Hi,
with the typical low gains required for audio there´s no compelling reason to go OPAmp style when designing with discrete components.
In the end we want to build an amplifier and not an integrator that is brute forced to ´amplify´.
As such especially linelevel stages can be build very simple and with low parts count and still good measurement values ... and best of all ... they also sound rather better than OPAs
jauu
Calvin
with the typical low gains required for audio there´s no compelling reason to go OPAmp style when designing with discrete components.
In the end we want to build an amplifier and not an integrator that is brute forced to ´amplify´.
As such especially linelevel stages can be build very simple and with low parts count and still good measurement values ... and best of all ... they also sound rather better than OPAs
jauu
Calvin
With > 1V out from from the usual sources even a buffer could be sufficient “gain”
So a simple design can be buffer - volume - buffer.
I am carefull going for volume - buffer only although many sources (incl the soekris dam1121) could drive the ladder volume control direct.
Next issue is that most buffers are inverting so using 2 buffers like “buffer - volume - buffer” avoid an inverting output.
I can imagine that a discrete buffer design could me my golden ticket for my requirements.
There are so many pre-amp designs available but most with some active amplification ( gain > 1 )
So a simple design can be buffer - volume - buffer.
I am carefull going for volume - buffer only although many sources (incl the soekris dam1121) could drive the ladder volume control direct.
Next issue is that most buffers are inverting so using 2 buffers like “buffer - volume - buffer” avoid an inverting output.
I can imagine that a discrete buffer design could me my golden ticket for my requirements.
There are so many pre-amp designs available but most with some active amplification ( gain > 1 )
I've not met an inverting buffer - the simplest kind I'd suggest would be a source follower, either JFET or MOSFET, loaded with a current source. These are non-inverting. To get a low output impedance use a FET with high gm - Toshiba SSM3K339R and SSM3J327R are the top of the pile in this respect. I use them a lot for driving cable. If you're buffering an input you might prefer a FET with lower parasitic capacitances - JFETs normally win on that score.
Thanks, unfortunate I am not knowledge enough to design a buffer even with your guidelines.
Which proven DIY circuit would you suggest ?
Which proven DIY circuit would you suggest ?
My own design of course 🙂 Here's the buffer I designed for the OPS of my multibit DAC. It has its own PCB, the gerbers are available on the thread : lingDAC - cost effective RBCD multibit DAC design
Nice clean circuit with low parts count 😀
As you need to drive higher currents into a headphone I understand the 2200 uF output capacitor.
With line level into >22k power amp I guess besides a smaller cap everything should stay the same or can I also use smaller Q1-2 / Q4-5 ?
PS I see the SSM3J327r are already SOT-23 style .... no need to go more “compact”
The text also described the adjustments for line level so this explains all 😀
Last edited:
I can think of one: it is easy to design a line or even a phono stage with no coupling caps using low offset opamps. Not so easy with discretes.
On top of that most good sounding caps require a substantial polarising voltage to perform at their best.
For me SS without dc coupling is pretty much pointless, i would rather use valves then.
If you're very concerned about parasitic capacitances (there will be some distortion for non-zero source impedances as the capacitance vs voltage characteristic isn't linear) then you could substitute a smaller MOSFET. There is one which is in a smaller package than SOT-23 which has good characteristics and for line-level duty you'll probably want to reduce the bias current anyway. The smaller package won't handle 50mA bias current but reduce it to 5-10mA and you'll be dandy.
The downside is its very fiddly to solder.
SSM3J56MFV | MOSFETs | Toshiba Electronic Devices & Storage Corporation | Asia-English
The downside is its very fiddly to solder.
SSM3J56MFV | MOSFETs | Toshiba Electronic Devices & Storage Corporation | Asia-English
SOT-723 is indeed really small ....
Thanks for your suggestions and I will let it sink in before I will a final decission
Thanks for your suggestions and I will let it sink in before I will a final decission
Food for thought ...
Think of the number of caps, opamps, what ever, that the signal/music passed through before you get a hold of it, to do your magic. Do you really think it is going to make a whole lot of difference? As was said before, just do not alter it, or you change it and you are good to go. Speakers have way more influence than a opamp vs a discrete design, Doug Self gave up on discrete after he demo'd a NE5534, when was that? have fun with a sot-723 🙂
Think of the number of caps, opamps, what ever, that the signal/music passed through before you get a hold of it, to do your magic. Do you really think it is going to make a whole lot of difference? As was said before, just do not alter it, or you change it and you are good to go. Speakers have way more influence than a opamp vs a discrete design, Doug Self gave up on discrete after he demo'd a NE5534, when was that? have fun with a sot-723 🙂
I suggest that you learn ltspice and some circuit design skills as a start.
More food...
I feel that it's fairly obvious at this point that what makes a component sound more 'musical' is some particular character and dynamic behavior of it's objective imperfections. That isn't a knock on subjectivists. I've certainly heard components which 'sounded' more musically convincing, more live sounding, than other components. For the reasons you state, objective perfection at the final stage of the playback chain, logically, isn't the reason some components SOUND better than others. Pouring dirty water in to a clean cup does not make the dirty water any cleaner.
What I would knock pure subjectivists over, however, is the common declaration that the sound of some favorite triode amplifier, for example, sounds more musical DESPITE it's distortion, and certainly not because of it. I feel that pure objectivists also deserve some knocking. They typically make the implicit assumption that because some component objectively measures better, it must, therefore, SOUND better. If any improvement in a given parameter is even audible at all. Which, is unlikely these days for any competently engineered component.
Since perceptual testing informs us that the measured performance of any competently engineered, purely electronic component is well below the threshold of human perception, any pure objectivist who's audio system isn't based around some mass market A/V receiver or integrated amp must logically explain why not. Logically, it seems to me, the reason some components SOUND more musical than others is due to some particular character of it's objective imperfection. Which fools the human ear/brain system in to better perceiving the reproduction as being closer to the experience of hearing live music. Just my two cents (or less) worth.
No, that's missing the point. To reproduce sound accurately it must sound just as _bad_ as the original source - no better, no worse. A well reproduced recording of fingernails down a blackboard _should_ sound bad - anything that makes it sound better is clearly distortion!!
Do you imagine Ozzy Osbourne or the other early heavy metal pioneers wanted their distorted over-driven guitars to "sound good"? Awesome perhaps, but not good. The function of sound-reproduction is to reproduce the signal as faithfully as possible, warts and all (but only the original warts!). Its not to impose a single adjective on all sound!
If you want a rose-tinted audio view of a recording, that's an effects unit you'll be wanting, not an amplifier, and that's fine, just don't mistake an effects box for a generic amplifier, they do different things.
For the time being, I will sidestep debating your assertions of what a home music entertainment system should and shouldn't aspire to achieve. That said, I don't believe I've taken a position above on what my own listening/system objectives are. I think that I gave fair criticism to both sides of the objectivist/subjectivist divide. However, I take from your comment that you consider yourself an objectivist. If that presumption is correct, please tell us about your mass-market based audio system. Any money spent for components beyond mass-market products does not seem a rational expenditure for a true objectivist, and would beg explanation.
Last edited:
The performance of Groner's composite is astounding, but given that it's a two pole compensated amplifier, it will not be stable in all of the circuits that will be stable with a single LF dominant pole compensated amplifier like a traditional op amp. So, it's not an easy drop in. Still, there are many fixed gain or limited gain range circuits that could be fitted with a two pole composite amplifier, given the right compensation. It's well worth reading his article - composite amps have a lot of promise!
Yes, as Groner well covers in his paper, the stage is conditionally stable, including usefully being unity gain stable.
- Status
- Not open for further replies.