pro said:
Hi Pro,,
thank you very much for posting your results and for your faith in the SPLIF amp.
Charles 🙂
darkfenriz said:
Hi darkfenriz,
thanks for posting your interesting simulations.
Do I understand right that you feed a pure 500Hz into the amps input and do a FFT on the amps output while feeding a 70 Hz wave into the output ?
I was not aware that there are simulations on distortion behaviour. Will those simulations look the same for different output mosfets ? Does the simulator also simulate distortion of the driver OPamp or is that part considered 'ideal'?
Thanks for input, Charles 🙂
Charles said:
hi
i fed the amp with 500Hz as input signal and 70Hz as the signal 'produced' by a speaker
different fequencies just to recognize which is which when analysing the spectrum.
you may count (THD)TotalHarmonicDistortion on FFT by definition:
as you feed the amp with a sine of frequency 'f' then harmonic distortion is :
[sqare root (sum of sqared amplitudes of 2*f,3*f,4*f)]/amplitude of f
IMD intermodulation distortion is when you feed the amp with sum of two tones f1 & f2 (completely different):
[sqare root (sum of sqared amplitudes of f1+f2, f1-f2, f1+2*f2, ... ,33*f1-17*f2, ....)]/sqare root (sum of sqared amplitudes of f1, f2)
of course there are different types of distortion that aren't so easily measured
i used 'sophisticated' ORCAD PSpice so I think it is very close to true behavious of different components.
But trying different FETS and/or opamps may take you whole life
what do you think of my 'compromised' topology????
cheers
Charles said:
No. Since damping factor is DEFINED as the ratio of load impedance over output impedance, what you suppose is not possible.
Jan Didden
Charles said:
Charles,
You are quite smart. The situation is clear: If the dummy output stage and load are identical to the real output stage and load, you would have an traditional 100% global feedback amp.
If you would take the feedback from the Vas, you would have an open loop zero feedback output stage. Since the output stages are the same, but the loads not, you have something in between: less than 100%, more than zero feedback. The closer your dummy load comes to the real load, the closer you get to 100% global feedback. So, you try to get away from that, but you also lose the benefit of the feedback from the dummy output stage.
Saying this is zero feedback from the speaker is simply misleading.
If that is how you want to play this, hey, you can do that on this forum. In fact, you can mislead and tell anything you want even if you know it isn't true on this forum all you want, as long as you don't use a four-letter word.
I think I had all I can take of this.
Jan Didden
Charles said:
Just to make sure, would you mind explaining where the speaker EMF comes from?
Rodolfo
darkfenriz said:
That's really interesting but I think I don't understand how this distortion simulation thing is working. If you would leave the speaker load out of the simulation would the Splif speaker output still have this higher distortion shown or would it be similar to the dummy load output (conventional feedbakck amp)?
Also interesting. I would have to try it out in order to really know what it does. From just looking at the resistor values the parallel connection of 2*20K with 2*0,47 in between does not make up much difference to the single 10K resistor.
Interesting that the Splif simulation still dampens the 70Hz from 500mV down to 65mV...
Would that mean, that the Splif amp simulation shows a output impedance of 1 Ohm ?
Charles 🙂
janneman said:
Hi Jan,,
I think this definition does not make sense for amplifiers without feedback...
Charles 🙂
janneman said:
Yes.
1) The real speaker will be at exactly 8 Ohm for 1/1000 of the time.
2) You can also play the dummy stage without any resistor load. -> infinite load impedance. The amp still plays great, it just has a slightly different signatur. With a dummy load added, the sound is a little fuller as the output impedance is reduced.
I don't agree.
You still seem not to understand.
Zero feedback means: Whatever the speaker does, the amp does not see it, and it does not react to it.
The drive of the output stage will always be the same, whether the speaker's momentary impedance be infinite or zero, positive or negative.
The output stage drive will always be the same.
If you would have never looked into this thread (I'm glad you did), we would'nt have got any feedback from you, would we ?
This is zero feedback, not looking, not reacting...
Charles 😉
If I understand the design correctly, I interpret it as
effectively having "zero feedback from the speaker"
but at the same time having feedback from a dummy
load resistor.
I think you need a more descriptive term than "zero
feedback amp".
effectively having "zero feedback from the speaker"
but at the same time having feedback from a dummy
load resistor.
I think you need a more descriptive term than "zero
feedback amp".
Hi Charles again
the more loads (speaker load and dummy) are similar to each other, the less distortion splif amp makes
...
of course providing, that speaker is ONLY a (passive) load.
the speaker EMF makes things more complex.
if there was no EMF, no 'microphone-like' ("speaking back") issue, no speaker voltage source
then janneman would be absolutely right: identical loads =>why not take feedback from speaker?
You may try 1 ohm or even 2 ohm instead of these 0,47 gate resistors
the value of 20K was just to have in paralel 10k
the gate resistors are the 'margin' for speaker doing what it wants
EXACTLY
the more loads (speaker load and dummy) are similar to each other, the less distortion splif amp makes
...
of course providing, that speaker is ONLY a (passive) load.
the speaker EMF makes things more complex.
if there was no EMF, no 'microphone-like' ("speaking back") issue, no speaker voltage source
then janneman would be absolutely right: identical loads =>why not take feedback from speaker?
You may try 1 ohm or even 2 ohm instead of these 0,47 gate resistors
the value of 20K was just to have in paralel 10k
the gate resistors are the 'margin' for speaker doing what it wants
EXACTLY
mikelm said:
Essentially, there was much more air between several the instruments. The sound is more fluid, exits more easy from the loudspeakers. The grain is more fine and the soundstage more depth. I have not noticed differences on the LF, perhaps because of the closed box of my woofers.
I am sure of this, because, having a couple of those Rotel for biamping , I modded the first, and then i have made a lot of listening test, before to modify the second one.
But the biggest difference, is been when, before selling these amplifiers, I have brought back they to their original state .
To me portable radio seemed to listen to one.
Ciao.
Nelson Pass said:
Hi Nelson,
yes, zero feedback is the wrong term for the SPLIF amp.
Zero feedback from the speaker is right.
Charles 🙂
Hi darkfenriz,,
when the SPLIF's output impedance is about 1 Ohm for voltages that are generated by speaker back EMF, which are also uncorrelated to any present input signal, I would suggest that the amp's output impedance for frequencies that match the actual amplifier's input signal is much lower ...
Would you agree ?
Charles 🙂
when the SPLIF's output impedance is about 1 Ohm for voltages that are generated by speaker back EMF, which are also uncorrelated to any present input signal, I would suggest that the amp's output impedance for frequencies that match the actual amplifier's input signal is much lower ...
Would you agree ?
Charles 🙂
ingrast said:
Hi Rodolfo,
as I see it, two things happen. Lets just take a simple speaker. The voice coil pole-piece assembly acts as an inductor. If the amplifier applies a voltage at the voice-coil, a back EMF voltage is generated immediately which delays the current flow through the coil which builds up the magnetic field. When the voltage is removed from the coil, a reverse voltage is generated in the coil, falling rapidly until the magnetic field is gone.
The second thing that happens takes into account the speakers mechanical properties. Exciting the speaker cone through application of a voltage at the voice coil, is the same as elongating a spring (here in form of the spider and cone surround). If you remove the voltage, the spring force will move back the cone resulting in a damped oscillation of the cone/voice-coil. This back movement is motivated only through spring pressure, but also generates a voltage, just like a dynamo.
A conventional amplifier reacts on this dynamic speaker behaviour in that it tries to hold the speaker output voltage always in gain proportion to the amp's input signal.
The SPLIF amplifier just ignores the speaker.
The philosophy in principle is that of a piano player. You just hit the key, the piano makes the tone...
Charles 🙂
Charles said:
Good, now by which means is electrical energy transferred to the air as moving sound pressure waves. I mean, how do you physically express and quantify this phenomena.
Rodolfo
- Status
- Not open for further replies.