John Curl's Blowtorch preamplifier part II

[john curl]

Unfortunately, you are correct, Wavebourn.

[fas42]

Originally Posted by john curl

We are still trying to know what is still unknown as to WHY some amps sound ok, yet measure lousy, and visa-versa. '-)

AFAIC, it's not the THD number that matters, but the spectrum makeup of the distortion. Tests have been done where staggering quantities of distortion are injected, in particular down at the bass end, and no-one's the wiser. One amp can have 100 times lower overall distortion than another, but it's particularly obnoxious distortion in character, and the ear has no trouble picking it up and saying it doesn't like it ...

Frank

[Kindhornman]

fas42,
It sounds like we would have to have a weighting system for harmonic distortion based on where in the 20 to 20Khz frequency curve we are. You are saying that at the lower frequencies that the distortion is not as important if I am following you here. At what point would that change to being an acute level that even at low percentage that we would clearly notice harmonic distortion between at least two separate tones.

[morinix]

Huge amounts of 2nd order harmonic distortion are nearly un-noticeable. In the RDH something like 3 or 5% THD is stated to be acceptable; That's 3 or 5% TUBE distortion though.

[fas42]

This was quite eye-opening, when it was first pointed out to me: Experimental Study : Distortion - Axiom Audio ...

Frank

[Wavebourn]

Originally Posted by fas42

This was quite eye-opening, when it was first pointed out to me: Experimental Study : Distortion - Axiom Audio ...

The next time they need better speakers for such research. At least, woofers are needed much better.

[fas42]

Originally Posted by Wavebourn

The next time they need better speakers for such research. At least, woofers are needed much better.

In what way precisely would "better" woofers have changed the results?

Frank

[RNMarsh]

researchers have determined that harmonics can be detected down to .05%. Masking from system distortion (speakers) prevented knowing the true lower limit of hearing ability to detect change due to harmonics. I dont recall the bass only data but it would be higher, of course. For the electronic evaluation of an amplifier we would use the threshold (<.05%) to be sure no one could hear the distortion we measure under any condition. I would go ten times lower as we dont actually know the 'real' distortion detection threshold is. In listening tests, a different critia could be applied; maybe a more loosely allowed level of distortion due to masking et al the issues that are Not electronic in nature.

JC has mentioned there are already proposed weightings to individual harmonics. I suggested doing same to FFT tests/plots... to weight the harmonics being shown. Several others have done work in this area in the past and show the amount of weighting to apply. -RNM

[Wavebourn]

Originally Posted by fas42

In what way precisely would "better" woofers have changed the results?

Distortions, obviously.

[RNMarsh]

Originally Posted by Wavebourn

Produced in high quantity things designed on computer and never breadboarded are very cheap, especially when they made as cut and paste from datasheet examples. Similarly, produced in high quantity copies of some known designs. But how much will cost say an unique headphone amp for unique headphones? A fortune. You never know how much time is needed to design something unique and exceptional. And it requires unique skills that are not cheap.

Definitely. :-)