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 94

post #1396 of 2194
Quote:
Originally Posted by ter1 View Post
Germanium, I put the 2 caps back, and these days I use a fan to blow the card directly, everything is okay, but when I disconnect the fan, the card disconnect itself. This lead me to think the self shutdown is really caused by the heat.

Only the regulator above the DSP is pretty warm,,, no other parts that warm,,,

Do you have advice how to get rid of this strange situation? the fan is noisy and it's just a workaround...
Someting is drawing too much current or it is too close to other heat generators. Mine gets pretty hot too but that is true for both the regulator & DSP unit if left on long enough. Mine never shuts down though. Neither of my elite pros did that & I have the breakout box connected all the time. I do have a video card above mine though with a healthy size Zalman fan though & nothing below that generates heat. The only other thing I can think of is a out of spec circuit breaker.

You could try asking bichi if he has any ideas as i'm running low on them myself. I know it must be frusterating. It would be to me also. I have none of that problem myself.
post #1397 of 2194
Quote:
Originally Posted by germanium View Post
You are very correct. The only things negative I found by shorting the coupling caps is a slight & I mean slight increase in second harmonic distortion. All other harmonic distortion the same or reduced. There is a slight increase in noise below 1KHz on the elite pro card as well but you will not likely see that on the lower cards. What you will see though on the lower cards but not on the elite pro card is high D.C. offset. About 180mv or more as opposed to less than 5mv on the elite pro.
Well, I never thought too much about DC offset of X-Fi, but I guess I'll measure it tonight. It is not currently connected to anything expensive, so I didn't really check and recheck and check again.
post #1398 of 2194
Quote:
Originally Posted by Maniac View Post
Well, I never thought too much about DC offset of X-Fi, but I guess I'll measure it tonight. It is not currently connected to anything expensive, so I didn't really check and recheck and check again.
High D.C. offset would only be a serious problem with D.C. coupled amplifiers (amplifiers that can amplify D.C.) as it would cause them to over heat & possably damage the speakers as well but these type of amps are not likely to be found in anything that is meant to be connected to the computer.

Direct coupled preamps can generate noise in thier volume pots as well with high D.C. offset feeding them. This is a scratchy sound when the volume is changed only. Digital colume controllers will pop at each change in level if direct coupled.
post #1399 of 2194
So after letting my card run for a couple of hours I did a listening test with it and the unmodified Fatal1ty, and I have to say I can't hear the slightest difference. Zip. I just have this feeling that the modded card has a slightly less "sharp" feeling, but there's no way I would be able to tell them apart in a double-blind test.
post #1400 of 2194
anyone know the output of the X-Fi digital out using a multimeter? I need to know so I can adjust my Ori dac accordingly. thanks
post #1401 of 2194
Quote:
Originally Posted by Maniac View Post
you don't need a cap there, just short it out.
Sorry for the noobish question, but HOW DO you short it out?
post #1402 of 2194
Quote:
Originally Posted by BigSurSpoon View Post
Sorry for the noobish question, but HOW DO you short it out?
Well, just solder a small piece of copper wire across the two legs of the capacitor, one per capacitor, and soldering on 4 pieces was all I did. :P
post #1403 of 2194
Quote:
Originally Posted by lutorm View Post
So after letting my card run for a couple of hours I did a listening test with it and the unmodified Fatal1ty, and I have to say I can't hear the slightest difference. Zip. I just have this feeling that the modded card has a slightly less "sharp" feeling, but there's no way I would be able to tell them apart in a double-blind test.
If your sound system isn't up to to the same level as the card you may not hear much difference. All my sound system has been modified with high resolution in mind & the diferences are actually pronounced on my system.

Check here for my sound system specshttp://www.head-fi.org/forums/member.php?u=43462
post #1404 of 2194
Have been working on Germanium's SB0550 and documented info that may be useful for other's modding/troubleshooting the Elite version...

SB0550 - X-FI Elite Pro only:
- (UPDATED) Nov 03, 2007

LINE-OUT:
U7: Line-Out opamp (NJM2114)
C45, C35, C58, C49: Line-OUT coupling capacitors to Cirrus Logic CS4398 DAC (Jamicon SS, 47uf @ 16vdc)
C45, C35 = Right; C58,C49 = Left
*BIAS: 0.023uvdc, measured across sleeve-ring and sleeve-tip, 47uf coupled
*BIAS: Coupling-caps shorted (100ohm): +3.1 mvdc Right, -2.2mvdc Left, measured across sleeve-ring and sleeve-tip (AD8599)
*BIAS: 2.5vdc @ 1.55ma, measured across coupling capacitor, post initialization (supplied after relays set, un-mute)


LINE-IN:
U3: Line-IN Opamp (I5532; equiv to NJM5532)
C27, C15: Line-IN coupling capacitors (Wincap NP, 10uf @ 16vdc)
U5, U2: Differential opamps (I5532; equiv to NJM5532) to AKM AK5394 ADC
C29, C14: U3 Line-IN opamp coupling capacitors to U2, U5 differential opamps (Jamicon 47uf @ 16vdc)
*BIAS: C27, C15: 1.9mvdc @ 0.02ua, measured across coupling capacitor
*BIAS: C29, C14: 2.45vdc @ 1.45ma, measured across coupling capacitor


ANALOG/AUDIO SIGNAL PATHS:
Line-OUT: J1 <- relays <- U7 opamp <- C45, C35, C58, C49 coupling caps <- CS4398 DAC
Line-IN: Flexijack -> relays -> C27, C15 -> U3 opamp -> C29, C14 -> U5, U2 opamps, differential -> AK5394 ADC

VOLTAGE REGULATORS:
VR1: 78M05; supplies Cirrus Logic CS4398 DAC (x4), +5vdc, Va, analog sections
VR2: 79L05A; supplies 74HCT4053 analog mux, -5vdc
VR3: AMS1117-3.3; supplies Cirrus Logic CS4398 DAC (x4), +3.3vdc, Vd, and AKM AK5394 ADC, Vd, +3.3vdc, digital sections
VR4: 78M05; supplies AKM AK5394 ADC, Va, +5vdc, analog section and +2.5vdc BIAS divider for U3, U5
U21: TPS54352; supplies X-FI CA20K1, +1.2vdc switching regulator

POWER FILTER CAPACITORS:
C104: +12vdc from PCI bus, (Jamicon SS, 220uf @ 25vdc)
C133: -12vdc from PCI bus, (Jamicon SS, 220uf @ 25vdc)
C154: +5vdc from PCI bus, (Jamicon SS, 22uf @ 16vdc)
C151: +1.2vdc from U21 (TI TPS54352) switching regulator to X-FI CA20K1 DSP (Jamicon WL, 220uf @ 16vdc)
*CAUTION: C151 should not be "ultra-low ESR" capacitor, ie., OSCON, etc. See TI TPS54352 datasheet, page 22

POWER BYPASS/DECOUPLING CAPACITORS:
C52: VR1 +12vdc input (Jamicon 22uf @ 16vdc)
C55: VR1 +5vdc output (Jamicon 47uf @ 16vdc)
C109: VR2 -5vdc output (Jamicon SS, 22uf @ 25vdc)
C131: VR3 +3.3vdc output (Jamicon SS, 22uf @ 25vdc)
C6: VR4 +5vdc output (Wincap 10uf @ 16vdc)
C149: CN3 to AD-Link +12vdc (Jamicon SS, 22uf @ 25vdc)
C19: +12vdc input for U3, U5, U2 Line-IN opamps (Wincap 10uf @ 16vdc)
C17: -12vdc input for U3, U5, U2 Line-IN opamps (Wincap 10uf @ 16vdc)
C193: +2.5vdc U5, U2 BIAS divider, from VR4 (Jamicon 22uf @ 16vdc)

POWER PATHS:
+12vdc PCI Bus -> C104 -> V+ Line-OUT/IN, DIFF opamps -> VR1 and VR4 +5vdc 78M05 regulators -> +5vdc CS4398 and AK5395 analog section
-12vdc PCI BuS -> C133 -> V- Line-OUT/IN, DIFF opamps -> VR2 -5vdc 79L05A -> 74HCT4053 analog mux
+5vdc PCI Bus -> C154 -> V+ 3.3vdc AMS1117 -> CS4398 (x4) and AK5394, Vd, digital sections
+5vdc PCI Bus -> U21 TPS54352 +1.2vdc switching regulator -> C151 -> X-FI CA20K1 DSP

CN3 +12vdc (ext power) -> Polyswitch F1 -> C146 -> AD Link
(note: external power does not supply any other section on the card)

OTHER CAPACITORS:
Cirrus Logic CS4398 DAC Filter Ref (FILT+) and Voltage Reference (VREF)
C40: Positive Voltage Reference (FILT+) to GND (G-Luxon SS, 100uf @ 25vdc)
C60: Voltage Reference, internal sampling (Jamicon SS, 47uf @ 16vdc)
*CAUTION: see page 19, CS4398 datasheet before modifying values

AKM AK 5394: Offset Calibration Capacitors: (VREFL +/-/R/L)
C47, C51, C7, C4: Offset Calibration Capacitors: (VREFL +/-/R/L) (Jamicon SS, 470uf @ 6.3vdc)
*CAUTION: see page 15, AKM AK5394 datasheet before modifying values
C48: Serial Interface Timing (SMODE) (Wincap 10uf @ 16vdc)
*CAUTION: see pages 4, 10, AKM AK5394 datasheet before modifying value

TEMPERATURES: (AMB: 72f/22c; 60mins; taken on "power-only" test fixture, no cooling fans)
U7: Line-OUT opamp: 127f/53c
U3, U5, U2: Line-IN opamps: 123f/51c
VR1: 95f/35c; VR2: 90f/32c; VR3: 95f/35c; VR4: 120f/49c; U21: 97f/36c
U4: ADC AK5395: 122f/50c

Reference Picture:
http://www.esnips.com/doc/c8b7ada7-b.../X-FI-MOD-084b

Key / Description: (in zipped TXT format)
http://www.esnips.com/doc/a397a894-6.../X-FI-MOD-085b

Add/Corrected:
a) X-FI LEDs are not powered by VR3, but supplied direct from +5vdc PCI BUS
b) added: audio signal paths
c) added: temps, power-only test fixture
d) added: DC voltages: Line-Out coupling-caps shorted

REF:
TI 78M05
http://www.ti.com/lit/gpn/ua78m05
TI 79L05
http://www.ti.com/lit/gpn/mc79l05
TI TPS54352
http://www.ti.com/lit/gpn/tps54352
AMS AM1117-3.3
http://www.advanced-monolithic.com/pdf/ds1117.pdf
Cirrus Logic CS4398
http://www.cirrus.com/en/pubs/proDat.../CS4398_F1.pdf
ASAHI KASEI AK5394
http://www.asahi-kasei.co.jp/akm/en/...5394a_f03e.pdf
Fairchild LCX245 (74LCX245)
http://www.fairchildsemi.com/ds/74/74LCX245.pdf
post #1405 of 2194
Quote:
Originally Posted by germanium View Post
"Someting is drawing too much current or it is too close to other heat generators. <xxxx> You could try asking bichi if he has any ideas as i'm running low on them myself..."
Germanium and Ter1,

- refer to above post for details

1) Given Line-OUT coupling capacitors have a 2.5vdc BIAS @ 1.55ma, shorting four will cause additional 6.2ma load on the VR1 regulator. Shorting all 16 Line-Out coupling caps would add 24ma. Each of the four CS4398's nominally consume 25ma, totaling 100ma. The 78M05 is rated at MAX 500ma and nominally would be loaded at under 125ma to remain "cool to the touch," given alotted PCB space. Also note measured temperatures of regulators.

2) The Polyswitch F1 only protects the supply connected to the external CN3 plug, +12vdc (ext power). It goes directly to
the AD plug. It does NOT supply power to any sections of the card itself and is not causing the card to "shutdown."

I'm still not sure which regulator is getting "hot," so maybe use the reference picture in post above and indicate by "VR" number?
post #1406 of 2194
Quote:
Originally Posted by bichi View Post
Germanium and Ter1,

- refer to above post for details

1) Given Line-OUT coupling capacitors have a 2.5vdc BIAS @ 1.55ma, shorting four will cause additional 6.2ma load on the VR1 regulator. Shorting all 16 Line-Out coupling caps would add 24ma. Each of the four CS4398's nominally consume 25ma, totaling 100ma. The 78M05 is rated at MAX 500ma and nominally would be loaded at under 125ma to remain "cool to the touch," given alotted PCB space. Also note measured temperatures of regulators.

2) The Polyswitch F1 only protects the supply connected to the external CN3 plug, +12vdc (ext power). It goes directly to
the AD plug. It does NOT supply power to any sections of the card itself and is not causing the card to "shutdown."

I'm still not sure which regulator is getting "hot," so maybe use the reference picture in post above and indicate by "VR" number?
bichi, Germanium, thanks for your detailed information!!

1. For my card, it's VR4 that is the hottest. I didn't notice VR1's temp since it's under my replaced caps of C55, C48 and C52. I shorted all the 16 caps plus the 2 line-in caps, I don't know how the shorting of line-in caps add current to VR4?

2. For C151, I changed it to BG 16v3300uf one, can you confirm whether this change could cause the shutdown? I don't know the card parameters of in TPS54352 datasheet, so I don't know whether 3300uf is too big for the TPS54352.

3. Besides the above change, I also changed the value of the following caps:
a) c104 and c133 to 50v1000uf
b) c47, c51, c7, c4 to 6.3v1000uf
Are the above value change making sense or good for sound quality? Can this cause self-protection? I don't think so for latter.

4. Besides the above changes, I changed all the caps on board to the same value better electrolyte ones. I know you'll suggest use different kinds of caps in digital and analog path, but that's too many words to be covered in this thread if you don't like to talk more on that!~

5. I changed all the 7 opamps to LME49860, I think the change draws more current to VR1 and VR4 that caused the heat/overcurrent protection?

6. I bypassed all the 4 c40 with 400v6.8uf SOLEN. I bypassed all the 4 c60 with 400v3.3uf SOLEN. I bypassed C104 and C133 and C151 with 400v22uf SOLEN. I think the bypass with cause more current through VR1, right? Will cause more current in VR4? Do you think the bypasses reasonable to sound quality?

7. Since you confirmed the board will not use the external power that just the breakout box will use it. Then do you suggest I disconnect the CN3 in order to make less psu noise in the card? My psu is not good enough and noisy.


If the self-shutdown after 30min didn't caused by the 3300uf BG DSP cap mod and did caused by the heat of VR4 or VR1 as per your examination, how to do with the 78M05? Can I change both/specific1 to another higher power ones? Because if I use a fan to blow the card, in the past 7 days, not shutdown, but if I remove the fan, shutdown in 30min!



Thanks much!!

Also,
1. Do you know what's the function of the 20-pin CI under U19?

2. For the 2 diodes above and to the up left corner of PolySwitch F1, do you know what're the functions for the 2 diodes? I occasionally removed the 2 and shorted them, no malfunctions so far. Anyway, I'll change back the 2 diodes when I can get the component.

In all, I like so much of the sound produced by the mods. Just the self-protection and the noisy added fan caused me mad, I really need your help, thanks again!
post #1407 of 2194
It seems to me after a little research, The X-Music's have the same DAC's and opamps as the Audigy 2 ZS? can anybody elaborate?
post #1408 of 2194
Ter1,

1) Main goal is to solve your "shutdown" problem first, sound quality disscussion later.
2) What is your ambient room and internal case temperatures? (should be Russian winter there, right?)
3) Does the blue X-FI LED stay lit when your card is in "shutdown" mode? (powered by VR3)
4) PSU make and model?
5) Solder type and flux type? (SnPb or SnAgCu with rosen flux)
6) Assumes your added capacitors are not defective and do not have "high leakage" problem.
7) Post a picture of your card installed in your PC?

1. For my card, it's VR4 that is the hottest. I didn't notice VR1's temp since it's under my replaced caps of C55, C48 and C52. I shorted all the 16 caps plus the 2 line-in caps, I don't know how the shorting of line-in caps add current to VR4?

ANS: Remove all coupling-cap shorts until "shutdown" problem is solved. VR1 supplies CS4398 (x4) +5vdc analog sections, which couple to Line-OUT opamps. Shorting will cause CS4398 output to "sink" additional 1.55ma per capacitor. Also note C48 controls AK5395 SMODE mode and should be left at original value. VR4? VR4 supplies AK5394 analog section....

2. For C151, I changed it to BG 16v3300uf one, can you confirm whether this change could cause the shutdown? I don't know the card parameters of in TPS54352 datasheet, so I don't know whether 3300uf is too big for the TPS54352.

ANS: It is not C151 value, but the capacitor ESR type, where "ultra-low-ESR" capacitors are not optimal. BG 3300uf should be OK here, unless it is defective and has high-leakage.

3. Besides the above change, I also changed the value of the following caps:
a) c104 and c133 to 50v1000uf
b) c47, c51, c7, c4 to 6.3v1000uf
Are the above value change making sense or good for sound quality? Can this cause self-protection? I don't think so for latter.

ANS:
a) C104 and C133 filter +/-12vdc from PCI bus and 1000uf should be OK.
b) C47, C51, C7, C4 control AK5394 offset calibration and should be returned to original values until "shutdown" problem is solved.
note: C48 controls AK5395 SMODE mode and should be left at original value.

4. Besides the above changes, I changed all the caps on board to the same value better electrolyte ones. I know you'll suggest use different kinds of caps in digital and analog path, but that's too many words to be covered in this thread if you don't like to talk more on that!~

ANS: Agreed. Solve "shutdown" problem first, then discuss "sound quality." Verify all electrolytics are installed with correct polarity.

5. I changed all the 7 opamps to LME49860, I think the change draws more current to VR1 and VR4 that caused the heat/overcurrent protection?

ANS: NJM2114=9ma, NJM2068=5ma, x5532=9ma, LME49860=10.5ma (typical) All the opamps draw current directly from +/-12vdc PCI bus, no regulator involved and should be OK, assuming new opamps are not defective.

6. I bypassed all the 4 c40 with 400v6.8uf SOLEN. I bypassed all the 4 c60 with 400v3.3uf SOLEN. I bypassed C104 and C133 and C151 with 400v22uf SOLEN. I think the bypass with cause more current through VR1, right? Will cause more current in VR4? Do you think the bypasses reasonable to sound quality?

ANS:
- C40 and C60 (and rest of purple dotted caps) control CS4398's FILT+ and VRef. Not sure if bypassing all 7 caps has changed CS4398 current demand. I would disconnect all the added bypass caps in this section.
- C104 and C133 are +/-12vdc filters from PCI bus and bypass caps should be OK.
- C151 is the +1.2vdc power capacitor from TPS54352 switching regulator and bypass cap would not add much to sound quality.

7. Since you confirmed the board will not use the external power that just the breakout box will use it. Then do you suggest I disconnect the CN3 in order to make less psu noise in the card? My psu is not good enough and noisy.

ANS: If you do not use AD Link, then disconnect external power plug until thermal shutdown problem is solved. Hard to say if noise from "noisy" PSU will get injected when AD-Link is not in use. If AD-Link is connected, then yes, excessive PSU noise might leak in via breakout box.

If the self-shutdown after 30min didn't caused by the 3300uf BG DSP cap mod and did caused by the heat of VR4 or VR1 as per your examination, how to do with the 78M05? Can I change both/specific1 to another higher power ones? Because if I use a fan to blow the card, in the past 7 days, not shutdown, but if I remove the fan, shutdown in 30min!

ANS: There are higher current 3-pin regulators, but the limit is "thermal dissipation" when soldered to PCB. (PCB acts as heatsink) Secondly, PCB trace current carrying capability is another limit. The regulators are tricky to remove, especially 78M05's.

OTHER:
1) Try installing card in another PC, to eliminate PC related problem.
2) You mentioned "3hr" recovery after "shutdown." Normally, 3-pin regulators recover within seconds, after removal of short or excessive load. There is a remote possiblity of a "hairline" PCB trace fracture, marginal solder bridge or cold solder joint, which may have longer thermal "recovery" latency.

Also,
1. Do you know what's the function of the 20-pin CI under U19?

ANS:
U20 = PCI bus data interface device. Fairchild LCX245 (74LCX245)
http://www.fairchildsemi.com/ds/74/74LCX245.pdf

2. For the 2 diodes above and to the up left corner of PolySwitch F1, do you know what're the functions for the 2 diodes? I occasionally removed the 2 and shorted them, no malfunctions so far. Anyway, I'll change back the 2 diodes when I can get the component.

ANS:
NEVER, NEVER SHORT DIODES!!
- if the diode is a rectifier, zener or clamp, it can short power supplies
- if the diode is a "snubber" it will cause short to control circuit (like D1, D10, D7, D9, snubber for relays on the X-FI Elite)
- if the diode is a small signal detector, it can cause amplifier stages to overload
- if the diode is a "logic buffer," it can cause logic device to behave erratically or latch

D8 and D11 are related to Xilinx XC9536XL. (complex programmable logic) It's +3.3vdc is supplied by VR3. Since this is programmed by Creative, it's specific function is unknown. Remove shorts NOW!

In all, I like so much of the sound produced by the mods. Just the self-protection and the noisy added fan caused me mad, I really need your help, thanks again!
post #1409 of 2194
Finally got Germanium's card operating again...
- started with Germanium accidentally shorting something near C15 and D9.
- repaired burnt trace near Flexijack/Line-OUT/Mic-IN.
- replaced SW2 relay, open contacts.
- replaced R35, failed "open." (+12vdc, 4.7ohm current limiting resistor for U3, U5 and U2)
- replaced U5 and U2, NJM5532.
- installed missing L12 ferrite bead.
- installed missing U19 with NJM5532. (stock = NJM2068)
- replaced AK5394 ADC. (blown audio-in channel)
- replaced VR4, 78M05 regulator. (blew during test after almost completing repairs)

Components installed for test:
- stock NJM2114 opamp for Line-OUT.
- replacement 47uf @ 16vdc, Line-OUT coupling capacitors, C45, C35, C58, C49. (originals not sent)
- stock NJM5532 Line-IN opamp. (original not sent)
- replacement 22uf @ 16vdc, Line-IN coupling capacitors, C27, C15. (originals not sent)

Other Modifications:
- U3 and U7 socketed.
- C104, C133: Panasonic FC, 220uf @ 25vdc from G-Luxon SM, 220uf @ 25vdc. (+/-12vdc from PCI bus)
- C154: Panasonic FC, 220uf @ 25vdc from Jamicon SS, 22uf @ 16vdc. (+5vdc from PCI bus)
- C151: Panasonic FK, SMT, 1500uf @ 16vdc from Jamicon WL, 220uf @ 16vdc. (+1.2vdc for CA20K1)
- various power de-coupling/bypass capacitors replaced with tantalum at stock values.

Off to snag dinner at neighbors, if anything is left.......

Picture:
http://www.esnips.com/doc/d7cc9964-e.../X-FI-MOD-086b

RMAA v6.0.6: (96/24 and 48/24)
DETAILED:
http://www.esnips.com/doc/b3c83c05-f.../X-FI-MOD-087b
post #1410 of 2194
Phone conversation with Germanium, discussing function of C29 and C14:
- C29, C14: U3 Line-IN opamp coupling capacitors to U2, U5 differential opamps. (Jamicon 47uf @ 16vdc)
- BIAS: Differential coupling caps, C29, C14: 2.53vdc @ 1.45ma, measured across capacitor.

In the process of tracing circuit, discoverd C14 measured 26.3uf and is out of spec for a 47uf.
- socketed C29 and C14.
- replaced both C29 and C14 with Panasonic SM, 47uf @ 16vdc.
- corrected Left-Right frequency response difference, between 0~200hz.

RMAA v6.0.6 test results with replaced differential coupling caps: (96/24 and 48/24)
SUMMARY: http://www.esnips.com/doc/cf594ffe-0.../X-FI-MOD-089b
DETAIL: http://www.esnips.com/doc/b45011d4-3.../X-FI-MOD-088b
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)