Which SACD player 2nd hand?

[stephensank]

When given abundant energy storage(large Nichicons) and HF filtering (polystryene film caps) IMMEDIATELY at each supply pin to ground, the OPA2134 is a better sounding opamp than the OPA2132, and better than any opamp I've yet heard. Tonally rich without coloration(put several 134 stages in series & they still sound great, which when done with 'super chips' like the OPA627 yields mush) and super smooth & detailed. The 2132 is analytical in comparison.
When you've been in the service/design/upgrade business for 30 years like I have, you learn not to trust ANY electrolytics but Nichicon & upper grades of ChemiCon. I can't count the number of Elna, Rubycon, Sanyo, Matsuhita/Panasonic & other brands of lytics that I've had to replace due to failure or chemical leakage(Panasonics & Elnas are GUARANTEED to leak within 5 to 10 years or so, which is unacceptable to me). I worked on a CDP-X33ES a little while back where ALL of the Elna Duorex caps had leaked & caused corrosion damage to large areas of circuit traces, and I've seen the same thing with Silmics & other Elnas. Only caps more guaranteed to FAIL are the Sanyo Oscons. The S9000ES uses a lot of really nice Nichicon caps, they're just not big enough for my taste, nor close enough to the stages they're supplying, and not bypassed with good enough nor big enough film caps, but it also uses quite a few Elna caps in the dac & output stages, which gotta go.
As for the unnecessary DC blocking caps, they are there mainly, I think, to make idiot insurance people happy. In the extraordinarily rare case of one of the opamps dying, a no-blocking-cap circuit might spit out some dc voltage to the preamp, which might, oh the horror, cause noise in your preamp's volume control or cause your amp to mute, if you use a passive preamp. Since no capacitor(or paralleled group of caps) is perfect at passing an audio signal, whenever I can eliminate a cap from the signal path without genuinely compromising safety or reliablility, I'll do it.
And even a stock 9000 should mop the floor with an 555ES, which is a cheap piece of crap by comparison, and with a notoriously unreliable laser.

[ClaveFremen]

Originally Posted by stephensank

When given abundant energy storage(large Nichicons) and HF filtering (polystryene film caps) IMMEDIATELY at each supply pin to ground, the OPA2134 is a better sounding opamp than the OPA2132, and better than any opamp I've yet heard.

Never tried in this config, unless for big you mean 100uF, so I must give you credit.

I imagine that OPA2132 was compared in the same circuit, right?

But for a simple opamp swap without psu mods OPA2132 is for sure better.

Originally Posted by stephensank

When you've been in the service/design/upgrade business for 30 years like I have, you learn not to trust ANY electrolytics but Nichicon & upper grades of ChemiCon. I can't count the number of Elna, Rubycon, Sanyo, Matsuhita/Panasonic & other brands of lytics that I've had to replace due to failure or chemical leakage(Panasonics & Elnas are GUARANTEED to leak within 5 to 10 years or so, which is unacceptable to me).

In my 8 years old SACD player Elnas are still good.

We well see in the next 2 years, ok?

Originally Posted by stephensank

As for the unnecessary DC blocking caps, they are there mainly, I think, to make idiot insurance people happy.
...
Since no capacitor(or paralleled group of caps) is perfect at passing an audio signal, whenever I can eliminate a cap from the signal path without genuinely compromising safety or reliablility, I'll do it.

I must be on the idiot's side...

But I agree that DC coupling sounds always better.

Originally Posted by stephensank

And even a stock 9000 should mop the floor with an 555ES, which is a cheap piece of crap by comparison, and with a notoriously unreliable laser.

Here I think you're wrong, my SCD-555ES is the European single drive version, not the much inferior SCD-C555ES cd changer sold in the USA.

SCD-555ES is the little brother of SCD1/SCD-777ES.

SCD-C555ES:



SCD-555ES:



SCD-777ES:

[ClaveFremen]

Originally Posted by stephensank

Only caps more guaranteed to FAIL are the Sanyo Oscons.

Mmh, the dozens of ten years old servers in h24 service I've seen are full of Sanyo OSCONS that still works...

It's strange we have such a different experience with those caps...

[stephensank]

You're quite right on the model number. I was not aware of the SCD-555ES and thought you were referrring to the horribly cheaply built five-disc changer of similar number. However, I do think the 9000ES has the advantage in that the dsd-proc/dac/audio board has it's dedicated power supply right on the same pcb, and has more room in the supply area for much larger Nichicon pre & post regulator filter caps & big fat film caps. So, stock unit to stock unit, probably the 555 + 777 might just beat the 9000 by a little bit, but I think, at least the way I approach upgrading, the 9000 has more potential, plus is an excellent dvd player.

[ClaveFremen]

Bigger Nichicons than these?

As you can see the PSU has all Nichicons of big value except the analog dual line that has BHC big caps and Silmics small ones.

Two R-Core transformers one for the transport and one for dac/analog, 7 regulated power lines with a diode bridge for each one, big caps (2200uf, 4700uF, 6800uF), all elcos bypassed with film caps.

Not bad for a PSU, or not?

[stephensank]

The 9000ES starts out with 3900uf Nichicons pre-regulator on +/-15V & +5V supplies, and 1000uf immediately post reg on each, plus nominal amounts of distributed smaller Nichicons in each of the stages, with fairly small film caps post reg(1uf max). When I upgrade the model, I use 10,000uf Nichicons pre-reg & 4700uf post reg, plus WAY larger Nichicons in each stage, and LOTS of film caps(at least 10uf of mixed film types immed. post reg). The room the 9000 has on board, especially in the power supply section, makes this very feasible. However, I do have to go 'under-board' to get lytics(470uf at least) & films as tightly close to each output stage chip supply pins as I like to have them. Also, the 9000 starts out with very nice, TO220 package rectifiers, but haven't found data to confirm they're soft recovery, and I like the IXYS soft recoveries best anyway, so I always install them.
In my view, there is no such thing as a power supply that is too big or too clean.

[ClaveFremen]

Originally Posted by stephensank

The 9000ES starts out with 3900uf Nichicons pre-regulator on +/-15V & +5V supplies, and 1000uf immediately post reg on each, plus nominal amounts of distributed smaller Nichicons in each of the stages, with fairly small film caps post reg(1uf max).

Well, comparing stock PSUs to me mine seems better.

Your PSU mod seems overkill but all that filtering power could be unnecessary with soft recovery diodes.

Your modded PSU improves if you swap diodes with soft recovery ones?

On my audio board there's further LM7XXX regulation for DAC and opamp/discrete regulation for the opamps analog stage.

Do you think that a stiffer PSU could help also in this case?

If you agree I could e-mail you the service manual so you can have a documented idea.

[stephensank]

Have emailed you the S9000 audio diagrams. It uses opamp/discrete regulators direct to the output stages, but derives +5/-7V supplies for the 'pulse-dac/LP filter' chips from those +/-12V supplies. Straight discrete reg for the +7V for the pre-DAC circuits, from which each section(clock, DSD proc, spdif driver, etc.) has it's own 78xx or similar reg ic for 5V or 3.3v supply. So, great design, quite well executed for a stock product, but ample room(physically & sonically) for significant improvement.
In my experience, no matter how elaborate/well done a supply is done, and no matter how well regulated, good soft recovery diodes(IXYS, IR, or at least Motorola/ONsemi) ALWAYS make some significant sonic improvement, as does going bigger & better on even pre-reg filtering, and most crucial is getting final supply filtering/energy-storage as physically close and large as possible to each opamp, dac chip, etc. Even with soft recovery diodes, there's no such thing as too much filtering. Each stage will benefit from having super clean, ample & instantaneously available current, and will also be that much less likely to negatively interact with other stages via supply modulation.

[ClaveFremen]

You have mail, Stephen

Ouch!

I've missed the regulator...