Op-amp Rolling Using Sparkos on Fosi V3 Mono

[olieb]

Measurements are important in helping us identify which products are engineered and designed well. Just don't tell me they all sound the same cause the measure the same.

If there is a difference between two electrical signals at the output this can be
1. one signal is stronger (level), this has to be taken care of by level matching
2. one signal has a different FR - this will be in the measurements
3. there is a difference in noise - this will be in the measurements
4. there is a difference in nonlinear distortion - this will be in the measurements

If all these differences are small enough [and well below audibility] then there is just no place for a signal sounding different but in your brain (from bias).
Yes, the science of electrical measurements is that good. And listening with proper control will show it.

 

[BobPM]

What we need in this hobby / profession is a cite dedicated to doing blind listening tests. I know it is easier said than done, but measurements have driven amp and DAC design to ultra low distortion levels, and is now debunking most tweaks. The only response has been, but I still hear a difference. I would love to be able to turn to a website with golden ears doing blind tests. Wouldn't it be great is industry funded such an effort. I could actually see some of the chinese companies doing it to overcome the inherent bias of many reviewers against inexpensive products.

 

[martinus]

I have to admit, I did change the original op-amp for a sparkle op-amp in the same way as in the review. I did change op-amp 2. And indeed I did not hear any difference. So the question is, when is it usefull to swapp the op-ampt ?

 

[JeremyFife]

I have to admit, I did change the original op-amp for a sparkle op-amp in the same way as in the review. I did change op-amp 2. And indeed I did not hear any difference. So the question is, when is it usefull to swapp the op-ampt ?

Based on the many threads and measurements on Opamp swapping - it is never useful to swap them. Just don't bother.

(Slight caveats if you are talking about something like a phono stage, or if you are re-designing the entire circuit)

 

[solderdude]

So the question is, when is it usefull to swapp the op-ampt ?

When it is broken (which in practice never happens) and even then a $ 1.- opamp is already more than good enough in this amp.

 

[amirm]

I would love to be able to turn to a website with golden ears doing blind tests.

The results of such tests will be dismissed as listeners being deaf, not having high enough resolution, etc.

 

[oleg87]

I have to admit, I did change the original op-amp for a sparkle op-amp in the same way as in the review. I did change op-amp 2. And indeed I did not hear any difference. So the question is, when is it usefull to swapp the op-ampt ?

Essentially, if you know what you're doing better than the designer (or aren't subject to the pressure to cost-cut to design to the bone), and understand the circuit it's going in.
Otherwise, the risk of creating problems is much higher than the likelihood of improving anything IMV.

 

[KL....]

Just to indicate that I know what I am talking about when I speak about op-amps. This is my current workbench:

View attachment 441591

I always write based on knowledge, design experience and production/testing experience. I do not attend "debates for debates" as a kind of chat, as a mere debater. A kind of respect would be appreciated.

@pma, thank you, respect and appreciation for your efforts and you are seeking the Truth of the matter (or at least that is my observation), aren't you?

To ask, if a device measures differently or (even) a little differently, at which frequency (or group of frequencys) do you match volume? This is certainly a conundrum, isn't it? why? They (may) sound they same at the matched frequency/s but what about the other (unmatched) frequency/s?

Also, FFT even though it is not an interpolation method, it can be used as a tool for interpolation, particularly in the frequency domain, can't it? Are the number of frequency points (or bins) used, for the frequency range (20-20K), generally 32, 36, 40 then a straight line drawn between them, to ask? I should also ask @amirm, shouldn't I, to ask?

 

[solderdude]

To ask, if a device measures differently or (even) a little differently, at which frequency (or group of frequencys) do you match volume? This is certainly a conundrum, isn't it? why? They (may) sound they same at the matched frequency/s but what about the other (unmatched) frequency/s?

That would only be applicable to transducers and acoustical measurements.
Electronics are usually dead flat between 100Hz and 10kHz so matching at any of those frequencies will be fine as long as the used meter is suited for the used frequency.
Matching should be done within 0.1dB = 1% accuracy.
Also note that it should be done for both channels. Sometimes volume controls can have channel imbalance over 0.1dB.

When measuring amplifiers under actual speaker load and the amplifiers have an unusual high output resistance opposite the other one it becomes a bit more complex.

 

[KL....]

Electronics are usually dead flat between 100Hz and 10kHz so matching at any of those frequencies will be fine as long as the used meter is suited for the used frequency.

Thank you, but if they are not matched below 100Hz, wouldn't the resulting (bass frequency/s) decay change the presentation, between 100Hz and 10kHz, even though (initially) matched between 100Hz and 10kHz?

 

[Scrivs]

Sure...

I think the goal of ASR is to post the technical side such as (basic) measurements and discuss these, but not the math and electronic design stuff for nerds/experts, so readers could benefit. Usually that does get discussed in the thread in some depth.

You can't win them all so I see no reason not to explain things anymore because a few of them don't (want to) get it.
You also can't please designers and electronic engineers.
You can't do a full suite of measurements under all conditions (including overload) in all situations either.

So reviews are limited in what they show.
Techie guys would prefer more measurements, others want more 'subjective babble', others want features explained/measured more or want different measurements.
It is what it is.

My point was never a question of the balance of content- It's just where it needs to be-something for everyone. The answer to almost any question vis a vis audio, or otherwise is usually minutes away via the search window.

Many people sign up on ASR, because they WANT TO BE TOLD WHAT TO BUY. That's ok, I guess There's rarely an attempt to at least gather information before the inevitable: "What Dac is the best?" Or some other lame, lazy braindead species of the same question.

Solderdude, I cinsider you one of the O.G. techs, and value you thoughts and contributuons on here. I've quietly read an learned a lot through your post and value your opinions.

This is what I love about ASR

--Apologies for the sidetrack

 

[totti1965]

@amirm - Allow me one last attempt to recognize any measurable differences (and to play the „advocatus diaboli“):

What is the second harmonic (k2, at 2 kHz) with the SPARKOS SS3602 in this picture?
You can't see it through the two graphs on top of each other!

SPARKOS SS3602 much better here?

 

[antcollinet]

Thank you, but if they are not matched below 100Hz, wouldn't the resulting (bass frequency/s) decay change the presentation, between 100Hz and 10kHz, even though (initially) matched between 100Hz and 10kHz?

then you are talking about an audibly non flat frequency response. Creating a genuine audible difference - also measurable.


But in this case, look at the frequency response for the two op amps. Absolutely identical.

 

[NTK]

@amirm - Allow me one last attempt to recognize any measurable differences (and to play the „advocatus diaboli“):

What is the second harmonic (k2, at 2 kHz) with the SPARKOS SS3602 in this picture?
You can't see it through the two graphs on top of each other!

SPARKOS SS3602 much better here?

View attachment 441935

Use this graph instead.

 

[Jimster480]

Close enough to where my money would be comfortably on sighted bias.
Even if it's plausible a very distortion-sensitive listener could discern them in a blind test, these would by no means be "clear" differences.

No need for audiologist visits, hearing illusory differences is the normal behavior of human hearing. That's why we go on and on about controls.

Of course it would be sighted bias with performance numbers like this. To me the best thing is that the sparkos OPAmps don't reduce quality in any way.

I'm here because I have an interest in the measurements.
What i have an issue with is being told that because the measurements are the same or similar, all dacs sound the same.
If that was the case why isn't everyone here using the same $100 dac that is perfectly transparent.
Why isnt everyone using the Topping L30ii Headphone Amp that is nearly the highest measuring amp on the chart and can put out 3w, enough to drive almost any pair of headphones on the planet.

Measurements are important in helping us identify which products are engineered and designed well. Just don't tell me they all sound the same cause the measure the same.

I can tell you I don't use those things because I like the features of the higher end stuff. I like having a remote, BT input (and several inputs that can be switched quickly). On my D70S I use both XLR and RCA outputs (also a feature of a higher end DAC). At one point I wanted MQA support as well because I had Tidal and lots of stuff on there was MQA.
I use a THX AAA 789 as my main amp and a Geshelli Archel 3 Pro as my 2nd amp... I have 3 headphones on my desk, so I can pick whichever one I want and listen to it without messing with the cables.
I'm sure I could use the cheapest setup possible and it would still work.... but it would be inconvenient to me. My other desk has a Topping EX5 and I have one computer on USB and the other on Coax and it has auto switching. I have both powered speakers & headphones, and I can select which one I want even using a remote. That is a reason to spend more money. I would also lose the tone controls of the Geshelli, which is why I bought it. I like being able to adjust bass & treble to my liking on the fly.

 

[DualTriode]

Are we splitting Fine Red Frog Hair?
or
Have we lost our way?

Can our human ears discern the difference between Op-Amps. The answer is almost never. This allows for only a few exceptions. Fill in the blank.

If you look at any ASR amplifier review there is a Eye Chart that ranks every amplifier by dB's SINAD. The current crop of amplifiers has improved as a result of the ASR ranking system.

95dB and above is Excellent
94dB to 61dB is Very Good / Good

80dB is Average

60dB and below is Poor

Over all Audio Quality correlates only to a small degree with SINAD. Sure you may discern a audio Quality difference between 120 dB SINAD and 58 dB SINAD.

Pavlov's human subject may be killed by Experimental Neurosis attempting to discern a difference between a SINAD of 110 dB's and 95 dB'S.

SINAD alone is a poor tool to rank Human perceived audio quality. For an example 2nd and 3rd HD's at even fairly high levels do not influence perceived audio quality. Even fairly low levels of higher order HD's 7, 8, 9 ..have a much higher influence on perceived audio quality.

Audio manufactures, you know the people that make cell phones, smart speakers, powered monitor speakers and automobile sound sound systems know and use all this (PEAQ measurements) Sound Quality stuff designing their products.

When was the last time you saw SINAD measurements for a power amplifier internal to a powered monitor speaker. Not even at ASR.

The same audio manufactures sell op-amp rolling as a sales gimmick for audiophiles.

Thanks DT

Check this out:

npr.org/sections/therecord/2015/06/02/411473508/how-well-can-you-hear-audio-quality

 

[amirm]

Please take generic rants about measurements to dedicated thread.

 

[solderdude]

Thank you, but if they are not matched below 100Hz, wouldn't the resulting (bass frequency/s) decay change the presentation, between 100Hz and 10kHz, even though (initially) matched between 100Hz and 10kHz?

Well, it could be that some amplifiers (not the op-amps they are by nature all flat from DC) have a coupling cap and may start to roll-off below 40Hz or 10Hz or so.
No one will ever level match on such low frequencies so it is moot.
Nor will someone ever level match at 10kHz, 20kHz or 100kHz anyway.
This needs to be done with a continuous tone so even if there was some inexplicable difference in decay it would not matter.
When you want to level match using a simple voltmeter (multimeter with ACV setting) then frequencies between 100Hz and 1kHz are fine for level matching.
Only the cheapest ones with the lowest ACV settings being 200VAC or so are not really suited for this task.
Amplifiers should be matched with a representative load connected to it.

 

[pma]

This needs to be done with a continuous tone so even if there was some inexplicable difference in decay it would not matter.
When you want to level match using a simple voltmeter (multimeter with ACV setting) then frequencies between 100Hz and 1kHz are fine for level matching.
Only the cheapest ones with the lowest ACV settings being 200VAC or so are not really suited for this task.
Amplifiers should be matched with a representative load connected to it.

Comparing 2 power amplifiers and matching the level, you do not consider interactions of power amp output with speakers, right? Because the single tone method would not tell you.

 

[amirm]

SPARKOS SS3602 much better here?

No. It is hard to make these FFT overlays be revealing. But here is a better shot:

I consider all of these differences negligible and subject to run to run variations.

The idea behind this FFT was to show broadband view which is quite similar and no oscillation is seen.