The Opamp thread

[Rroff]

The input - should be a single dual opamp - is the gain stage and generally has the most impact on sound - the output stage should be 2x dual opamps. If you literally just want clear and balanced sound then LM4562 for the input and 2x RC4580 for the output would give that but musically probably a bit cold.

The OPA132 probably has the most impact in the configuration above though the combination probably comes from both as the OPA2111KP IIRC is often described as a bit tubey.

 

[raoultrifan]

These are not originals at all, these have been added after the modification for sure. O2 has, in the original design, the very low-noise NJM2068 (also used now in the extremely low-noise JDS Atom) was used in the input/gain stage and the buffers were NJM4556 paralleled.

I strongly suggest you to revert to the original opamps, at least on the output buffers that are probably oscillating and overheating, not mentioning that OPA211KP are able to deliver +/-10mA vs. the max. 150mA delivered by the NJM4556.

However, if you'll never use planars or hard to drive cans with your O2, you might install MUSES8920 if you really don't like the original NJM4556 opamps, although...besides a lower output DC of 1mV vs. 3mV I wasn't able to find other highlights.

 

[Rroff]

Yeah was wondering about that with the OPA2111 only able to supply 10ma - was hoping it was actually in the gain stage as terrible choice for O2 buffers.

The NJM4556 are pretty decent if you have headphones that need a lot of current and work well in that configuration but I'm not as convinced by them for overall use especially with high impedance, low current headphones - hence I prefer the RC4580 as a straight replacement if sticking with the spirit of the original amp.

Personally my O2 variant at the moment is running an OPA1692 gain stage and OPA2209 buffers but that isn't very DIY friendly.

OP275 is also quite good with this setup.

 

[raoultrifan]

The input - should be a single dual opamp - is the gain stage and generally has the most impact on sound - the output stage should be 2x dual opamps. If you literally just want clear and balanced sound then LM4562 for the input and 2x RC4580 for the output would give that but musically probably a bit cold.
[...]

Interesting opamp: http://www.ti.com/lit/ds/symlink/rc4580.pdf, thanks for sharing @Rroff. I don't remember of anyone testing this, and I think I had the entire diyaudio forum on the O2 thread, but definitely I missed one page for sure. Happen to know if someone tested this under a scope?

I'll purchase a couple of these to test them, thanks again.

 

[wolkegeist]

The amp is highly modified by the guy I know, same schematic but different pcb design. Here’s the picture

 

[Rroff]

Interesting opamp: http://www.ti.com/lit/ds/symlink/rc4580.pdf, thanks for sharing @Rroff. I don't remember of anyone testing this, and I think I had the entire diyaudio forum on the O2 thread, but definitely I missed one page for sure. Happen to know if someone tested this under a scope?

I'll purchase a couple of these to test them, thanks again.

As per the datasheet it is a drop in replacement for the NJM4580 and very similar to the NJM4556 but slightly less current capability but with a few tweaks by TI that in my opinion are an improvement on the original and it is designed around audio use - it is largely transparent possibly a touch raw sounding - it is no detail monster though sufficient but I've heard more revealing opamps.

 

[raoultrifan]

It has a lower output current, but I'll give it a try sometimes for sure.

 

[Rroff]

That is very modified from the original and the part numbers aren't clear enough on the opamps to know what the setup is so I'm not 100% sure on the best recommendations. The original design had a bank of 480uf caps in the power stage and 220uf feeding the opamps for a reason.

 

[wolkegeist]

That is very modified from the original and the part numbers aren't clear enough on the opamps to know what the setup is so I'm not 100% sure on the best recommendations. The original design had a bank of 480uf caps in the power stage and 220uf feeding the opamps for a reason.

It’s for desk use only, so it’s wall wart powered, but I also have a different battery designed for it, by the guy I bought it from. I need two single op amps to replace the stock OPA132s tho. I think they are the reason for it to sound tubey. The stock buffer chips are JRC4556s but I replace them with OPA2111KP because I only have low impedance headphones like Grado SR80 and Etymotic ER4S. It’s very tubey with the 2111s and a little. V-shaped with the 4556 and the OPA132s. Coupling caps are Sprague 680uF 25V. The two op amps near the stereo jacks are dual OPA132, the op amps beside the big orange caps are two OPA2111KP

 

[carlmart]

But D50 is already measuring about perfect? Why fell like changing anything inside this DAC?

Who said "measuring perfect" equals or is the same as "sounding very good"? There's no perfect in subjective listening.

Only inexperienced or stubborn electronics engineers would say that. Pick someone like Walt Jung, Bob Cordell or Nelson Pass, who are electronics engineers that actually LISTEN to audio equipment, and you may get a much better opinion on what an opamp or a power supply can do to audio quality,

You may want to change something when you think it may be improved, in this case anything on the audio quality. Or do you think an opamp doesn't have an audio signature?

Why increasing the voltage on the opamps will change anything in D50?

Because specs will improve when you get closer to the maximum voltage an opamp was designed for.

 

[raoultrifan]

An audio gear that is measuring very well in regard with several audio aspects is all it counts for me, because subjective listenings are mostly based on audio memory, which I don't trust, or incorrect A/B tests most of the times (it should be less than 1ms latency between swapping inputs sources which many times is not happening). However, I really appreciate a correct A/B test when I see it and I do trust these reviewers, especially if there are measurements to backup their assertions involved! Meanwhile, if I'll ever want to change the sound coming from my cans I'll just easily grab a DSP plugin, because I find a DSP plugin being cheaper than a new amp or DAC and with lot more flexibility in choosing the output sound.

The above engineers are picking up opamps inside their circuits based on the datasheet and audio measurements, not based on their ears, although odd harmonics do sound more fuller and better to our ears and some manufacturers are indeed "dialling-in" more odd harmonics by adjusting the bias-level or negative feedback, just to make the device to sound nicer, despite the increased THD. But indeed, listening is the most important part, so this is why less THD, less noise, less IMD, low mains hum, perfect freq. response and no phase delay means better audio performance...in all cases, no matter our subjective listening tests. After all, amplifiers are designed by engineers and not by musicians, even that both parties are involved in the final listening.

I'm not EE, nor a musician, I'm just an objective audiophile that realises that measurements are the first we need to read when picking up audio gear, and later the listening part, because different listening conditions will result in different opinions. A good reading about choosing audio gear could be found here: https://benchmarkmedia.com/blogs/application_notes/124929287-how-should-i-buy-an-audio-system. Worth mentioning that my "audiogram" shows that I can hear -85dBFS in a silent room with closed-back cans, I can distinguish a volume difference of less than 0.5dB and I do hear harmonics if worse than -70dB (THD of 0.03%) on speakers and only if worse than -60dB (a THD of only 0.1%) on headphones.

What I'm looking myself when listening for the first time to a new device would be a black background, lack of mains hum noise, no scratchy potentiometer, no audible distortions. Later, I look for a good match between the amp and my cans, because sometimes this is not happening due to different specs and needs (different designs, technical limitations etc.).

BTW, specs should be related to the measurements after all, and increasing the voltage inside D50 will not lower its harmonics below the existing -115dB (https://www.audiosciencereview.com/...ments/topping-d50-dac-measurements-png.19810/), and if it will, I'm not sure how will you be measuring this. Also, harmonics below -90dB are kinda impossible to hear.

 

[carlmart]

I don't discard measurements, either from the datasheets or made for the tests. But I do think they are far from enough.

If eventually I had to choose between a review mostly based on measurements, like those from amirm from www.audiosciencereview.com, and another done by a serious reviewer, respected because of previous tests, like some people from Stereophile, I'd definitely go with the latter. You would apparently do the opposite, and that's a personal choice.

Those engineers I mentioned take datasheets as starting pointers, but they do extensive listening. The question is that there are many things that the human ear preceives that aren't or can't be quantified, for now at least.

Serious engineers that do reviews many times devise ways to measure what the ear perceives and the instrument do not show. Walt Jung is one of them.

You should read old reviews made in old Audio Electronics issues or in Linear Audio, and you will find many comments about how things sound, not just measure.

My thing is speakers, not headphones. But I did use Beyer DT48s in my film & TV audio engineer, going from locations to studios, recording all kind of things. You do develop that audio memory you don't seem to trust.

Let me tell you that you are completely wrong in not trusting audio memory, as you can educate your ear and your brain to serve you as the best tool there is, much superior to any instrument.

Recreating the original audio sound, as listened by your naked ear, was my job as location engineer. At the studios you had the best speakers and amplifiers money could buy, so the task was to reproduce that original sound as close as possible to what it was. A good reviewer does exactly that.

There's not such a thing as an "objective audiophile". By definition, and audiophile will always be subjective, never objective. If you do any measurements and know how to do them, that part would be the objective one. But correlating that to good or bad quality sound would be false. There was a time when electronics engineers said that if two amplifiers measured the same, they would sound the same. Some probably still do, but many are not that sure of that now.

Perhaps you should do the test of increasing power supply voltage to a good quality opamp, and listen to how it sounds different as you go higher in voltage. And if the measurements do not show any differences, then the measurements are faulty, not the listening. And for a trained ear, that anyone can achieve, there's not such a thing as "imagined" results. BTW, blind tests have proven nothing.

 

[raoultrifan]

@carlmart I totally appreciate your comment and constructive critique, so thank you for your detailed answer!

I would love to see headfiers like this guy here that is actually proving that he is right, but using his own recording studio. I'm pretty sure different harmonics and phase-shifts are changing the way we hear the sound, but as en engineer I think there is a good way to find what measurements we need to do to measure the sound.

BTW, for a better espresso we use a TDP-meter, which is far away from reality to tell you about the differences from a good and a perfect espresso, but will definitely tell you if you're done a bad espresso. So doing measurements prior to listening it's like a phone-screen when dealing with a job-interview: you just need to ensure the guy really knows what's written on its CV, nothing else, the actual jobs interviews are done onsite anyway.

 

[Spektykles]

Hi, I found an offer online for a pair of used AD797SQ (ceramic DIP version) for a very cheap if not suspicious price. Is there any fake ceramic chip out there? And quick dipping test with acetone still valid for today's fake chip or Chinese fakers learned something new?

 

[LaughMoreDaily]

Hi, somebody offered me a pair of used AD797SQ (ceramic DIP version) for a very cheap if not suspicious price. Is there any fake ceramic chip out there? And quick dipping test with acetone still valid for today's fake chip or Chinese fakers learned something new?

Everything can be faked even ex-girlfriends. I know about that last one all too well.