Sound Quality Vs. Measurements

[dvv]

Originally Posted by jacco vermeulen

Unfortunately, Siemens hasn't done semi's in a dozen years, and Infineon does not manufacture them.
NXP still has the Philips BF720/721 versions on active duty, afaig.

That works for me. The only version of those transistors which were a bit odd were by Zetex, they gave me a hard time, I think their output capacitance was quite bit above what they say and what's in the Data Sheet, but who cares, I got rid of them.

[a.wayne]

Whats the difference's here in relation to micro-watt distortion and that all important first watt ..... ?

[simon7000]

Well a bit on topic is how to measure amplifier distortion below the noise floor. Although you can try and average out the noise, you might loose something. (I don't want to get into sampling theory and finite limits so a bit of simplification).

A better technique is to use two identical amplifiers and measure their outputs and then do a cross correlation. Now you can do lots of averages this way and get to what is actually the distortion or other artifacts common to the design and reduce the noise errors.

But then I am giving away secrets on another project!

[nigel pearson]

Spray locally with liquid nitrogen would be invalid ? Assuming DC conditions don't change .

I have often thought of it as I have access to it . I could never imagine where it could be used .

[simon7000]

Originally Posted by nigel pearson

Spray locally with liquid nitrogen would be invalid ? Assuming DC conditions don't change .

I have often thought of it as I have access to it . I could never imagine where it could be used .

Well you could try to start drinking cold beer.

[a.wayne]

Interesting List, The first i agree with , especially after having much of those at one time or the other ..



The Ten Most Significant Amplifiers of All Time | AVguide

[dvv]

Originally Posted by a.wayne

Interesting List, The first i agree with , especially after having much of those at one time or the other ..



The Ten Most Significant Amplifiers of All Time | AVguide

I would question only the Phase Linear amp. I agree it made NOMINAL (but not real) power affordable, and that is its only clame to fame. In terms of sound, the instant the load dropped to 7.9 Ohms, it was in trouble. And I found its sound to be so-so at best.

Carver could do much, MUCH better, and did, later on. I never doubted his savvy, just his financial motives.

Instead of it, I'd include Harman/Kardon's Citation XII on the list. It must rate as one of the most copied amps in the world ever, and even many years after its demise, people like Nelson Pass were tuning them up with MOSFET output devices and so forth.

Also, I note that their Citation XX, the most elusive power amplifier of them all, is missing from the list. Not many were produced which could deliver 500 Amps of current, even if in impulses only. And with so little overall NFB. Anyone who thinks that's no big deal and is irrelevant, please feel free to show us your construction bettering it.

[dvv]

Originally Posted by simon7000

Well you could try to start drinking cold beer.

Yeah, your slew rate improves no end.

[scott wurcer]

Originally Posted by simon7000

Well a bit on topic is how to measure amplifier distortion below the noise floor. Although you can try and average out the noise, you might loose something. (I don't want to get into sampling theory and finite limits so a bit of simplification).

A better technique is to use two identical amplifiers and measure their outputs and then do a cross correlation. Now you can do lots of averages this way and get to what is actually the distortion or other artifacts common to the design and reduce the noise errors.

But then I am giving away secrets on another project!

I doubt that the distortion characteristics at that low a level correlate that well. I would prefer 4 million point FFT's.

[simon7000]

Originally Posted by scott wurcer

I doubt that the distortion characteristics at that low a level correlate that well. I would prefer 4 million point FFT's.

I stole the technique from NIST! They used it to measure better than -200 db re 1 V! (100,000 averages.)