Rotation of the output signal shows the characteristics of the amplifier. I want to find a method to assess the amplifier without complex and expensive equipment. Oscilloscope is sufficient. Find a method that corresponds to what is heard. Think that 5-6 ° at 100kHz, the threshold to be starting to hear.
Whoah, it turns out I understood you all along. Surprised myself there.
So, to develop a method... I don't think we can go by phase shift alone, BUT, it could add meaningfully to the variables already established.
So we're looking, specifically, for the correlation between phase shift and sound. To go at this directly we would need to set up a whole ABX experiment type thing, but there are other ways we can attack it indirectly.
We know (unless I'm mistaken) that while harmonic distortion is tolerable in any well designed amplifier, nonlinear distortions such as IMD are the most unpleasant.
So I think the best route right now is to see what correlations there are between phase shift and IMD.
What are the causes of IMD/nonlinear distortions? how do these same causes affect phase shift?
I suspect that in order to use phase shift as a meaningful measure we will have to combine it with another variable, such as Vdif. I don't think phase shift alone is enough, but if we make it an important part it may be revealing. If we are successful, we will end up with a measure that closely parallels ABX tests.
- keantoken
So, to develop a method... I don't think we can go by phase shift alone, BUT, it could add meaningfully to the variables already established.
So we're looking, specifically, for the correlation between phase shift and sound. To go at this directly we would need to set up a whole ABX experiment type thing, but there are other ways we can attack it indirectly.
We know (unless I'm mistaken) that while harmonic distortion is tolerable in any well designed amplifier, nonlinear distortions such as IMD are the most unpleasant.
So I think the best route right now is to see what correlations there are between phase shift and IMD.
What are the causes of IMD/nonlinear distortions? how do these same causes affect phase shift?
I suspect that in order to use phase shift as a meaningful measure we will have to combine it with another variable, such as Vdif. I don't think phase shift alone is enough, but if we make it an important part it may be revealing. If we are successful, we will end up with a measure that closely parallels ABX tests.
- keantoken
Me on the phase shift brought Vdif.
For sinus:
Vdif = input voltage * (1-cosine phase shift)
Vdif = V(a) * (1-cos phase shift)
(1-cos phase shift) indicates the profit margin AU*- AU, while decreasing the profit margin grows and grows Vdif phase shift.
Output voltage = Vdif * AU*
V(c) = Vdif * AU*
AU* = gain without feedback
AU = Amplification Amplifiers
In the phase shift is included AU*, AU, SR and V(c)
When Vdif in the non-linear, there is IMD. Something I wrote here and here
Last edited:
Okay, so this gives the perfect AC Vdif, all linear distortions removed.
So say we make a plot:
X=frequency
Y=phase shift
And then regress said plot into a model. Differences between the model and the measurements indicate nonlinear distortion.
So it seems you've got all the math worked out, but need a way to measure in real life?
- keantoken
No such relation exists. You do not distinguish linear and non-linear distortions. Phase shift is a linear distortion, though THD is a non-linear distortion. Linear distortion does not create new harmonic components. Non-linear distortion creates new harmonic components, on multiples of basic frequency, that were not present in the original signal. Your permanent mixing of phase angle and THD is just confusing for those who are not experienced in the subject.
Tone controls introduce substantial phase shift... far more than 6 degrees and within the audio band to but it is not THD 🙂
Similarly linear tape recording has significant phase shift due to azimuth issues with the replay head but again it is not "distortion" in the way you seem to mean.
The ears on your head introduce phase shift, it's partly how we locate sound direction, but it's not introducing "distortion"
Wait, how about
X=Phase shift
y=Vdif
Problem with using phase shift as a variable is that it is caused by 2 factors:
1: Cdom. This is not as dangerous as
2: Nonlinear/diode capacitances.
[PS: I thought linear distortion was harmonic distortion and nonlinear distortion was IMD. I suppose I should call IMD nonharmonic distortion. oops]
Okay, I think I should ask the question:
Is "transistor sound" caused by nonlinear distortion or nonharmonic distortion?
- keantoken
X=Phase shift
y=Vdif
Problem with using phase shift as a variable is that it is caused by 2 factors:
1: Cdom. This is not as dangerous as
2: Nonlinear/diode capacitances.
[PS: I thought linear distortion was harmonic distortion and nonlinear distortion was IMD. I suppose I should call IMD nonharmonic distortion. oops]
Okay, I think I should ask the question:
Is "transistor sound" caused by nonlinear distortion or nonharmonic distortion?
- keantoken
No.
Total harmonic distortion - Wikipedia, the free encyclopedia
or here:
http://www.rane.com/note145.html
They are both nonlinear distortions. In case you have a transfer function of the circuit (electronic circuit = nonlinear circuit), then there is a direct mathematical relation between THD and IMD.
No such relation exists between phase shift and THD (or IMD).
And, phase shift in a non-linear circuit is non-linear as well (i.e. depends also on amplitude).
Last edited:
I was working this out for a good few hours, and it seems so.
However, there is a relationship between phase shift and the rise of THD20; Higher Cdom increases THD20 as well as phase shift (while it won't affect lower frequency distortion), and this seems to be Frederman's point.
EDIT: Maybe I'm wrong, I though Frederman said something but I can't find it anymore.
- keantoken
Last edited:
There is a relationship between decreasing loopgain (feedback) at higher frequencies and THD20.
Phase shift at higher frequencies is related to amplitude rolloff at high frequencies. Yes, there is a direct relationship between amplitude and phase. But any attempts of claims like this:
http://www.diyaudio.com/forums/soli...listening-characteristics-11.html#post2018826
are extremely confusing.
Phase shift at higher frequencies is related to amplitude rolloff at high frequencies. Yes, there is a direct relationship between amplitude and phase. But any attempts of claims like this:
http://www.diyaudio.com/forums/soli...listening-characteristics-11.html#post2018826
are extremely confusing.
Last edited:
hmm, I have to consult Mr. Stee? 😕
Just trying to find correlations, maybe our new web spider @ Diyaudio Mr. tiefbassuebertr can help us? 😕
Cheers Michael
Just trying to find correlations, maybe our new web spider @ Diyaudio Mr. tiefbassuebertr can help us? 😕
Cheers Michael
2 Federmann
Real amplifier with output phase shift at 100kHz about 25 degrees (should have according your theory much more than 1%THD) measured THD at 20kHz about 0,005%, DIM19+20 kHz about the same..
How would you explain this?
Real amplifier with output phase shift at 100kHz about 25 degrees (should have according your theory much more than 1%THD) measured THD at 20kHz about 0,005%, DIM19+20 kHz about the same..
How would you explain this?
Last edited:
Please go ahead and bring any mathematical proof of this statement to clarify relationship between linear distortion (phase shift, amplitude change) and non-linear distortion (THD). Many of us are waiting for a scientific proof of this speculation.
- Status
- Not open for further replies.