The primary focus of this mod is the eight resistors and two capacitors that comprise the filter following the “current pulse†D/A converters. The existence of this filter is unusual. It’s typical practice to run the output of a current-output DAC directly into an op amp using it as a transimpedance amp (or current to voltage amp). This practice gives the op amp a bit of a challenge since it is being fed a high-speed pulse train and faster op amps tend to sound better when used this way.But Sony runs the DAC output into the virtual ground of an op amp through an R/C filter – effectively accomplishing the I/V conversion passively. The op amp sees a smoothed analog signal, making its job easier. This passive I/V conversion may be one of the reasons the Sony high-end SACD players sound as natural as they do, even when playing CDs.
Sony apparently designed its current-output DACs (and the DAC they designed with Burr-Brown -- the DSD1700) to tolerate a voltage burden. The discontinued NPC SM5864 is the only other DAC I know of that was designed specifically for passive I/V conversion. Other discontinued current-output DACs (the Burr-Brown PCM63 and Philips TDA1541/3) were found to be capable of handling a voltage burden although the app notes made no mention of it. And apparently the Burr-Brown PCM1738 (the DAC used in the XA777ES) can handle a voltage burden -- cause Sony puts a resistor between the virtual grounds of the op amps and this DAC.
So, the resistors and capacitors in the filter after the D/A converters are probably the most important passive components in the player – and they could be better. The caps appear to be dipped Shizuki polyester film caps and the resistors are surface-mount carbon film MELF resistors (that’s Metal Electrode Leadless Face). At least the resistors don’t have steel end caps like most MELFs.
Finding suitable caps to replace the stock caps is easy as Rel-Cap offers several choices in the .068 f size. I decided not to spare any expense and bought TFT Exotica Teflon caps. I mounted them vertically which isn’t great for physical symmetry – but horizontal space is limited.
The resistors proved to be more of a challenge. Sony used ½ watt Rikens in the direct signal path, but used MELFs in the legs to ground – presumably to save space and cost. But for reasons not clear to me – the MELFs found their way to the most critical location on the audio board.
I replaced the MELFs with Caddock flip chips – a serendipitous choice. I wanted to use Caddock resistors because I thought they would be most compatible with Sony’s “analogy†voicing – which I happen to like. Flip chips are the only type of surface mount resistor Caddock makes – a product with which I was not familiar. The resistor’s metal film faces toward the board and mounting pads are applied directly to the film. This results in super low inductance – a characteristic that seemed good to me. As it turns out – it’s a very, very good, but more about that later.
Replacing the MELFs requires skills and equipment not possessed by the average tweaker – including me – so I made another trip to my local prototyping and repair shop. The solder and glue holding the MELFs in place must be removed completely because the flip chips have virtually no vertical clearance. The solder pads are not accessible from the side of the chips (they are solder bump type components). And if you still need to be convinced not to attack with your Radio Shack iron – the lands on the board aren’t really the right size for the flip chips.
I also replaced the four resistors in the feedback loops of the OPA2604 op amps with Caddock MK132 resistors. The OPA2604s are responsible for just about all of the voltage gain – so the resistors in the feedback loop are critical. I like the Rikens a lot and I’m keeping all of the rest of them, but this location calls for a resistor with less noise, more transparency and greater precision.
When I listen to real music, it always sounds bigger and more dynamic to me than recorded music –and it seldom sounds more detailed. So my first objective in making this mod was not to lose the big, weighty, slightly mellow sound of the Riken resistors. I’m happy to report that nothing was lost. In fact, the sound is now substantially bigger and weightier – getting closer to the real thing.
I’ve never thought the reproduction of high frequencies on digital players has compared well to live music either. The SCD-777ES has been a godsend because its treble performance is quite tolerable – especially on SACDs. In substituting metal film resistors for carbon films I was afraid that the treble performance might take a turn toward “steelynessâ€. But – miraculously -- the treble performance is vastly improved. In fact, it’s like nothing I’ve heard from digital electronics. It’s very natural sounding and there is great extension with no “edge†at all – even on CDs!
And of course there is the expected improvement in detail retrieval. Although metal film and foil resistors can have faults, detail retrieval is not one of them. The improved resolution also improves sound staging that was already excellent – especially on SACDs. The greater precision of the new resistors may also be contributing to this improvement.
The biggest surprise – and the thing about the mod that reaches out and grabs you – is the improvement in dynamics. Percussion is awesome and the realism with which individual instruments enter an ensemble performance is shocking. And when the music “gets busy†there is virtually no congealing of the sound.
So what’s going on here? It’s a little hard to say exactly since I changed resistors and capacitors. However, the effects I described were immediately evident when the caps were still in their virginal state (not broken in). So I believe that I am hearing more of the change in resistors than in capacitors. I would hypothesize that most the change is due to the ultra low inductance of the Caddock flip chips – perhaps aided by their superior power handling capabilities (they’re rate at 3/4 watt vs. 1/8 watt for the MELFs).
This mod may even explain why Sony 1-bit players (both CD and SACD) have always had a reputation of sounding a little “soft†compared with their multi-bit counterparts. But now I’ll challenge any multi-bit player to a dynamics contest.
Things are sounding good enough that I’m losing motivation to continue. I think I could listen to the player in its current state indefinitely. But I want to try one more power supply mod -- and then there’s the clock.
Note for DVP-S9000ES and XA777ES Owners: The good news for DVP-S9000ES owners is that the filter resistors are thru-hole types and can be easily replaced. The bad news is they’re thru-hole types and ultra low inductance will remain a dream. The XA-777ES uses MELFs – in fact all of the resistors in the audio signal path are MELFs. However, the value of the resistors is 47 ohms – and 10 ohms is the highest value Caddock makes in a flip chip (they’re marketed for current sense applications). Vishay makes metal foil flip chips in larger values (like the VFC1206) – and maybe you can find some.Other SCD-1/777ES Mods and Information
AC Board:
Removed choke on analog power supply
http://www.audioasylum.com/forums/hirez/messages/32357.htmlReversed AC polarity
(Not posted)Output Board
Replaced RCA jacks with Vampire OFC jacks
Replaced output wire with Kimber AGSS
Revised grounding creating a star ground for audio board
http://www.audioasylum.com/forums/hirez/messages/41691.htmlAudio Board
Replaced resistors and capacitors in I/V conversion stage
(This post)Replaced AD712s in the output buffer stage with hi-spec OPA2132s
http://www.audioasylum.com/forums/hirez/messages/78259.htmlSet filters to “customâ€
General Description
Analog Signal Path
http://www.audioasylum.com/forums/hirez/messages/33130.htmlAccessories
Electraglide Reference Glide power cord
Chang Lightspeed HT1000 power conditioner (w/upgraded capacitors)
AudioDesk Systeme CD Lathe
Aurio Pro Media Isolation Bearings w/optional tungsten carbide balls
This post is made possible by the generous support of people like you and our sponsors:
Topic - SCD-1/777ES I/V Conversion Stage Mod - jhietbrink 19:42:37 08/27/02 (10)
- Re: SCD-1/777ES I/V Conversion Stage Mod - fmak 00:17:08 08/29/02 (1)
- Re: SCD-1/777ES I/V Conversion Stage Mod - jhietbrink 19:53:50 08/29/02 (0)
- Re: SCD-1/777ES I/V Conversion Stage Mod - PhopsonNY 06:17:05 08/28/02 (3)
- Re: SCD-1/777ES I/V Conversion Stage Mod - jhietbrink 19:30:13 08/28/02 (2)
- Re: SCD-1/777ES I/V Conversion Stage Mod - PhopsonNY 07:44:16 08/30/02 (1)
- Re: SCD-1/777ES I/V Conversion Stage Mod - jhietbrink 18:43:21 08/30/02 (0)
- Reversed AC polarity? - jeromelang 02:09:21 08/28/02 (1)
- Re: Reversed AC polarity? - jhietbrink 18:29:56 08/28/02 (0)
- Re: local prototyping and repair shop.. - la6 20:49:03 08/27/02 (1)
- Re: local prototyping and repair shop.. - jhietbrink 18:59:27 08/28/02 (0)