OpAmp replacement on PC soundcard (and maybe the ClockGen.)

[Foxer]

N.B - I did not read the whole thread; I saw enough of the usual "this is all audiphoolery" and that the OP doesn't feel his questions have been answered.

...
There has been a trend in the past few years of modding sound cards by replacing the OpAmps, capacitors, and even the quartz clock generator with an ultra-precision TCXO 1ppm (or even 0.3ppm) of the same frequency as the original (in this case 24.576Mhz). And all of this was said to have VERY good audible results.

Generally speaking, jitter - or cycle by cycle phase skew - of the clock is more important than absolute frequency accuracy, especially for delta-sigma (aka "1 bit") DACs. Fortunately, most crystal oscillators have exceptionally low jitter; unfortunately, most DACs get their master clock via a fractional-N PLL, which tends to dramatically increase the jitter. Thus changing the crystal oscillator is unlikely to have any effect whatsoever on sound fidelity.

See Maxim App Note 5477 for more on clock jitter.

Problem is, after a while some of the cards started experiencing problems.. ranging from crackling noises in the speakers, to just plain dying....

Crackling noises are usually from vibration causing "piezoelectricity" in high-K ceramic dielectric capacitors, particularly in local IC bypass and coupling applications.

What i'm mainly interested in, is changing the 4 original dual OpAmps (1x NJM2114 for front + 3x NJM2068 for surround), with 4x OPA1602AID dual OpAmps. Some also suggested to put a 0.1uF ceramic capacitor on the V+/V- of each of the 4 opamps as a little backup power or something, as it would help normalise the sound.

It has been my experience - and I am speaking as a bit of an audiophile (who at age 43 can still hear up to 16.5kHz) - that replacing really crappy op-amps like, say, an LM358, will improve sound quality (lower distortion, better power supply rejection ratio, less phase shift w/ frequency due to higher gain-bandwidth product, etc.) but once you are at the level of an NE5532 or an LM833 there isn't much benefit to be had (exception - if driving headphones directly then an op-amp with higher than usual output current capability helps - I like STMicro's TSH22).

The OPA1604 is better on paper in most respects than either of the two NJM op-amps, but I honestly doubt you'll notice the difference. One aspect of the OPA1604 that isn't usually good for audio applications is its rail-to-rail output - this output configuration tends to have higher distortion as the signal level approaches either rail (of course, if the signal is at or exceeds the rail it will be clipped, and that results in huge amounts of odd order harmonics that sound downright harsh).

Bottom line: I'd leave the NJM op-amps in.

...
Voltage-wise the OPA1602AID seems ok ..i'm thinking nothing on that sound card is going to go above 12 volts, so i'm well within limits.. but more than that, i have no clue. So please, i really need some help on weather or not this swap us fully compatible and is not going to kill my card after some time.

The OPA1604 is a fine op-amp and will certainly not be the limiting factor in how well a soundcard sounds (even though it is a rail-to-rail type) but as I already said I don't think you'll actually notice an improvement. Sadly, the things that most affect sound quality in a soundcard are difficult or impossible to change such as board layout, power supply noise/ripple and the filtering/decoupling thereof, and EMI/RFI from all that digital crap, both on the soundcard itself and inside the computer.

As for capacitors, it always helps to make sure all aluminum electrolytic types are good quality Panasonic, United Chemicon or Nichicon brands intended for switching regulators (look for "high ripple current" and "low impedance" attributes), and replacing any high-K ceramic surface mount capacitors with X7R types rated for at least 2x the voltage they will experience, where possible, will reduce "piezoelectric microphonics" as well as phase shift errors from capacitance change due to temperature and bias/signal voltage (only a problem with coupling capacitors, not those used for bypass/decoupling).

Thank you MagicSmoker for the good info!