The best sounding audio integrated opamps

[Andrea:]

I think you're talking by hyperboles, as usual


BTW, how funny. Whenever from time to time I go back to listening to my CD player (Cambridge Audio D300SE, tweaked), I keep being pleasantly surprised by its warm, luminous, soothing musicality...

I'm starting to question the musicality of these new and "super" DAC chips. My CD player has a CS4338, a tiny 8-pin 24/96 DAC which "only" has 92 dB of S/N ratio and -88 dB of THD. Oh and on its output (hidden by a metallic shielding cover) there must've always been a 5532.

[Andrea:]

I think I'd enjoy trying that NG98 with the AD1852, as I remember this chip to (funnily since it's not even the same brand) produce a similar sound to my CDP...still a little more bland, but, possibly, more similar than my 2 DACs equipped with CS4398. Then I can play with the opamps to tailor the sound further

That or the Poppulse...curious about the BB DAC chips.

[Salas]

JE990 that they use in studios, anyone?

[Andrea:]

Originally Posted by Salas

JE990 that they use in studios, anyone?

Huh? Google says "discrete opamp"..


Anyway... I like this too

Welcome to Audiophilechina

And it costs just 119 euro shipped. I like the PCM1702 + OPA134 = 123479, which promises a good combination with my amp's JRC 4570.

Yepp... the numbers are the key, as usual. My CDP has CS4338 + NE5532, a lot of 3's and 5's, i.e. what for my ears are the most 'music friendly numbers'.

[Andrea:]

Anyway my Super Pro has CS4398 + OPA1611 = 134689, 6 different numbers, and once again a good pairing with the JRC 4570. And it hears, in the form of a very neutral and impartial sound.

[Salas]

Originally Posted by Andrea:

Huh? Google says "discrete opamp"..

Yes, Burson concept is popular enough, isn't it?

[Mooly]

Some real world examples of substituting opamps without optimising anything. The circuit is a simple inverting configuration gain of 1 buffer with 100k feedback and input resistors. The circuit is on a proper PCB.
The output was at 2 volts peak to peak

Hover cursor over each for details

[Mooly]

The NE5532 with compensation, again 30 khz and 2 volts pk pk using first 3pf then 5pf and finally 8pf compensation cap.

[Mooly]

Now at 20khz and 4 volts pk pk, first no compensation, then 3pf and then 5 pf. Note the sharp non linearity now with 5pf and a higher output voltage.

This is a detail of the leading edge.

Notice how only 2pf extra leads to that non linearity.

[Mooly]

So I hope you can see now why just swapping devices without any measurement or basic optimisation is a little misleading,

and to finish guess what this is, first shot uncompensated, second with 3pf