So after watching the Hifinow DAC thread, I thought to do a little op amp rolling of my own. Care package arrived from RS: TI LME49720 and LM4562, and BB OPA2134.
I listened to several hours of alternately Black Sabbath and Imogen Heap (if you can do old skool rock and very highly produced female vocals...)
Difference is very very slight -- I tried to do an AB-X "Double Stupid" test (I wrote my subjective feelings on a list, and then blindly wrote new ones when I changed, assuming my kak memory would not remember which op amp was which).
I felt there was a definite improvement to clarity and focus going to the LME49720. Unfortunately I could not back to back with the original NE5532 as I very, very carefully let all the smoke out of it.
LME 49720 - bought lots of detail and timbre, with slightly subdued but very crisp base. certain clearing up of the high end - the original 5532 was a little buzzy in character.
LM4562 - had a great midrange, almost bursting at you. Bass was more pronounced but slightly fat.
OPA2134 - extraordinary clarity and focus ... almost disconcerting focus in the sound stage, very very revealing. Almost too much so, had a feeling there was a slightly metallic element.
Now this is obviously very subjective -- and swapping an op amp over takes a good 20-30 seconds so plenty of time for the brain to play tricks.
Looking at the specs, the new op amps are hgher spec than the 5532 -- slew rate of 9V/us compared to 20+. Cant find THD+N graphs for 5532, but it's vanishingly small anyway. Interestingly the (TI) Burr Brown distortion rises quite quickly ? from nothing bits of a nothingness to 0.001% at 10k and rising. But basically looking at thousands of a percent.
The OPA2134 has higher input noise than the other two, and worse CMRR and PSRR (-100dB vs -120dB), but excellent channel separation (135 vs 12).
So interesting that an op amp with seemingly not amazing specs (OPA2134) is so distinctly different in quality from a cheapie (NE5532).
I?m leaving the OPA2134 for a couple weeks (days) to get used to it, and then roll them again.
LME49720
THD+N .0.00003 (.00009 limit) -- almost constant to 00005 at10k
Input Noise .34uV
Noise density 6.4-2.7nV/sqHz
slew rate: 20V/us
CMRR 120dB
PSRR 120dB
Vos .1mV
Crosstalk-120dB
Burr Brown OPA2134
THD+N .00008 (heads rapidly to 0.001at 10kHz)
Input Noise 1.2uV
Noise density 8nV/sqHz
slew rate: 20V/us
CMRR 100dB (hits -60--80 at1kHz)
PSRR 106dB
Vos .5mV
Channel separation 135dB
TI LM4562
THD+N .0.00003 (.00009 limit)
Input Noise .34uV
Noise density 6.4-2.7nV/sqHz
slew rate: 20V/us
CMRR 120dB
PSRR 120dB (105 by 10k)
Vos .1mV
Crosstalk -120dB
TI NE5532P
THD+N .0.00003 (.00009 limit)
Input Noise .34uV
Noise density 8-5nV/sqHz
slew rate: 9V/us
CMRR 100dB
PSRR 100dB
Vos .5mV
Crosstalk 110dB
I listened to several hours of alternately Black Sabbath and Imogen Heap (if you can do old skool rock and very highly produced female vocals...)
Difference is very very slight -- I tried to do an AB-X "Double Stupid" test (I wrote my subjective feelings on a list, and then blindly wrote new ones when I changed, assuming my kak memory would not remember which op amp was which).
I felt there was a definite improvement to clarity and focus going to the LME49720. Unfortunately I could not back to back with the original NE5532 as I very, very carefully let all the smoke out of it.
LME 49720 - bought lots of detail and timbre, with slightly subdued but very crisp base. certain clearing up of the high end - the original 5532 was a little buzzy in character.
LM4562 - had a great midrange, almost bursting at you. Bass was more pronounced but slightly fat.
OPA2134 - extraordinary clarity and focus ... almost disconcerting focus in the sound stage, very very revealing. Almost too much so, had a feeling there was a slightly metallic element.
Now this is obviously very subjective -- and swapping an op amp over takes a good 20-30 seconds so plenty of time for the brain to play tricks.
Looking at the specs, the new op amps are hgher spec than the 5532 -- slew rate of 9V/us compared to 20+. Cant find THD+N graphs for 5532, but it's vanishingly small anyway. Interestingly the (TI) Burr Brown distortion rises quite quickly ? from nothing bits of a nothingness to 0.001% at 10k and rising. But basically looking at thousands of a percent.
The OPA2134 has higher input noise than the other two, and worse CMRR and PSRR (-100dB vs -120dB), but excellent channel separation (135 vs 12).
So interesting that an op amp with seemingly not amazing specs (OPA2134) is so distinctly different in quality from a cheapie (NE5532).
I?m leaving the OPA2134 for a couple weeks (days) to get used to it, and then roll them again.
LME49720
THD+N .0.00003 (.00009 limit) -- almost constant to 00005 at10k
Input Noise .34uV
Noise density 6.4-2.7nV/sqHz
slew rate: 20V/us
CMRR 120dB
PSRR 120dB
Vos .1mV
Crosstalk-120dB
Burr Brown OPA2134
THD+N .00008 (heads rapidly to 0.001at 10kHz)
Input Noise 1.2uV
Noise density 8nV/sqHz
slew rate: 20V/us
CMRR 100dB (hits -60--80 at1kHz)
PSRR 106dB
Vos .5mV
Channel separation 135dB
TI LM4562
THD+N .0.00003 (.00009 limit)
Input Noise .34uV
Noise density 6.4-2.7nV/sqHz
slew rate: 20V/us
CMRR 120dB
PSRR 120dB (105 by 10k)
Vos .1mV
Crosstalk -120dB
TI NE5532P
THD+N .0.00003 (.00009 limit)
Input Noise .34uV
Noise density 8-5nV/sqHz
slew rate: 9V/us
CMRR 100dB
PSRR 100dB
Vos .5mV
Crosstalk 110dB
