or Connect
Head-Fi.org › Forums › Equipment Forums › Computer Audio › Hotrodding the X-Fi: A Layman's Guide (No 56k)
New Posts  All Forums:Forum Nav:

Hotrodding the X-Fi: A Layman's Guide (No 56k) - Page 78

post #1156 of 2194
Decided not to install AD8599 this round, but installed LME49720, to compare against previous LME49860.
- curious to see if any compromises were made, given higher Vs voltage range. (+/-17 vs., +/-22)
- appears to have the same "excellent" sonic quality.
- RMAA results show no obvious deviations.

Will have to order SOIC socket/flipchip, this week.

Eyeballing NJM2144 for possible installtion...
- just so I can win a prime-rib dinner from "GermaniumShortedByPassBoi."
- so our gear is more "apples-to-apples," his with caps stapled on and my tants... (LOL)

Change Status: ** Sep 04, 2007 **
- LME49720 from NJM4556 (Line-Out) (<-- new, was LME49860)
- LME49860 from TI M33078 (Line-In)
- (C177) Panasonic FK, 1500uf @ 16vdc from Jamicon WL 220uf @ 16vdc
- (C16) Panasonic FC, 220uf @ 25vdc from Jamicon SS 100uf @ 16vdc (replaced C46 with original Jamicon 100uf)
- (C107) 47uf @ 20vdc TANT from G-LUXCON, 47uf @ 16vdc (<- new add)
- (C72) 10uf @ 16vdc TANT from Jamicon SS 22uf @ 16vdc
- (C91, C119, C108) 4.7uf @35vdc TANT from Wincap 10uf @ 16vdc (<- new add C108)

Hardware Pic:
http://www.esnips.com/doc/f8a6eac3-9.../X-FI-MOD-029b

RMAA Results Summary:
http://www.esnips.com/doc/fae858ca-8.../X-FI-MOD-030b

RMAA Results Detail:
http://www.esnips.com/doc/aa2e8487-3.../X-FI-MOD-031b
post #1157 of 2194
Quote:
Originally Posted by germanium View Post
Here is link to my tests on Esnips http://www.esnips.com/web/X-FiRMAAtests/
Nice!
- gonna be hard to "whup" yer bypassed, shorted hinder...
- got me beat by 10db in noise, at the low end!

And now that you're all set on eSnips, post a picture of yer, umm, umm, "thing..."
post #1158 of 2194
Quote:
Originally Posted by bichi View Post
Nice!
- gonna be hard to "whup" yer bypassed, shorted hinder...
- got me beat by 10db in noise, at the low end!

And now that you're all set on eSnips, post a picture of yer, umm, umm, "thing..."
Shoot, I even beat Creatives own tests by 3db in the noise catagory. Check out the IMD swept distortion. Mine is way better at the top end.
post #1159 of 2194
I'll play this game, and whoop you all.

Be right back with 24bit 96khz tests.
post #1160 of 2194
Here's what I got so far.

Will see how it does with a new opamp, and once the new caps go through a few power cycles.

EDIT: Also bichi, me and you at a big disadvantage since germanium has a better DACs.

Our big improvement can come from sound detail and such from using high slew rate, low settling time opamps.

By picking something much faster than a slow 16v/us.
post #1161 of 2194
Quote:
Originally Posted by LawnGnome View Post
I'll play this game, and whoop you all.
Be right back with 24bit 96khz tests.
Quote:
Originally Posted by LawnGnome View Post
EDIT: "...Also bichi, me and you at a big disadvantage since germanium has a better DACs...."
cool!
- the fun is comparing DC-coupled and parallel-bypassing against AC-coupled, op-amp swap and tantalum bypassing.
- plan for me was to complete tant bypass stuff, then test direct coupling, as the final step, using Germanium's excellent results, as a reference. (yes, well aware his SB0550 is a superior card)
- all this prodded by Rod Elliott's and Cyril Bateman's papers and caused me to re-think design habits. (ref below)

So we can see each other's complete RMAA results:
- both "[DirectSound] SB X-Fi Audio [7C00]-xxx" folder and "[DirectSound] SB X-Fi Audio [7C00]-xxx-set.htm," zipped.
- also, Germanium provides excellent analysis of RMAA detailed results too...
- open an eSnips account? - (free) http://www.esnips.com/
- pictures would be nice too...

You are the one with "recapped MB and PSU," right?

Germanium:
OS: Vista Ultimate 64b; XP sp2
X-FI Driver: 2_13_0012
AUD: X-FI Elite Pro (SB0550)
TRICK: metalized bypass; DCouple; stock NJM2114
RMAA: RMAA 6.0.5
MB: Abit AB9Pro, Core2Duo E6600 overclocked to 3.006 GHZ & 4GB memory
PSU: Pure power 480
Link to pics and Results: http://www.esnips.com/web/X-FiRMAAtests/

bichi
OS: XP SP2
X-FI Driver: 5.12.6.1187
AUD: X-FI F1 (SB0460)
TRICK: tantalum bypass; LME49720 L-Out; LME49860 L-In
RMAA: RMAA 6.0.5 (demo)
MB: ASUS P5WDH, Q6700 @ 3.0, 2/4G RAM @ 500mhz (PC-8800)
PSU: Seasonic S12+ 550
Link to pics and Results: http://www.esnips.com/web/X-FIOpAmpandCapModification

REF:
Capacitor Characteristics; Ron Elliot, (Audio Applications)
http://sound.westhost.com/articles/capacitors.htm
post #1162 of 2194
Quote:
Originally Posted by bichi View Post
- plan for me was to complete tant bypass stuff, then test direct coupling, as the final step, using Germanium's excellent results, as a reference. (yes, well aware his SB0550 is a superior card)
Interesting. Why tantalum? I've heard of it being preferenced in other threads, but never really sure why. Germanium seems to prefer film. I'm thinking of tantalum as well, just cause it's easier to squeeze an SMD part in parallel and tantalum is easier to obtain.
post #1163 of 2194
Quote:
Originally Posted by holland View Post
Interesting. Why tantalum? I've heard of it being preferenced in other threads, but never really sure why. Germanium seems to prefer film. I'm thinking of tantalum as well, just cause it's easier to squeeze an SMD part in parallel and tantalum is easier to obtain.
umm, short answer:
- tantalums have very low-dissipation and ESR figures, with much better stability/drift, compared to electrolytics.
- to your point, they are compact and are space efficient.
- downside: are sensitive to voltage abuse and are more expensive than AL electrolytics.
- (but MUCH cheaper than AuriCap <-- points at Germanium...)
- since tantalums are constructed with a "solid slug" electrolyte and are not "wound," do not have the same inductive characteristics. (subject of high controversy, see Ron Elliot's paper, figure 11)
- Creative (SB0460) uses "parallel power bypass" with an electrolytic and small-value chip cap for the CS4382 DAC power pins. (VA, VD, VLS, VLC: see Cirrus Logic CS4382 datasheet, figure 6)
- much, much more technicals on tantalum with Google.....

My "casual" theory is, given the very noisy PC environment (ghz hash), the DAC's digital nature, tantalum's ability to shunt effectively (hash in as well as out, given close proximity to audio op amps) energy density (power transient) and space on the card, tantalum would be as effective or maybe more effective than adding "external" bypassing caps.

In the current state of mods, listening to "Finding Neverland" track (from movie DVD), there is a section where theme song is playing, characters are moving about in a playhouse. A bell "tinkles" and an "electrical" special-effect is present.
The damn "tinkle" sounded like it was "INSIDE" my head, I felt like I was sitting IN the balcony and the "electrical" special-effect was amazing..... Positive I never heard it the same way with stock card. (never thought such high-quality, with precision phase and dynamics could come from PC audio.....) Playback is not anything special, just a pair of M-Audio BX-5's....

REF:
Electromagnetism 1 - Ivor Catt
http://www.ivorcatt.com/em.htm
post #1164 of 2194
Oh, sorry, didn't mean relative to electrolytic. I meant relative to film (Germanium's preference). I guess that's answered by your second bullet (size)?

Yep, saw the CS spec some time ago. I haven't checked to see what's on the X-Fi board though, so I'm just leeching.
post #1165 of 2194
Quote:
Originally Posted by bichi View Post
umm, short answer:
- tantalums have very low-dissipation and ESR figures, with much better stability/drift, compared to electrolytics.
- to your point, they are compact and are space efficient.
- downside: are sensitive to voltage abuse and are more expensive than AL electrolytics.
- (but MUCH cheaper than AuriCap <-- points at Germanium...)
- since tantalums are constructed with a "solid slug" electrolyte and are not "wound," do not have the same inductive characteristics. (subject of high controversy, see Ron Elliot's paper, figure 11)
- Creative (SB0460) uses "parallel power bypass" with an electrolytic and small-value chip cap for the CS4382 DAC power pins. (VA, VD, VLS, VLC: see Cirrus Logic CS4382 datasheet, figure 6)
- much, much more technicals on tantalum with Google.....

My "casual" theory is, given the very noisy PC environment (ghz hash), the DAC's digital nature, tantalum's ability to shunt effectively (hash in as well as out, given close proximity to audio op amps) energy density (power transient) and space on the card, tantalum would be as effective or maybe more effective than adding "external" bypassing caps.

In the current state of mods, listening to "Finding Neverland" track (from movie DVD), there is a section where theme song is playing, characters are moving about in a playhouse. A bell "tinkles" and an "electrical" special-effect is present.
The damn "tinkle" sounded like it was "INSIDE" my head, I felt like I was sitting IN the balcony and the "electrical" special-effect was amazing..... Positive I never heard it the same way with stock card. (never thought such high-quality, with precision phase and dynamics could come from PC audio.....) Playback is not anything special, just a pair of M-Audio BX-5's....

REF:
Electromagnetism 1 - Ivor Catt
http://www.ivorcatt.com/em.htm
Not using Auricap, Just a cheap no name but good polyester metalyzed film cap. I see no purpose in going to epensive ones unless you realy need large caps such as 200mf which used to & may still be made by Solen. Solen makes realy good caps but are expensive & really only worth it if you can't get a film cap of your size requirement any where else.
post #1166 of 2194
Quote:
Originally Posted by holland View Post
That's about average. 30-60 is the range of most opamps. My reading was he wanted to drive it with authority, meaning you would need a couple hundred mA. The opamp was also mentioned early on, as an alternative to the LM4562, if driving headphones.

LawnGnome, if you're going to try and run a finicky opamp, you need to figure out the entire opamp circuit. There's a decent chance you will oscillate.
Actually not looking for massive current just good current with no loss of performance with a load resistance less than 100ohms is all bu I'm finding that very hard to find. Most opamps seem to like 600 or more ohms with 2000 oms being optimal for them. Unfortunately 600 ohms & above is not optimal for driving headphone or capacitative cabling as values that high leave the sound somewhat muffled if you have to run any lenth of even modestly capacitative cable. That is why I make my own cables for connecting everything else but the PC, to keep the capacitance below 50pf. If it gets up to anywhere near 100pf i start hearing a slightly lifeless modestly muffled sound. The higher the value of the load resistor the more muffled the sound for a given capacitance.

The reason I haven't made my own PC interconnects is the load resistance hence output impedance is small enough the even moderately capacitive cables don't take the life out of the music. There is still enough output impedance though to protect the amp from oscillating in the presence of capacitative cabling. Some amps by the way are not stable into capacitative cable & will oscilate like crazy if exposed to the capacitance with no resistance between the amp and the cabling.
post #1167 of 2194
By the way I relooked at the specs of the lm4562 & I have to say I was dead wrong on the voltage where it performed best. It's 4-5 volts instead of 10 volts. My bad memory getting me in trouble again I guess, anyway I just wanted to correct that.

4-5 volts is actually A reasonable test voltage as it is not very much worse at 2 volts but would still like to see 2 volts as a standard for testing the opamps but still post the graphs for those that need the higher output voltage such as some car stereo preamp outs.
post #1168 of 2194
Quote:
Originally Posted by LawnGnome View Post

EDIT: Also bichi, me and you at a big disadvantage since germanium has a better DACs.
It's not that my DACs are that much better as there should only be 6db noise difference in the DACs. It is the ADCs that are really handicapping you. The noise spec for the DAC in the lower X-Fi cards is 114db & mine is 120db. My ADC has a noise spec from the factory of 123db which is better than the DACs 120db. The ADC on the lower cards is rated only 102db S/N.

http://www.wolfsonmicro.com/uploads/.../en/WM8775.pdf
post #1169 of 2194
Quote:
Originally Posted by germanium View Post
http://www.wolfsonmicro.com/uploads/.../en/WM8775.pdf
It's not that my DACs are that much better as there should only be 6db noise difference in the DACs. It is the ADCs that are really handicapping you. The noise spec for the DAC in the lower X-Fi cards is 114db & mine is 120db. My ADC has a noise spec from the factory of 123db which is better than the DACs 120db. The ADC on the lower cards is rated only 102db S/N.
http://www.wolfsonmicro.com/uploads/.../en/WM8775.pdf
So that means we get a 6db noise handicap.... (weeeeeee!)

Stopped by tech surplus and might have found a cheap, low-friction, low-pull SOIC op amp socket solution.
- a bit of careful cutting and it appears I can make a pair of sockets and plugs (chip carrier) per strip.
- about $2.75 USD for each part
- back in a bit, after installing and dinner...

pic: http://premium1.uploadit.org/bichi00...t-hack-001.jpg
post #1170 of 2194
Quote:
Originally Posted by bichi View Post
So that means we get a 6db noise handicap.... (weeeeeee!)

Stopped by tech surplus and might have found a cheap, low-friction, low-pull SOIC op amp socket solution.
- a bit of careful cutting and it appears I can make a pair of sockets and plugs (chip carrier) per strip.
- about $2.75 USD for each part
- back in a bit, after installing and dinner...

pic: http://premium1.uploadit.org/bichi00...t-hack-001.jpg
More like 18db handicap. 120db S/N DAC compared to 102db S/N ADC best case scenerio noise figures. Only problem there is enough noise in the ADC to completely obliterate any differences in the analog output from the DACs & opamps unless you make a really serious design error. If you go by creatives specs for my card which is 116db then it is a 14db handicap
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Computer Audio
Head-Fi.org › Forums › Equipment Forums › Computer Audio › Hotrodding the X-Fi: A Layman's Guide (No 56k)