I've now have my own SD-3990
player based on evaluation of a borrowed unit. In my modest system, I
hear a very coherent,, detailed and very natural balance of sound.
One of the people, I infected was a person who grew up on
planar speakers (Quad ESL-57 and Apogee Scintillas) and he thought
it was comparable to $1000 units and much better than his Rotel 1072.
There is an interesting thread on a new transport
called the Memory Player that dispenses with the conventional
Reed-Solomon error correction and basically re-reads the data
into a memory buffer until "correct". Could it be that some
of the issues with CD playback have to do with the digital
data stream from reading pits? Along this theme, I asked the question
no one question the digitized sound encoded onto LP
from Telarc? Were the Philips/Sony consortium right about 44.1
kHz being enough? This pre-CD digital acted as a springboard for the
question of "how much
better the sound could get if the digital bits could be heard directly
The belt driven CEC TL-51XZ
(itself an upgrade from the 2001 model TL-51Z) has been upgraded to an updated
TL-51XR version to apparently
good result. This is the transport that the well received
Ack! dAck was voiced with.
October 23, 2005
What started as an experiment by Gordon Rankin with a USB
DAC at the 2003 CES has now evolved to general acknowledgement of
improved sonics from computer sources when used as "transports". There
remain questions as to why this should be so and in an effort to
clarify some of the issues and to lay down my own thoughts on this,
I've created a section called Last
Frontier (Transport Jitter)
Looking for a good no nonsense DVD player for all your
various digital storage formats (DVD, CDs, CD-ROM, JPG), then look no
further than the Harmon Kardon DVD-31
There has been a flurry of activity over the last 9 months
with reliable and cost effective distribution of CD-PRO2
based DIY kits...arguably one of the most trusted transports
used in high-end consumer transports.
May 15, 2004
Put together a DIY
Roller Block set of feet using just pop cans. Lateral and
rotational energy are effectively damped by the units own moment of
March 10, 2004
The killer transport
has arrived in the form of the Toshiba SD-3950. It has transformed the
sound of my DACs and may do the same for you....on second thought...bet
on it. The unit has been recommended by user Rosconey
on AA. Two users with decent DACs in the form of
Vinnie R. and John
Swenson picked up on it and added significant fuel to the
"3950" fire. The rest you can read on AA.
A Chain is as strong as its weakest link. Cables,
amplifiers, and DACs all have a bearing on the final sound produced. My
experience has been that some of the brightness/hardness in the system
can be minimized by judicious selection of these components. I have
found that Cardas 2x21.5 gauge twinax interconnect cables between the
CD output and the preamp to make an immense difference to the
reproduced sound in terms of warmth and lack of hardness. Line filters
have also been used to great effect to isolate the effects of line
noise (RFI on the grid and common mode noise from refrigerators, air
conditioners, hair dryers, and PCs). In a world that is getting
increasingly transistorized, the signal to noise levels drop due to
lower supply voltages. Transistors are very sensitive to millivolt
levls of signal AND NOISE! Tubes offer a level of immunity to noise at
this level by virtue of the much higher operating voltages (hundreds of
volts versus tens of volts) and by virtue of the need to transformer
couple outputs. These transformers are great RFI filters.
What if we put a transformer on the very most frontend
portion of the read/playback system of a CD player/transport ? With its
assortment of motors and control circuits to maintain constant angular
velocity, lateral tracking servo and focus servos, it makes sense to
isolate the power coming into the transport as best as possible. Two
ideas on this front are by Jon Risch and Reference Audio Mods in the
form of Jon's SQ&D; (Super Quick and
Dirty) Isolation filter and Mod's DMF
(Digital Mains filter). I would recommend these changes first
before replacing clocks, supply capacitor blanks with black gates, and
February 22, 2003
Good power does not come cheap. When you can find different
individuals producing like product approaches...it usually means an
agreed upon approach to wringing out the best power. The best power
seems to come from generating two anti-phase 1/2 line voltages of 60
Volts (i.e. +/-60V). One of these connects to the "hot" (small
rectangular slot) on your power chord and the other connects to
"neutral" (smaller slot of power chord). The safety "ground" is
optionally connected/floated. For more information:
What the Art DI/O community needs now is a good
inexpensive transport with a co-axial output that will complement the
the DI/O. As early as 1993, Robert Hartley has reported
that the sound of CD players is heavily influenced by distribution of
jitter in frequency or the jitter transfer function. Many well heeled
audiophiles who have jumped to the DI/O DAC from the more expensive
kilo-buck breathen, based upon sound quality alone, have the transport
problem solved partly by using a kilo-buck transport already in their
possession. Some good names mentioned in various audio forums are:
Chuo Denki Company, better known as C.E.C. was a
company that put out some fine belt-driven turntables. They use this
expertise to come late into the CD playback scene but recoginzed the
major failings in the architecture of servo controlled CD playback
outlined in their patent.
Disparity of electrical activity of
motors when they are on versus when they are off leading to electrical
cross talk and EMI.
Varying rotational speed can change from 600 rpm
(inside/1st track) to 200rpm (outside/last track). In a sense, CD
playback was designed to continually hunt.
The use of higher torque motors than needed to
realized a overly tight grip (high rate of correction/bandwith via high
feedback) on the varying rotational speed requirement
What you see in CEC products such as the TL
0 transport offered by C.E.C are paralleled in the re-newed
turntable industry today:
Reduce torque motors similar to
reduce voltage to turntable motors once startup is achieved
The use of nature's free flywheel effect to achieve
a temporal constant rotational playback conceopt for a system that is
traditionally viewed as "variable rotational". The result is less motor
activity and interference.
One question is what exactly do these well regarded
units have in common?
When I began writing these pages (4 years, earlier), it was
generally agreed that the DAC was the prime determinining factor in
digital playback. It is now agreed better results could be
inexpensively, if the internal DAC is of sufficient quality and that
the SP/DIF inteface not be used due to the possible introduction of
additional jitter and the SP/DIF receiver. I would say that
Toshiba have a couple of engineers that really know what and how to
listen to their products. Similar to how Marantz has operated
without the traditional and well-earned markup. How they manage to
incorporate these core values at entry level prices that keep
sprialling downward is a minor miracle.
1.1 Background to Transport Mod
Doug Blackburn, formerly of Soundstage, wrote an article on modifying a
DV-525. In it he theorized that supply regulation, vibration,
RFI, and internal reflections had a lot to due with the sound of a
transport. On the commercial front, Modwright will update the a
transport with a standard set of mods
for succeeding generations of DVD players (DV-414, DV-525, DV-333,
DV343 ) used as CD transports. The mods offered cover the following:
Products Pioneer Transport Mods for Use With an External DAC This mod
optimizes the unit for use as a transport, but in no way disables the
video function of the player or its ability to play music as a
standalone player. Below is a list of mods offered by Modwright.
Level I Transport Mods:
Upgrade rectifiers to ultrafast soft-recovery diodes
Add an IEC socket to allow the use of upgrade power
Modify and improve the power supply filtering with the
use of Panasonic capacitors, bypassed with Solen poly caps
Upgrade stock digital out jack to a high-quality Cardas
rhodium plated RCA
Modify digital out circuit to include the addition of a
high-quality 75ohm pulse transformer and upgraded digital out resistor
to Caddock precision metal film resistor
The digital out circuit path is simplified and the
output filter cap is upgraded to a Black Gate.
Level II Transport Mods (Including the above mods and the
Key digital decoupling caps are upgraded to Black Gate
FK and Nx-HiQ caps
Upgrade main AC filter cap from a Panasonic capacitor
to Black Gate's best WK "Power Tank" cap, bypassed with an Auricap Film
cap instead of Solen poly capacitors
Upgrade IEC socket to a filtered IEC socket
One unit worth mentioning is the JVC XL-Z232
CD player which earned a garnering of the title "giant
killer". in a TNT review from September 2000. The player
offered a compelling overall high end sound due to a decent transport,
DAC and a decent output filter. The latter two are not needed by DI/O
types. What is needed is a co-axial output which this unit
unfortunately does not have. For the owners of a DI/O, the basic
High quality/low jitter output buffer with 75 ohm pulse
Upgraded power supplies for the buffer via a
combination of schottky diodes, buffer bypasses,
The CD player is still on sale today in the JVC catalogue
despite the fact that combined CD/DVD/MP3/CR-RW/DTS compatible DVD
decks will probably soon obsolete it based upon the limited real estate
found on stereo component racks. ...hands up those of you who have a
dedicated CD player alongside your DVD player.
The above mentioned Pioneer/JVC units are no longer available and I
have not heard or read about current product. The one thing that has
driven low jitter transports are the developements in HDTV and large screen format televisions
that magnfify jitter issues during the decode of "vertical lines".
These have low frequency (cumulative) jitter issues that take
form of wavy line artifcacts: Sorting out these issues is part of the
reason (I think) that DVD players serve as such excellent
Redbook CD playback units.
Transport Related Links
The links below are a first attempt at organizing information that will
form the basis for eventual "giant_killer" transport derived from some
existing player. Keep your suggestions coming in.
E3 CD player has an instrumentation grade clock
The Short Transport List
I for one am truly dissapointed in the agenda of reviewers and
manufacturers in promoting the next format as the answer to our prayers
of good sound reprodcution from digial sources at the turn of the
millenium. I can only put down here what I perceive as good "vibe"
reproducing machines and hope that competing manufacturers listen to
these units to see what half-way decent sound is like.
CD playback continues to evolve. The sore point for years was thought
to be (in rough chronological order):
Brick wall filter
limited resolution of 16 bits
Beating of multiple digital clocks
Reading capability of laser pickup
Higher RF clocks demodulating RF noise down to
SD-3990 (Jan 28, 2007):
What I brought as "lark" purchase has now earned its way into the top
of the list based upon the performance and very low cost. It gets most
of what matters...right!
A separate section,
below, contains links, listening impressions, partial schematics,
reliability concerns and potential modifications for these units.
It seems as if the Toshiba corporation has been paying
to these threads based upon correction of a number of quirks discussed
within the "Asylum" forums. Amongst these are:
Disabling the automatic power-down mode. Simply
CD in the tray and the unit will remain on. With a DVD video, it will
power down like before.
Improved output stage that was often bypassed or
upgraded. The present stage is an esoteric design with NO coupling caps
There are some new issues related to gapless
to recorded digital audio where a continuous/seamless playback of
live events is desired. Stock and right out of the box, it is
definite improvement upon the SD-3950/SD-3960 that so many have raved
Ayre Cx-7 (Jan 29,
From the ears of reviewers that I respect, it seems as if Charles
Hansen has struck the middleground between the hyper-detail of early
upsampling converters such as Theta. Reviewers are
increasingly aware of the importance of "rhythm" and "pace" and long
term listen through. Perhaps the best
writeup was by none other than John Atkinson in his Dec 2004 follow-up.
As Mr. Atkinson makes note, the rise of external DAC converters was a
backward step in that the SP/DIF
interface itself could compromise any gains in the DAC converter chip
and analog gain stages of the
external DAC. The great thing about all-in-one units is that you get to
remove one layer of compatability
problems and instead replace it with the convenience of what the
designer, in this case...Charles Hansen,
wrought. This is definitely a player to put on my short list of
auditions and yours.
P.S. I do think that external converters did advance the state of the
art by evolving the plethora of DAC
chips and receivers. The AKM chipset used in the Ayre, first came to
big notice in the ART DI/O converter.
This DAC when modified in the appropriate hands is a true gem despite
the low cost and simple power supply
(Positive Feedback ) and predecessor TL-51XZ
(Stereotimes 10/05). Transports from the folks that actually make
sub-assemblies for Oracle(turntables) and SME (tonearms).
Apollo (Jan 21, 2006):
I have not heard this new player (Jan 2006) but it apparently comes
from the company that continues to understand the segment of the audio
community that appreciates and understands the fundamentals of good two
channel playbackat a fair price point.
Rega has been forced to face the eventuality of having their supply of
dissappear. Read about this in this one page excerpt recounting the story behind the Rega Apollo.
What is there to like about the new player, plenty:
Removal of magnetic puck and its rotating magnetic
field. The "3 ball" collet pioneered in portable CD players is used to
Class A discrete Output Stage
Co-axial digital output to support your favourite
outboard DAC. The Jupiter was often used as a decent Transport by
owners of outboard DACs that I'm currently using. This is just great
Optimized read calibration. I'm not sure if this
is has any relation to the pause-access
experiments from AA.
Saturn (StereoTimes, 9/06)...a Apollo with boutique
components taken to the next level...and clever marketing
Kardon DVD-31 (June 3, 2005):
is priced a bit higher that bottom of the rung DVD players available
but you get bonafide "good" sound. In some ways, it follows the Toshiba 3950 legacy where demanding
large screen video quality build paved the way to improved audio. The
article above compares the unit to entry level offerings from Panasonic
(DVD-S27) and there is just NO
comparison. Why ? I posted some comments
on the Digital Drive forum on AA that you may be interested in.
DVD-3910 (May 2005 entry)
When I heard this unit at my friends "really" high end setup, it just
sounded more like music than anything I've heard on good ole fashion
CDs...not that it does not have everything (including the kitchen sink)
included in terms of format support (universal) and the latest video
output standards...just slightly ahead of our time...and what my wallet
perceives as a bargain. But what a sound it had in reproducing the
clash of cymbals being struck.
Toshiba SD-3950 (April 2004 entry)
Taking a look at the PDF
document of the specs would not lead you to believe that this
sub $100 dollar unit could generate the interest that it did in
February 2004 ...see below. A
classic design where every dollar was spent to realized better video
and audio...there are some questions about build quality
DVD-S25/DVD-S35/DVD-S47 (March 2004 entry)
These units are promoted by Rick Schultz of "tweakaudio" fame. The
motors are quieter than the Toshiba 3950 and some say that running
stock favours the pannies.
Take a look at Rick's links for more information. Funny thing is that
these units were promoted early on in the year around the same time as
the Toshiba Craze ...but never caught on.
This unit is the lastest generation progressive scan unit that has a
new tracking mechanism called Precision
Drive. The unit has had a few posts within Audio
Asylum comparing it to other highly acclaimed two box
transport/DAC units made of one part Theta DS Pro DAC. Even challenging
the likes of Classe
CAP-150 and DI/O gear. More information can be found here
18, 2005: The word out Vinnie
is that the sonic replacement for the 3950/3960 now comes in a third
revision of the SD-3980 with an all metal case. The power supply board
is slightly different. case
Nov 14/2005 Update: The
is increasingly suspect. David Pogue is an
experienced audiophile who
found his second returned unit to also be problematic. I've
put my own thoughts
about this being the extreme low cost
manufacture leading to fundamental short cut being taken. Unit
February 24/2006 Update : 2006
has seen the introduction of the a replacement to the 3980 in the form
of the SD-3990. Take a look at some of the early comments on Asylum.
Also take a look at the 2006 SD-3990section
where most of the reliability issues have been sorted. Toshiba offers a
very genenerous no question asked 12 month exchange policy.
The unit has been recommended by user Rosconey
on AA. Two users with decent DACs in the form of
Vinnie R. and John
Swenson picked up on it and added significant fuel to the
"3950" fire. This initial buzz increased to a frenzy on Audio Asylum as
some claimed that this $60 clearenace item was causing people to
consider replacing their CEC belt drive transports, Theta
Basic II transport(s),
I picked up my unit from Walmart for about $100 in
"black". I have been experimenting with some 8x CD- ROM drives, NEC 6x
CD-ROMS in an effort to unlock more of the sound of my new Ack! dAck
DAC (just had a dAck attack). The skinny from me would be use this unit
unmodified as a transport.
also has progressive scan (TV), supports Kodak Picture format (2000
JPEG images on a CD-ROM), and has a great image with an extended black
range to mine detail in the shadows. It has ended my 3 year search for
a decent low cost transport.
internal DAC was given a quick listen after a day of
burn-in...not too shabby. More on this later.
Before I put down how it performed in my system, here is some more
reading for those with an itch to know a bit about this player.
affects the 3950 ...just like turntables. Add some DIY rollerballs described further
down on this page.
Add a heavy hardcover textbook on top of the unit to
kill vibration from the top.
Be sure to read the instruction manual with regard to
"DRC" and "PCM Downsampling". The following works for me:
PCM Downsampling: OFF
Some claim different.
different settings are preferred for using the internal DAC versus the
jitter heritage from SD-9500 The jitter (RMS) is claimed to
be about 40 ps for the top line 9500 unit. In the review from Home
Cinema, they measure 170 picoseconds (peak) value for the SD-9500
which probably corresonds to the 3 signa value if the two reported
values are too be close. The RMS value would be about 170/3
Jitter measurements Home Cinema has been specifying this
parameter since August 1999. So blame the Toshiba's performance on the
I guess with my purchase of my Toshiba 3950, I come to an
obligatory review of its sound in my system. I tried it with two sets
of speakers (Linaem LX-5 and Axiom M22-ti) and two sets of speaker
cabling. The improvements were on the order of the opamp change but
maybe more stunning in their refinement. The vocal midrange became lush
with tone and warmth. I heard deep tuneful bass and my liliputtin LX-5
(5" woofer in a cubic foot box) seemingly going full range and the
Linaeum tweeter working new magic on me. Bottom line, it has somehow
created more music in my little room with my system that I have ever
2.1.2 Toshiba Mods
There are two primary sets of mods for the Toshiba that require
removing the top to swap out components. As oppose to "tweaks", mods
require the ability to solder and some familiarity with electrical
components. The Swenson mod is more the case of less is more and goes
about its business by removing the active output filter stage/op-amp.
The other mod by Vinnie upgrades supply decoupling caps, op-amp, and
signal coupling caps. :w
Mod...the Swenson mod uses the fact that the DAC raw output
is just about right at 1.3V RMS.
Swenson Internal Jack Mod: Back when the Toshiba craze took
off in April 04, I always thought the Swenson mod a little arkward
with the long traces snaking out of the unit and the need for new RCA
I now have a modified version of the mod whereby I use the existing RCA
jacks and do not require external coupling wires to be added. Read
through the Swenson mod first and then follow the instrcutions below.If
you can change the oil in your car...you can do this mod!!!
Internal Swenson 3950 MOD...click for larger
Prior to soldering these connections, take a
scewdriver and scrape the point where you are going to solder on the
jack until you have a 4mmx4mm patch of bare metal. Tin this area with
Tin the bare ends of yellow 2.2uF Auricaps ("red
This mod also requires that the op-amps are
disconnected from the jacks by removing two additional output coupling
caps (C918 and C920). These are
shown by the "red arrows" in the photo below.
Like the Swenson mod, the outputs from the DAC (BB
1751) is taken from the "-" lead of the two input coupling caps (C910
and C912). These input caps positions are shown by
the "purple" arrows in the photo below.
Once the caps are all removed (4 of them), solder
the "red" wire of the 2.2uF cap to the RCA jack
The black lead ends are too large to fit into the
mounting hold connection of C910/C912. Strip this lead and remove about
2/3 of the strands to reduce the diameter. Tin the remainder. Attach to
the point shown in the photo
Congratulations...re-assemble and give it a listen
Update: I came across two sites
that have a slight variation on the above internal mode. One site
likes the removal of a 5.2K series resistor between
the DAC output and the large coupling cap. This resistor apprently
causes a loss of output but I do not see any isses as I measured a
healthy 1.8V peak (1.3V RMS output). There is also discussion about
removal of a a small yellow shunting the RCA jack to ground near the
RCA jacks. These caps made be responsible for the lack of staircase
artifacts that I saw on 2 KHz and 20 KHz sine waves. The other
post (actually two) seems to be the original source of these
ideas and is authored by John Day. One futher note on C903/C905.
If you try this and the above procedure, PM me.
2.1.3 Toshiba SD-3990
This Toshiba player has not enjoyed the wide spread crazed popularity
of it's predecessors (3950,3960). It may be due to the quality control
issues of the the earlier unit (3980) but based upon recent listening
comments of a few ears that I trust, this unit promises potentially better
sound than the 3950/3960 (even modded units) and there are
relatively few reports about the units dying unexpectedly. This unit
plays DivX ...I
think this means that it can play the AVI/MP4 files that are not
generated within digital cameras.
Last, the new unit now apparently comes in a metal case. This unit is
also sold as the SD-K760 (think CostCo/Sams) which I've been told has
identical innards with boards labeled 3990. Probably something along
the lines of Michelin tires rebranded as store exclusive so as to not
annoy the likes of BestBuy with price matching policies.
Doom9 Forum, March-Sept 2006)compatability
discussion. Some operational quirks regarding the lack of "resume" from
power down. Video centric comments from MetaReview
(Tom Larson, Asylum, Nov
24 2006 ) at sound but not the gapless encode replay
Head-Fi, Oct 2006)The most thorough set of
pics and info on this unit. The DAC is contained in a large ASIC by
Zoran. An external buffer op-amp (LM4588) is used with no
input coupling caps (to the op-amp) or output coupling caps (to the
jacks) that were used in the SD-3950. The good news is that a Swenson
mod is not necessary. No techies have jumped in yet.
July 2006)with an emphasis on video decode
and video bandwidth/linearity.
Yahoo Photos, 2006)of a unit
modified with Rope Caulking material
Generational Sound(Ed Kobesky, Positive-Feedback,
April 2006) To understand how Tohsiba
consistently gets good sound from its architechture, take a look at
this review in "Positive Feedback" on the carousel player SD-6915 2005
model. Also see the Stereophile comments for the 3950
in Sept 2004.
I ended up purchasing a unit for a friend who wanted me to evaulate the
sound of the unit on November 12, 2006.
1st listen (Nov 12): I played
Norah Jone's second album called "Feels Like Home"...in particular: Cut
3 ("Those Sweet Words"), Cut 6("Be Here to Love Me"), Cut 9("Humble
Me") and Cut 12("The Prettiest Thing").
Cold out of the box and resting on Rollerballs
and with the co-axial out terminated with a 75 ohm RCA plug, I would
say "not too shabby". Mechanical noise was low. The back layout was
different from the 3950 with connection now centrally located (instead
of on the right) and the cord connection on the right (instead of on
ExperimentsWhen the termination was moved to the
"composite video" output (from the co-axial), the pace seemed to
quicken to bad effect. Vocals were sharper and upper treble more
present. The odd thing is that as the unit warmed up, its out of the
box "goodness" seemed to progressively be replaced by a hardness that
was not present at all at the start(???).
ComparisonPutting back in the Auricapped SD-3950
was a welcome relief. The soudstage widen considerably, the pace became
naturally more relaxed, and Norah's voice felt comfotable, extended,
and just great.
Nov 16 Listen: The unit was
setup this morning to reapeatedly play the XLO burn-in disc. The
S-Video cable was plugged in and the unit was moved to the floor. The
isolation were rollerballs comprised of inverted tea cups.
Norah Jones(Feels like Home), Neil Young
(Decade), Over the Rhine (Good Dog Bad Dog) and Miles Davis (Kind of
Blue) showed that this unit had improved a lot. It always sounded
unstrained and in control. The LM 4558 op-amp is a derivative of the
741 and sells for about 50 cents. Despite the use of this component,
the designers at Toshiba manage to wring out a decent sound that is
detail, fast and great on voices and sounds very un-digital. As
impressed as I was with the unit, I think that seperation and
soundstage can be improved in both width and depth. Did I say maybe a
bit more "SLAM". The simplicity and very small layout size of the stage
must have something to due with very controlled and dynamic sound that
encourages you to turn up the volume.
This unit does not "turn off automagically" like
its predecessors did after amount 15 minutes of inactivity. This should
make the burn-in a less prolonged affair and/or save wear and tear on
the drive itself. I left the unit on overnight and played back
"Flamenco Sketches" this morning. Not much sonically different from
yesterday and I can say that the sound seems to have stabilized. How a
track is accessed seems to
have some influence but I'm still trying to sort it out. Using an
external DAC may be the order of business but my feeling is that the
signal ahead of the op-amp might be something really special and
revealed by an op-amp upgrade. For now, the 50 cent LM 4558 does a good
enough job that I think your musician son/daughter will enjoy cranking
up the unit in their bedroom.
Nov 19 Listen: CD playback is a
difficult nut to crack. Their are/were fundamental issues to solve in
improving it to compete with the Phono/LP playback. Mechanical
sensitivity to different cartridges, mats, suspension type, bass
extension, tonearm type and the like also meant that these parameters
were at disposal of the motivated audiophiles. Introduce something like
CD playback and you have very little to muck around with other than
cables and mats for some enterprising souls with the right type of
tray. The Toshiba SD=3990 was setup side to side with the 3950 on a
coffee table to ease comparisons. The 3950 used "pop-can" based
rollerballs while the 3990 used "inverted tea cups". Cables used were
the Ack Industries "Analog Connect". After listening to a handful of
cuts, I "thought" I had the 3990 pegged as having reduced soundstage
with little image height. Norah Jones seemed to stand about 5 feet 10"
in my living room with a somewhat muted high-end. Most of this was
compared to the 3950 sitting pretty on the "rollerballs". The 3950 seem
to have Norah about 4 feet higher up on a stage with instruments spread
out in beside and behind to her side. Her voice sounded naturally
sharper than when played back on the 3990. Could it be the "tea cup"
supports? Short answer: Yes. Still the "pop can" supports begin to move
the sound toward the "auricapped 3950" but was still behind. In the
course of the weekend, this is what I found:
The chassis screws that were removed need to be
secured back on to the chassis before doing detail sonic comparisons.
The mass of the book on top of the unit matters
with the lightweight chassis of the 3990. Too much and things can get
The type of "ball" for the rollerballs has an
influence. I found a rubber ball used on the backleft, only, to work
well with the center and right balls made out of metal.
I won't get into the LP controversy, but CD setup with this Toshiba
unit into a Art SLA-1 (with Belden power cord) driving HD-14 (correct
directionality) into Axiom M22 to be very musical. The most interesting
things about all the components is that they are all units do not have
any internal upgrades. This is coming from a person who has literally
replaced every cap in a NAD 3020 (and then some). The most costly item
would be to rent me and to pay the price of a plane ticket to have me
fly over and tweak your system personally. I've heard a lot of
expensive reviewed equipment (TAS, Stereophile, UHF approved systems)
with wanting sound at shows due to bad setup but realize that it is not
easy task to sort out room nodes, noisy AC, and strange floor
construction. They have my sympathies.
Hopefully you will
give the 3990 a chance. It burns in easily and can be left on to
facilitate this. I found one CD-R that it had trouble tracking with and
will have to see if similar issues are on the 3950. You will pay twice
what you do at Walmart for their $30 player but you will be rewarded
with a far superior picture and audio that is not laced with RF hash
and hum. Do the following:
Terminate co-axial out with 75 ohm RCA plug.
Terminate composite video out with 75 ohm RCA
Leave the unit powered on for a few days. Does not
have to be playing but a CD must be in the tray.
Consider upgrading the op-amp at some later point.
Mechanicals and Reliability: The
case is metal with the tray off to the left (versus in the center on
the 3950). The old remote works with the new unit but there are some
quirks. You can index the beyond the 9th cut but in a different manner
from the earlier units: The 10+ button is no longer
needed to access cut 11. Now you just press "1", twice, and the the
PLAY button. When a song is paused within the first 5 seconds of the
start of a track, pressing "PREVIOUS" will skip to the previous track.
If you press pause after the first 5 seconds, PREVIOUS will take you
back to the start of the track. The numeric display cannot show you
elapsed time as on the 3950 and now comes in green.
The DVD drive making up the transport is no longer
suspended from "rubber grommets" as was the 3950 drive. This may
explain some of the changes in sound as the case was deadened with rope
caulking report by some users.
It is early, but weakest link of this amp may well be
the op-amps used. Lasts, unit to unit variation is possible as I found
in 2004 with the SD-3950. The purchased unit sounded quite different
from the one my friend purchased and they were built in the same month.
often a sore point with Toshiba DVD players. This might be a result of
getting what you pay for in terms of design, assembly, and warranty.
The unit may be suspectable to damage from static electricity either
due to weakness in the chips or initial damage incurred during assembly.
This weakening/damage can lead to something called "infant mortality"
where the unit cannot withstand the rigors of everyday usage and fails
early on in its life...given the chance. The best safeguard against
this is to just screen it out is by leaving the unit powered on
long enough to evaluate the weakness. If it passes this period, you can
be assured years of reliable service.
The SD-3990 no longer powers down
by itself after 10 minutes of non-usage with CDs. This was a major issue
before as you really need something like a week's worth of continuous
powering to screen these faulty units. Do not actually play
a CD at
this time but just load it into the tray. Mechanical wear
is another issue that you'll be trading off against should you decide
you also like to spin the CD during this break-in period. . Note: With
DVDs in the tray, the unit will power down after 10 minutes. When
viewing that are short in portrait/vertical orientation, the 3990
correctly displays the thumbnails which the 3050 did not. Lost however
is the ability to remember changes to the orientation which the 3950
had as a hidden feature, so that you could sort out the orientation of
individual JPEGS (ahead ot time) and then run a slideshow, seamlessly.
Below are some
"buyer posts" outlining their experience with the 3990.
SD-3990 board/signal trace ...click for larger
SD-3990 Output Buffer Schematic ...click for larger
The differential output signal handling from the Zoran processing chip
would have been classified as very esoteric in the days of the kilobuck
Theta DAC. Because we mostly live in a single-ended world,
the differential beast is converted from It its native differential
nature via a third order filter also doing single ended conversion.
From Figure 2, the equivalent 1/2 circuit shows the first two stages to
be totally passive (and differential). Poles are at (1/2*Ra*Ca) and
(1/(2*R1*C2). This is then fed into an active inverting gain stage with
a first order filter. The gain is RC5/R3 with a pole at 1/(RC5*C5). Not
shown at the output of the op-amp are an output RC section with a
muting transistor after the resistor and in parallel to the capacitor
to ground. These components are in the bottom right corner of the board
image. The op-amp if fed from +/-11V supplies with the signal directly
coupled from the DAC to the output. This means that there are no input
coupling caps (neat!) or output coupling caps (even better!!) to the
output due to the inherently low DC offset of the ciruit configuration.
High end and esoteric indeed...No boutique coupling caps (teflon
anyone?) needed here!
MODS: Unlike the 3950/3960
generation, the output from the Toshiba is differential and not single
ended. This means that a Swenson type passive mode is not
possible...unless you happen to have Balanced inputs on your pre-amp,
in which case you may implement a balanced version of the Swenson mod.
Region Free mod
With the tray open and the unit powered up: Type 24039609 . Your unit is now region free. More info here
Upgrading the op-amp (similar to Vinnie's mod) is one
possible path. One recomendation would be the LT1361
in the appropriate S8 surface mount package. The first thing to do
might be to use higher quality components for the supply bypass. The
other thing to consider is whether the surface mount components can be
upgraded to boutique ones favoured by many modders. (eg. Dale or
It is possible to use a more refined Sallen and Key
filter arrangement on this unit similar to the Art
DI/O which pioneered the use of integrated differentail
outputs on low cost DACs back a mere 4 years ago.
If you stayed with me till now take a look at some
late breaking easter eggs...
These devices have been purported to improve anything from
speakers to turntables. The following has been key to keeping my DVD
player from becoming a paperweight. More info here.
Rollerball supporting Toshiba 3950
Block"/Rollerball Isolation The idea for this implementation
came from a pop can based home made campstove project I built using the
ideas of Lamar
Kirby. Basically, each roller ball support is made up of from
the bottoms of two pop cans splice together. The picture below shows
two completed units.
Pop Can Rollerball with Ball Bearing and Mouse
The extremely thin
lightweight nature of pop cans means that little energy will be stored
in them. They are virtually airtight and this dampens any
self-vibration which (I think) leads to them being a good
transmitter/sink of any motor vibration within the CDP/DVD unit. To get
an idea of how they operate, I put together this animation
of them in action as they come to a stop from
being jarred sideways.
2.4.1 DIY Roller Block Notes
These are constructional notes from having put together a number of
these roller blocks. Give yourself about 10 minutes to put together one
of these when you get good at it. Accident prone types might give
themselves about 30 minutes for the first effort. My first few resulted
in vertical slits forming from too much uneven pressure during assembly
of the two halves making up the support. Putting a inward curve using a
spoon (Step 7, below) solved the problem. It may take you 3 tries (like
me) before you get the first one right. Good Luck
Two cans (six altogether for three feet)
1/2" masking tape
Metal Cutting shears
1/2" Ball bearing
The cans are getting progressively thinner these days.
Some aluminum cans (older) are made from heavier stock and are
preferred. Despite this fact, pushing the two halves of the can
together will result in a very strong structure that does not ring.
Step 1: Take two empty pop cans Note: Click on each image for enlarged view.
Mark a series of dots 1 inch above the bottom of the can using a ruler.
Step 3: Use masking tape to
define a cut line over the 1 inch marks.
Step 4: Cut the can in 1/2
starting from the top opeining
Step 5: Cut a 45 degree line that
just touches the masking tape and then use your shears to trim along
the masking tape edge
Step 6: Completed bottom section
Step 7: Use a spoon to curve the
edges toward the middle of the can so that the top section can slide
over. The edges are curved by adding thumb pressure on the can edge
against the inside of the spoon. Rotate can about 10 degrees and repeat
until entire edge is curved slightly inward.
Step 8: Repeat steps 2 to 6 (skip
step 7) to form the top. Skipping step 7 should result in the diameter
being slightly larger than the bottom so that you can slide the top
over the bottom. After fitting top over bottom, slowly push the two
halfs together. The top edge should come close to the rounder portion
of the bottom (about an 1/8 of an inch).
You can try different size bearings and materials to get
the best sound. I have used plastic balls and mouse balls in different
combinations with the metal ball bearings to get the CDP sounding just
[2.5] Last Frontier: Transport Jitter Increasingly, reports about "ripping" CDs to
computers and then using the digital stream (USB, Firewire, I2C) to
non-SP/DIF converters being better than 5 figure d cost no object
conversters have surfaced. The mightly Mark Levinson reference DAC
essentially not being necessarily providing you get rid of the SP/DIF
interface. Why is this so ?
[2.5.1] Background: 1st Generation Redbook
In the decade preceding the release of the first CD players (1982),the
Personal Computer had captured the imagaination of the public and of
large companies such as IBM which released it's first version at about
the same time. Encoding 16 bits of information for two stereo channels
at 44.1 kHz meant that a bit stream of 2 million bits a second: A
farily high bandwidth of data for the time. This prevented the
possiblity of having the transport detect errors being read and
resending correct data but required data to be corrected on the fly by
an algorithim pioneered by Reed-Solomon.
[2.5.2] USB/Hardrives: Lower Jitter via "Hurry up
[2.5.3] Why DVD make for better transports
[126.96.36.199] 8x Oversampling and Transports slaved to
[188.8.131.52] Low frequency jitter and Big screens =
Audible differences can be heard dependent on how
long the tray was open/closed since playing the present
track...I can only ask you to try and report back. I hear something
(positive) to the method/madness.
More madness...now the opening/closing of the tray is involved.
Email me if you can confirm or deny this. Below is a list of positive
results from different users: