I took a brief survey of solid-state power amplifiers as reviewed by Stereophile. I chose one point, THD+N at 20KHz into 4 ohms at roughly 20V. These are demanding conditions. Stereophile's measurements are:
The distortion measured in commercial amplifiers is roughly an order of magnitude higher than the simulated distortion numbers seen on this board. I can think of several reasons why simulation may be optimistic:
Some commercial designs have enough distortion to be audible, but I don't think they are paying less attention to distortion than hobbyist designs. I also don't think the simulation will be off by an order of magnitude if given realistic conditions.
What do you think? Are we living in a bubble?
Ed
Code:
THD+N (%) at 20KHz at 12.67-28.3V into 4 ohms
---------------------------------------------
.003 Electrocompaniet AW 800M (*)
.008 Accuphase A-300 (*)
.012 EMM Labs MTRS (*)
.020 Krell KMA-i800
.030 Naim NAP 250
.050 Ayre VX-8
.060 Karan POWERa Mono (*)
.070 Esoteric M1X
.100 Dan D'Agostino M400 MxV
.150 JMF HQS 7001
.200 LSA Warp-1
.230 Plinius A-150
.250 Burmester 216
.300 Rogue DragoN (*)
.500 Infigo Method-3
(*) THD+N decreases at 20KHz likely due to the output inductor.The distortion measured in commercial amplifiers is roughly an order of magnitude higher than the simulated distortion numbers seen on this board. I can think of several reasons why simulation may be optimistic:
- Test conditions are easy: 1KHz where the loop gain is high and 8 ohm load.
- Complementary devices are perfectly matched.
- Bias current is optimal. This ignores that the amplifier will be hot.
Some commercial designs have enough distortion to be audible, but I don't think they are paying less attention to distortion than hobbyist designs. I also don't think the simulation will be off by an order of magnitude if given realistic conditions.
What do you think? Are we living in a bubble?
Ed
I have 2 reasons:
1: normal listener don't feel distortion untill it is really unpleasant. I myself am unable to listen, say, 5% distortion because my ears are not traied in music.
2: the market demmands high power (PMPO?), no low distortion.
Most music is distortioned intentionally. Rap, trap, pop and most modern music use distortion musically.
1: normal listener don't feel distortion untill it is really unpleasant. I myself am unable to listen, say, 5% distortion because my ears are not traied in music.
2: the market demmands high power (PMPO?), no low distortion.
Most music is distortioned intentionally. Rap, trap, pop and most modern music use distortion musically.
Last edited:
Hello Ed,
I would like to briefly discuss the audibility of these harmonics.
THD20k_4Ohm means:
40kHz, 60kHz, 80kHz, 100kHz, 120kHz, 140kHz, 160kHz, 180kHz, ...
all these components are no longer audible. But they certainly provide modulation products and this influence should be investigated and the question should be asked what use the THD20k_4Ohm specification has if it is in a vacuum, I always miss a lot of additional information with these detached specifications.
But basically, the real world always looks a little less good than the model-based calculation values, algebraically solved equation systems / matrices.
But that's not really tragic.
Greetings,
HBt.
(13V)^2/4Ohm = 42.25W
(28V)^2/4Ohm = 196W
Which natural instrument requires this electrical power of its fundamental vibration 20000 1/sec ?
Hello HBt,
I picked THD at 20KHz because it is a measure of the amplifier's linearity without a lot negative feedback.
While the harmonics above 20KHz are inaudible, the same non-linearity produces intermodulation products that will be audible. I would not expect the amplifiers with more than 0.1% THD to sound clean. In practice, they may never be played loudly enough to notice.
Ed
I picked THD at 20KHz because it is a measure of the amplifier's linearity without a lot negative feedback.
While the harmonics above 20KHz are inaudible, the same non-linearity produces intermodulation products that will be audible. I would not expect the amplifiers with more than 0.1% THD to sound clean. In practice, they may never be played loudly enough to notice.
Ed
A more or less useless number. As soon as the spectrum is collapsed into a single number, all the important information is lost.
Your study is only of acedemic interest, telling us little to nothing about the amplifier (unless the numbers are really high or really low — that often indicates a problem)
dave
Then pay attention top the measures that mean something, instead of stuff that the marketing deptment uses.
dave
Changing thread title accordingly would gain more participants.
I believe that you understand the difference. Is your goal, perhaps, to quantify difference under specified conditions?
bohrok2610 - The 20V amplitude is sufficiently high to put the harmonics above the noise floor.
planet10 - I think Stereophile's measurement methods are sound. Interpretation is up to the reader.
tombo56 - I think I understand most of the reasons why simulation and measurement can differ. I am looking for others.
Ed
planet10 - I think Stereophile's measurement methods are sound. Interpretation is up to the reader.
tombo56 - I think I understand most of the reasons why simulation and measurement can differ. I am looking for others.
Ed
I don't know if it is an important parameter or not. But common people don't care.
Example: can those boomers some guy use in cars that cause strong vibration of the doors (and destroy the point welding of the structures) can be considered as distortion? If it is true, why do one may ask for 0.00000000001% distortion in the amp itself?
I agree. This is the reason to start this thread:
https://www.diyaudio.com/community/...st-possible-thd-n-really-the-best-way.367692/
The same question should not only be asked in relation to integrated OP amp ICs (as asked in this thread), but also in general - i.e. also in relation to power amplifiers.
If compare of various power amps the character of the distortion itself (fundamental notched out - images in post #2 below) it is easy to understand, why high values of THD not in all cases does represent a disadvantage during a listening test of various models - like the Aleph-series (Single Ended) from Pass-Labs with sine wave character of the residual distortion.
Those guys dont care about .0000000001% THD. But they (and I) would prefer .05% vs 10% at X watts, because the latter is already clipping hard at X watts, and the former will physically play louder.
Because it shows its linear. As an amplifier should be. Taming distortion in mechanical devices much harder so let's leave the "art" there
My experiences show that people think that lower distortion in high frequency and high slew rate is sound better. I designed many amplifiers with several different concept. My amplifiers that have lower distortion in high frequency and high slew rate compared to commercial amplifier with 2x ~ 3x price. The people who compared think my amplifier sound better.
If you can make lower distortion without increase the price or only increase the price a little bit, than why not? Of course, it is different story if you cannot reduce the distortion or you don't understand what the cause of distortion.
Sometimes people already pessimistic and have no gut to compare it. Make your lowest distortion amplifier that you can build and compare it. It does not matter if they will sound same for you when you compare to higher distortion amplifier. It is DIY Audio that should we enjoy to build many different amplifiers.
If you can make lower distortion without increase the price or only increase the price a little bit, than why not? Of course, it is different story if you cannot reduce the distortion or you don't understand what the cause of distortion.
Sometimes people already pessimistic and have no gut to compare it. Make your lowest distortion amplifier that you can build and compare it. It does not matter if they will sound same for you when you compare to higher distortion amplifier. It is DIY Audio that should we enjoy to build many different amplifiers.