The Sound

I auditioned a number of CD-Players, before buying my CD-67 SE. I felt that of all the CD-Players in question (which shall remain unnamed) it offered the most music for the money. Having lived with the "stock" unit for several month, I'd describe the sound as "save". This is an eminently listenable CD-Player, which does smooth over some of the rough Edges. I found the player however ultimately lacking in resolution and "air-guitar" factor.
The CD-67 SE certainly does the M-thing (musicality) very well, but as Warren G. say's, it's a G-Thang too.
The Cd-67 SE's sound favored acoustic Jazz and classic, but I felt that my tweaked ultra-cheap Philips CD-720 offered more get-up'n-go and certainly did more for me when I played Rock or Dance Music.

At the same time, the typical digital nastiness that is displayed by so many "budget" CD-Players (and some quite expensive ones too), is absent, although there is still a notable exaggeration of sibilants.

Rhythmically a bit slow and with an overly smooth presentation, the Marantz CD-67 SE was often blunting a bit the impact of the music. Also, with classical Music, while the CD-67 SE's smoother presentation stopped the violins from screeching, it did not reveal as much detail as the Philips CD-720.

All in all however, this is a CD-Player that is certainly way better than average and it fully deserves all the (almost) cult following, it has achieved among music lovers worldwide (and even among magazine reviewers).

Given that I found the Player to have some flaw's which I could not live with in the long run, let's have look how it is made up.

Technical Realisation

Before we start into the "meat", may I humbly suggest that you read A FUNDAMENTAL INTRODUCTION TO THE COMPACT DISC PLAYER by Grant Erickson. This rather good paper covers a lot of the things you need to know about how a CD-Player works.

The Marantz CD-67 SE has got a long history, going back over five years to the CD-52. As Marantz is a 100% Daughter of Philips/Matsushita, it is hardly surprising that much of the CD-52 came direct out of the OEM Part's Catalog of Philips. In the factory-tweaked CD-52 SE Version, this Player won a lot of fan's and critical acclaim.

However, time moves on and when the original CD-63 was introduced, some marked Changes took place. First, Marantz ditched the Philips Bitstream DAC in favor of a part from NPC (an Epson/Seiko company). They also changed over to the linear-tracking CDM-12.1 Mechanism (still Philips) and implemented a different analog filter and Output Circuitry.

One of the (technical) Reasons I decided to buy the CD-67 SE (with a view to future modifications), was the NPC DAC used. This is a very interesting part, which seems to be a copy of the "original" Delta/Sigma DAC, the Crystal CS4303, a pretty old part that still nevertheless outperforms most current Bitstream DAC's.

ThanX to the friendly people at Farnell Electronic Components, who photocopied me the Datasheet of the NPC SM5872 DAC from their Databook (yup the whole 26 pages). This DAC Chip is very well designed. While it combines a crystal-oscillator, a digital interpolation filter (8 times Oversampling) and two Delta-Sigma DAC's on one Chip, all these components have separated power-supply and ground connections run out from the chip.

The actual DAC's put out a purely digital, differential pulsewidth modulated Signal. While this (unlike the Crystal CS4328 DAC) necessitates a relatively complex external analog stage, it also allows the Designer to optimse the analog filtering and such, instead of offering one level of quality due to build in Op-Amp's of NE5532 caliber (as is the case with most Philips DAC's).

The separated PSU and ground-lines also allow optimised Layouts and decoupling, to achieve the highest possible quality. Marantz uses the on-board crystal oscillator of the DAC to generate the system-clock and has taken great care with the PCB-layout.

The DAC is followed by a fourth Order analog filter build around the NJM2114 Op-Amp. This is essentially a "supercharged" NE5532 and not really bad (it is still used in the CD-17 KI Player) but than again, nothing really good either.

The Filter is followed by a discrete buffer called for no particular reason HDAM. After that we have two muting transistors per channel (Yuk). Finally the Output is AC-Coupled via two electrolytic Capacitors in a "back-to-back" arrangement.

In the SE many of the electrolytic capacitors are special Elna Units (Silmic/Starget/Cerafine). The normal CD-63/7 has fewer of them and the CD-53/7 (which for all other purposes but the HDAM Buffer uses the same chassis) uses standard capacitors.

The SE unit also uses many Film-Capacitors (though only Mylar) in critical locations.

The Power-Supply arrangement for the CD-63/7 family is quite good. Separate regulators feed the Digital Circuits and the Transport. Completely separated windings on the transformer are used to power the analog section.

All in all, quite competent engineering, if not outstanding. So why does it not work all that well?

The CD-67 SE revealed

Beware, there is a seriously X-rated CD-Player Picture shown below. We see a CD-67 SE stripping, though it cares to keep some of it's more intersting parts hidden.

So You know what is where, here is the lowdown:

1. Chassis Cross-brace (SE and KI only)
2. Location of the PCB Positions for Mains Filtering (hidden by 1)
3. Location of the analog Supplies (hidden by 1)
4. Location of the "secret" HDAM's (Hyper Dynamic Amplifier Modules) (undercover by 1)
5. Location of the Filter-Circuit (NJM 2114 Op-Amps + PSU note the red electrolytic capacitors - they are Elna Starget/Cerafine types)
6. Location of the DAC (underneath the PCB)
7. Location of the main clock-crystal (note the rubber-grommet around it)
8. Location of the two digital Supplies (one for the Transport and one for the Signal Circuits)
9. Location of the system controlling Microprocessor
10. Location of the Decoder/Servo-driver Chip (underneath the PCB, will be different for the 53/63 Series)

So, now we had a look inside, let's see what ails the CD-67 SE. I do not have direct access to the expensive measurement equipment needed for jitter measurements and the like, thus I had to rely on Circuit Simulations in Pspice and technical reviews of the CD-63/7 Series to gather evidence.

In his Article in the November 1997 issue of HiFi-New & record Review (HFNRR) Ben Duncan notes that for accurate reproduction of 16-Bit resolution, the amount of jitter presented to the DAC MUST BE below 100ps (pico-seconds). For 20-Bit resolution by the way, it must be below 6ps. Thus any CD-Player (or transport/DAC Combo) that has jiiter-level higher than 100ps is incapable of extracting all the information , potentially presented by the CD replayed.

So how does the CD-63 KI stack up? Crummy! HFNRR measured the CD-63 KI at 687ps Jiiter. That is more 14-Bit performance than 16-Bit. On the plus-side, stereo-separation is excellent and distortion levels are low, as is noise.

So, is the CD-63/7 Sound jitter related? It seems like it. Most of the Jitter is located at frequencies of around 500-1000 Hz. Let's check out the experience Paul had when he accidentally introduced notable amounts of jitter to his CD-Player.

I shall quote: "Another funny thing (that will drive the audiophiles crazy) was that by accident, I threw about 500 ps jitter into the clock (the jitter freq. was about 250 Hz). I listened to the sound afterwards and thought just how nice warm and smooth it sounded, figuring this was it for best sound." ..." Once I fixed that up, the sound turned cold and quite clinical. The detail was greater, but it lacked the musicality the jitter condition had!"

What else? Well many people on the Net reported that replacing the NJM 2114's with something better (Burr Brown OPA2604, Analog Devices/PMI OP275 or OP249) did yield very good results. Someone did put a Tricord Clock 2 unit and reported exceptional improvements (easy to see why, the Clock will feed the DAC directly and reduce jitter massively).

Looking around inside the CD-67 SE with a 'scope, I noticed that despite quite serious attempts by Marantz to decouple the different PSU-Lines in the digital section, there was plenty of wideband noise present which easily got through to the Crystal Oscillator and likely modulated it. Could that be source? If you read on, you will find that partially it was.

Another thing I noticed, was that Ken Ishiwata had gone through a great length (specifically on the CD-17 KI) to improve the mechanical integrity of the CD-Player's Chassis for units he tweaked.

So I decided to start whipping the Cover of my CD-67 SE and do things. The following observations on my modifications are amalgamated from various post's I made to the Soundpractices E-Mail-list (Joe-Net).

Much of my electronic Mod's owe a lot to postings Pete Goudreau made a while back in the rec.audio.high-end Newsgroup on modifying the DAC-1 from Sonic Frontiers. Have a look in Deja-News and do a Search for him as Author.

During the modification process (and at other times) I had some stimulating (e-mail) conversations with David Bardes, Frank Deutschman, Carlos Martinez and Guido Tent all of which where able to help me with various bit's of information. ThanX to all of you and many others which shall remain unnamed due to space and time restrictions.

This article and the really exceptional improvements I achieved with my CD-67 SE would not have been possible without You.

Mechanical Improvements for the CD-57 / 67 / 67SE

Simple Improvements to the analogue Electronic of the CD-57 / 67 / 67SE

Improvements to the Digital Electronic for the CD-57 / 67 / 67SE

Improving the Power-Supply for the 57 / 67 / 67SE

Soundz of Muzik

Where do I go from here?