[audio-heaven]

Ah and there I was thinking of painting my NC400's bright pink with the words SUPER TURBO written across them, would that harm the performance? I hope not I'm really REALLY keen on making them look SUPER COOL!

Please don't laugh but I have a theory that the color pink is the very best color "in the quantum universe" at suppressing RF and EMI emissions at the atomic and sub atomic level....discuss.

P.S If you would like to start using my pink paint on your Ncore's Bruno I have patent pending so I'm afraid it's going to cost you

ADVERTISMENT: Quantum Pink paint (patent still pending) with free SUPER TURBO sticker pack ONLY $299.99 per fluid ounce. I'm almost certain that it will improve anything beyond your wildest dreams. STRICTLY NO REFUNDS ACCEPTED

[cab]

I see I have so much to learn from the experts who really know what matters: big caps and fancy materials, the sort of things that are simply over the head of the poor bloke who merely got the thing to work.

Still, from this humble position I would like to offer some interesting tidbits:
*the power rails are LC filtered. Not much PWM current gets out through the supply wires, it's mostly just the average (audio frequency) component. It's called "decoupling". As in "making sure circuits are no longer coupled".
*The electrolytic capacitor bank was designed to have a damped impulse response. Also, the series inductance of the capacitor bank as measured at the power stage works out as <1nH. The large film capacitors shown in this thread have an inductance that's fully determined by their size and attached wiring, and are probably more like 100nH. They won't be very effective bypass caps. That's putting it mildly. To get low inductance, get a small loop area which is not compatible with big components.
*DA (soaking) is irrelevant in DC applications. Besides, it isn't in any way improved by adding a small low DA cap across it...
*FET linearity is not dependent of heatsinking. FETs just aren't linear at all. Besides, Ncore is a class D amp... Not much sense in making FETs more linear then.

Give me some credit. There's still some wisdom left in this old brain.

There you have it from the horse's mouth.....Thanks for taking the time to clear the air a bit....

[poseidonsvoice]

Bruno,

Unfortunately, even with the warnings that have been repeated over and over on diyaudio, the modifiers continue to encroach upon your work with a very limited understanding of the ramifications of doing so. Moreover, they are willing to go against a young 39 old who has quite a few patents to his name especially within the field he is working in (unlike other high end audio pundits)!

All I can say is...thank you for doing what you do. We are nearly of the same age and I respect a man who not only provides his designs at such low cost to the diy community but also defends his product without pride nor prejudice, just pure facts.

Best,
Anand.

[jmbulg]

Bruno,

Unfortunately, even with the warnings that have been repeated over and over on diyaudio, the modifiers continue to encroach upon your work with a very limited understanding of the ramifications of doing so. Moreover, they are willing to go against a young 39 old who has quite a few patents to his name especially within the field he is working in (unlike other high end audio pundits)!

All I can say is...thank you for doing what you do. We are nearly of the same age and I respect a man who not only provides his designs at such low cost to the diy community but also defends his product without pride nor prejudice, just pure facts.

Best,
Anand.

+1 

[serengetiplains]

Bruno, I'm sorry you feel disrespected.  From what I see, you're a genius in electronics design, and on the very cutting edge of class D development, which I think is the cutting edge of audio advance.  It looks to me that (and please bear with my language) digital is the future for audio generally, and class D, being a digital-friendly form of amplification, is the future of amplification.  I really can't see it otherwise.  Looking into the future, I sense that tubes and class A will remain as defined niches, but their advance, from what I can see, is now limited, as it has for some time, to component improvement---better amplifying devices, better passive components, but nothing fundamentally new on the horizon and therefore only incremental advance. 

And yes, that post-modern levelling is finished.  People are getting with that by moving into a systems awareness (the Libor collusion, among many other things, will put the last nail in the PM program).  In audio, that means a complexity amenable to system-based understanding that incorporates all usable, successful discoveries of the previous class A and A/B era in a leap to something new at a more fundamental level.

So I really think you're on the cutting edge and respect your knowledge, ability and genius.

I am nowhere near a genius in electronics soldering let alone design.  But I have other experiences I have come to trust and that have worked for me in my similarly dedicated 30 years of this pursuit.  I'm just putting together some pieces here the most of which couldn't realistically be incorporated into a commercial design.  There is, in any event of that, no need to imply that my or others' experience is simply valueless.

So far as large capacitors are concerned, my speculation is that these capacitors operate at audio frequencies.  Lead inductance seems to me to ensure these capacitors will operate only such.  I find using them to be an overall improvement.  Among other reasons, there is only (from brief memory) 250uF? proximal to each channel's output.  I have heard obvious effects of using film bypasses across much greater uF electrolytics.  I also heard these effects on the Ncores.  Speaker driver control improved and distortion decreased allowing me to listen at full volume while thinking I could turn the volume up considerably more.  "That's all?  Too bad."

I agree with you that DA is irrelevant to DC.  But capacitors only operate on AC, and DA is relevant to any capacitor.  And just like there is no digital amp (music is analogue), there is no pure DC.  AC's everywhere.

In any event, thank you for your work and your brilliance.

[serengetiplains]

And forging at the cutting edge, you obviously have and will experience attacks from many angles.  That's not what I'm doing here.  I'm not, and I don't intend by implication to be so doing.

Looking at DA, which seems not fully or perhaps even poorly understood, if one uses the Cole-Cole method of calculation, wouldn't the effective frequency of DA release be much lower than the trigging signal to which a given release relates?  If that's true, good capacitors operating at audio frequencies+ may be just the thing for a significant DA cleaning of the rails.  J'speculate.

[serengetiplains]

Oh yes, and class D, to me, differs from class A and A/B by being what I like to call a dual technology.  Not to start an endless semantics loop, but to make a point to say something I find interesting, class D is to me analogue and digital.  It is digital insofar as it, for one part of its operation, renders the analogue signal a PWM pulse train.  From Bruno's patent:

Pulse width modulation is a technique that can be used for converting an analogue signal into a binary signal ...

 

At that place of the binary codification of the analogue signal, one has latitude to, among other things, apply digital algorithms, so by definition a class D amp has access to a range of possible enhancements and solutions unavailable to the mono-technologies of classes A and A/B, C perhaps included, whatever the hell C is.

Dual thinking is quantum thinking: both.  Both is always already a systems view that manifests, in audio, imho, as combining boths where they may lay: digital-analogue (detail-in-wave), electrical-magnetic (ditto), etc.

And quantum thinking/viewing/science is simply more powerful than classical-only push-pull energetics.  No more post-modern levelling: more powerful is more powerful.

There's my 50000 foot view to supplement my DA and other ramblings. (both )

[Rclark]

Serengeti said it himself a few pages back: "I barely understand how this amp works".

That said, he is enjoying the hobby at a level I never will, nor will most here. He is a gentleman of the highest order, and is merely playing in his sandbox, enjoying himself. He is messing with them simply because he can.

 It seems Bruno has changed the entire paradigm rather violently, in every way possible. I've made arrangements for a fully built pair and look forward to them, and they will be my second ever hifi amp.

 But I recognized the achievement instantly: $2000 built, modules the size of a coaster delivering up to 1200 watts@4 ohms bridged with supreme alpha dog measurements, sipping a few watts at idle, beating and matching far more expensive amps. Beware Ncore and the eventual Ncore II, etc. Thing's done gone' and 'diffrent now.

 The vast majority of the reviews so far are in high end systems and the owners positively laud this amp. People with decades of experience and owners of truly expensive gear.

 I count myself a fan.

[serengetiplains]

Thank you, R.  I think these amps must be practically unchallenged in what they do well.  Really good stuff.

[bruno]

@Serengetiplains don't worry I wasn't feeling you were disrespecting me. I was responding to several posts at once. It was someone else who irked me a bit by his style of delivery (can't remember the name). And indeed people should not feel stopped by theory to try out things. The important thing about theory is that it shows what is known, what is likely and what is unlikely. Once you find something completely unexpected by experiment you can try questioning theoretical assumptions. That's what experiments are for.

Usually they're not so much fundamental things but merely overlooked consequences of established theory. So far I found most audio "mysteries" to be mysterious only when theory is oversimplified. For instance, people who expect wiring to be of no importance will stay baffled until they stop confusing actual theory with simplifying assumptions. The converse doesn't help either of course, people who latch on to the first "explanation" they find and who then attach their whole world view on that (the wire metallurgy brigade springs to mind).

So when you say you expect adding caps to do something at audio frequencies, I cannot rule that out. What theory does say is that the soaking hypothesis (not sure who brought that up) would be an unlikely explanation for any audible changes. Some of the anecdotal reports suggest that many caps add the same character whether used as coupling or as decoupling. What that indicates is that soaking might not be the principal source of "cap sound". Whenever I hear soaking touted as "the answer" I go "think, boys, think"...

[serengetiplains]

It was I who raised the issue of capacitor soakage (Pease's term).  Whether soakage or DA accounts for the principal source of capacitor distortion is I think open to question, but what can be said is capacitor dielectric effects---a more general umbrella term for capacitor distortion differences---show up consistently for different dielectrics in test results, showing characteristic distortion spectra for the various dielectrics in question. 

As to series/shunt uses of capacitors, Bateman measured distortion differences between series vs. shunt uses and found those differences to be minimal.  An example of this is shown below, with two diagrams also appended to show the test setup he used.  In that setup, Bateman simply switched the position of the series resistor and DUT to change the position of the DUT from series to shunt.

[serengetiplains]

Notice that Bateman specifies film/foil polystyrene capacitors as the coupling caps in his test setup.  He found that the resolution of his measurement instrument was limited by the distortion (resolution) of these capacitors.  That makes sense.  One cannot measure something that is beyond the measurement capabilities of the machine in question.  Analogously, one cannot photograph anything smaller than the highest usable frequency of the light used in the photographic setup.  Quantum physicists know this problem intimately.

Here is a quote from Bateman discussing this issue, followed by his ruminations on parts quality:

I started by carrying out some initial capacitor intermodulation tests. Experiments involving simple harmonic distortion testing revealed easily interpreted differences when testing less good capacitors. Testing good capacitors however confirmed that my existing signal generators introduced far too much distortion.

[gnomon]

@Serengetiplains don't worry I wasn't feeling you were disrespecting me. I was responding to several posts at once. It was someone else who irked me a bit by his style of delivery (can't remember the name). And indeed people should not feel stopped by theory to try out things.

bruno, serengetiplains - what a treat to find such cordiality in a discussion on such a contentious topic.  I last saw the like almost a decade ago.

Thank you both.

[serengetiplains]

Tip of the hat to you, gnomon.

I was trying to imagine how DA might manifest in a simplified circumstance.  By DA, I mean capacitor distortion attributable solely to the workings of the dielectric in question.  Other forms of capacitor distortion or non-linearity---ESL, ESR, DF, microphonics, etc---of course manifest without rest, but here I want to focus on this matter of charge retention.

DA measures the rate at which a capacitor regains its charge after having been discharged.  That charge recovery is to my thinking a quasi-active source of spurious voltage that adds to or subtracts from the capacitor's output voltage---whether a signal voltage in the case of signal transmission, or DC stability in the case of output power rails.  The capacitor output will modulate at frequencies of the dielectric's characteristic voltage recovery as that voltage appears progressively on the capacitor's output.  From what I can discern, DA voltage frequencies, as it were, are much lower than the original voltage charge to which that recovering DA voltage refers.

Here's something I find interesting.  DA is usually described as a charge recovery.  I suspect that the mechanism responsible for charge recovery---a kind of dielectric response-sluggishness---operates in both the charging and discharging process.  If this is true, it implies that a capacitor having a given DA will not only discharge spurious voltages after a charge, it will also not receive the charge's instantaneous value.  I think the latter implies that charge peaks are to the extent of dielectric sluggishness clipped.  The dielectric simply hasn't the quantum-molecular speed to absorb the full charge before that charge changes---the more for a fast-moving transient.  High DA capacitors have always sounded dull to me, electrolytics being the worst given their reigning status as the highest DA capacitor type.  I think dielectric sluggishness, working in both the charge-receiving and -discharging functions of the capacitor, accounts for at least part of this.  Slow in and muddied out, and where did the subtle information go?

I look to more than listening impressions and theoretical implication to substantiate this idea of charge-receiving sluggishness.  I performed tests on capacitors where I charged them for increasing times and measured the recovery voltage.  In this test, I charged a 3.3uF PIO capacitor for successively longer periods of time, increasing the total charging time by 5 minutes for each next charge.  Interestingly, the DA recovery voltage hit its highest level at 70 (!) minutes.  This, to me, implies that the capacitor only slowly receives (imprints with?) the uppermost level of the original charge.  A 70 minute transient is not what I'd call a transient. 

In looking at how I conducted the experiment, I could have given greater time between successive charging cycles, but DA recovery to peak recovery voltage was fairly quick, and that recovery voltage then subsided just as quickly (at most, tens of seconds).

Also, on the numbers, peak recovery voltage represented almost 2% of the original charging voltage.  This number should be increased given voltage leakage of my measurement setup.

Here are the results:

[serengetiplains]

I was looking at the data of DA tests I performed.  From this data, I noticed two variables in respect of DA recovery voltage.  The first was the peak recovery voltage attained.  The second was the time it took for that peak to be reached.  Looking at the data through squinted eyes, I see the following patterns:

1) peak recovery voltage is higher the higher the capacitance

2) peak recovery voltage is higher the lower the DCV rating of the capacitor

3) peak recovery voltage is lower the lower the dielectric constant

4) time of peak is higher the higher the capacitance

5) time of peak is higher the higher the dielectric constant

All other things being equal, then, the lower the dielectric constant, the higher the dielectric DCV rating and the lower the capacitance, the lower the peak recovery voltage and the time peak is reached.

1) shows that DA recovery voltage magnitude relates directly to the capacitance of the capacitor.  This implies that DA recovery voltage is based on area, as greater capacitance means greater area.  Greater capacitance also means greater volume of dielectric being acted upon.  But area seems the true measure.  Bateman explains that DA recovery voltage is greater the greater the stress placed on the dielectric as measured in volts per micron of dielectric thickness.  Volts per micron decreases when dielectric thickness increases.  I found this in 2) above.

DA recovery voltage therefore looks to be a product of the k of the dielectric, the area of dielectric involved and the dielectric thickness, something like:

                  (k x area)/thickness

Time of peak seems likewise a product of the above equation.  Though I don't have enough data to see a reliable pattern with the thickness element of the equation, dielectric thickness seems inversely correlated with time of peak.  Time of peak otherwise correlates with k and area.

One outlier in the data concerned capacitors with combined liquid/solid dielectrics.  Typical oil capacitors tend to display large peak recovery voltage but a comparable time of peak comparable to other dry capacitors of the same capacitance.  Thus a Jensen 3.3uF PIO capacitor had a recovery voltage (~485mV) 750 times that of an RTX 3uF polystyrene (0.646mV), but the two reached that peak in approximately the same time (55 seconds for the Jensen, 45 seconds for the RTX styrene).

The oil used in oil capacitors, not to mention the paper, typically has quite a high k (4-8 depending on the paper or oil used).  The increased DA recovery voltage for high k dielectrics shows in my data.

The worst type of capacitor is a large electrolytic.  I tested an expensive Black Gate 1,000uF non-polar FK 50V capacitor.  It's recovery voltage was about 3X that of the Jensen, but its recovery time was 60X that of the Jensen.

Bateman speaks to the better distortion performance of non-polar electrolytics over that of polarized types.  A 100uF WKZ 500V Black Gate gave a peak recovery voltage 5X that of the 1,000uF FK (time of peak was not recorded).

It seems to me that both the absolute peak of recovery voltage attained and the time it is attained affect a capacitor's audible performance.  The best capacitors, to my ear, perform best in both respects: they demonstrated both lower peak V and lower recovery time. 

[serengetiplains]

I think the above gives a decent view of DA.  My own experiments, limited as they were, follow results predicted for DA (Cole & Cole) and largely conform to experiments Bateman and others performed to test DA across differing dielectric types.  Generally speaking, DA is worse the larger the capacitor and the higher dielectric constant.

How might this understanding apply to the Ncore?  Here's my hunch.  The Ncore draws a series of pulses from the power supply.  Those pulses induce, among other disturbances, pulse-like voltage variations on the supply's complex dynamic impedance.  The current pulses are for their part square wave composites of the amplified analogue waveform + switching harmonics.  The psu voltage variations they induce will be a distorted version of those pulses.  It is these variations, among others (like mains noise etc), that the electrolytic capacitors are placed to filter.  Anything not filtered passes directly into the speaker.  Because psu variations follow the music signal to some degree, and in other ways bounce around in some complex fashion (= a fashion that, imo, is sufficiently high to escape characterization, therefore not random), their presence in the audio band will muddy and distort the reproduction, quite audibly so.

I know the Ncore's feedback, in its limited way, addresses part of the distortion generated by these supply variations; so, too, does differential cancellation (PSRR).  Because these are limited remedies, and because feedback introduces problems like higher-order distortion, the more one can eliminate noise from the rails before feedback and differential cancellation the better.  I follow the rule that any node in a circuit is in its most general sense a voltage divider.  The less noise appearing at any given node, the less is passed through to whatever next stage (or speaker) that node feeds.

All other things being equal, it seems to me that audio-band supply noise is perhaps the worst for class D amps.  In addition to the normal range of supply disturbances affecting audible frequencies, class D amps also have the following:

• Switching noise modulating audio and audio+ frequencies.  I maintain that frequencies beyond 20KHz are audible in some form of sensory perception.  I can personally hear these frequencies as something of a pressure affecting my hearing, particularly where those frequencies are incoherent with the musical signal.  I hear this effect with SACD and PWM amplifiers, and I suspect the use of feedback complicates the type of noise present (reference Pass' article on feedback-related IM effects).

• DA noise.  My tentative theory concerning DA noise in class D amps is this: because DA recovery voltages are fractions of the original inciting frequency---the UHF pulses---this noise looks undoubtedly to reach downward into the audible (including audio+) frequencies.

It seems to me that part of what I hear wrong with class D HFs---their opacity, lack of sparkle and fineness, etc---traces to the above sources of noise, which are largely unique to class D design.  The level of this noise looks to me to trouble just those HF subtleties I find missing in class D, including Ncore, reproduction.

And fwiw, and apart from DA, Ncore output capacitance looks to me insufficient to filter any such audio-frequency voltage variations.  At 100Hz, 250uF has a capacitive reactance of about 6.4 ohms.  Depending on the speaker's impedance, noise appearing at or near that frequency will virtually halve between the capacitors and the speaker.  I'm uncertain what capacitance the Hypex smps has after its output, but from memory it seems that most of that capacitance resides one stage back in that supply.

______________________________

There's my long answer---the many posts above, including this one---to why I prefer to bypass output capacitors in this amp.

[Andre2]

I'm glad you did not bring up tensor theory on curvilinear coordinates on the above...

[cab]

You might find a more receptive audience for your thesis on diyaudio.com

[serengetiplains]

Shall I take a poll?

[cab]

Maybe I should have said "informed" instead of receptive.....