Hi Mauro,
I too look forwards to hearing some more details of the amplifier that produced trace 2.
Thank you for passing on your recent test findings. It would be great to find some test format that we can relate directly to reproduction quality.
You have several amplifiers and testing facilities, thus I wonder if you might be able to run some reverse tests at 10kHz; this as a substitue for upper crossover induced back-EMF.
Run say 1V.rms and 10V.rms in series with a nominal eight ohm non-inductive resistor into the output terminal of each amplifier, with amplitude and phase of error developed at the output terminals measured with respect to generator output.
Of course this is merely examining the simple damping factor at 10kHz, but few amplifiers actually maintain reverse phase coherence, and the angle of closed loop damping is hardly ever specified.
The amplifier generated error becomes part of the audio waveform that is fed to the loudspeaker, and that which is not in phase will distort tweeter drive, thus I wonder if you might be able to do this test, and possibly correlate findings with your experience of the amplifiers.
Cheers ......... Graham.
I too look forwards to hearing some more details of the amplifier that produced trace 2.
Thank you for passing on your recent test findings. It would be great to find some test format that we can relate directly to reproduction quality.
You have several amplifiers and testing facilities, thus I wonder if you might be able to run some reverse tests at 10kHz; this as a substitue for upper crossover induced back-EMF.
Run say 1V.rms and 10V.rms in series with a nominal eight ohm non-inductive resistor into the output terminal of each amplifier, with amplitude and phase of error developed at the output terminals measured with respect to generator output.
Of course this is merely examining the simple damping factor at 10kHz, but few amplifiers actually maintain reverse phase coherence, and the angle of closed loop damping is hardly ever specified.
The amplifier generated error becomes part of the audio waveform that is fed to the loudspeaker, and that which is not in phase will distort tweeter drive, thus I wonder if you might be able to do this test, and possibly correlate findings with your experience of the amplifiers.
Cheers ......... Graham.
Correct questions, Lumanauw. I use 2 way speakers: ( Kef 102 Ref. & Dynaudio audience 52 ) 2 loads diametrically opposite, but 2 good sonic formulations.
To the first approach, the graph with loaded resistive and real am equal, for which I have ometted the analyses with loading reality ( for non-being thrown out of from house! ) When will have more time will start again verify this thing.
Ciao
Mauro
Hi, Graham.
I share much the hypothesis that the "non linear" behaviour in phase and frequency, over all in the comparisons of the back-EMF, is to the base of the acoustic differences in the amplifiers. Is capitate me to see graph of analysis of intermodulation on the loudspeakers, it is me am always asked in that percentage the amplifier used for the tests contribute to that IMD.
Doesn't know if succeed to reproduce a setup in degree to isolate this problem, even because this test develops it at home, and I doesn't dispose of tool does sophisticated. Surely I will try.
Thanks for the cues.
Ciao
Mauro
I share much the hypothesis that the "non linear" behaviour in phase and frequency, over all in the comparisons of the back-EMF, is to the base of the acoustic differences in the amplifiers. Is capitate me to see graph of analysis of intermodulation on the loudspeakers, it is me am always asked in that percentage the amplifier used for the tests contribute to that IMD.
Doesn't know if succeed to reproduce a setup in degree to isolate this problem, even because this test develops it at home, and I doesn't dispose of tool does sophisticated. Surely I will try.
Thanks for the cues.
Ciao
Mauro
I think that key issue here is open loop damping factor to get low intermodulation with signal aplied to output. To achieve this I would intentionally lower the output impedance of VAS and use tripple darlington output stage, maybe with paralleled output devices. Taking feedback directly from emitters (leaving emitter resistors out of global feedback) also can improve things a little bit because impedance at emitters is lower than at output (usually twice or 3 times).
regards
regards
I have performed in provisional form some test on suggestion of Graham.
Have connected a 8 ohms load on the output of 2 amplifiers ( technique of measure of the dumping factor ), but on a test amplifier has connected a 10Khz sine 3Vrms and on a amplifier of " EMF injection " a 10Khz sine 3Vrms out of phase of 90 degrees ( to get greater contrast of measure ).
The results are to confirmation of the graph that I have published with the other measures; The amplifiers with greater open loop Zout ( my_ref, JLH ) have ( for evident reasons ) a greater excursion of phase when they are solicited by the out of phase signal.
The amplifiers with low Zout O.L. ( Load Inv.) it's am poorly interested by this phenomenon. To the excursion of phase it even corresponds the relative variation of amplitude of the useful signal. If put together the picked data in this thread, it may risk a very approximate analysis, but that is to the base of the motivations that have pushed me to build the my_ref: An amplifier with current output stage conditions in minor way the dynamicses of mechanical shading of the loudspeakers on account of the minor dumping in the comparisons of the back EMF. The variations of phase( and amplitude ) show that the loudspeaker goes into the transfer funtion of the amplifier, with all the aftermathes (good and bad) that derive.
From the purely subjective point of view, the listenings stretch to prefer this topology, but is not a discounted thing ( the general quality does always the difference ).
Be pleasant other opinions and supplements of analysis
Ciao
Mauro
Have connected a 8 ohms load on the output of 2 amplifiers ( technique of measure of the dumping factor ), but on a test amplifier has connected a 10Khz sine 3Vrms and on a amplifier of " EMF injection " a 10Khz sine 3Vrms out of phase of 90 degrees ( to get greater contrast of measure ).
The results are to confirmation of the graph that I have published with the other measures; The amplifiers with greater open loop Zout ( my_ref, JLH ) have ( for evident reasons ) a greater excursion of phase when they are solicited by the out of phase signal.
The amplifiers with low Zout O.L. ( Load Inv.) it's am poorly interested by this phenomenon. To the excursion of phase it even corresponds the relative variation of amplitude of the useful signal. If put together the picked data in this thread, it may risk a very approximate analysis, but that is to the base of the motivations that have pushed me to build the my_ref: An amplifier with current output stage conditions in minor way the dynamicses of mechanical shading of the loudspeakers on account of the minor dumping in the comparisons of the back EMF. The variations of phase( and amplitude ) show that the loudspeaker goes into the transfer funtion of the amplifier, with all the aftermathes (good and bad) that derive.
From the purely subjective point of view, the listenings stretch to prefer this topology, but is not a discounted thing ( the general quality does always the difference ).
Be pleasant other opinions and supplements of analysis
Ciao
Mauro
In this thread I have set the hypothesis ( founded on any fast test ) that with analysis it founds on EMF solicitations himself is able somehow it reproduce a "sonical" imprint of an amplifier.
Naturally to extrapolate un coherent data wants an evaluation more scientific, but the idea of founding it he likes. The first test are founded on phenomenon of pure intermodulation ( injection 1:1 50Hz ) and relative FFT observation. Are underlined of the objective difference among every amplifier on a test basis, is in amplitude of intermodulation that on time of decadence. Datum that I know perfectly the sound of every amplifier make a will, have been able to draw any considerations " to warm ".
Graham Maynard has suggested me to try to associate my empirically evaluations even to 10Khz, with analysis on the variations of phase it provokes from EMF.
I have started to do some simple test, and have illustrated the first impressions. In practice the intermodulation amplitude in the preceding FFT is comparable with the variations of measurable phase ( +amplitude = +phase modulation ).
To this has added the mysubjective impressions to complete the picture (provisional) of the attempt to associate the things.
I hope of is me open...
Ciao
Mauro
Naturally to extrapolate un coherent data wants an evaluation more scientific, but the idea of founding it he likes. The first test are founded on phenomenon of pure intermodulation ( injection 1:1 50Hz ) and relative FFT observation. Are underlined of the objective difference among every amplifier on a test basis, is in amplitude of intermodulation that on time of decadence. Datum that I know perfectly the sound of every amplifier make a will, have been able to draw any considerations " to warm ".
Graham Maynard has suggested me to try to associate my empirically evaluations even to 10Khz, with analysis on the variations of phase it provokes from EMF.
I have started to do some simple test, and have illustrated the first impressions. In practice the intermodulation amplitude in the preceding FFT is comparable with the variations of measurable phase ( +amplitude = +phase modulation ).
To this has added the mysubjective impressions to complete the picture (provisional) of the attempt to associate the things.
I hope of is me open...
Ciao
Mauro
Hi Lumanauw, Absolutely of accord.
Sincerely he is stimulated a lot of by the matters that you open in this forum. Perhaps is the only thing that interests me of the audio world. To interest me it reads and the phenomenona that are to the base of the audio reproduction. They wait technicians that turn in this forum me annoy. Sincerely, believe that a middle planner disposes of the resources necessary to define by himself the parameters of work ( and topologic structures ) of an amplifier.
I wrung a confession of a project is always in the quality of the philosophy that has produced it. In the scientific disciplines is always an explanation to all. Enough knows it ( and want it ) look for. My regret is that exist little spaces and he has many competent persons for this discussions.
To the next question...
Mauro
Sincerely he is stimulated a lot of by the matters that you open in this forum. Perhaps is the only thing that interests me of the audio world. To interest me it reads and the phenomenona that are to the base of the audio reproduction. They wait technicians that turn in this forum me annoy. Sincerely, believe that a middle planner disposes of the resources necessary to define by himself the parameters of work ( and topologic structures ) of an amplifier.
I wrung a confession of a project is always in the quality of the philosophy that has produced it. In the scientific disciplines is always an explanation to all. Enough knows it ( and want it ) look for. My regret is that exist little spaces and he has many competent persons for this discussions.
To the next question...
Mauro
I hoped that some did of the comments with regard to this methods of investigation ( to improve them ) and on the problem of the " Dump back-EMF " yes or no. Other point, I hoped that some were prepared for sharing the really material on this themes.
Have not more newses in connection with the Hiraga test ( scale factor and details of the constructive topology of the amplifiers in test ).
If this things are do "secrets" by cannot be integrated in the forums ( or on the magazines ) the propose have not sense . A demonstrative experiment has value exclusively if is reproducible...
Ciao
Mauro
Have not more newses in connection with the Hiraga test ( scale factor and details of the constructive topology of the amplifiers in test ).
If this things are do "secrets" by cannot be integrated in the forums ( or on the magazines ) the propose have not sense . A demonstrative experiment has value exclusively if is reproducible...
Ciao
Mauro
Apologies for sending this as Post#35 to the RIAA string by clicking wrong link in my Inbox.
Hi Mauro,
I am not in a position to make any real-world reverse testing measurements or specrutm anslysis.
I feel that reverse (back EMF) injection can cause a NFB loop induced problem that will modulate forward amplified signal, and that this will show on a spectrum analyser with say 10kHz creating additional interference to say a 1kHz signal input; similar to your dumping factor test set-up in Post#70.
I assume that the other 'Results (preliminary)' set-up was to check for internal damping response phase shift induced error.
Different word usage between our native languages prevents me from fully understanding your last posts, and I did not realise that you were waiting for comment. I was waiting to see if you had any findings from those tests.
I wish you luck in completing those tests.
Cheers ......... Graham.
Hi Mauro,
I am not in a position to make any real-world reverse testing measurements or specrutm anslysis.
I feel that reverse (back EMF) injection can cause a NFB loop induced problem that will modulate forward amplified signal, and that this will show on a spectrum analyser with say 10kHz creating additional interference to say a 1kHz signal input; similar to your dumping factor test set-up in Post#70.
I assume that the other 'Results (preliminary)' set-up was to check for internal damping response phase shift induced error.
Different word usage between our native languages prevents me from fully understanding your last posts, and I did not realise that you were waiting for comment. I was waiting to see if you had any findings from those tests.
I wish you luck in completing those tests.
Cheers ......... Graham.
mauropenasa said:
Mauro:
I can understand you may feel disapointed for lack of feedback, may be even suspect unspoken interests are behind this. While some of the latter may be true to a measure, you should also be ready to notice a couple of points
1. You should be ready to acknowledge observations. I can remember a couple of posts from myself, and several others in the sense of noting your topology starts with a current mode output stage but ends in a voltage mode overall operation due to global voltage feedback.
While this configuration may be sonicaly good, it is not clear this is the result of a purpose or design. You failed to at least voice your oppinion with respect to this issue.
2. On the Hiraga test, basically what it does is exercise the amplifier under test output impedance. Here there are basically 2 possible extreme situations.
- A extremely low output impedance attained through large open loop gain and global feedback should provide good results. Note this implies the amplifier must be able to overcome a virtual short circuit at the back injected frequency, that is what low output impedance means in the end. If it is not capable of performing, then the whole chain gets disrupted.
- A no negative feedback design may perform equally well as long as the back injected signal does not intermodulate in the output stage. Again how much this is an issue depends on the output stage design and drive capability.
A final comment, I myself want to encourage your drive to experiment and share results. From this you benefit as much as all the rest of us. You may be certain to get feedback, questioning and encouragement as long as you are responsive.
Chau !!!
Rodolfo
I have seen that the tests with FFT underlines better the problems of IMD caused by Back EMF, and the variation to 10 Khz that you suggest has to be + selective. The measures found on the phase "coherence" among the input signal and the Back-EMF are native by a my elaboration of your suggestions and show the limits of the NFB (global) in the comparisons of a "no-coherent" signal with Input.
Exactly. I would add even that this characteristic would be able be exploited for Current Feed-forward corrections.
You are right. In fact have already mentioned the reasons that have convinced me to use this technique ( already used by other encircled to solid state ), but as love say you, an account is the theory and an account is the reality. Much hide after the " sound better ", I prefer attend to have finds instrumental to confirm my hypothesis.
From the technical point of view I am in pursuance of your hypothesis of methods to increase the immunity to the Back- EMF. The problem is that I am stop to the first step:
Datum that much of the considered very good amplifiers by the "musical" point of view has a "opposite" structure to what we hypothesize to improve the problem ( es: all the single-ended ), in that direction move?
Ciao
Mauro
Hi Mauro,
My take on this is -
The nature of back-EMF induced distortion relates more to the number and type of stages and thus the phase change within the global NFB loop at frequencies higher than input where the amplifier is obliged to attempt correction, and Mosfets introduce more phase change than bipolars. That which is often termed 'current feedback' ( I disagree with that term ) effectively places the first device non-linearity outside of the NFB loop; this simultaneously removes the bandwidth influencing phase change due to that first stage of amplification from the closed loop, whilst still allowing its gain to usefully contribute.
In attempting to acheive accuracy plus musicality, back-EMF immunity can be improved by;-
reducing natural (non-global NFB generated) output impedance, as with higher biasing or class-A (Nelson), or increased local output stage feedback (Susan Parker);
applying emitter/source/cathode based NFB at a first stage that self-limits hf input banwidth and thus the subsequent necessity for error correction; (some recent triode buffer stages / DAC circuitry 'musically' curtail the upper limit);
ensuring that the open loop phase change remains low, because it is this that leads to the generation of a phase shift induced interface error from loudspeaker generated back-EMF once the loop is closed; it is the use of NFB to reduce amplitude distortion, upon an amplifier that has a poor natural phase response, that causes that amplifier to behave in a 'tiring' manner when it attempts to counter phase shifted, and occasionally considerable, loudspeaker currents in order to maintain closed loop monitored voltage constancy.
You mentioned that there might be an advantageous use of current feed-forward correction, but I find this hard to understand. The amplifier would still have a form of controlling feedback, and thus respond to phase shifted loudspeaker generated back-EMF (current); ie. whether responding to current or voltage sensed error, it would still react/amplify differently with different types of loudspeaker.
Cheers ......... Graham.
My take on this is -
The nature of back-EMF induced distortion relates more to the number and type of stages and thus the phase change within the global NFB loop at frequencies higher than input where the amplifier is obliged to attempt correction, and Mosfets introduce more phase change than bipolars. That which is often termed 'current feedback' ( I disagree with that term ) effectively places the first device non-linearity outside of the NFB loop; this simultaneously removes the bandwidth influencing phase change due to that first stage of amplification from the closed loop, whilst still allowing its gain to usefully contribute.
In attempting to acheive accuracy plus musicality, back-EMF immunity can be improved by;-
reducing natural (non-global NFB generated) output impedance, as with higher biasing or class-A (Nelson), or increased local output stage feedback (Susan Parker);
applying emitter/source/cathode based NFB at a first stage that self-limits hf input banwidth and thus the subsequent necessity for error correction; (some recent triode buffer stages / DAC circuitry 'musically' curtail the upper limit);
ensuring that the open loop phase change remains low, because it is this that leads to the generation of a phase shift induced interface error from loudspeaker generated back-EMF once the loop is closed; it is the use of NFB to reduce amplitude distortion, upon an amplifier that has a poor natural phase response, that causes that amplifier to behave in a 'tiring' manner when it attempts to counter phase shifted, and occasionally considerable, loudspeaker currents in order to maintain closed loop monitored voltage constancy.
You mentioned that there might be an advantageous use of current feed-forward correction, but I find this hard to understand. The amplifier would still have a form of controlling feedback, and thus respond to phase shifted loudspeaker generated back-EMF (current); ie. whether responding to current or voltage sensed error, it would still react/amplify differently with different types of loudspeaker.
Cheers ......... Graham.
Hi Graham, thanks for the interest.
hmm..., The Pass new tendencies (single-ended) has not to guarantee a good immunity to the given Back-EMF, because this topology is resulted the more tender to this problem. This the have verified on the JLH ( John Linsey-Hood 10W class A bipolar single-ended).
For the rest, I am of accord.
I creed that the "secret" is in the exact nature ( and composition ) of the back-EMF, given that the loudspeaker is not only a reactive element but overall a mechanical resonator ( and the Back_EMF come produced from this resonant dynamicses )....
Ciao
Mauro
hmm..., The Pass new tendencies (single-ended) has not to guarantee a good immunity to the given Back-EMF, because this topology is resulted the more tender to this problem. This the have verified on the JLH ( John Linsey-Hood 10W class A bipolar single-ended).
For the rest, I am of accord.
I creed that the "secret" is in the exact nature ( and composition ) of the back-EMF, given that the loudspeaker is not only a reactive element but overall a mechanical resonator ( and the Back_EMF come produced from this resonant dynamicses )....
Ciao
Mauro
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.