• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

please help me bettter understand amplifier measurments

wadec22

Senior Member
Joined
Oct 9, 2018
Messages
319
Likes
239
why do all amplifier measurements show a downward sloping line until they clip? often the distortion at 1mw is significantly higher than close to 1w. I have a tough time wrapping my head around it, as I wouldn't think distortion would actually set in, until the amplifier was boosting the signal beyond it's means.

as a follow up, with headphones that are extremely sensitive - 108db/mw and 25 ohm - does that make the overall SINAD score of an amplifier moot to me? Should I only be concerned about 1mw and below - as I know I'm never listening at 108db!

thanks for any insight guys. I love learning about this stuff!
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
5,580
Likes
9,873
Location
Monument, CO
Most all analyzers actually plot the signal-to-noise-and-distortion (SINAD) ratio, not just THD. It is hard to isolate the actual distortion from the noise at very low power levels so noise and distortion get lumped together. Noise is roughly constant, so as the signal level (power output) drops, the SINAD decreases -- same noise, smaller signal, so the plots rise because noise is included and the ratio of signal to noise goes down. Again, smaller signal, same noise, so signal (down) to noise (same) means the overall curve goes up at low power. Similarly at higher power the noise becomes less a fraction of the overall measurement and the curves flatten out because noise and distortion are roughly the same for a bit. Sometimes you'll see a curve flatten for a while until the signal significantly exceeds the noise and distortion starts to dominate, then at high power it is all distortion as the amplifier begins to clip and the curve rockets up.

HTH - Don
 
Last edited:

RayDunzl

Master Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,506
Likes
12,132
Location
Riverview FL
The same thing happens measuring speakers in-room.

At low SPL, measured "distortion" is high.

As the level of the speakers increase the measured distortion decreases.

At some level, actual distortion exceeds the noise floor, and the measured distortion rises.

For my main speakers the reversal in the direction of the distortion measurement occurs at about 90dB SPL in room.

See https://www.audiosciencereview.com/...-and-dac-measurements.5734/page-3#post-128094
 

RayDunzl

Master Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,506
Likes
12,132
Location
Riverview FL
with headphones that are extremely sensitive - 108db/mw and 25 ohm - does that make the overall SINAD score of an amplifier moot to me? Should I only be concerned about 1mw and below - as I know I'm never listening at 108db!

Seems to me the answer might be "yes" to the second part.

Small numbers there:

1546648519995.png
 
OP
wadec22

wadec22

Senior Member
Joined
Oct 9, 2018
Messages
319
Likes
239
Most all analyzers actually plot the signal-to-noise-and-distortion (SINAD) ratio, not just THD. It is hard to isolate the actual distortion from the noise at very low power levels so noise and distortion get lumped together. Noise is roughly constant, so as the signal level (power output) drops, the SINAD decreases -- same noise, smaller signal, so the plots rise because noise is included and the ratio of signal to noise goes down. Again, smaller signal, same noise, so signal (down) to noise (same) means the overall curve goes up at low power. Similarly at higher power the noise becomes less a fraction of the overall measurement and the curves flatten out because noise and distortion are roughly the same for a bit. Sometimes you'll see a curve flatten for a while until the signal significantly exceeds the noise and distortion starts to dominate, then at high power it is all distortion as the amplifier begins to clip and the curve rockets up.

HTH - Don
This was a damn good explanation, thanks!
 

SplitTime

Member
Forum Donor
Joined
Dec 31, 2018
Messages
62
Likes
80
Seems to me the answer might be "yes" to the second part.

Small numbers there:

View attachment 19915

Maybe I’m missing something with the “overall” SINAD score... but several devices have SINAD plots with different noise levels. As a result the noise dominated portion of the plot can be lower for some devices than others... at a given power level. So, I think (??) the plots are still useful for giving you an indication of which devices would still sound better (e.g. have less noise at your desired power level).
 

RayDunzl

Master Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,506
Likes
12,132
Location
Riverview FL
why do all amplifier measurements show a downward sloping line until they clip?


They don't.

2, 4, and 8 ohms load top to bottom. Aug 10, 2000

1546652817111.png
 
Last edited:

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
5,580
Likes
9,873
Location
Monument, CO
They don't.

2, 4, and 8 ohms load top to bottom. Aug 10, 2000

View attachment 19920

Actually, I'd bet the downward slope at the left is noise-dominated, then distortion kicks in around the 1 W level (4/8 ohm loads), then we see normal'ish distortion changing with bias and such until clipping. I suspect this is a balanced amplifier design as they tend to have that sort of "humped" distortion plot, and furthermore guess it has very high SNR (and perhaps relatively low feedback) given the low point at which it crosses from noise to distortion in the plots.
 

RayDunzl

Master Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,506
Likes
12,132
Location
Riverview FL
Actually, I'd bet the downward slope at the left is noise-dominated, then distortion kicks in around the 1 W level (4/8 ohm loads), then we see normal'ish distortion changing with bias and such until clipping. I suspect this is a balanced amplifier design as they tend to have that sort of "humped" distortion plot, and furthermore guess it has very high SNR (and perhaps relatively low feedback) given the low point at which it crosses from noise to distortion in the plots.

Noise dominated on left - agree

Bias - is (theoretically) Class A, with (maybe) seven increasing levels - they call them "plateaus". On mine, next newer model (still old now), from the wall, I see levels of 100W (idle), 150W, 250W,, 400W (more or less, and stable give or take a watt or three of music power) and not ventured (measured) above that point, though I have popped the 15A breaker (two amps plus everything else on one outlet). Higher signal level "trips" higher bias levels, if no other "trip' in 30 seconds or so, it backs down to a lower bias level.

Balanced - yes, differential output, no single ended input

Very high SNR - 118dB SNR A-weighted quoted

Feedback - unknown

For the posted graph - https://www.stereophile.com/content...nced-350mc-monoblock-amplifier-specifications

Mine are a slightly dehanced greener version - Krell FPB 350 mcx - specs at the very end
 
Last edited:
Top Bottom