The best sounding op amps for audio work.

[catman]

G'day all, I guess that this sounds suspiciously like the subject of 'op amp rolling, but actually it isn't.  As much as I think that op amps do have a 'sound', in recent days I've been revisiting the LM4562 and sonically wise, to my ears it doesn't really have that 'typical' op amp signature, it just sounds clean, quiet and with particularly good and strong bass. 

 

Curiously enough it seems to sound its best in combination with the OPA2134.  In some respects I think of the LM4562 as a somewhat upgraded NE5532.  I have used the OPA2134 as the first gain stage in a number of high quality DIY  phono stages, with the LM4562 in the second overall gain stage.  I have done this in both the ESP P06 and Muffsy DIY  phono stages.  Doing this is also reputed to improve the overall S/N ratio, however to my ears it sounds particularly nice and pleasant!  A bit of a mystery yes but a very audible one....to my ears anyway.  

Edited March 1, 2021 by catman
Punctuation.

[roger777]

Thanks for the information.  I confess I wasn't even aware of the existence of LM4562 op-amps.  I've been using one of those Jaycar kit phono pre-amps that uses an LM833. The RIAA gain circuit is clearly straight off the LM833 datasheet. I found there to be too much distortion and to fix it I increased the ground resistor in the RIAA gain circuit from 390 ohm to 1k ohm.  The disadvantage of this was is a 7dB reduction in gain at 1kHz. It seems the equalisation is insignificantly affected.

I think the problem is that the open loop gain of LM833 may be too low above 1kHz, per the spec sheet dropping from 82dB at 1kHz to 62db at 10kHz. While the RIAA gain circuit only demands 32dB at 1kHz, I wonder whether somehow there is insufficient open-loop gain and whether the LM833 I bought is not performing to spec.

The LM4562 datasheet indicates it has at least 12dB greater open loop gain than LM833 at all audio frequencies.

I've been researching and designing possible solutions around LM833 op-amps. I note that the LM833 and LM4562 datasheets contain identical RIAA phono circuits in their application notes. I conclude that I should obtain an LM4562 and drop it into the socket once I have reinstated a 390 ohm resistor per the original circuit design.

I assume that to be more assured that I am obtaining the real McCoy, I should purchase a branded IC rather than one of those unbranded IC's from China, which I guess could actually be anything.  Do you have any thoughts on purchasing an LM4562?

Edited March 1, 2021 by roger777
Correction of open loop gain comparison.

[catman]

G'day mate, I've built that Jaycar kit with no issues at all.  Interesting!  I think that the LM4562 is an incredibly impressive dual audio op amp.  Try it out!  Regards, Felix.    

[mbz]

NJM2043D for phone stage.
LM4562 if you are wanting accuracy, nice tones
OPA2134 if you want more body/timbre, nice tones.
Choice depends on the equipment (signature) of the gear you are
connecting to.

 

Not a huge fan of jcay but then again it's about learning and having some fun.

2c...

[mbz]

2 hours ago, roger777 said:

Do you have any thoughts on purchasing an LM4562?

Buy from a reliable source though I understand shipping to NZ is an issue. Preferred sources are mouser and digikey in the US, typically free shipping to AU on orders over $60-. RS stocks the LM4562 but they want you to buy a minimum of 5, https://au.rs-online.com/web/p/op-amps/1977299

I wouldn't buy semi's from jcay, certainly not off the bay.

[stereo coffee]

Yes the LM4562 dual is excellent for audio use. I have a pair of these in a marantz CD52SEmk2 , and it is by far the best device to my ears, in that CD player. Other  op amps, I regularly use are opa2107.  Discrete circuits where you can implement them using jfets ,  are perhaps the best for audio. 

 

The LM833 is still a good op amp. I think National Semi got its design correct,   it even has its own thread :

 

 

[betocool]

Just because of measurements I made, I prefer the 5532 over the 2134, that being said, if you pointed a gun at my head and asked me which one is "better" I'd probably give you a random answer.

 

But, on all my projects, from headphone amps to phono stages, the 5532 is my choice. I use it on my ESP Phono and I can't complain. Detailed and a nice punchy bass.

 

Cheers,

 

Alberto

[roger777]

On 01/03/2021 at 9:36 PM, mbz said:

RS stocks the LM4562 but they want you to buy a minimum of 5,

Thanks for your advice Mike. RS also has a dedicated NZ site. Apparently I could obtain them from there cheaply and rapidly. Unfortunately for me I had already rushed ahead and used "the Bay" to order some National Semiconductor branded LM4562NA/NOPB units from an outfit in Florida at significantly greater cost.  With partial hindsight, I am now concerned about them being National Semiconductor branded whereas I assume contemporary units should be branded TI.

[Ian F]

The new breed of TI J-Fet opamps offer superior performance to that old soldier the OPA2134.The OPA827 for example has superior load drive,much lower distortion and miniscule DC offset.Close to zero DC mv at its output would enable removal of the dreaded electrolytic coupling caps which will definitely lower colouration and improve sound quality. 

Edited March 12, 2021 by Ian F
update

[Guest Mr Thorens]

I’m a bit underwhelmed by the LM833. I find the TI version of NE5532 better at revealing detail and timbre and with better stereo imaging. 
 

OPA2134 has good tone and detail when in single mono mode ie with two op amps one per channel. When used as a single dual stereo op amp I find the soundstage a bit flat and a thick sound. 
 

LM4562 is another level altogether. Agree with previous post, fast, detailed, revealing and neutral. Just have to watch that it doesn’t oscillate in circuits not designed for such a high speed unit. 

[Ian F]

The LM833 is a bit of a shocker performance wise.In Samuel Groner's mammoth test of opamps he found the LM833 had 2-3% distortion at higher frequencies and the heavy loading typical of phono preamp feedback networks.

Edited March 12, 2021 by Ian F

[muon*]

I subjectively preferred the sound of the 2604 to the 2134, but preferred the Burson hybrid V5i over both, except maybe the 2604 had a nicer mid range, but over all the V5i was preferable.

 

I don't like the 5532 or the 833.

 

Edith: I tossed my 5532's but think there are a pair of 833's in the parts box for trouble shooting, and gave away the 2604's and 2134's.

Currently making a 2SK389V jfet output stage so if that works out well I'll sell the V5i's as they have a little value compared to the others that are a few bucks to 10 bucks each.

Edited March 12, 2021 by muon*

[Ian F]

In my experience any audible difference between well set up opamps is fairly subtle.There's a series of uploads on YouTube featuring most of the popular opamps.The sound quality is quite reasonable unlike most YouTube audio demos.Even using quality headphones it's a challenge to pick any significant differences let alone the night and day improvements some folk are proposing..

 

Edited March 12, 2021 by Ian F

[muon*]

It's debatable IMO, as every person is different and the perception of sound through the processing of individual brains can differ from one to another.

 

Not every system or circuit is the same either.

 

I never hold much cred in recorded evaluations as that is another process that can either heighten or lessen the ability of the subjects performance in discerning the differences.

 

Just my personal views

[Steffen]

52 minutes ago, Ian F said:

In my experience any audible difference between well set up opamps is fairly subtle.There's a series of uploads on YouTube featuring most of the popular opamps.Even using quality headphones it's a challenge to pick any significant differences.

 

 

I agree with the sentiment (unless you pick an opamp unsuited for the job), but sound comparisons via Youtube video have got to be the most pointless thing people have come up with in recent times. 

[Galileo77]

In my personal experience, I find differences between op amps dependent on the listening device.

Testing couple of models on a PC with Asus STX and:

- $150 speakers - no difference almost at all, unless you focus hard to catch it

- HD600 - could definitely hear, as they're bright, quite easy to catch in upper registers in particular

- Focal Clear - differences were significant to the point that I could almost immediately tell that op amps are changed after listening to 30s of practically any track that I knew well. Also, some op amps started to become irritating to listen to and didn't last longer than an hour or so.

 

Out of all (reasonably priced) op amps that I tried, personal favourite is OPA2107, but I'm keen to test either Burson V6 or SparkOS soon

 

[andyr]

On 12/03/2021 at 2:05 PM, Ian F said:

 

The new breed of TI J-Fet opamps offer superior performance to that old soldier the OPA2134.  The OPA827 for example has superior load drive, much lower distortion and miniscule DC offset.

 

Close to zero DC mv at its output would enable removal of the dreaded electrolytic coupling caps which will definitely lower colouration and improve sound quality. 

 

 

While close-to-zero DC offset is a great thing ... who on earth uses electrolytics as signal coupling caps???

 

Andy

 

[Ian F]

27 minutes ago, andyr said:

 

While close-to-zero DC offset is a great thing ... who on earth uses electrolytics as signal coupling caps???

 

Andy

 

Just about any main stream manufacturer of affordable gear like Marantz.Their amps are full  of 'em albeit nicer sounding ones like Silmic.

[andyr]

10 hours ago, Ian F said:

 

Just about any mainstream manufacturer of affordable gear like Marantz.  Their amps are full  of 'em albeit nicer sounding ones like Silmic.

 

 

So all of those electrolytics we can see in the pic are coupling caps?  WOW!  

 

Andy

 

[Ian F]

As an example the Marantz PM6003 has six coupling electros in the phono stage alone including the feedback cap.A 220uf film cap aint gonna fit!  Using supposedly superior film caps is just another urban myth. I've tried most of them including the megabuck variety.Unfortunately ALL  caps have colouration just  different flavoured varieties thereof.A simple straight wire bypass test easily verifies that.

Edited March 19, 2021 by Ian F

[catman]

G'day all, this has been an interesting thread, so thank you.  On electrolytic capacitors I find that I used to be very anti electrolytic capacitors, and yes they do have some limitations and they do need to be configured correctly in a circuit, but having said that a couple of my best sounding phono stages actually have electrolytic capacitors in signal coupling parts of the circuit.  Interesting indeed!  Regards, Felix.   

[andyr]

2 hours ago, Ian F said:

 

As an example the Marantz PM6003 has six coupling electros in the phono stage alone including the feedback cap.A 220uf film cap aint gonna fit!

 

 

 

 

Perhaps you might be kind enough to identify these 6 coupling caps in the Marantz phono stage, Ian (as I can only see 4 caps that are in the signal path - and therefore are 'coupling'; C117 & C118, and C105 & C16 ... so 2 in each channel?

 

There are, similarly, 2 coupling caps in each channel, in my own Muse phono stage - though of much less value (given, approximately, by the equation: Desired Roll-off Frequ = 160,000 / (Zin * uF):

• 15nF between the 1st gain stage and the 2nd gain stage
• and between 1uF and 3.3uF on the output - depending on what the Zin of the linestage the phono is going to be feeding.

 

2 hours ago, Ian F said:

 

.ALL  caps have colouration just  different flavoured varieties thereof.  A simple straight wire bypass test easily verifies that.

 

 

I agree with you completely!    However, the words I would use are "all caps in series degrade the signal - easily verified by a simple straight wire bypass test; some caps degrade it less than others".  

 

Andy

 

Edited March 19, 2021 by andyr

[Ian F]

Andy you love splitting hairs don't you? There are eight electrolytics in the PM6003 stereo phono preamp including two PS decouplers C121,C122.Your count has omitted caps C109 and C110 -the feedback coupling caps-just as critical if not more so and clearly audible when an upgraded cap such as Silmic or Blackgate is used.As to your question "who on earth uses electrolytics?"Answer-just about everybody unfortunately-there's little alternative in most SS circuits short of going down the DC servo path which has its own set of compromises .In my experience simpler circuity with a minimum of coupling caps brings improved SQ with greater immediacy and dynamic expression plus long term listening satisfaction.

Edited March 19, 2021 by Ian F

[andyr]

Thank you, Ian.  

 

Andy

 

[Addicted to music]

On 18/03/2021 at 9:50 PM, andyr said:

 

While close-to-zero DC offset is a great thing ... who on earth uses electrolytics as signal coupling caps???

 

Andy

 

Most manufacturers will, it’s to provide circuit isolation incase DC offset increases.   It’s acts like a DC isolator.   Even some amplifier manufacturers use it one on the output of there preamp and then one on the input of the power power amp.  In this arrangement if there equipment is connected to other manufacturers products they guarantee stability and reliability.   Even the Magtech has it on the input of the RCA, hence the reason I use XLR.