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 71

post #1051 of 2194
I read through the thread and saw you can't do it on a XtremeGamer, so I've found someone to trade with for an XtremeMusic. Now no one has done an in depth comparison of the LM4562 vs AD8599... So if I want a sound that is maximized for detail, not musicality (Think an anemic sound that reveals every nuance of each instrument) would I want the LM4562 or AD8599? I'm thinking the AD8599 would be best? I'll be using my LDII++ so I don't need tons of coloring in the opamp.
post #1052 of 2194
I tried to mod my card and at first it didn't work (no sound coming in on the headphones, although windows showed it was sending sound.

When I went to fix it, I tore 3 of the pads off of the card and couldn't fix it.

Be forewarned, if you are not good with soldering very small electronics (or have the right tools) you can ruin your hardware.
post #1053 of 2194
Deleted
post #1054 of 2194
shorted the line in coupling caps as well as the other mods & got these results in windows vista.

24 bit 96Khz
Frequency response +.01,-.04
Noise level -116.2
Dynamic Range -116.0
THD .0007%
IMD+noise .0010%
Crosstalk -111.2
IMD+noise (swept freq).0008%

Very very good results I would say. Noise floor lowered a bit in windows vista when I shorted the line in coupling caps & distortion went down slightly. Stereo seperation improved as well. Mine is totally clean on the noise test, Only one low frequency spike no higher than -124db Otherwise all noise is below -130db. All noise above 200Hz is below-140db & at 3KHz is -151.43!!!!!
post #1055 of 2194
Quote:
Originally Posted by ter1 View Post


Thank you for your very detailed info!! I'm now want to have a summary w/ picture....

1. The red-dotted caps need to be shorted.
2. The blue-dotted op is for line-in, need to be replace with LME49860.
3. The 4 caps with blue "100" be bypassed with 2 3.3uf metalized film caps in parallel.
4. The 4 cpas with "47" be bypassed with 1 3.3uf metalized film cap.
5. The blue-line 4 caps can be replaced to 1000uf ones for ADC.

Are the above right? If so, I think lots of modder will follow...

but is the 220uf power supply cap as you stated "I also bypassed the 220uf power supply cap with1-3.3uf metalized film cap" the Jamicon near the DSP?

Also, I have some questions about your mod,
1. Do you suggest to replace the power caps (both "100" and "47") with more capacitance BG ones? If with more capacitance ones, the capacitance for film caps need to be changed too, right? --------or--------------
Just replace them with the same capacitance 100uf/25v and 47uf/16v BG ones with your suggested bypassed caps,
which will be the better choice?

2. After your mod, how about the RMAA frequency response line for 96kHz? Still sharp drop after 20kHz?

3. I'm thinking the caps of 4,5,6,7,8,9,10,11 are for the ADC, what's your suggestions of them? Replace with better BG ones? How about the capacitances? Bigger?

4. What's the functions of cap 1,2 and 3? What's your suggestions?

5. What metalized film caps are you using? Suggest more better ones?

I know it's quite a lot of questions, but we all want to make elite pro even better.... thx...


It's better you take a shot of your mod!~~
The above description is correct except I did not change the 470uf caps to 1000 but did read in the design sheet that that could be done to get very slightly better low frequency distortion. I also only did the front channels so I only bypassed the 100uf closest to the outside of the board & the second 47uf cap in the back row. There is not enough room to put 12 of these mettalized film caps that it would take to do all channels. I only use stereo anyway.


1. No do not change those caps size as that is what Cirrus logic calls for as far as the 100uf one, the design sheet calls for 33uf where the 47uf cap is but I doubt that that one is critical but 100uf one is probably critical.

2 excellent response to 43KHz. -.12db. The recording setting setting settings must have been set wrong. That is why other people were getting the sound chopped at 20KHz. Set correctly it should go all the way to 44KHz with almost no rolloff.

3. cap 4&5 are coupling caps & can be shorted. The others I'm pretty sure are powersupply. I would leave these at their stock size & maybe use BG caps. Shorting 4&5 improved performance in several areas with no drawbacks.

4.caps 1&2 are +&-12volt powersupply for the analog output opamps. I'm not sure of th number 3 cap.

5.Didn't see a brand name but they are polyester 3.3@100volt. Higher voltage ones may sound better but take up much more room & would stress the solder connections.

no pics as Headfi won't accept my pics as they are to large.
post #1056 of 2194
Quote:
Originally Posted by Gautama View Post
I read through the thread and saw you can't do it on a XtremeGamer, so I've found someone to trade with for an XtremeMusic. Now no one has done an in depth comparison of the LM4562 vs AD8599... So if I want a sound that is maximized for detail, not musicality (Think an anemic sound that reveals every nuance of each instrument) would I want the LM4562 or AD8599? I'm thinking the AD8599 would be best? I'll be using my LDII++ so I don't need tons of coloring in the opamp.
I've used both and even though I'm not going to do an in-depth review about opamps right now, I can tell you that neither of them sound anemic, which would be a lack of bass response and 'gusto'/impact, as far as i know. Both opamps do just fine in that department. So if your amp is too bassy then neither of the opamps will change that. In my experience with the two, the AD8599 has a bit more detail and separation to it than the LM4562.
Hope that helps.
post #1057 of 2194
Quote:
Originally Posted by phalanger View Post
I've used both and even though I'm not going to do an in-depth review about opamps right now, I can tell you that neither of them sound anemic, which would be a lack of bass response and 'gusto'/impact, as far as i know. Both opamps do just fine in that department. So if your amp is too bassy then neither of the opamps will change that. In my experience with the two, the AD8599 has a bit more detail and separation to it than the LM4562.
Hope that helps.
By anemic I meant bland and a "flat" sound. So good bass was wanted

I'll probably grab the AD8599.
post #1058 of 2194
try em both if you can though swapping opamps too often increases risk of problems ofcourse. You can also use both if youre mostly after 2 channel sound and have a 7.1 cable, thats what i did. It comes out a stereo plug for both of them after that cable. Then with the mixer panel you can route where the channels are mapped. If you have to go for one of them then yeah the AD is probably your best bet. It has amazing clarity and detail.
post #1059 of 2194
Quote:
Originally Posted by germanium View Post
The above description is correct except I did not change the 470uf caps to 1000 but did read in the design sheet that that could be done to get very slightly better low frequency distortion. I also only did the front channels so I only bypassed the 100uf closest to the outside of the board & the second 47uf cap in the back row. There is not enough room to put 12 of these mettalized film caps that it would take to do all channels. I only use stereo anyway.


1. No do not change those caps size as that is what Cirrus logic calls for as far as the 100uf one, the design sheet calls for 33uf where the 47uf cap is but I doubt that that one is critical but 100uf one is probably critical.

2 excellent response to 43KHz. -.12db. The recording setting setting settings must have been set wrong. That is why other people were getting the sound chopped at 20KHz. Set correctly it should go all the way to 44KHz with almost no rolloff.

3. cap 4&5 are coupling caps & can be shorted. The others I'm pretty sure are powersupply. I would leave these at their stock size & maybe use BG caps. Shorting 4&5 improved performance in several areas with no drawbacks.

4.caps 1&2 are +&-12volt powersupply for the analog output opamps. I'm not sure of th number 3 cap.

5.Didn't see a brand name but they are polyester 3.3@100volt. Higher voltage ones may sound better but take up much more room & would stress the solder connections.

no pics as Headfi won't accept my pics as they are to large.

Thanks so much for your detailed info!!

but is the 220uf power supply cap as you stated "I also bypassed the 220uf power supply cap with1-3.3uf metalized film cap" is cap 1 and 2 in the picture or the 220uf Jamicon near the DSP?

If it's 1&2, do I need to bypass both with 1 polyester 3.3@100v for each?

How about the main powersupply cap near DSP? Is there a need to bypass with 1 polyester 3.3@100v also?
post #1060 of 2194
Quote:
Originally Posted by ter1 View Post
Thanks so much for your detailed info!!

but is the 220uf power supply cap as you stated "I also bypassed the 220uf power supply cap with1-3.3uf metalized film cap" is cap 1 and 2 in the picture or the 220uf Jamicon near the DSP?

If it's 1&2, do I need to bypass both with 1 polyester 3.3@100v for each?

How about the main powersupply cap near DSP? Is there a need to bypass with 1 polyester 3.3@100v also?
Yes the one near the DSP. I found that the combined total if you also bypass the ones near the opamp makes it seem a little too bright.
post #1061 of 2194
Shorting the caps used to provide enough current for the opamp to function properly is ridiculous.

It makes absolutely no sense at all from a design view. And they aren't in the signal path, so there goes that theory.

Provide real information to support your claims instead of "i think it sounds better"

EDIT:

Also, if those capacitors actually were coupling capacitors, wouldn't shorting them let the DC signal pass along with the AC signal result in very bad results?

Also, think about it, if the capacitors only do harm to the sound quality, and do no good for the circuit, why would creative put them there? Why would a company which is notorious for being cheap, unnecessarily increase its product costs?
post #1062 of 2194
X-FI OP AMP, CAP MOD - SOLDERING PASTE METHOD:
- a bit wordy, but after a bit of practice, much easier to execute than with "wire" solder

Link to pictures:
http://www.esnips.com/web/X-FIOpAmpandCapModification

PREP:
- secure the board to prevent movement while work is in progress
(ie., PCB vice or masking-tape board to kitchen type cutting board)
(imagine trying to solder, hold parts with board skating across desk...)
(admit it, even experienced pukes have done this..)
- eye-loupe or magnifying glass
- sleep
- no triple-java-double-pump vente coffee or 5 cans of Jolt cola
- proper tools (no hardware store wire-cutters, pliers, prying with screwdrivers)
- practice on scrap, if needed

STANDARD Sn/Pb Leaded Solder vs. Lead-Free SnAgCu (ROHS)
- melting point of "Lead-Free" solder is higher, about 218c/424f, as opposed to Sn/Pb, 183c/362f, depending on brand
- as of July 1, 2006, lead was to be eliminated from electronics manufacturing


OP AMP REMOVAL:
Clipping leads of NJM4556 preferred over heating joints and "pulling" part off board. (can cause solder pads to be pulled off)
- carefully clip NJM4556 leads with appropriate "small tip" cutter and remove body
- heat each remaining lead and "lift/slide" off with tip of soldering iron or small tipped tool (ie., fine tipped tweezer, toothpick)
- remove remaining solder with solder wick or "vacuum-pump" based desolder tool
- clean pad area with acetone, denatured alcohol or other suitable "flux remover"

INSTALLING NEW OP AMP:
- place small amount of solder paste on each solder pad on board with fine tipped tool
- carefully place new op amp in correct orientation/position (dot indicates pin 1)
- wipe soldering iron tip to remove any residual solder (wet sponge preferred method) (repeat this, before each lead or set of leads)
- hold new op amp in place with fine tipped tool and "touch" clean soldering iron tip to one joint
- solder paste should go molten very quickly (less than 1 sec)
- remember to wipe tip before next lead!
- check for correct lead/pad position before soldering remaining leads
- (soldering one lead will hold the part in correct position while soldering remaining leads)
- (soldering one lead, makes for easy removal and repositioning, if op amp shifted out of correct position)
- clean soldered area with acetone, denatured alcohol or other suitable "flux remover"
- inspect solder joints with eye-loupe for solder bridges, gaps and other soldering defects

CAPACITOR REMOVAL:
- as stated in this thread, removal of the 220uf filter capacitor requires different technique
- mass of PCB ground plane acts as a "heatsink" and takes a bit of time for joint to "soften" enough for safe removal
- pulling the capacitor lead, if solder joint is not "soft/molten" state, can lead to pulling the PCB VIA out
- overheating the joint can cause copper VIA to oxidize, become brittle, intermittent or disintigrate
- techniques I use:
1) heat the joint with both soldering iron and desolder tool and when solder is "molten" state, activate pump
2) clip the capacitor out, leaving just leads and desolder, removing lead stubs
3) if capacitor is close to PCB, then "cut capacitor body" until only the leads are left, then desolder, removing lead stubs
4) "heat" and push-out with pin is effective, but attention to "solder softness" of joints, prior to "push-pull" is key
5) easy way out: clip out capacitor, leaving enough of "stub" length to solder new capacitor

Link to pictures:
http://www.esnips.com/web/X-FIOpAmpandCapModification

SOLDER PASTE:
http://www.digikey.com/scripts/DkSea...e=SMD291SNL-ND
http://webtronics.stores.yahoo.net/noclsopas29.html

SOLDERING TOOLS:
Hakko 454 25W soldering iron: http://shop2.outpost.com/product/155...H:MAIN_RSLT_PG
Hakko 808 desolder tool: http://shop2.outpost.com/product/276...H:MAIN_RSLT_PG

SIDE CUTTER:
Xcelite 170M: http://shop2.outpost.com/product/191...H:MAIN_RSLT_PG

PARTS USED:
Opamp: National Semiconductor: LME49860 - http://www.national.com/pf/LM/LME49860.html
Press Release: http://www.edn.com/pressRelease/2140057957.html
C177 Capacitor: Panasonic: FK, Low-ESR Series, SMT, 1500uf @ 16vdc (best fit part on hand)


Two RMAA v6.0.5 test results
- 16bit / 44khz and 24bit / 96khz
post #1063 of 2194
Quote:
Originally Posted by LawnGnome View Post
Shorting the caps used to provide enough current for the opamp to function properly is ridiculous.

It makes absolutely no sense at all from a design view. And they aren't in the signal path, so there goes that theory.

Provide real information to support your claims instead of "i think it sounds better"

EDIT:

Also, if those capacitors actually were coupling capacitors, wouldn't shorting them let the DC signal pass along with the AC signal result in very bad results?

Also, think about it, if the capacitors only do harm to the sound quality, and do no good for the circuit, why would creative put them there? Why would a company which is notorious for being cheap, unnecessarily increase its product costs?
I'm not shorting the power supply caps. I'm bypassing them with metalyzed film caps of relative high value. I have already posted my RMAA results to show no loss of performance. RMAA results however cannot predict actual sound. The sound from the modded card is substantially clearer than the stock card yet the frequency response measures exactly the same,THD only very very slightly worse & IMD is actually better in the modded card by a good margin.

Yes shorting the coupling caps does result in a small amount of D.C. at the output but in most situations it is of little or no consequence. It is many cases more expensive to direct couple as you have to use better active parts & do more work to ensure that the D.C. from the DACs input into the output opamps is perfectly balanced so the D.C. output is within spec. Creative has in fact produced D.C. coupled soundcards. The Audigy 2 series is a good example. In many ways the Audigy 2 sounded clearer & cleaner than the stock X-Fi cards even though technical performance was worse. D.C. coupling the X-Fi brings them up to snuff with & improves to above the Audigy 2 series in sound quality as X-Fi uses better DACs.

In just about every transistor design the power supply is in signal path as the signal almost always has to return to the powersupply when it is finished doing it's work. Precious few designs get around this with transistors.Push-pull class A1 transformer coupled tube amps the power supply is not in any way in the signal path as the total power though the output transformer is not modulated at all until you reach near max power and that is only due to the nonlinearity of the tubes at that power level. Below that they are very linear.
post #1064 of 2194
Somebiody shoud ASAP fix the first picture missing about the removing caps... It is not there and no way to get it from the cursed Imagetrash thing...
See:
http://img232.imageshack.us/my.php?i...labeledal7.jpg
post #1065 of 2194
HI, i did this mod, but i only did one opamp to start. I was about to order more opamps for the rest of the mod, when i found this: LME49720

I was wondering if it is also a drop in replacement, because it is better than the other one which is recommended in the tutorial.
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)