something different than
many other licensees based on their long term experiences in researching
digital sound codec. I started my mods trusting Sony's tons of opamps and
FETs will work fine with appropriate mods (not only as a marketing gag),
and then the performance will get closer to what Sony guys heard in their
lab, or at least, in their mind.
----
I started with DAC/Analog-Output PCB, locates right hand side in above
picture.
Fig.4 DAC & single-end analog output PCB (original)
Upper half of the PCB is for R channel and lower is L.
Two DACs, CXD2562Q, are mounted on the pattern surface of the PCB, at two
white printed frames rounded by blue ceramic capacitors.
Each digital power line are decoupled by Nichemi
AVF grade 100uF electrolytics (total five, right hand side of PCB), and
analog powers for DAC are decoupled by Panasonic(?) Purism grade 100uF
electrolytics (five for each DAC chip. One is with copper shield).
All of DAC power pins are separately bypassed (nine blue cera caps for
each DAC chip. Only eight are visible in the above picture). This is one
of the good things on this board, I think. Xtal for DAC locates very close
to DAC chips and this is also good.
(See Fig.5 for close-up)
+5V for digital circuitry is made on this board, from +15V for analog circuitry.
I am not sure but the NE5532P at the center of the PCB, near the big electrolytic,
may work for making +5V. +5V is decoupled by this big electrolytic capacitor
- Nichemi 1200uF/63V - then used for DAC clock via Nichemi AWF 100uF, and
for DAC and other digital circuitry.
One DAC chip has eight analog voltage output signals - means four parallel
differential outputs. Four are broken into two set of parallel, where one
parallel set is delayed to reduce switching? noise. On this board, eight
outputs go down to four with some LPF, then go into the 1st NE5532P which
works just as simple inverted buffer. Buffered differential signals are
also used for the balanced output circuitry on the other PCB.
On this main PCB, differential signal from the 1st NE5532P go into the
2nd NE5532P and converted to single ended signal.
Single ended signal go into the combination of 3rd NE5532P and AD712, which
may work as high order LPF.
LPF'ed signal seem to go to Tr driver which drives output FET (four green
and black ICs on the right). Final opamp - OPA27GP - works as DC servo
with 2uF film capacitor (biggest black lump at the left hand side of the
above picture. I will mention 2nd size lump (total four) later).
Fig.5 Decoupling & bypassing caps for DAC
----
While power supply lines for DAC are very well supported as above, power
supply for analog circuitry, opamps and FETs, are questionable. +15V and
-15V are regulated on PS PCB (upper center in Fig.3) by power transisters.
These are decoupled by 1200uF/63V electrolytics (same type capacitor as
+5V decoupling capacitor on main PCB) on PS PCB, and then go to the bus-bar
for opamp power and the FETs, separate lines for each from PS PCB to main
PCB. On main PCB, +/-15V for FETs and opamps are bypassed by 1uF film capacitors
(total four 2nd large black lumps), and there is no electrolytics on main
PCB to decouple +/-15V.
Both FET output buffer and tons of opamps, both L & R, all powers for
these devices are only decoupled by a single 1200uF electrolytics on PS
PCB far away from devices. It is very contrasty to the Arcam case where
many regulator ICs and electrolytics locates close to the devices. Simply
from my experience with Arcam, I felt something odd on this archtecture.
But I am not familier with bus-bar solution, I started modding, leaving
the bus-bar as the key component for opamps' power supply.
----
So first, based on my Arcam case where DAC PS decoupling was fatal, I decided
to sweep away all the capacitors around DAC and digital devices on the
main PCB, and replace them by the best popular hifi components.
Fig.6 Eighteen ceramic caps and ten Purism elec caps were swept away
Near the GND metal (copper), Soshin's
audio quality mica caps and xtal for DAC clock (master clock?).
Some of Nichemi AVF caps for digital +5V were already replaced by 220uF
OS-CONs.
My choice here was a famous BlackGate
for analog power lines, and OS-CON for digital power lines. Though there
are some negative comments on BlackGate, I have had nothing bad around
BlackGate when I modded Arcam. Maybe it just fit to my preference.
Fig.7 BlackGates and OS-CONs
All bypassing ceramic caps were replaced by BlackGate NX HiQ 47uF/6.3V.
Purism 100uF/50V electrolytics to BlackGate NX 220uF/6.3V. AVF to OS-CON
220uF, and I replaced two ceramic caps by 10uF OS-CONs ,and reinstall ceramic
caps on the pattern surface of the PCB. Bulky 1200uF/63V Nichemi custom-like
audio electrolytic was here replaced by BlackGate NX 2200uF/6.3V. Nichemi
AWF for DAC clock was to 470uF OS-CON.
----
I was not confident how I could improve power supply situation for opamps.
Anyway, I replaced two 1uF film caps by BlackGate NX 680uF/35V, and reinstall
film caps, ERO MKC .33uF, on the pattern surface.
Fig.8 Support bus-bar for opamps power
All opamps are now in sockets. Types of opamps shown in this picture are
not final.
White bar below BlackGate NX is Bus-Bar. It is made by three thin OFC board,
-15V, GND and +15V.
BlackGate FK 220uF at the center of this picture was originally Nichemi
AWF 100uF (same type as DAC clock decoupling), seems to be used to make
+5V.
----
I do not like many black bulky lumps on PCB, I replaced these by fancy
ERO MKC.
Fig.9 Lump and ERO
Fig.10 Fancy PCB (I am using Kodak LS753 digital camera. Thanks to Kodak
Color!)
This is not an exact picture at this stage. Pale green film capacitors
for LPF, replacing pale blue ones (see Fig.8), are installed after my 1st
trial failed. Will be mentioned later.
----
At this moment, though I decided to rely on bus-bar about opamps PS, I
thought I should do something on FET power supply.
I found a pretty good sprague electrolytics (470uF/16V) at Akihabara and
installed them for decoupling FET power supply.
Fig.11 Sprague
Upper left, two ERO MKC .33uF bypassing bus-bar
Above all are the first mods on the main PCB. Then I modded the analog
power supply PCB and the digital power supply PCB.
----
First, I changed rectifiers on analog PS PCB to SBDs.
Fig.12 Original rectifier (left) and SBD (right)
Electrolytics on analog PS PCB:
Two Nichemi AWF 1200uF/63V -> BlackGate FK 2200uF/35V
Six Nichemi AWF/AVF 100uF~220uF -> Rubycon YK 2200uF and Nichicon MUSE-FXs
Add some ERO MKC and MKP for high frequency stability.
Fig.13 New electrolytics and SBDs
Fig.14 ERO film capacitors added
----
For digital PS, I also replaced rectifiers by SBDs.
Fig.15 Four SBDs
Fig.16 Two AVFs to OS-CONs, and one AVF to MUSE
----
After I made all the above expensive mods on main/power PCBs, I finally
build up PCBs into the player, and listen to the music.
The result was very disappointing, unexpectedly.
Cont'd to part2.
Oct.14 2004
chiaki014@mac.com