Sound Quality Vs. Measurements

[Pano]

Thanks for the great info chaps, sorry i could not get back here earlier. now to pull the discussion back to thermal stability:

Originally Posted by john curl

...real designers make sure that the THERMAL CAPACITANCE is very high by close mounting the output devices on good insulators (if necessary) properly torqued down and used with thermal grease on a substantial heatsink . This removes most of the short term gain changes due to temperature.

Yes John, that pretty much what I was trying to express. Since Class-A runs a such at high idle current and idle temperature, just how much will short term bursts of music change that point? I.E., if the dynamic peaks are only equal to the idle current, the change in current flow won't be high in relation to idle current. Temperature should not change much. Thorsten seems to say that it will, but I ask how much as compared to a low bias amp?

Tubes by their nature stay pretty warm, even in class AB. How much will the dynamic peaks of music change that temperature? And what differences will thermal stability make?

[SoNic_real_one]

If the total measured THD of an audio chain (to the amplifier output posts) is below 0.05% there is no point in splitting the hair into odd-even harmonics. That level is low enough not to be a problem.
Limited bandwidth of a 44.1kHz sources is clearly audible on the sounds like in the original post. I found out that DSD or PCM 96kHz are needed for a "life-like" sound. Bandwidth and step/impuls response is clearly measurable.

Now, the speakers will never be in that range. I was amazed how much difference is between my speakers and a pair of decent headphones.
Measuring speaker's THD is a little more "delicate" to do because of their nature (pressure devices) and nelinear dependence of the actual level of electrical (input) signal.

As for thermal on tubes - anodes will heat up based on the current is flowing thru them (number of electrons hitting the anode). That will affect efficiency, noise...

[a.wayne]

Originally Posted by Pano

Thanks for the great info chaps, sorry i could not get back here earlier. now to pull the discussion back to thermal stability:


Yes John, that pretty much what I was trying to express. Since Class-A runs a such at high idle current and idle temperature, just how much will short term bursts of music change that point? I.E., if the dynamic peaks are only equal to the idle current, the change in current flow won't be high in relation to idle current. Temperature should not change much. Thorsten seems to say that it will, but I ask how much as compared to a low bias amp?

Tubes by their nature stay pretty warm, even in class AB. How much will the dynamic peaks of music change that temperature? And what differences will thermal stability make?

agree................ and my original point !!!

Originally Posted by SoNic_real_one

If the total measured THD of an audio chain (to the amplifier output posts) is below 0.05% there is no point in splitting the hair into odd-even harmonics. That level is low enough not to be a problem.
Limited bandwidth of a 44.1kHz sources is clearly audible on the sounds like in the original post. I found out that DSD or PCM 96kHz are needed for a "life-like" sound. Bandwidth and step/impuls response is clearly measurable.

Now, the speakers will never be in that range. I was amazed how much difference is between my speakers and a pair of decent headphones. Measuring speaker's THD is a little more "delicate" to do because of their nature (pressure devices) and nelinear dependence of the actual level of electrical (input) signal.

As for thermal on tubes - anodes will heat up based on the current is flowing thru them (number of electrons hitting the anode). That will affect efficiency, noise...

You should get better speakers, when correctly voiced they should sound pretty close to the headphones.

[Wavebourn]

Originally Posted by Wavebourn

Traditional means static measurements using sinewave generators, notch filters, AC voltmeters. However, spectrum analyzer is the way to observe dynamic behavior, but you can't name and calibrate all possible dynamic flaws of all possible topologies, to define a new tradition of measurements.
And even if you do, how many would agree with your new standards? At least, Stereophile always has own opinions, independent on engineers'.

Can't edit: I can use signal generator, DVM, oscilloscope, CD player, to observe behavior of the amp I design and to find potential flaws and directions where to go, but I would not call it traditional measurement. I would rather call it a specialized, topology - dependent test.

[SY]

Originally Posted by a.wayne

could you elaborate more on your setup ..? what is an a typical PC setup.?

M-Audio 192 card in a cheap HP PC, which allows both balanced and unbalanced I/O as well as SPDIF I/O. Street price is ~$150. Bought it on Amazon. AudioTester, REW, Rightmark, HOLM, Soundeasy software. AT was about $40, the others were free.

Two interface boxes, one is Pete Millett's, the other is my own design (BUF03s, noisy but superbly low distortion and source Z). My box is set up to use standard (1x, 10x) scope probes.

I have conventional gear as well for backup and verification, but that seems to get used less and less. I'd love to have an AP2, but with a little elbow grease and one-off jigs, I can get just about all the measurements that an AP2 is capable of with the PCI card.

[SY]

Originally Posted by Wavebourn

Traditional means static measurements using sinewave generators, notch filters, AC voltmeters.

IOW, using your grandpa's gear.

A tacit point that you've made is that the ability to determine which measurement to make, what the error sources are, and how to interpret the measurement is thousands of times more important than the brand of equipment.

[vanderpol4]

John,
I am sure you balance steady state harmonics perfectly, given the topology and noise constraints.
What I pointed out was that given time series of "pink noise" will have a fourier transform; after amplification that fourier transform will be modified; against known human preception of that harmonic content of the transform S(w)amp/S(w)o there may be a measurable FOM that sorts ampifiers by their percieved quality.

In summary The point is pretty simple; why can't there be an accepted measurable quantity that can be used to gauge audible "fidelity"? If I look at THD and IMD S/N for:
a 1970's yamaha reciever
a 1970's kenwood 40WpC integrated amp
a 1960 HK Cit II
a Carver"the receiver"
a 1990's denon pre-amp
a pioneer super tuner car radio
A pair of PP monoblocks
Any modern amplifier "high" or "low" end
I can't "see" what I hear or have heard over the years . Further the feedback topology of the HK CITII suggests you can't "hear" the tubes, however I certainly know when the splitter pentode is aging, ( not sure if there is a measurable THD IMD with this though).
I also know a single ended triode preamp sound is recognizable, as this is DYI audio what I really want to know is ( so I clean up my preamp!):
a) why does a high NFB class A preamp sound "sterile" what measurement is sterile to musical?
b) what do I measure and minimize in the triode class A preamp to have my cake and eat it to?; that is a musical combination and no syrup colorations.
Do I move the B+ up? drop the 12AU7 stage? move the S.E. 12Ax7 final stage to SRPP? Do I take gain and feedback? How much? How do I know? What do I measure? What figure of Merit do I use?
What if I wanted to switch preamp? what FOM do I look for? That one I know: there is no published marketing info that tells me, not slew rate, not THD, not IMD. It is all word of mouth and audio trial and error.
c) Does a different topology change the sound/noise e.g. if I totem pole either the 12au7 or the 12ax7 stage will I "improve" the amplifier fidelity?
That is what I am looking for, a measurable way to gauge my amplifier mod or review known topologies and compare the fidelity of each.
Lou

[Wavebourn]

Originally Posted by SY

IOW, using your grandpa's gear.

A tacit point that you've made is that the ability to determine which measurement to make, what the error sources are, and how to interpret the measurement is thousands of times more important than the brand of equipment.

Exactly. But we still read here about THD and DF comparisons

[SY]

Only from people who believe that nothing has advanced since 1968.

[a.wayne]

Originally Posted by SY

M-Audio 192 card in a cheap HP PC, which allows both balanced and unbalanced I/O as well as SPDIF I/O. Street price is ~$150. Bought it on Amazon. AudioTester, REW, Rightmark, HOLM, Soundeasy software. AT was about $40, the others were free.

Two interface boxes, one is Pete Millett's, the other is my own design (BUF03s, noisy but superbly low distortion and source Z). My box is set up to use standard (1x, 10x) scope probes.

I have conventional gear as well for backup and verification, but that seems to get used less and less. I'd love to have an AP2, but with a little elbow grease and one-off jigs, I can get just about all the measurements that an AP2 is capable of with the PCI card.

Thanks i have a spare PC , this has opened up some options ....

http://www.amazon.com/dp/B00064ABSG/...l_6zsc1krie4_b