Objective Guide to Op-Amp Rolling - Part 1

Jump to Latest
#2
:up:
Nice 😉

... so why pay more than the 1 usd for the LM833? 🙂

Maybe I would include 2 more parameters in the comparison, noise performance and CMRR ...

.. and it will be interesting to see the possible differences in more realistic loads than the 100k

Thanks for sharing 🙂
 
#4
Distortion measurements with hardly any load aren't particularly illuminating(*) - I'd suggest using the "standard" 2k load, and also 500 ohms (some opamps will not handle that so well).


NE5532 is worth adding as a baseline, and perhaps some of the "rail to rail" opamps like the OPA1562 (low power consumption, JFET, looks pretty capable).




(*) with very light loading the output stage of the opamp may be in class A for instance.
 
#6
If I could also request test of:

OPA1642 ~ 2 Euro
OPA1656 ~ 2.5 Euro
OP275 ~ 3 Euro
OPA1612 ~ 6 Euro

According to specs I would think the OPA1642 would be one of the best compromises cost vs performance wise. (Normally one would probably point to LM4562 as being the best compromise, but I think the Quiescent Current of 10mA is very high)
OPA1656 has even lover distortion.
OPA1612 too expensive
OP275 ... an older type ... the only one I think I have heard sounding a bit different (more sweet) .... but I'm really doubting that I or anyone would be able to tell the differnce between any of these opamps in contolled ABX test 😉

Great work and thanks for sharing
KR Baldin
 
Last edited:
#10
orchardaudio said:
Next time I order something from Digikey / mouser I will get a NE5532.
Click to expand...
Make sure you DO NOT order the T.I. NE5532---it has been shown to have worse distortion and sound than the others. Best would be to get an original Signetics or Philips, as those are the ones that won its reputation with all the pro audio console manufacturers.
It also would be nice to use one of Rod Elliott's ideas:
Sound Impairment Monitor (SIM) - Is This The Answer?
and see how they measure up. Basically it is an arrangement that subtracts the input from the output of a device, leaving only whatever artifacts that the device introduces. It seems that this would be a good attempt to quantify an IC's "sound signature".
 
Last edited:
#11
I think it’s important to include TI’s version of the NE5532. It’s the most accessible one on the market, so I should think it’s the one DIYers will settle upon if the genuine Signetics are more difficult to find.

Would be nice to see the OPA1692 compared against the others.
 
#14
00940 said:
It really would be a shame to limit yourself to DIP opamps. TI has done quite a lot of work on their audio opamps lineup in the last years but the newer opamps only comes in smd packages.

Frankly, the opa134 line has become irrelevant in my experience since the release of the opa1641 line.
Click to expand...

If somebody here wants to send me these SMT opamps all ready to go in DIP sockets converters I will test them.
 
#15
Hello,
In 2010, in the french version of Elektor, Douglas Self described a 64 X NE5532
power amplifier.
In this paper, he also made some measurements on this OP to know which brand he should use.
Between TI, ON Semi & Fairchild, he found that the best one was the Fairchild.
Maybe you can still find the english version of this paper in your archives ??
 
#16
If somebody here wants to send me these SMT opamps all ready to go in DIP sockets converters I will test them.
Click to expand...
I can offer you some pcb adapters I designed, just pay for the shipping, PM me if interested. You can order your own parts and solder them up. I'll even throw in some of the fine no-clean solder I use since buying a roll is a life times supply in my case. If you want me to solder & supply the parts then you are into more $ since I do not usually do this for others.

Cheers
Rick
 
Last edited:
#17
alayn91 said:
Hello,

In this paper, he also made some measurements on this OP to know which brand he should use.
Between TI, ON Semi & Fairchild, he found that the best one was the Fairchild.
Maybe you can still find the english version of this paper in your archives ??
Click to expand...
If D Self's "The Design of Active Crossovers" (2nd Ed), chapter 16 ("Opamps for Active Crossovers"), there's a section titled "Which 5532" which discusses this briefly.

Using an inverting unity gain circuit (shunt feedback) with 2k resistors he measured the TI chips has having higher 20kHz distortion (0.0008%--0.0012% range as opposed to <= 0.0007% for other manufacturers). Below 10kHz all devices were at/below the measurement floor for an Audio Precision SYS-2702.

In the ultrasonic range the discrepancy between TI and other versions increases, being "3 times worse at 30kHz"...

You may or may not worry about a rise from 0.0007% to 0.0012% at 20kHz, its hardly a show-stopper for audio and most people only quote the 1kHz value anyway...


Above 10kHz no harmonics are audible, so it only going to be relevant for intermodulation effects.
 
Last edited:
#18
alayn91 said:
Between TI, ON Semi & Fairchild, he found that the best one was the Fairchild.
Click to expand...
But Fairchild is owned by ON. AFAIK, ON and TI are the only current manufacturers of NE5532s; and only TI still offer them in PDIP-8.
I have also heard horror stories about clients being very upset after Philips/Signetics NE5532s being replaced by TIs, and made the tech re-do the repairs. And then there is also the scary story about a tech having to replace over 800 chips in an SSL console because SSL had used Raytheon chips instead of Signetics. Both of those clients insisted that they could hear the difference.
 
Last edited:
You must log in or register to reply here.
Top Bottom