[Jedi]

Examining a plot of a typical amplifier's THD percent versus power output, typically shows a sharp increase in distortion at both ends of the power spectrum -- steeply rising distortion at low power levels of less than one watt, and again at the upper end of the amplifiers power output capability. Is the reason for the rise in distortion at the lower power end to do with the noise level floor becoming significant as a percent of the output signal, or is there some other explanation? I'm told most listening, in home systems with moderately efficient speakers, occurs at levels of around one watt per channel or less. This being the case, should one be concerned about an amplifier's distortion behavior below one watt. Should this be a critical point to focus upon when comparing specifications across amplifiers and receivers, or is it generally inaudible?

 

[Jedi]

bump

 

[tubeguy44]

yes.... noise floor does intrude at extremely low volume levels.....


generally, it is not an issue because of enviromental factors.....


unless your room is all hardwood floors, hard walls, lots of windows and with little furniture and the furniture you do have is leather..... (a room that would sound terrible btw)..... you will not hear that noise floor.....


not to even mention refrigerators, furnaces running, air conditioners.... heck - even projectors for projection tv have fans built in...... the enviroment can be rather noisy....


a room with the correct treatments and/or carpet and the more typical furniture should have no problem in "losing" that low level of noise floor....



cheers!

 

[M Code]

Quote:

Originally posted by Jedi Examining a plot of a typical amplifier's THD percent versus power output, typically shows a sharp increase in distortion at both ends of the power spectrum -- steeply rising distortion at low power levels of less than one watt, and again at the upper end of the amplifiers power output capability. Is the reason for the rise in distortion at the lower power end to do with the noise level floor becoming significant as a percent of the output signal, or is there some other explanation? I'm told most listening, in home systems with moderately efficient speakers, occurs at levels of around one watt per channel or less. This being the case, should one be concerned about an amplifier's distortion behavior below one watt. Should this be a critical point to focus upon when comparing specifications across amplifiers and receivers, or is it generally inaudible?

You are correct THD can be very high at low power (

 

[sfogg]

"-- steeply rising distortion at low power levels of less than one watt, and again at the upper end of the amplifiers power output capability."


Have you actually examined this or going from Audiophile lore?


All of the pics below are into an 8 ohm load testing at 50hz.


Here is the distortion of an amp at 0.1w:

http://forums.klipsch.com/idealbb/files/amp5-0.1w.jpg


second harmonic is 0.000434v.


Same amp at 1w:

http://forums.klipsch.com/idealbb/files/amp5-1w.jpg


Second harmonic is: 0.00104v


Same amp at 10w:

http://forums.klipsch.com/idealbb/files/amp5-10w.jpg


Second harmonic is: 0.00438v.


Here is a different amp at 0.1w

http://forums.klipsch.com/idealbb/files/amp6-0.1w.jpg


Second order was .000296v


This amp at 1w

http://forums.klipsch.com/idealbb/files/amp6-1w.jpg


Second order was 0.00121v


Same amp at 10w:

http://forums.klipsch.com/idealbb/files/amp6-10w.jpg


Second harmonic .43v

3rd - 0.371

4th - 0.225

5th - 0.126

6th - 0.0313

7th - 0.0122

8th - 0.04v

9th - 0.0303

10th - 0.0207



The first set of tests listed here is for a 'digital' amp... a Crown K2. The second set of measurements is for a zero feedback Single Ended Class 'A' Monarchy SM-70.


I've done these tests on a few other amps as well.


Shawn

 

[Jedi]

SFOGG....I'm just a hobbiest, but I find these plots interesting and educational. If I understand the screen shots, is the idea to sum the peak signal levels of each of harmonics and express this total as a percent of the reference signal peak? Do you measure peak to peak or area to area? Do you have to measure this by eyesight, or are the readings and math handled by software? On the plots would 0.1% distortion be represented by a level of 3 major gridlines below the reference level bar height. Why is the 10W distortion so high in the final plot? Do amps usually give similar results at different frequencies?

 

[sfogg]

"Do you have to measure this by eyesight, or are the readings and math handled by software?"


The hardware gives you the voltage readings of each peak if you put the marker on it. For example look at the second picture (as it is easy to see) and look at the main peak. That is the fundamental. The square box on the top is the marker. If you look at the bottom left of the screen it is telling you the marker is at 50hz and the voltage is 2.83v. For the voltage readings of the other harmonics I just moved the marker to each harmonic and wrote down the voltage level.


The vertical axis is voltage, not % so the grid lines don't correspond like that. And it is a log. scale not linear so eyeballing it can be a little misleading if you don't take that into account.


" Why is the 10W distortion so high in the final plot?"


The amp is starting to clip. I think it is rated to 20w or so but distortion is obviously starting to rise before that point.


Shawn

 

[keepinitcool]

You might want to read the definition on Rane's website.

http://www.rane.com/par-t.html