Bonsai, I am not talking about IC op-amps pre 5534. I am talking about the ones Jacco mentions and others.
I once went to the trouble of designing some AD811 power supplies to drive other IC op-amps. Better sounding than a three terminal series regulator, you bet. But then a simple discrete shunt improved on that again and was a lot cheaper.
The grounds were clean. Never run a small signal common on a large signal (dirty) common return. Scoped a lot of that out.
Oh BTW, those silver mica caps, I stopped using them a long time ago. The ones I used sounded very iritating after a while, cornell dubilier IIRC. Again a grainy coarse sound. Although the leads were made out of steel so I expect that contributed a little.
Give some Wima polypropylenes a go. Film and foil.
I once went to the trouble of designing some AD811 power supplies to drive other IC op-amps. Better sounding than a three terminal series regulator, you bet. But then a simple discrete shunt improved on that again and was a lot cheaper.
The grounds were clean. Never run a small signal common on a large signal (dirty) common return. Scoped a lot of that out.
Oh BTW, those silver mica caps, I stopped using them a long time ago. The ones I used sounded very iritating after a while, cornell dubilier IIRC. Again a grainy coarse sound. Although the leads were made out of steel so I expect that contributed a little.
Give some Wima polypropylenes a go. Film and foil.
Hard to figure out - agreed. I've given up trying.
Other than good engineering practice, I tend to disregard most of my education and experience in electronics and the semi industry when it comes to audio. I build something, optimise/fiddle about with it to get a "better" sound then listen to it for a few weeks. If I go back to my reference and don't miss the new item then I regard it as sounding worse. I will probably try again a few weeks later, maybe with different ancilliary components to see if I come to the same conclusion. I don't make an active decision over which is best I just use the one that my subconscious chooses.
The persuit of ppm distortion and ultra low noise figures is pointless. Indeed I notice that many of my favourite recordings have noticeable hiss on them.
My subconscious has chosen my JE-990 variant over all of the IC gain stages I have built +
Salas shunt/B1
SKA pre
Naim gain stage - with various mods
Grounded grid
Aikido
The Daniel Vacuum Tube Preamp (what a mouthful! and easily beaten by NE5534s)
nothing (i.e. passive pot)
The IC amps, in fact al the amps, sound very different depending on the PS and decoupling used. The best figures don't always give the best sound.
Stupid hobby really
Nothing stupid about pursuing the love of music and faithful reproduction IMO.
You are one of the first to mention EMI. This is an area where many IC's fall short. It is not their loop area but the lack of ability to reject EMI coming in on extended leads. It is often needed to add a small LC filter on the input in uses such as a preamp. Inside gear such as CD players FET imput devices often work better even with the other specifications being worse!
As to resistors, if a resistor is actually being used to carry signal the quality may be important depending on how much actual resistance modulation occurs.
It is certainly possible to design a monolithic opamp to equal a discrete design. The last attempt I am aware of fell short in EMI suppression.
Then there is the controversial issue of global feedback. There may be some actual advantages to feedback around each gain stage, this is difficult to do well with poly silicon resistors.
Last edited:
No, but it's a question of degree, and the obsessive and unrealistic pursuit of faithful reproduction obscuring the love of music and the love of music being used to justify that obsession. We already have reproduction perfectly adequate for the enjoyment of music widely available at very reasonable cost. Building your own opamps is fine, if that's what floats your boat, but people who love music listen to music. People who build opamps build opamps because they love building opamps, and it won't be 5 minutes before they're building another opamp.
Myself, I love to play. I love to design and build electronics too, but I don't design and build equipment because I'm dissatisfied with the sound of what I've got to listen to, or in the expectation of a marked improvement in fidelity. I expect at most minor differences when changing anything but speakers or headphones (and microphones) and anyone who expects anything different is simply being unrealistic unless the electronics in question is exceeding poor. The satisfaction that I get is in the physical demonstration that the understanding that permitted the design was good.
Build opamps if you wish, but don't expect any satisfaction other than to be able to say, 'I built that opamp'.
w
Won't find many of those in linear processes, mostly thin film. Pretty good too for only a few hundred atoms thick. The mass of the resistive matierial compared to an RN60 would be shocking, yet we still make ppm accurate circuits.
Any time you want to disclose all the fun little details of your processes I will be more than glad to listen! If you ever have a chance to play and want to make 10 of those matched thin film resistors for me to test I'll have a field day.
Then there are some other parts you get to use that are not available as discretes that would be a lot of fun to get.
So what does a used very small fab cost?
So you don't design and build to fix problems in the sound you are hearing. Do you do it only for the fun of building then? Me, I do it for fun and out of curiosity - understanding the hows and whys of what I hear. Without the hope of some improvement then at least half the interest would be gone.
Sound advice - expectation effects will be sure to mess up listening tests.
If I told you how precious a thin film target is you would be surprised. There is actually a little magic involved and some are thrown out. We had a shaman of the process who could put the TC at exactly -15ppm and voltage coefficient at an unmeasurable level, you would love it a real curmuddgeon. One target does thousands of wafers and you just hope to be on a roll.
Who decides when that degree has been exceeded? If there is an advancement made is it rejected because it was done out of obssesiveness and not love?
How was this decided? Was it democratically, or by central government mandate, perception by a single individual? Or extraterrestial being.
Not many facts here, whole lot of generalisations and stereotypical views.
Another attempt to force ones viewpoint and actions after trying to discredit others experiences and findings.
Why, is that your experience? It is certainly not mine.
I don't build discrete op-amps for serious audio reproduction or recording anymore, unless a design cannot be implemented without doing it that way.
They were a stage of progression from IC opamps.
Built a discrete/hybrid open loop input stage that had bandwidth and distortion that would put IC op-amps to shame. Primary objective was to use a topolgy combination that would minimise individual semiconductor and thermionic vacuum tube parasitic effects on the signal. It just so happened that the end result was something that also had extremley wide bandwidth.
When you can distinguish between a well-designed chip circuit and a well-designed discrete circuit by ear alone (that means no peeking, true blind testing), then maybe you can make these assertions with any meaningfulness. Lots of people make lots of claims about audibility with no actual evidence.
Yes, you are correct Scott - the ploysilicon R's are generally not used for linear - my apologies.
As for the earlier comment about inability of IC's to reject incoming EMI, that's simply not correct. You have EXACTLY the same issues to deal with in discrete circuits. If you are referring to JFET vs bipolar ability to reject RFI, then again, this applies to discretes as much as it does to opamps.
So far wrt opamps we have debunked just about everything thrown at them that are supposed to make them 'inferior'. We've agreed that for some specific ultra low noise applications, high output drive current (although this is easily fixed in my view) and high (i.e. > +-15V), discretes offer advantages.
Anything else we can look into?
Indeed, all sorts of claims of audibility are routinely made, no matter how lacking in support or even plausibility.
With all the trolling I for a moment got deflected from the true purpose of this thread Back on target now...
IC opamps these days don't break a sweat at sub -120dB distortion. So your open loop stage exceeded this distortion by accident or by design? If by design do you consider such low distortion to be important to sound quality?
Back on target now...IC opamps these days don't break a sweat at sub -120dB distortion. So your open loop stage exceeded this distortion by accident or by design? If by design do you consider such low distortion to be important to sound quality?
"Built a discrete/hybrid open loop input stage that had bandwidth and distortion that would put IC op-amps to shame."
Come on Craig. Making a statment like that says you know more about small signal design than a whole segment of the semi industry, including some leading lights around here. Be a little more magnanimous about this whole thing and a little less black and white. Very few things in audio are as certain as you seem to want to believe.
Come on Craig. Making a statment like that says you know more about small signal design than a whole segment of the semi industry, including some leading lights around here. Be a little more magnanimous about this whole thing and a little less black and white. Very few things in audio are as certain as you seem to want to believe.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.