I don't think the design of the basic NAP series amplifier is the issue. It's the tweaks and pooges to it that give it a unique sound character, a composite of several sound effects mostly with delightful results but occasionally with unpleasant consequences. That is to say, it's unlikely that you'll have everything sounding right all the time for everyone when you go about improving sound quality by mods which deliberately set a certain profile of distortion harmonics at listening levels.
Shrillness is a sensation I feel when listening with very low distortion gear. I don't know why this should be but it was the main reason for dropping ULD and blameless designs from my list some years ago when I went crazy with a QA400 distortion measurement system. I did think that it was a psychological response - an addition or compensation for what we don't hear but may be present in the sound we are accustomed to with our usual audio system. Either that or we have to consider our system is crap
We've been over the circuit features in the original design that emphasise specific types of distortion many times in this thread so I'll leave the digging up and and reviewing to anyone interested.
Shrillness is a sensation I feel when listening with very low distortion gear. I don't know why this should be but it was the main reason for dropping ULD and blameless designs from my list some years ago when I went crazy with a QA400 distortion measurement system. I did think that it was a psychological response - an addition or compensation for what we don't hear but may be present in the sound we are accustomed to with our usual audio system. Either that or we have to consider our system is crap
We've been over the circuit features in the original design that emphasise specific types of distortion many times in this thread so I'll leave the digging up and and reviewing to anyone interested.
I have never seen the real NAP in Spice (due to VAS transistor that i have never wanted to acquire) but I know that it is not easy to make it right. More than 90% will be wrong. I don't know whether JV knew about the secret but Traderbam has confirmed that the original has this issue, and i have heard them too (and i could feel/heard the issue).
I understand what you mean that with so many compromises it is hard to get one that will sound right every time in every aspects. But i can tell you that i rarely looked at fft to see the distortion profile, never care about THD, never tried to improve some trivial aspects and parameters, because there is something that when done right, everything will fall into the right places.
I think you are in the same camp with Aksa regarding distortion profile, but in my experience, distortion profile is only a side effect. It can be 'engineered' to look 'good' while actually it is still wrong.
Very low distortion gears use NFB to get to that distortion level but they use NFB incorrectly. That is why.
The ratio of open-loop gain to closed-loop gain, which varies with frequency of course.
I agree.
I agree.When we change the NFB, or any random resistors, there are more than one parameter changed also in the circuit. If only people know which parameter that is responsible with the change in sound when they change NFB (increase or decrease) then they will be more capable of creating good sounding audio systems.
Yep. Note that loudspeakers have THD in the order of 1% anyhow.
The system is crap.Shrillness is a sensation I feel when listening with very low distortion gear. I don't know why this should be but it was the main reason for dropping ULD and blameless designs from my list some years ago when I went crazy with a QA400 distortion measurement system. I did think that it was a psychological response - an addition or compensation for what we don't hear but may be present in the sound we are accustomed to with our usual audio system. Either that or we have to consider our system is crap
I don't have any time for "experts" who use THD as their God of sound quality, especially those profiting by selling this approach. I think it is both lazy and arrogant to blame the listener for not appreciating the "perfection" of a low THD design, unless your customer is a spectrum analyzer.
Good on you Ian for trusting your ears.
Dear Gareth, I once again draw your attention to the fact that all questions on the sharpness and brightness of the high frequencies will stop bothering you if you consider the simple formula that: sound NAIM = NAP + NAC.
I tried to remove the sharpness of the high frequencies, putting on the input of the amplifier - a tube preamplifier, which was supposed to make the sound "warm" or just another ...
But the presentation of the sound has not changed - the bright accent of the sound of the NAP amplifier continued to be felt. Only after switching on the NAC preamplifier in the amplification circuit, the familiar NAIM sound returned without irritation.
I’m not afraid to say that only five additional transistors (the amplifying part of the preamplifier circuit) connected to the input of the NAP power amplifier constitute the very magic of the NAIM sound.
In the circuit of the preamplifier itself, a large role in the final result will be provided by a capacitor in the 47 MkF feedback circuit. Which I advise you to choose for your hearing preferences.
In the complete set, the NAIM sound may well compete with the tube amplifier, which is famous for the absence of any unpleasant sensations for your ears.
Try it and maybe you will find happiness!
I tried to remove the sharpness of the high frequencies, putting on the input of the amplifier - a tube preamplifier, which was supposed to make the sound "warm" or just another ...
But the presentation of the sound has not changed - the bright accent of the sound of the NAP amplifier continued to be felt. Only after switching on the NAC preamplifier in the amplification circuit, the familiar NAIM sound returned without irritation.
I’m not afraid to say that only five additional transistors (the amplifying part of the preamplifier circuit) connected to the input of the NAP power amplifier constitute the very magic of the NAIM sound.
In the circuit of the preamplifier itself, a large role in the final result will be provided by a capacitor in the 47 MkF feedback circuit. Which I advise you to choose for your hearing preferences.
In the complete set, the NAIM sound may well compete with the tube amplifier, which is famous for the absence of any unpleasant sensations for your ears.
Try it and maybe you will find happiness!
Member
Joined 2009
Paid Member
Member
Joined 2009
Paid Member
I have to say that I don't fully understand how all the different design aspects of an amplifier affect the subjective performance. I do use Spice and I do look at the simulated FFT which I find provides useful feedback during the design process on the impact of changes. I would measure the FFT if I had the equipment. I have a new (very old LeCroy) scope to try maybe it has an FFT capability - will check.
I am very suspicious when I read that NFB is to 'blame' for ill effects. My experience has been that NFB is a superb tool. As you say, it has to be used properly. I suspect that too much 2nd harmonic distortion is a cause of flabby bass and although it sweetens up the mid range it's best to avoid too much unless you are bi-amping and have a clean amp for the bass. I suspect that the worse kind of distortion is in the treble, where it causes most fatigue. Why ? - firstly, most NFB amps have a declining feedback factor above 1kHz due to single dominant pole compensation; secondly, higher frequency harmonics are where the character of musical instruments live, it's what makes the difference between a flute and oboe when they are playing the same fundamental note so any distortion here messes things up royally; thirdly, the earbrian works to diminish the impact of higher order harmonics providing the low order harmonics are dominant so adding more distortion at higher frequencies falls foul of this process. So the trick with NFB for me is in how well you control the upper harmonics / upper frequencies. My best sounding amp uses two pole compensation to maintain a high feedback factor to the upper treble. No feedback amps leave the treble un-molested. It's the mid feedback factor amps that screw it up.
A positive effect is obtained from the amplifier cascade. Cascade frequency filter - this effect does not produce.
At the local forum of fans of the company NAIM, I also met exactly the same opinion when the owner of the equipment tried to exclude from the path the amplifying part of the circuit - the "brightness" remained in the sound.
The amplifying part of the circuit works on this positive effect. Although the gain itself, especially in our time, when most signal sources have a level of at least 2V RMS - and it seems absurd.
At the local forum of fans of the company NAIM, I also met exactly the same opinion when the owner of the equipment tried to exclude from the path the amplifying part of the circuit - the "brightness" remained in the sound.
The amplifying part of the circuit works on this positive effect. Although the gain itself, especially in our time, when most signal sources have a level of at least 2V RMS - and it seems absurd.
Last edited:
Member
Joined 2009
Paid Member
Huggygood, I am sincerely glad that the two of us are devoted to the secret of receiving the sound of the NAIM system. On the one hand, it is so simple, but on the other, it is really incomprehensible: how much the impression of the sound changes, while maintaining all the original advantages of the sound of this brand! Increased "brightness" at high frequencies disappears, and thin and quiet sounds at these frequencies absolutely do not lose their airiness and intelligibility.
On the one hand, it is so simple, but on the other, it is really incomprehensible: how much the impression of the sound changes, while maintaining all the original advantages of the sound of this brand! Increased "brightness" at high frequencies disappears, and thin and quiet sounds at these frequencies absolutely do not lose their airiness and intelligibility.
Member
Joined 2009
Paid Member
Attachments
A 12K negative feedback circuit can be reduced to 5.1K. A further decrease may lead to an increase in distortion. This information is from those who tried to reduce the gain. In my version used the standard denomination.
The convenience of adjusting the volume is obtained by connecting a resistor in series with a volume control (in the upper arm of the divider).
Perhaps this is not the most optimal solution, since the input circuit is obtained with increased resistance (75K + 50K DACT). Possible increase in noise at the entrance.
The convenience of adjusting the volume is obtained by connecting a resistor in series with a volume control (in the upper arm of the divider).
Perhaps this is not the most optimal solution, since the input circuit is obtained with increased resistance (75K + 50K DACT). Possible increase in noise at the entrance.
Last edited:
An interesting feature of the early NAC preamps is the 24VDC supply used for all of them, based on a standard 7824 regulator circuit. That simple supply was later outmoded by the introduction of SNAPS external power supplies at unbelievable extra cost and complexity, considering how much improvement was really worthwhile and and that it could easily have been incorporated in most existing models. That's where the "backwards compatibility" concept went haywire, I think. Now we have people designing and selling "super regs" and boards with multistage circuits that allegedly turn the Naim and Avondale clone preamps into amazing sounding devices. I can only wonder how bad the noise and regulation in the DC supply must be to need such extreme levels of regulation and filtering or maybe its really about impressing people with specifications and letting the numbers do the convincing.
We can improve the linearity of discrete circuits with higher supply voltages like 30V, 48V or +/-15V but a single 24V regulated supply isn't too bad for small signal levels and is easier to implement. Another advantage is that you can use up those low voltage, low noise transistors you've had lying around for years - I have lots of them that I've collected over many years but they're obsolete and otherwise have few uses.
There are lots of suitable low voltage, low noise types but I'm thinking of all those inexpensive BC549C/559C transistors that are still widely available and many of us will have stored away somewhere.
We can improve the linearity of discrete circuits with higher supply voltages like 30V, 48V or +/-15V but a single 24V regulated supply isn't too bad for small signal levels and is easier to implement. Another advantage is that you can use up those low voltage, low noise transistors you've had lying around for years - I have lots of them that I've collected over many years but they're obsolete and otherwise have few uses.
There are lots of suitable low voltage, low noise types but I'm thinking of all those inexpensive BC549C/559C transistors that are still widely available and many of us will have stored away somewhere.
I fully agree on all the premises, but with absolute respect for your opinion, based on all my experiences and tests made over years of circuit modifications, especially on the application of the quantity and frequency range of the Negative Feedback Factor, I am of the opinion and conviction exactly opposite to yours: ....... Feedback on the mid-high frequencies (even accepting percentage values of THD Harmonic Distortion significantly higher) must remain as low and contained as possible.
I am in fact convinced that the excessive "brilliance and brightness" of the highs in this type of NAIM circuitry, is accompanied and caused not so much by a high Harmonic Distortion - THD, but rather and much more heavily by the Dynamic Intermodulation Distortion - IMD.
Dynamic Intermodulation Distortion - IMD occurs when a signal with a higher slew rate is applied to the input of an amplifier than the system can handle and unfortunately increases and has a relationship of proportionality with the feedback rate applied to an amplifier.
This phenomenon has been described by Matti Otala, who not only demonstrated its existence, but has also measured the threshold of audibility with a blind test, which has been attested around 0.5% for the most sensitive subjects.
The same subjects showed instead a threshold of audibility to the Harmonic Distortion - THD that was attested on the 4/5% (!!) and therefore of a superior order, confirming that the human ear and the psychoacoustic are clearly more sensitive to the Dynamic Intermodulation Distortion - TIM than to the Harmonic Distortion - THD.
To all this we must also add the consideration that a high Feedback factor, especially on the high frequencies of the audio band, completely skips the delicate, fundamental and indispensable Phase and Group Coherency between the fundamental Tones and the relative Harmonics of any musical instrument, creating precisely that unconscious sense of unnaturalness, fatigue and listening difficulties that in part everyone perceives and defines as excessive "brightness and brilliance".
And it is no coincidence that, among the very first tests I performed years ago on the NAIM clone, precisely that of shifting the feedback rate (by means of an RC network of variable value) on the mid-high frequencies confirmed to me, after intense weeks of listening, all these premises.
So much better to take into account acceptably high values of Harmonic Distortion - THD than to try to clean it up with high negative feedback factors, especially on high frequencies.
Therefore, I think it is very important, in order to understand correctly the musicality of an amplifier and all its limits, to reflect on the consideration that in the case of audio electronics, the type and nature of a Distortion is therefore almost more important than its quantitative entity.
Last edited:
@nikola
I do not know the secret of the nap / nac couple but on the other hand, I can immediately make the difference
when the couple is formed or when the couple is not.
I tried several combinations nap amp with other preamp and nac preamp with other amp (I'm lucky to have a lot of amp and preamp available) and I always returned to the pair nac / nap.
after, thanks to this thread and especially to "those who know" I learned what was important in the circuit of nap and nac, so I particularly focused on the selection of components and on the EXTREMELY important quality of the power supply in terms of both component and layout.
as i have said many times, i have not made a nap140 but rather something between the 90 and the 110 and this set gives me a smile everytime i use it and it's for me what is the most important.
In addition, I rarely had a set as "silent" as this one
I think the therm used is "black wall" and it corresponds well to what I feel.
a short time ago, I repaired a tube amp renamed SE 845
and I found (all proportions kept) this type of soft and robust sound, clear and fine at the same time.
on the amp in SE 845 was much better on all the register
but the plot was the same.
finally, I think we should not take this circuit in the lightly and that to get good results, we must work on it, do a lot of research and select its components.
Those who want to assemble a clone without asking the right questions will get to have an amp that makes sound, but certainly not music.
I do not know the secret of the nap / nac couple but on the other hand, I can immediately make the difference
when the couple is formed or when the couple is not.
I tried several combinations nap amp with other preamp and nac preamp with other amp (I'm lucky to have a lot of amp and preamp available) and I always returned to the pair nac / nap.
after, thanks to this thread and especially to "those who know" I learned what was important in the circuit of nap and nac, so I particularly focused on the selection of components and on the EXTREMELY important quality of the power supply in terms of both component and layout.
as i have said many times, i have not made a nap140 but rather something between the 90 and the 110 and this set gives me a smile everytime i use it and it's for me what is the most important.
In addition, I rarely had a set as "silent" as this one
I think the therm used is "black wall" and it corresponds well to what I feel.
a short time ago, I repaired a tube amp renamed SE 845
and I found (all proportions kept) this type of soft and robust sound, clear and fine at the same time.
on the amp in SE 845 was much better on all the register
but the plot was the same.
finally, I think we should not take this circuit in the lightly and that to get good results, we must work on it, do a lot of research and select its components.
Those who want to assemble a clone without asking the right questions will get to have an amp that makes sound, but certainly not music.
Power supply is important for almost any amplifier. But unlike common circuits, compensations ala Naim is intolerant of power supply variation. The question is how did you dial in the right voltage? Is it exactly as shown in the schematic?
Member
Joined 2009
Paid Member
I relied on a Nait3 that I had at that moment.
I picked up the tension and then I went digging into my power tranformer pool.
I ended up with a voltage (no load) a little higher than born, so I decided to adapt the rest to get closer to a 110 but without making a 110 because I have output devices in TO220.
from good quality photos of nap110, I could read the main resistances that make the difference with the 140 and of 90, I then proceeded by trial / error until the sound I like and that Offset and bias are stable and in good values.
he has not aged at all or changed values since all these years
but it is in a fully 2U aluminum case with the compensation transistor mounted below the pcb and in contact with the case.
my two cards are placed on the opposite side of the output device, on CD player optical block dampers because for some unknown reason, it is even quieter when the pcb are not fixed firmly.
I guess it's microphony but I can not explain it.