I was talking to a guy in denmark.
A new AP analyzer would cost 8000$. So i think i will buy one for the company.
Hopefully, í will be able to back som of my designs and comments up with real measurements.
If a discrete circuit will sound better? or equal?
A discrete circuit will not suffer from heat infuced distortion from one circuit part to another if designed correctly.
It will be a lot harder to prevent with an opamp design.
- Sonny
A new AP analyzer would cost 8000$. So i think i will buy one for the company.
Hopefully, í will be able to back som of my designs and comments up with real measurements.
If a discrete circuit will sound better? or equal?
A discrete circuit will not suffer from heat infuced distortion from one circuit part to another if designed correctly.
It will be a lot harder to prevent with an opamp design.
- Sonny
It's cheaper to build several versions and audition them. Would surprise me AP measurements will give you the best sounding.
Please excuse me if I take again an old 3D and maybe my following question has been already answered in the same long 3D (by the way ... are not that easy not to get lost in this long 3Ds)
Anyway I am referring to this message
I have an old preamp (bryston) based actually on an discrete op-amp
The sound is clean but quite hard in the midrange
I already changed some local bypass caps with better quality but i want to replace them with top quality Panasonic or Os-con
Could i get any improvement replacing the feedback resistor with something top quality ?
I am very interesting in this topic indeed. Very.
Thank you very much indeed for any suggestion.
Kindest regards,
gino
Anyway I am referring to this message
I have an old preamp (bryston) based actually on an discrete op-amp
The sound is clean but quite hard in the midrange
I already changed some local bypass caps with better quality but i want to replace them with top quality Panasonic or Os-con
Could i get any improvement replacing the feedback resistor with something top quality ?
I am very interesting in this topic indeed. Very.
Thank you very much indeed for any suggestion.
Kindest regards,
gino
Last edited:
I think DF96 is simply commenting that opamps are generally so resolving now, even Discrete if well done, that the main distortion/noise components left are from the composition/type, resistance value, tempco, voltage coefficient and by extension matching of said variables is what is left to create a 'sound signature'
Yes, the whole point of an opamp is that you don't hear it. If you can hear it then something is wrong; most likely the way you have used it.
Resistors do not need to be 'top quality', especially in preamps where signal levels are low. Decent quality, properly specified, is sufficient. Many circuits where 'tweaking' appears to change things have design errors.
Some local bypass caps need to be low quality, so they damp RF resonances. 'Upgrading' can be a serious downgrade!
Resistors do not need to be 'top quality', especially in preamps where signal levels are low. Decent quality, properly specified, is sufficient. Many circuits where 'tweaking' appears to change things have design errors.
Some local bypass caps need to be low quality, so they damp RF resonances. 'Upgrading' can be a serious downgrade!
Thank you for the voice of reason.
Op amp audio circuits fall short when
1) We try to use them too close to their design parameters. The most obvious (but not the only) example is for example requiring 35 mA output current from a 40 mA op amp. (The Pimenta circuit is the classic example of this.)
2) Topology is less than ideal (partly related to 1). (This includes supporting circuitry like power supply.)
3) Inferior parts or the "wrong" parts being employed in the circuitry.
4) Layout of physical circuit board, including grounds.
There are a lot of things that op amps won't do, and for that we have discrete circuits.
This is coming from a guy that started out with tubes because transistors were hardly taken seriously by "audiophiles" at the time. I use op amps (mostly 5532
) all the time now. I am presently prototyping a simple "preamp" (really an audio control unit) that employs multiple parallel iterations of the 5532 (like Doug Self's design I think?) . So far I have met 100% of my design objectives on the first try. (It's not my first time around the block though. ) So far I have incorporated zero discrete transistors in my design, but I am toying with the idea of discrete shunt power supply regulators.As an example of how even "cheap" chips can be optimised, I did an exercise a few years back with a "no name" amplifier board out of a "hi-fi" tabletop stereo.
First I hooked it up with no mods. It sounded thin with flimsy veiled bass. Turn it up and it ran out of steam really quick.
I changed the feedback and input caps for poly caps from Rat Shack. Then I took the power supply off board, used two (cheap salvage) transformers monoblock style, used capacitor banks on the PS board (quadrupled capacitance over stock), and replaced the PS caps on board with 330 uF "voicing" caps. At this point it was sonically indistinguishable at "moderate" volumes from my Nakamichi Stasis amp. Steady state power was slightly more than doubled. The subjective difference was like night and day. I put it in a box and now it's my second TV amp.
My point is that objectives can be met with quality parts and logical design, even with "cheap" chips. I don't even know what type of chip the amp uses (no markings) and yet I was able to optimise it to a very large degree by employing tried and true tweaks.
First I hooked it up with no mods. It sounded thin with flimsy veiled bass. Turn it up and it ran out of steam really quick.
I changed the feedback and input caps for poly caps from Rat Shack. Then I took the power supply off board, used two (cheap salvage) transformers monoblock style, used capacitor banks on the PS board (quadrupled capacitance over stock), and replaced the PS caps on board with 330 uF "voicing" caps. At this point it was sonically indistinguishable at "moderate" volumes from my Nakamichi Stasis amp. Steady state power was slightly more than doubled. The subjective difference was like night and day. I put it in a box and now it's my second TV amp.
My point is that objectives can be met with quality parts and logical design, even with "cheap" chips. I don't even know what type of chip the amp uses (no markings) and yet I was able to optimise it to a very large degree by employing tried and true tweaks.
Thank you sincerely to all. I think that finally i am starting to understand better
So many op-amps are almost perfect, and measurements reported in datasheets confirm
The implementation is not an easy task at all, and very unfortunately but here is the challenge
I am starting to understand better with your kind and extremely valuable advice
Thank you again !
Kindest regards,
gino
So many op-amps are almost perfect, and measurements reported in datasheets confirm
The implementation is not an easy task at all, and very unfortunately but here is the challenge
I am starting to understand better with your kind and extremely valuable advice
Thank you again !
Kindest regards,
gino
Hello !
I see. But are discrete op-amps less sensitive to implementation than monolithic op-amps ?
Thanks again
gino
They are if you design them to be.
The only practical reason I see for using discrete op amps is if there is a specific parameter that can't be met with practical available op amps. Examples are higher voltage output, class A output, or high current/power output. Perhaps you could come up with a design that is more linear than available op amps, but there are tradeoffs involved here too.
If you're designing a solid state power amp, it is practical (for several reasons) to employ a topology similar to "discrete" op amps. But it isn't mandatory (although many would disagree with me).
Very interesting and thanks a lot !
I think i will stick with a tested kit ever. I feel that the implementation (not to talk even to design) it is absolutely beyond my possibilities.
I am finding the commercial offerings, apart the very expensive ones, somewhat limited.
They have quite weak transformers and small caps in particular.
This is a very bad start.
Thanks a lot again.
Regards,
bg
You mean consumer grade audio components, or op amps? Because there are a whole lot of different op amps available; including high voltage and op amps and high current unity gain integrated buffers. There are low noise op amps, DC optimised op amps, super fast op amps, ultra high input impedance op amps, and probably more.
Consumer grade stuff can be tweaked too.
I like some of the old stuff (which I can repair and upgrade). But if you know what you're doing and have the money, you can build stuff that can go toe to toe with "high end" stuff. I think Krell started out as guy that wasn't satisfied with consumer grade equipment.Hi ! i am speaking in general
There are not needed features and bad parts where it counts
When i see at those tiny ps caps i want to cry. They are even silly
And ps caps are so important
no op-amps are not an issue because are very very cheap indeed
And if well implemented they are good. This i have understood.
I was talking about other parts where the savings impact the sound quality.
Consumer grade stuff can be tweaked too.
But if you know what you're doing and have the money, you can build stuff that can go toe to toe with "high end" stuff.
I think Krell started out as guy that wasn't satisfied with consumer grade equipment.
And this would be my aim. But they study the lay-out in a way that just replacing ps caps with something bigger and better is almost impossible.
Just look here
I would like to replace the ps caps with something bigger ... there is absolutely no space. Every thing is so packed that there is no space left.
I think that this is done intentionally
This thing with just two Mallorys would gain a lot. I am sure about this.
What a pity !
Regards,
gino
I am not sure, I would need to evaluate the outcome before reaching any conclusion.
Are you really sure?
Hi ! sure not ... of course
But if you replace a ps cap with another one with:
- lower ESR
- higher ripple current
- lower impedance
could it be worse ?
And looking at the datasheet when you increase uf for a same brand and series you get always better specs. Always.
A Epcos Sikorel 20.000uF/63V is better than a Epcos Sikorel 10.000uF/63V
Some brands offer even a cap upgrades (uF increase same type usually) for their amps
I made very rough experiments increasing uF in the PS
I have always heard benefits.
I am pretty convinced on this ... because they are more expensive usually
It is the same old story ... the better the more expensive
Thanks and regards,
gino
No really, but you are probably throwing away your money.Hi ! sure not ... of course
But if you replace a ps cap with another one with:
- lower ESR
- higher ripple current
- lower impedance
could it be worse ?
Assuming the experiment was done in the right way then this is a clear signal that the PS design was flawedAnd looking at the datasheet when you increase uf for a same brand and series you get always better specs. Always.
A Epcos Sikorel 20.000uF/63V is better than a Epcos Sikorel 10.000uF/63V
Some brands offer even a cap upgrades (uF increase same type usually) for their amps
I made very rough experiments increasing uF in the PS
I have always heard benefits.
I am pretty convinced on this ... because they are more expensive usually
It is the same old story ... the better the more expensive
No really, but you are probably throwing away your money.
Assuming the experiment was done in the right way then this is a clear signal that the PS design was flawed
It could be the case easily ... i usually "play" with cheap amps not ML or Krell
For me flawed means that it does not function properly
I cannot say that a nad c352 is "flawed"
I am pretty sure that with higher speced and more expensive parts could be better (transformer, caps, volume pot, etc.)
But i would not call it "flawed"
Regards,
gino
100´s of opamps in a typical signal chain, that´s an urban legend which pops up now and then...
How´s about this: Rupert Neve Designs 5088 Standard Mixer - 16x8x2
"For signal amplification and control, all new, high voltage, discrete op-amp cards have been developed especially for the 5088 that eliminate crossover distortion while offering extended headroom, dynamic range, and frequency response. ..."
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.