Big test of DIP dual op-amps for audio: LM4562, NE5532, OPA2134, TL072 and LM1458

[dreamth]

One has c- mos input the other is bipolar...normal results.

 

[torgeirs]

Yes. Wanted to lilustrate just that voltage noise thing.
And FET input is better at very high inpedance.( Current noise).
But at 2 to 10 kohm, the cheap good ones are sufficient and the distortion gain drowns inn the noise anyway.(ok, a little dogmatic)
Don’t think it is clear to everybody, hence the opamp swapping in circuits with 2 k resistors.

 

[torgeirs]

@pma. You probably already know about the stepped responce in REW.
Can show seperate harmonics and noise at down to 0.1 dB intervals. Takes forewer, but the graph is very informative.
Each messurement can get repeated and high Fft bin count so the noise pr bin gets down. Then the distortion gets through nice and clear. Example, took about 6 hours, so let it go overnight

SMSL DL200 the best ever value DAC/AMP?

If you mean whether you can control the hardware volume from the USB host, no, that option has not appeared. Yes, that was my question! Thank you for your answer!

www.audiosciencereview.com

 

[patrickjchase]

We can see that the LM1458 is simply unusable for audio, even at low output voltage and that we can assume it will have its own sound signature. Both JFET opamps, OPA2134 and TL072 have higher distortion than LM4562 and NE5532, in all 3 tests. The result of the old NE5532, that is slightly better than LM4562, is a small surprise to me, I have expected the opposite result.

Why is this a susprise? You've narrowly focused on one parameter (THD in a low-gain configuration) and ignored all of the other equally or more meaningful areas where the 4562 dominates the 5532:

• It has 5x the gain*bandwidth (which would have become apparent if you'd tested a higher-gain configuration as seen in, say, a preamplifier)
• It has a small fraction of the input offset
• If has half the noise (~2.7 nV/sqrt(hz) vs 5)
• CMRR (120 vs 100 dB)
• etc

Your "big test" here is like comparing a Camry and a McLaren based on wheelbase. Yeah, sure, they're "unexpectedly similar" if that's the only thing you care about.

If all you want to do is +6 dB gain then by all means use the 5532. It's cheaper and more than good enough for that use case. The 4562 is for more demanding applications.

If you looked at +60 dB instead (i.e. solidly in mic preamp territory) the 5532 would be noisier and down -3dB at 10kHz, while the 4562 would be flat out to 50k.

 

[BassDude]

OPA1611/12

 

[patrickjchase]

OPA1611/12

Oh wow, I was not aware of that part. Spectral noise density is 1.1 nV/sqrt(hz), so less than half as noisy as the 4562 and less than a quarter as noisy as the 5532. Also an extremely low distortion spec to match.

The "headline spec" for gain*bandwidth is only 40 MHz (vs 55 for the 4562) but when you actually dive into the datasheet it has an additional rating for 80 MHz GBP when gain >= 100. For audio you only need tens of MHz GBP when the gain is over 1000 (60+ dB), so it's got the bandwidth exactly where it needs it. The 27 V/uS slew rate spec is also about right for a modern lowish-voltage 80 MHz part (4562 is 20 V/uS, 5532 is 9 V/uS)

CMRR and PSRR are the same as the 4562 (120 dB typ, 110 dB min).

The only drawback that leaps out of the datasheet is that the open-loop gain is on the low side at 120 dB, but that's a secondary concern and should be plenty for a audio.

I've already added a few 1612s to my next Digikey cart to play with.