10uf silmic II on both I/V and the Buffer sound really good. Very tube like. Very warm. I would not mind a bit more definition in the treble though. So here are the experiments to see if this sound can be improved:
here are the test tracks I used:
1. Arianna Savall – Preghiera This tests sibyllance, 3d and locality.
2. Gazpacho – Valeries Friend A lot of cymbals, Prominent Vocals. Lesser Caps turn it into a screechy awfulness.
3. thePhantom – Freeze Deep Bass, synth cymbals
4. Yosi Horikawa – Bubbles Test coherence. Lesser capacitors become confused and it’s a mess.
Silmic II
By itself is good enough. It is so deep and rich, I cant get away from my computer. The sound is thick and rich and full and enveloping and powerful. I feel the bass in my chest. The soundstage is deep, 3 dimensional, every instrument is locked into place, rock solid and separate from each other.
Silmic II bypassed with 10nf MKP 1837
So I try the MKP 1837. Usually very good at cleaning up and “etching” the top end.
In this case it is just harsh. Just the Silmic leave natural sibilance natural sounding, just the SSSS that is produced with certain words. The MKP 1837 make it a grating CHCHCHCHCH. Arianna Savall is too sibilant and Gazpacho is intolerable. This is not a successful bypass. Also it made the sound stage feel much less 3D. So I had to take those off.
Silmic II bypassed with Silver Mica
My impression? I don’t know... tough to pinpoint. Only way i can describe it is that it sounds “lighter” in the treble area. Its not crisper, there is no sibilance, nothing etched but the top end sounds lighter in a soft way. Its not harming the soundstage much. It is deep, with all the instruments and sounds floating in their proper location but the “softness” causes a slight loss of coherence. This becomes evident with songs like “Bubbles” from Yosi Horikawa. Its as though there is a soft halo around the treble.
This is a very very nice sound. I could live with this.
The only sensible next step would be Teflon. So lets try those.
Silmic II bypassed with Teflon
The soundstage is absolutely correct. There is no sibilance. The teflons seem to have brought that very last 2% of clarity to everything. Could I reliably A-B test and pick it out though? I don’t know. It is not harming the feel of the silmics. There is warmth and body. It preserves everything that the silmics do right but add clarity to the whole picture. But this is a very subtle addition. It is just.. right. This is very very good.
All the micro-details are there, every shimmer from the cymbals is there. Gazpacho is great. The bass is fast. The voice is not screechy, The cymbals don't become a mess. Very dynamic. This is very very good.
Elna Cerafine
not as good as silmic. There is magic missing. The clarity is nice, the tone seems correct. It’s just not as awesome. Its more fatiguing than the Silmics. The cerafines are not producing details that the silmics are not. It just seems accentuated. This is not better and I am missing the Silmics. Don't get me wrong, these are excellent sounding caps but just missing some weight and "magic"
The difference can be described as a 60 lb kid on a trampoline jumping 5 feet in the air. He goes up and comes down and up and down. Now imagine a 300 lb dude jumping 5 feet in the air on the trampoline. They both move the same distance but the impact felt by the trampoline is different. That is how this feels. The notes are there but its not taking you and forcing you to rock and tap to the rhythm.... it's optional.
Panasonic SEPF 390uf
So I thought it would be good to find out at what voltage the sockets were running at before using a 20V capacitor. My Xonar is at 16.7v from – to + so I am ok with these caps.
There is more information in the music than with Silmic II. It is almost sibilant but it isn’t. It is just very very fast. You can hear every shimmer of cymbals. You can hear where on the cymbal the drum stick hits. All the micro detail is there. The bass is full and taught. It does not become confused.
But.. that fat, warm, luscious bass from the Silmic.. I miss it. Perhaps it is colored but I want it.
A perfectly cooked steak is delicious, but adding seasoned butter is good too.
It is so clean it is addictive. It sounds like the silmic when bypassed with the teflons.
Impression: Very clear, very clean, wide open, neutral. This is very very good.
How about the buffer?
Current Setup
AD797BRZ I/V with Panasonic SEPF
LME49990 Buffer with Silmic II
Fantastic. Amazing. I cant turn it off to test anything else. The bass is right. The mids are right. The treble is right. It is soooo fast but not etched. You are retrieving all the information there is to be had in a recording.
SEPF on the I/V and the Buffer.
This presentation is missing magic. Ariana sounds correct, but im not melting into the music, im not marveling. It is more 2D. It is slightly sibilant. This combo is not as good.
also, the LME49990 was definitely getting warmer.
Next up, SEPF on lme49990 buffer and Silmic on the AD797 I/V:
Not sibilant. Warm and musical. But I feel like im missing the extra detail from the SEPF on the I/V stage:
Next up: SEPF back on the AD797 and a bigger silmic on the LME49990 buffer.
390 uf for the I/V is not hurting anything. Lets see if 470uf on the buffer is ok.
Result: it makes no difference except that it is huge and looks stupid. No benefit over 10uf.
Next up LT1028A in buffer. No Cap: Bass is overblown, slightly flabby. It does not sound as liquid as the LME49990.
There are resonances that are not quite right. Good detail retrieval. Chips are warm. need to add the Silmic
With Silmic: Brings it all together. Bass is tighter. Smooths it out. Does not have the same sound stage as the LME49990. Seems more fatiguing. Not better.
Next up: LME49720HA with Silmic. Really good bass. Not sibilant. Treble is well defined. Actually a very good opamp but not as good as LME49990.
Current setup:
going back to Silmic with Teflon bypass on AD797 in I/V and Silmic on the LME49990 in the buffer.
This is very very good. The sound stage is phenomenal. The musicality is phenomenal, the prat is phenomenal.
my current recommendation:
if you have a xonar, put AD797B in the I/V stage and bypass it with 10uf Silmic II and a little Teflon. Alternatively, use a Panasonic SEPF. I have some 22uf ordered and will let you know if the smaller capacity makes a difference. Based on my experiment with the 470uf Silmic, I am guessing not.
Put a LME49990 in the buffer stage and bypass it with 10uf Silmic II.
This combo is really very very good.
If anyone tries this out, let us know your impressions.