Member
Joined 2009
Paid Member
A tube amp without negative feedback, somewhat traditional design, is likely dominated by 2nd harmonic. Can't really compare with SS amp using high feedback factor. You might be able to hear clear differences between the two, even on Youtube but I would also agree with Ian that identifying the proper sound character of amplifiers through Youtube (and drawing conclusions that an amplifier is 'metalic') has very doubtful credibility indeed.
Thanx Ian, ill follow the link later.
When youtube sound of a Naim is good, then it must be good (or better if still dont get it) when listened thru good system.
Yes, many clones sound wrong, but you cant underestimate people like Nigel who still keep the clone. It is true that many people fool themselves (their ears), especially when the amps measure better they cant be honest with their own ears.
But i have listened to a clone (i think it was made in US) with friends who do not know what Naim is. They are impressed and liked it. I didnt like it but i think i understand what quality that they like from it.
Those are old Naim like the Nap250 which should have much higher thd than the Naits. Yes, the thd of old naims are subjectively worse when listened thru youtube but i can predict how it would be heard when heard off line.
To be honest, i dont trust almost everyone credibility when they talk about sound, even when they listen to it in a good system and room. So if you doubt the credibility of youtube sound analysis, thats more than normal.
But observe what im saying and what not. Nigel used the words "metallic" and i think he has heard the clone and the original. Most probably it is the old Naim not Nait.
I told him that this "metallic" sound IS distortion. Im telling how a distortion is perceived by ears.
DDG,
Hmmm, if there is good distortion (tube), and bad distortion (SS) how do you distinguish between the two THD figures? It sounds a bit like good and bad cholesterol.......
HD
Hmmm, if there is good distortion (tube), and bad distortion (SS) how do you distinguish between the two THD figures? It sounds a bit like good and bad cholesterol.......
HD
Easy. One is tube amp's and the other one is SS.
Details is everything. 100kHz/1kHz/10kHz, max power / listening spl, low order domination or not.
And says who distortion is the only variable? If beauty is determined by hair, nose, mouth, breast, leg then it is fine if somebody wants a minimum of 36B. It doesnt mean that he will agree that one with 36B is always more beautiful than one with 24b. But when we talk only about certain specific metric or variable then it is definitive.
I read that but i have different unferstanding.
0.002% is related to the source, a CD with 1kHz sinewave.
Absolute minimum thd of Nait2 is 0.0165% (7-7.5mv across Re). Slightly higher with 6-6.5mv. So around 0.02% with a Nait. Nap250?
Bias of 400mA (90mv across Re) was tried and there was only little improvement (thd was 0.0135%) and there was thermal runaway.
I have come to similar design issue. I tried the highest bias (lowest thd) but where thermal runaway didnt occur, but the sound was not better with higher bias (lower thd). I believe this could be true with Naim amps too.
But most tube amps lack your most important criteria of drive 😕
Please could you give us a oouple of examples of amps that meet your criteria for excellent sound.
Fascinating:
- most instruments produce harmonics, so how can you differentiate between sound produced by the instrument and the amp?
- What does weak damping sound like (because I don't hear any problems with my KT88 PP amp compared to my ss amps)?
- some of my favourite recordings have noticeable hiss, why would amp white noise (at a much lower level) be a problem?
- group delay (at LF)... oh dear, don't get me started on this one. Anyone with ported speakers might as well just give up!!!
Last edited:
I wouldn't draw conclusions about what a Naim amp sounds like. I maintain that there is no such thing as a Naim clone unless manufacturered by Naim. The schematic might be cloned but invariably the parts used and the physical implementation will be different from Naim's. So the sound comparison will be no more valid than between a real Naim and any homebrew based on the RCA circuit.
Do you have a local Naim dealer in your area? Probably few and far between.
There are notes and instruments that produce dissonant harmonics. But musician dont play these notes and they dont use cheap instrument either. So whatever the instruments, they dont produce "distortion" and this should be true with most CD even high quality mp3.
Distortions produced by amps are different in that they are fatiguing.
Thats why i dont use tube amps. For some, the midrange magic is a good trade off for the weak drive. For me, SS can produce midrange magic too even tho not as seducing as tube amps, but more important is what it can do that tube amps in general cant.
Check "your" music collection. Mostly it is "girl with a guitar". Low damping amp cannot produce LF with high speed. Even rock music is not its match.
I didnt mean to say that white noise is a problem. But noise can be modulated and the resulting sound wont be as nice as pure white noise.
Thats why i dont use vented. Do you think JV will use vented? You know the answer. If only vented and sealed produce similar LF extension, no one will use vented unless they cant hear the difference.
Given two recordings, one is Naim and the other one is P3A, do you think you can differentiate which one is which in a blind test? I think it would be easy. Even topology has its unique sound. Thats because topology limits what a designer can do.
There is no magic in audio. Its Science and Math.
Not really. People often say mostly second harmonic. What they mean is hasn't got rid of the second harmonic using Push Push cancellation. A equally nasty trait of the high power valve design which sound, slow, nasal, lumpy, and wrong.
The NAP 140 we talk of puts the second harmonic back. It doesn't inject it from a pallet of tones. The VAS also provides some when a simple type ( good ).
The ear produces 30 % THD with an exponential cascade of harmonics ( University of Oxford Radcliffe hearing lab ). The shape of the inner ear dictates this. Conjecure is the brain knows this distortion and extracts the reality. The second conjecture is if the distortion we hear has this exponential harmonics the brain will say. " Yes, that's good ". However to remove the second harmonic will say " that's hi fi ". Remove everything up to the fifth and he brain will say if it exceeds 0.1%." That's nasty ". However to have the fifth at 0.1% when the second might be near 1% is OK. Further conjecture says 0% is " exactly " equaly to 1% exponential. Anything else is not as good. However 0.1% exponential will be very good as the trained ear will detect less cumulative tonal colour. However to some they will hear a lack of saturation of tonality. As speakers seldom give less than 1% there is little danger of owning a " less saturated sound ". The eye also likes this small exageration of colour saturation.
Anyone who hasn't heard how a good valve amp sounds ( of which there are very very few ) will never be able to understand the insights they give. Often the rhythm is not 100% and this is a major defect. On medium pace music rhythm is very good and and might still win the day as it has more detail in the thumping sounds made. The compromise need not be as bad as most port loaded speakers. The insights are much better than real life. One could never be as close as a microphone gets when lets say more than two performers ( Thanks Bob Stewart, his concept circa 1976 ). This sound picture is like when the people of Mauritius dream of snow and suspect it is not unlike the stuff in the fridge. They know the real thing is a different experiance, they infer it. It's so much better to have something closer to reality rather than the inside of the fridge and a VHS tape. Sound with OK distortion where as few devices as possible got involved is like the very best artificial snow. As we said in the 1970's " Photocopying losses of complex circuits ( Bob again ) ". If you never heard it you can only guess.
One has to understand that a testing method that singled out the 5% bad ones from the 1% OK ones in 1937 has no reality in the modern world. It is curious and interesting to keep improving this aspect of performance and perhasp says something good going on. It means almost nothing if the reality of the" machine that is us " is factored in. If lets say the very best 32 bit recording was made of any sound ( music if you like ). Then a before and after error reading could be made. Up to a point this is interesting. Then put in the speakers. We would then be forced to find out the very real truth, we would need a " perfect " microphe to do this. Only then might we find some truth? It would be a lifetimes work I suspect. THD would then be no more than taking the blood presure of someone before a blood test. I suggest there is no blood tests in hi fi, let alone a scan.
I have to say anyone who thinks they hear the distortion a Naim amplifier has is completely bonkers. Anyone who thinks they hear the mechanism I will begin to believe, THD is the ghostly suggestions of a crime. THD is a steady state measurement that is nothing like music. Like letters it has the components that make words, it's better than nothing. It's like the the infinite monkey theorem . That is, a monkey hitting the keys at random on a PC ( to be modern )for an infinite amount of time will produce any text, the complete works of William Shakespeare being favourite." It nearly that mad.
My speakers roll off at 10Hz with a 1st order rolloff. Effectively zero group delay. I only use the main amp above 100Hz - do I still need "drive"?
It is like bad cop worse cop! Instead of investing in high quality equipment, it costs nothing to listen to some other guys esoteric equipment on YouTube instead.
Hopefully the guy is honest and don't post pictures of different equipment to that he is auditioning, then it would be confusing to judge the equipment accurately.
For example the sound is from an SS amp and he shows pictures of a valve amp.
But guess what dedegogo would know without a doubt
Here is something I can't prove but feel is right. Two amplifiers as near as is possible identical ( null process perhaps ) except they have different ways of setting stability limits. One perhaps Cdom of Douglas Self and other like JLH. I bet you anything you like the sound is very very different. The NAP140 clone might be useable if only Nyquist stability was the question. If that's right why are we pushing Camels through the eyes of needles in these debates? I suspect we might have to simplify the amp design to get good proof.
LTP with 1K-22K
High gain VAS with low Cobo
CCS to LTP and VAS
Bias that is stable
BD139/140 drivers
30 MHz NPN/PNP outputs.
No correction networks.
0R22 in place of choke
Zobel to suit.
+/- 30 V > 250 VA 2 x 10 000 uF.
I would suggest the output stage the Self prefers that allows the outputs to switch off faster. This looks different to the ready made Darlingtons. Although not a Naim clone this should still sound very similar.
LTP with 1K-22K
High gain VAS with low Cobo
CCS to LTP and VAS
Bias that is stable
BD139/140 drivers
30 MHz NPN/PNP outputs.
No correction networks.
0R22 in place of choke
Zobel to suit.
+/- 30 V > 250 VA 2 x 10 000 uF.
I would suggest the output stage the Self prefers that allows the outputs to switch off faster. This looks different to the ready made Darlingtons. Although not a Naim clone this should still sound very similar.
Member
Joined 2009
Paid Member
We know that the compensation can have a profound impact on the sound. My own experience is mostly with Cdom on a single-ended single-device VAS like the AKSA 55 amplifier. However, there's no way that this is the whole story. You can try and simplify all you like but there are many factors to take into account beyond the compensation scheme. Especially the perception of bass.
Attached: founds some parts in my junk box to warrant firing up the iron 🙂
Attachments
That reminds me. I know you don't have pcb slots but I would suggest strapping 10k resistors across your supply rails to discharge the caps. In the past I've accidentally blown parts when reworking boards I have just turned off.
Member
Joined 2009
Paid Member
The methodology is wrong. What do you want to prove? That different compensation will yield different sound? Of course they will.
Stability is not really an issue. There are other things to be achieved simultaneously (incl. Stability of course). Different cct (such as different transistor) calls for different compensation.
There are around 5 metrics to achieve, and because the compensation parts affect each other, there must be a minimum threshold for each of the metrics. For example, regarding stability metrics, what is the methods? Nyquist? Capacitive loading? What is the threshold? 50/12 for PM/GM?
The first compensation part to design is the base stopper! Once this is found/calculated, the next process is much simplified. After the base stopper then the input filter. Input filter is the most effective one. It can fix what others cant. And input filter has almost no drawback. Find the maximum input capacitance based on your thresholds such as hf slew rate or whatever.
Im not going to tell everything tho. The metrics is based on known Physics and the thresholds is based on experiments requiring good ears.
Cdom is about the only easy area of debate with the Naim design that could be looked at without changing the qualities of the Naim design too much.
I thought of a single input transistor and how JLH made it work with no output capacitor. I even made a design in the past that proved to me it was workable. 50 mV offset by calculation alone. Was less when adjusted.
Here is an idea I doubt you have seen. Douglas Self shows how two pole VAS compenstion can be either to the 0V or power rail. The obvious is the 0V as it should be quieter. Here is is to the output which is an equally valid low impedance anchor. However it also allows output stage negative feedback. I found with Self typical Complimentary feedback pairs the improvement at 50 kHz was questionable. For other class AB types there was a noticable improvement. The designs I worked on didn't have the correction circuits of the Naim so I have doubts for NAP 140. With simple MOS FET's a high degree of perfection was had with a very simple single VAS. I was able to use very small values of capacitor with 100R to output. Regardless of if useful I hope this interests you. I was very surprised how little trouble I had with it. It's hard to say if the bootstraping negates that it is two pole. Who cares if it works. The 50 kHz was to exagerate the results. It is certain to help lower down.
I thought of a single input transistor and how JLH made it work with no output capacitor. I even made a design in the past that proved to me it was workable. 50 mV offset by calculation alone. Was less when adjusted.
Here is an idea I doubt you have seen. Douglas Self shows how two pole VAS compenstion can be either to the 0V or power rail. The obvious is the 0V as it should be quieter. Here is is to the output which is an equally valid low impedance anchor. However it also allows output stage negative feedback. I found with Self typical Complimentary feedback pairs the improvement at 50 kHz was questionable. For other class AB types there was a noticable improvement. The designs I worked on didn't have the correction circuits of the Naim so I have doubts for NAP 140. With simple MOS FET's a high degree of perfection was had with a very simple single VAS. I was able to use very small values of capacitor with 100R to output. Regardless of if useful I hope this interests you. I was very surprised how little trouble I had with it. It's hard to say if the bootstraping negates that it is two pole. Who cares if it works. The 50 kHz was to exagerate the results. It is certain to help lower down.
- Status
- Not open for further replies.