Digital Drive

Posted by mark_nz (A) on February 19, 2004 at 17:13:18

I have posted this on both the Digital Drive and Hi-Rez Highway forums since it contains relevant information for both forums.

BACKGROUND

I related my previous experiences with both the CDP-XA555ES CD Player and the SCD-777ES CD/SACD player in the following post:

http://db.audioasylum.com/cgi/t.mpl?f=hirez&m=62993

At that stage I found while the SCD-777ES benefited from extensive run-in and upgrading the I/V opamps to LM6172, ultimately the CDP-XA555ES still had the superior sound quality. My conjecture was that a redesign of the analog output stage of the SCD-777ES would transform the sound quality.

I have now built an analog output stage that has transformed the SCD-777ES performance for me.

ABOUT THE CDP-XA555ES

The CDP-XA555ES was Sony’s top of range ES model before the SCD-1 & SCD-777ES were introduced. It replaced the CDP-XA50ES that in turn replaced the CDP-XA5ES that was based on the famous CDP-XA7ES.
The overall circuit of the CDP-XA555ES is an improvement (new digital filter VC24, better opamps, better passive components, more extensive power supply, same transport) on the CDP-XA7ES, but does lack balanced outputs.

THE SYSTEM

The rest of the system consists of: Sony TAP-9000ES Preamp->Kimber Interconnects-> Marantz PS-17SA AV Amplifier (preamp section bypassed) -> Audioquest Forest Speaker Cable -> Jamo D830 (Concert 8) loudspeakers. The room is a dedicated listening room treated for sound quality with no regard for domestic acceptability. Power is applied via a dedicated mains line and a DIY mains filter.

THE JOURNEY

SCD-777ES: I mentioned in my previous post that I replaced the I/V opamps with LM6172s. What I found is that while it provided greater detail, it also added an unwelcome edge that was wearing on extended listening sessions. I put the original OPA2604 opamps back in which removed the edge but also made the player boring to listen to. I also mentioned in my previous post that there was ‘something not quite right’ about the sound quality of the stock player. The sound quality improved dramatically with run-in but what ever bothered me never totally disappeared. It is difficult to describe what this was, the best I can do is say that it manifested itself as slight ‘hoodiness’ with voices or ‘plonky’ piano sound! This was evident on both CD and SACD, but more so with CD.

CDP-XA555ES: At that time I still preferred the CDP-XA555ES to the SCD-777ES. I decided to tweak the CDP-XA555ES to eke more performance out of it. I bypassed the output stage coupling electrolytic capacitors and disabled the variable output. This seemed to result in a slightly cleaner presentation, but without an A-B comparison difficult to quantify. After reading all the positive feedback about re-clocking, I then installed the LC Audio LClock X02. This made a stunning difference. There was more detail, better timing, greater dynamics, better sound-staging, quieter background. What made me the most excited was how much more realistic musical instruments and voices sounded. I then tried to install the LC Audio Zapfilter. I could not get this to work as an I-V converter. I found a post on www.diyaudio.com that described a similar problem that seemed to be related specifically to the Sony Current Pulse DACs (These do not seem to be DACs in the usual definition, but rather Voltage to Current converters that clean up the output from the VC24 Digital Filter/DAC). The proposed solution was to use the Zapfilter with voltage input mode connected to the outputs of the VC24 Digital Filter/DAC, effectively bypassing the Current Pulse DACs. This worked, but I felt that was an overall backward step in sound quality. (Sony must presumably have a good reason for using the Pulse DACs). The Zapfilter was removed.

SCD-777ES: Curious to see whether the LClock X02 would wring the same magic out of the SCD-777ES, I uninstalled the clock from the 555 and installed it in the SCD-777ES. (Both players use the same digital filters and DACs, so the same clock frequency applies.) Yes, overall this improved the sound quality, but not to the same extent as it did to CDP-XA555ES. It also did not fix the one aspect of the SCD-777ES sound quality that always bothered me. Out of frustration I re-installed the LM6172 as I/V opamps. Again an unwelcome edge returned to the music. I conjectured that the LM6172 was possible oscillating due to lack of proper bypassing. Rather that muck further with the SCD-777ES analog output stage, I decided to build a replacement. I used two LM6172 dual opamps per channel; first pair used for the positive & negative I/V, the 2nd pair for summing and output buffer. I did not want to take a chance with the SACD high –frequency noise damaging amplifiers or loudspeakers, so I applied aggressive Bessel filter curve (-3dB at 20Khz, -20db at 50kHz and
–50dB at 100Khz). I built it on fiberglass prototype board with generic MF resistors and Wima polypropylene capacitors. I applied extensive bypassing to the opamps, using point-to-point wiring where possible to reduce connection lengths.
Wow, what a difference. Most noticeable difference was how smooth the sound became. There is no hint of the edge I had when I just replaced the I/V opamps with the LM6172. The whole soundstage opened up, providing a seamless wall-to-wall view. There is greater dynamic contrast due to the quieter background. Interestingly this did not take that long to run in, the positive sound qualities evident from early on. Only down side was a slight lack of sparkle in top-end and reduced stage height, both possibly due to the aggressive filtering. Both CD and SACD sound quality was improved, and the superiority of the SACD is still evident despite aggressive high frequency filtering.

CDP-XA555ES: Without the LC Audio X02 clock, this player could not compete with the modified SCD-777ES - it was comprehensively bettered in every respect except for a bit more sparkle on top-end, and (interestedly) PACE. I have always found the CDP-XA555ES to possess superior PACE relative to other players I compared it to, including the SCD-777ES. To redress the imbalance in sound quality, I threw caution to the wind and replaced the surface mount OPA2132 opamps on the CDP-XA555ES output stage with LM6172s. An interesting design aspect of the CDP-XA555ES I/V stage is the addition of a buffer transistor inside the feedback loop. I removed this transistor from the circuit to ensure that changing the opamp to the LM6172 would not induce instability. I added additional power supply filtering using 0.1uF polyester capacitors. These changes brought a welcome lift to performance; the most noticeable was improved transparency and smoother treble. Unfortunately without the LClock X02, the performance was still bettered in all respects, except in PACE, by the SCD-777ES.

CONCLUSIONS

CDP-XA555ES: Re-clocking made a huge difference to sound quality in every respect – highly recommended. Tweaking analog output stage and replacing opamps brought welcome lift in transparency and smoother treble – recommended.

SCD-777ES: Replacing output stage with custom design transformed the sound quality – highly recommended. Reclocking made useful difference – recommended, but not without addressing the output stage.