Success Rolling Opamps in EE Minimax DAC

Cheap amps like NE5532 (in my Benchmark DAC1) are so much worse than current best amps like LM4562 or OPA1612 but only if you include DC spects. For AC NE5532 is simply amazing in spite of its age.

Douglas_schroeder, Audio was ignored for a long time in op-amp design NE5532 being the first designed specifically for the audio. I'd like to try the newest OPA1612 that according to technology progress supposed to be light years ahead but I already read post claiming that in comparison its sound is too sweet and too laid back. My Benchmark DAC1 opamps are SMT but it is OK - I can replace them quickly but Benchmark's 5 years warranty speaks against it.

I finally replaced opamps in Benchmark DAC1 with OPA1612 and after initial brighter performance sounded a little mushy. There was a LOT of energy (slam) in the bass and extra extension but midrange became muddy.

Today I replaced them again with LME49860 (+/-22V version of LM4562) and sound went from mushy and bassy to detailed and clean. Voices are still free of sibilants but midrange is very transparent, soundstage expanded and cymbals sound pronounced and clean.

Both amplifiers are in 40-50MHz bandwidth range, both have THD in order of 0.00003% plus extremely low noise (OPA1612 has 1nV/SQRTHz) but sound is very different. Original amplifier NE5532 was very close to LME49860 (slightly less dynamic, clear and extended). If I could only cut them in half and attach bass portion of OPA1612 to LME49860 - well, my neighbors might disagree.

Be careful with ESR. If we talking DAC then we have most likely volatge regulators. Many voltage regulators don't tolerate low ESR very well. They tend to oscillate when ESR is too low or too high. Also supply current is more even going mostly to cover op-amps bias currents. Momentary currents are very small because of high impedances and easily covered from ceramic decoupling caps. I can see bigger problem in interaction between these ceramic caps and inductance of power supply electrolytic caps - creating parallel resonant circuit that tends to ring.

Mrtennis - OPA1612 (variation of OPA211) is what you looking for (was too warm and bassy for me) but it comes only in SMT (SO8). There are adapters and soldering of SO8 is quite easy.

National Semi amps like LM49720 or one I used LM49860 are more detailed and open/brighter sounding than NE5532. LM49720 is the same amp as LM4562 named differently for marketing. LM49860, I use, is again the same LM4562 with extended supply range to +/-22V. Some reported better sound (possibly die change) but it might be only perception.

NE5532 used to be thinner sounding until 2000/2001 when Signetics factory burned down and they stopped making it. Texas Instr. bought license from Signetics redesigned die (larger) and produced fuller sounding amps. If you have NE5532 check the logo on them. NE5532 are cheap ($.70) while OPA1612 are quite expensive ($8.50). National runs at about $4.