Not a single graph in that presentation had anything to do with 1/f noise, or did I miss something.
I suppose you have heard about 47 Laboratories Gaincard. The main argument thy used to explain good sound was extremely short feedback path(the feedback resistor soldered directly to the chip pins).
It hurt my head a bit, but I noticed something about some noise being down 10dB. But no, nothing about 1/f that registered for me.
I never heard anything about that 47 Lab.Gaincard ,
but they can claim anything they want trying to explain some particular good sound from that particular AF product ,
Fb res.soldered directly on the chip pins , hm ....I`m very suspicious on that sort of argument,
actually IMHO if that AF circuit works&sounds good only if that Fb loop must be so short than that basic design is completely wrong .
Last edited:
I think its safe to say that was boutique audio to the max though, with liberal splashes of marketing frou frou.
Page 45 : Nach dieser Messung ist von einer Unterdrückung sehr hochfrequenter Signale auszugehen und von keinerlei Veränderung im Audiobereich.
Page 46 : Man kann eine zwar geringe, aber nachweisbare Verbesserung darstellen. Diese ist durchaus hörbar!
-96.5dB = 0.001496 %
-97.5dB = 0.001333 %
THD+ Noise difference : <0.00017 %
Meiner Meinung nach : Totaler Quatsch !
Last edited:
No they saw no evidence of the claimed performance so they made something else up. As JC said he's not surprised neither am I.
I have a copy also. They talk about the stripe and outside foil and other static effects, but then being the true believer that you are unquoted anecdotal evidence is fine when it supports your beliefs.
And that would be due to piezoelectric capacitors, not the issue.
Now for those who admit they didn't look at the slides here is my take...
"A PDF copy of these slides (<2 MB) is available by sending your email address to BRUCEH@AP.COM"
First he shows basic Taylor series and how to model many distortion mechanisms using an emitter follower. This is very instructive as it shows even this topology used in virtually all power amplifier output stages adds distortion.
He then gets into resistor distortion which he points out the thermal part is frequency dependent. Now this was covered in this thread back around post 5000, also in an article by some fool in Linear Audio Volume 1. (It also mentions Scott W who constantly brings up the typical -120 dB resistor distortion figures at 1,000 hertz, ignoring that it increase as frequency decreases.) This is a more in depth thermal model that has been show previously. He also notes an issue with the bulk metal resistors. How ever he only sort of touches on the phase of the distortion produced by this thermal mechanism.
(But his bit does offer a reason for naked Vishays to sound different from standard ones. Who knew? Oh yeah the golden ears folks....)
He then goes into voltage coefficient of resistors which does not model well with Taylor series.
He of course mentions that proper circuit topography reduces the effects of resistor distortion. That is followed by a survey of resistor types. No surprise here he gets the same results as mentioned here before. Thin film types have the best temperature coefficients and a voltage coefficient of .1 to 1 ppm. Mentioned also is that all resistors that can contribute significant distortion should be used at a fraction of rated power. What I didn't see was a mention that wirewound resistors have an issue with how the leads are connected. That was covered here way back, there can be thermocouple effects along with nonlinear termination issues.
Now in the bit on wirewound resistors he mentions modulation noise, but this is one of the elephants in the room that most folks try to ignore. Covered is when using multiple resistors on a common substrate there may be cross coupling of the thermal issues.
The presentation then goes on to capacitors, listing all the common types and their issues. Here he does get into the nonlinear ESR termination issues. As most of these use a metalization spray our local expert on this is Pavel.
All ceramics except NPO are cautioned against along with using too low a voltage rating. He totally diss's piezoelectric ceramics. (Not even a mild surprise to the ears only folks.)
To me the good part is where he gets into microphonics caused by DC voltages and Delta C. This was also covered in this thread way back by the same resistor article guy who did a piece on that effect that ended up in Elektor.
The really interesting part is not just about voltage coefficient and why to use a much higher voltage rating, but where he compares NPO ceramics to a film and foil part.
Now I'll pretty much skip his bit on inductors except to mention that he does show the distortion caused by running current near magnetic materials and how aluminum can be an effective shield. But of course the Golden Ear brigade has long claimed aluminum is the way to go. J.C. has even mentioned his insistence on aluminum is some case parts.
The section of opamps is something best left to the reader.. Although he does mention the common audio technique of forcing class A outputs. He does mention power supply coupling issues, but I didn't see anything about RF sensitivity. (Also covered in this thread.)
Next up is the noise issue and he shows how to calculate a complete system's noise with an excel spreadsheet.
Now the bit on layout covers keep heat sensitive parts away from sources of heat and pay attention to things like trace resistance, inductance and coupling. Marce can carry on about these issues. The next bit about cross coupling and power supply busses has been covered hy by R.M.N. including his bit about using a single cap across power busses to localize current loops at the source.
The heart of the presentation concludes with some layout examples that would make J.N. proud.
The conclusion is that if you aren't aware of all these tweak issues then analog design can become a lost art.
Now what I think is missing is what Wayne points out. Measuring what an AP can give you is only the start of finding high end parts. The current AP systems are FFT based. Gizmos like the Shibasoku use a different method to measure harmonic distortion that can actualyy go below FFT based units. However harmonic distortion and even I.M.D. only tell part of the story. The bit about noise modulation (also a nice picture of that in this thread...somewhere) is at least a hint there probably is more to audio quality perception.
But the great part is that it is not 75,000+ posts!
Are you honestly trying to say microphonics in capacitors has no reference in the literature? DC voltage and delta C sounds like a condenser microphone to me.
And you continually measure resistors at a significant portion of their rated power and ignore the applications where this is simply not so. The effect goes as V^2.
None of this stuff is unknown to folks who have been designing discrete modules for 50yr. Now get Bruce to say the feedback goes around and around and I'll listen.
Last edited:
No Scott, it is you that goes round and round... But really why do you even mention feedback fallacies other that to imply that is an issue I have raised...
Someone recently mentioned the gain card that attributes it's sound to the tight circuit wiring, not that they use a carbon comp 1/4 W feedback resistor.
Peak musical energy occurs around 50 Hz. That alone gives 26 dB more thermal distortion than my 1,000 hertz measurements. I measured the distortion at 1/4 power. Yes there are amplifier designers that do rate the feedback resistor to just handle the maximum signal. That would add 12 dB more distortion. So from my measurements a signal KOA carbon film resistor would distort at -76 db. third harmonic distortion just by itself. I suspect most amplifiers use more than one resistor and in phase distortion adds.
Last edited:
- Status
- Not open for further replies.