Sparkos SS3602 Opamp Rolling In Fosi P4

[Old_School_Brad]

Let's just rename it to Audio Pedantic Review, then I could stop reading and half-expecting to find a discussion about technology and audio (by the way, half-expecting means 50% or 0.5, or 1/2 or 2x lower, or 2^-1, for avoidance of doubt)

At the very least, if you want to rename it, the science part should be marked with asterisks if you're not using technically accurate terminology.

Is a 10p coin really 1/11 a 1 pound coin ?

1/10
0.1

Anyway, there are plenty of those "slip of the tongue" expressions that aren’t technically accurate, but people still get the gist of what’s being said.

• "Heat rises." -Not quite. It’s the hot air that rises, due to being less dense.
• "Energy is lost as heat." -Technically, energy isn’t lost it’s just converted.
• The common confusion between power and energy -like mixing up kW and kWh.

But I’ll stop myself here and pass the “pedantry torch” on to someone else..

 

[amirm]

I'm British, so I don't lessons from USAicans such as the person you originally quoted.

But the article you linked to is written by an American educated person:

"About thebettereditor
Chris holds a BA degree in history from the University of Virginia and a Master of Fine Arts (MFA) Degree in writing from the University of Southern Maine (Stonecoast). "

 

[amirm]

No, I’d actually argue that it makes the text much clearer to read, precisely because it removes ambiguity.

If you want to play that game, it goes both ways. What is two times more than X? It could be 2X and 2X+X = 3X.

In my case, I provided the value of X so no ambiguity should have been there.

 

[pma]

Oh boy, what an “issue”!

 

[Old_School_Brad]

If you want to play that game, it goes both ways. What is two times more than X? It could be 2X and 2X+X = 3X.

In my case, I provided the value of X so no ambiguity should have been there.

"What is two times more than X?" is unclear.

If you want to describe a value that's double X, you should say "two times X," "2X," or "double X." Leave out the "more than" along with "two times" -it’s not technically correct. It’s a mix of multiplication ("two times") and addition ("more than"). -Which makes it ambiguous, as you say so yourself.

That said, it’s commonly used in everyday language without much confusion. If you provide the value of X.

 

[staticV3]

Sorry for the wait.

I only analyzed the first video (ZA3 Stereo Stock vs ZA3 Stereo Sparkos).

The second video did not play the complete track twice, once Stock once with Sparkos, but instead played each track once and switched between configs every 5 seconds or so.
That makes Null analysis impossible.

Anyway, here are the results from the first video:
View attachment 442509 View attachment 442508 View attachment 442507 View attachment 442506

No malicious file manipulation from what I can tell, but definite differences above ~500Hz.
Whether these are caused by an imperfect recording setup or represent that Sparkos magic, I can't say.

In any case, the real challenge starts now:
Attached here are the phase- and level-matched recordings exported directly from Deltawave. I anonymized the files but took note which one is which.

Challenge 1:
Using ABX Comparator, can you consistently tell the two apart by sound alone? (Please attach the .txt Log file below)

Challenge 2.
If you can tell them apart, tell me which one you think is the Sparkos and I'll reveal the answer

Any updates on this? @Bartez2000

 

[dpuopolo]

NE5532 is an audiophile's worst nightmare. Something as venerable and performant can't possibly be so ubiquitous and affordable.

I've been using 5532s since 1978. They have always sounded fine to me. The newer ones have better specs then the originals.

 

[Jimster480]

NE5532 is an audiophile's worst nightmare. Something as venerable and performant can't possibly be so ubiquitous and affordable.

Some things just are "solved". Once the solution is available the price drops over time. Its just like how dongle DAC's today have better performance than most DAC's on the market 10 years ago and yet they are cheaper...

 

[French Meloman]

Thank you Amir for demonstrating through these tests that there is no such thing as an audiophile operational amplifier. The recurrence of Amir's attempts, which all point in the same direction, ends the debate. All discussions on this subject are reduced, to me, to useless chatter. "Stupidity and nonsense" than to spend one's money on expensive components that bring absolutely nothing, Molière could have said.

 

[Jimster480]

Thank you Amir for demonstrating through these tests that there is no such thing as an audiophile operational amplifier. The recurrence of Amir's attempts, which all point in the same direction, ends the debate. All discussions on this subject are reduced, to me, to useless chatter. "Stupidity and nonsense" than to spend one's money on expensive components that bring absolutely nothing, Molière could have said.

Personally, I like the newer OPA op-amp instead. I don't exactly remember the number though. I'll have to look and see.
From a technical perspective, the TI OPA has a much better slew rate and the overall paper specs are providing better performance. So if there was a concern with the 5532, this would be completely erased with the OPA model I'm thinking of. Maybe it's like a OPA1632 or something. There's also an LME one that I believe starts with 45 and that one is also better in paper specs.

I did very extensive op-amp rolling testing several years ago using a topping A30 as well as a objective two amp and in neither scenario could I detect a difference even with my most complex music test tracks at any volume level. So I left the ones with the best paper specs, because those would surely eliminate Any problem which could have arisen from a theoretical standpoint using the 5532.
Outside of those amps which had a slightly lower SINAD performance in the first place, I have NEVER found a reason to even care to check an op amp in another amplifier which already measures far beyond transparent because obviously the op amps chosen are already totally transparent.


Edit: The OP Amp is LM4562NA also NJM4562D. They are both the same from what I understood (when I tested both, they were the same in my tests).
Upon Checking the OPA1632 was indeed one of the ones I was talking about. Although LM4562 was the one I used the most when I "rolled" OP Amps.
I also used OPA2134 in rolling, but again as I mentioned 4562 was the one I settled on most of the time.

Originally how I did testing was acquiring 2x of the same Amp and rolling one and keeping the other stock, then using a switcher into my headphones so that I can instant A/B between both..... then if I heard too many differences... I would ask my wife to go do the switching for me so its blind to eliminate any possible sighted bias. Which would then eliminate all the differences typically.... I did hear differences from the TI OPA2227 to the LM4562 or the OPA1632. However these are in extreme cases only, high resolution track with high frequency fast tone music like a violin or a specific cymbal at the highest frequency; there was a slight harshness / artifact that would be able to be heard in this very specific scenario.

It is worth noting that the designer of the original O2 Amp was the biggest fan of the LM4562 or NJM4562 (a cheaper version of the same part). He also mentioned that with low gain you can't beat the NE5532.

I suggest anyone interested in exploring Op Amps read NwAVGuy's blog which has tons of great information.

Op Amp Measurements

INTRO: In last week’s Op Amps: Myths & Facts article I covered the background for op amps, why they’re important, and some common audiophi...

nwavguy.blogspot.com

 

[KSTR]

Edit: The OP Amp is LM4562NA also NJM4562D. They are both the same from what I understood (when I tested both, they were the same in my tests).

No. The NJM4562 from New Japan Radio Corp. just happens to have the same number in its name but it is a completely different older and much lesser design.

Most importantly, it isn't even unity-gain stable, rather it requires a noise gain >= 10x, otherwise it will be oscillating (or on the edge of oscillating).

 

[Jimster480]

No. The NJM4562 from New Japan Radio Corp. just happens to have the same number in its name but it is a completely different older and much lesser design.

Most importantly, it isn't even unity-gain stable, rather it requires a noise gain >= 10x, otherwise it will be oscillating (or on the edge of oscillating).

Yes, it requires higher gain. Otherwise according to NwAvGuy it was performing similar to a LM4562NA.

It just happened to be that the amp I tested it in had this much gain. However I only tested it one time in 2017.
LM4562NA is the king, I guess Fosi didn't use it since it is 6x more expensive than the NE5532.

 

[antcollinet]

Yes, it requires higher gain. Otherwise according to NwAvGuy it was performing similar to a LM4562NA.

It just happened to be that the amp I tested it in had this much gain. However I only tested it one time in 2017.
LM4562NA is the king, I guess Fosi didn't use it since it is 6x more expensive than the NE5532.

Combined with it not offering any benefit in the application.

 

[dpuopolo]

This is an example of why you can't just throw ICs into circuits Helter Skelter. The NJM 4562 offers excellent performance at a great price (34 cents in 100 quantities) in a property designed circuit. It can also become an oscillator in a circuit NOT designed to maintain a gain of ten or higher.

 

[Jimster480]

This is an example of why you can't just throw ICs into circuits Helter Skelter. The NJM 4562 offers excellent performance at a great price (34 cents in 100 quantities) in a property designed circuit. It can also become an oscillator in a circuit NOT designed to maintain a gain of ten or higher.

Of course, this is why all the Op Amps have design sheets that come with them and specifications that explain how to use them.

Combined with it not offering any benefit in the application.

It would be because they designed around the NE5532. The interesting thing is that the Fosi Page for the V3 in specific mentions Op Amps to "roll" to but I don't really see why they would suggest that considering that if they designed around the NE5532 to the point that the LM4562 (which can operate within the same electrical properties) cannot create any benefit.... the rest of the Op Amps would really be a waste of money. They must have done it for marketing or something... but even then they don't sell those Op Amps nor do they sell versions with them.... so the whole thing is just very strange to me.

I think most of the Op Amp rolling stuff came from the time prior to the release of the NE5532 and its adoption in audio products in general. If you just read through NwAvGuys blog; he talks about some of this. There were many "worse" Op Amps and newer Op Amps kept coming out that were much better than the prior ones and people would "roll" to try to upgrade their amps.

 

[dpuopolo]

Of course, this is why all the Op Amps have design sheets that come with them and specifications that explain how to use them.


It would be because they designed around the NE5532. The interesting thing is that the Fosi Page for the V3 in specific mentions Op Amps to "roll" to but I don't really see why they would suggest that considering that if they designed around the NE5532 to the point that the LM4562 (which can operate within the same electrical properties) cannot create any benefit.... the rest of the Op Amps would really be a waste of money. They must have done it for marketing or something... but even then they don't sell those Op Amps nor do they sell versions with them.... so the whole thing is just very strange to me.

I think most of the Op Amp rolling stuff came from the time prior to the release of the NE5532 and its adoption in audio products in general. If you just read through NwAvGuys blog; he talks about some of this. There were many "worse" Op Amps and newer Op Amps kept coming out that were much better than the prior ones and people would "roll" to try to upgrade their amps.

The LM4562 offers one notable advantage over the 5532 - it has 120 db of common mode rejection compared with the mid 90s for the 5532. I remember Amir's original test of the Aiyima A07 showed a 60 Hz spike about 90 db down. That would not exist had an LM4562 been put into that amplifier. In fact, replacing the 5532 with an LM4562 is an Aiyima supported upgrade.
By the way, the single version of the 5532 is the 5534, which is only stable at gains of three or greater and has both a higher GBP and slew rate. It can be made unity gain stable (With 5532 specs) by adding a single compensation capacitor.

 

[dpuopolo]

Thank you Amir for demonstrating through these tests that there is no such thing as an audiophile operational amplifier. The recurrence of Amir's attempts, which all point in the same direction, ends the debate. All discussions on this subject are reduced, to me, to useless chatter. "Stupidity and nonsense" than to spend one's money on expensive components that bring absolutely nothing, Molière could have said.

WRONG! Most class AB audio power amplifiers ARE simply discrete power opampsI Indeed the Boulder amplifiers are a Jensen 990 opamp on steroids! Look at the schematic of most power amps and compare them to the schematic of the average IC opamp. They are near identical.

 

[dpuopolo]

Yes, it requires higher gain. Otherwise according to NwAvGuy it was performing similar to a LM4562NA.

It just happened to be that the amp I tested it in had this much gain. However I only tested it one time in 2017.
LM4562NA is the king, I guess Fosi didn't use it since it is 6x more expensive than the NE5532.

Fosi uses the LME49720 opamp in their ZD3 DAC. It is virtually the same as the LM4562.

​

 

[amirm]

WRONG! Most class AB audio power amplifiers ARE simply discrete power opampsI Indeed the Boulder amplifiers are a Jensen 990 opamp on steroids! Look at the schematic of most power amps and compare them to the schematic of the average IC opamp. They are near identical.

They are? Not the last time I looked. Unless you mean at very general level. IC op-amps have many transistor constructs that are not even available in discrete form. They also use far more transistors than a standard discrete amp because they can. Any caps in there is usually very small value unlike discrete amplifiers.

Here is NE5532P:

Show me a power amp with dual collector and dual emitter transistors in them. Or so may transistor blocks. It doesn't exist.

Yes, you can say a sports car and family car are the same because they both have wheels and engines but that is not a proper comparison.

 

[kyuu]

I wouldn't say "10 times shorter," because that phrase is inherently contradictory, ambiguous and nonsensical, "times" implies multiplication, not reduction.

What? Implication of multiplication does not similarly imply the result will be a bigger number. I'm assuming we all know that multiplication can result in a smaller number, just as division can result in a larger number?