stfoth,
Early on in the project (years) you can see the THD drop to low levels by making a specific part of the circuit take over the handling of the input signal.
Once the distortion hits the noise floor - two possibilities exist. If the noise is seen on a spectrum analyzer is -100db then the distortion is either -100 as well (or less). Listening tests at that point reveal a caliber of resolution that can be associated with those numbers. After that point any further improvements (hearing more resolution/less distortion) is happening below -100db. If you monitor it over time you can see the random nature of the noise will occasionally drop below -100 for a brief time. When this happens it exposes the harmonic measurement which if it was -100 then even in the absence of noise it would still read -100.
Instead it measures -130db. This quick peak indicates that the distortion is still being driven to deeper levels and you can verify by what you hear that you are going in the right direction. After that point the THD analyzer is of no use. Now we take over with the math to determine how far down it wold be. Because of the phase lock kicking in and stopping it from being able to generate harmonics - you now can use the degree of resolution as a gauge going forward. As I continue to raise the sensitivity of the detector by X amount - it translates directly into an increase in resolution. It is obvious when resolution goes up - other instruments that you never heard before are now "visible". Again, continuing to raise the detector output gives you the numbers needed to calculate the degree of lock presented to the fundamental. At this point it no longer can distort at least as seen by harmonic output. Now we are down to how much bandwidth are we limiting the fundamental to vary. As the detector continued to improve I can calculate how much phase shift is now allowable as far as deviation from the fundamental frequency. Since I know that the detector sensitivity can be triggered by as little as a few micro-degrees, It ensures that the smaller the allowable phase shift - the tighter the focus. Remember the detector drives the auto-focus.
The correction uses phase shift countermeasures that are extremely tiny and guarantees the fundamental is now the only thing that can exit the stage. The red shift / blue shift torque is held at a balanced point by a hair trigger which is the velocity detector. All of the correction as I mentioned is done on the horizontal axis. (time domain line). To do this the detector has to be rotated by 90 degrees so that it is literally seen by the music signal as a path that it must travel through to reach the output. (unlike monitoring the voltage or amplitude levels as in the vertical axis).
Trying to use negative feedback driven by amplitude measurements is no way to accurately correct anything. As you may know by the time you get this "sample" of output to use as a countermeasure its too late. That "piece" of music already left the circuit as distortion.
By using 90 degree phase detection capable of seeing a micro-degree of phase shift you have plenty of time to fix the problem in real time. Once this process takes over and we know the sensitivity of the detector we know the max deviation from the true fundamental. At this point it is right in the ballpark of the same phase shift you would expect sound to experience traveling through air. (virtually zero).
At this level of phase purity we have emulated the linear property of air.
The sound of the unit now seems more like a hole in the wall or portal through which sound waves are allowed to pass through unaltered.
The hardware is cloaked.
Hope this helps.
Early on in the project (years) you can see the THD drop to low levels by making a specific part of the circuit take over the handling of the input signal.
Once the distortion hits the noise floor - two possibilities exist. If the noise is seen on a spectrum analyzer is -100db then the distortion is either -100 as well (or less). Listening tests at that point reveal a caliber of resolution that can be associated with those numbers. After that point any further improvements (hearing more resolution/less distortion) is happening below -100db. If you monitor it over time you can see the random nature of the noise will occasionally drop below -100 for a brief time. When this happens it exposes the harmonic measurement which if it was -100 then even in the absence of noise it would still read -100.
Instead it measures -130db. This quick peak indicates that the distortion is still being driven to deeper levels and you can verify by what you hear that you are going in the right direction. After that point the THD analyzer is of no use. Now we take over with the math to determine how far down it wold be. Because of the phase lock kicking in and stopping it from being able to generate harmonics - you now can use the degree of resolution as a gauge going forward. As I continue to raise the sensitivity of the detector by X amount - it translates directly into an increase in resolution. It is obvious when resolution goes up - other instruments that you never heard before are now "visible". Again, continuing to raise the detector output gives you the numbers needed to calculate the degree of lock presented to the fundamental. At this point it no longer can distort at least as seen by harmonic output. Now we are down to how much bandwidth are we limiting the fundamental to vary. As the detector continued to improve I can calculate how much phase shift is now allowable as far as deviation from the fundamental frequency. Since I know that the detector sensitivity can be triggered by as little as a few micro-degrees, It ensures that the smaller the allowable phase shift - the tighter the focus. Remember the detector drives the auto-focus.
The correction uses phase shift countermeasures that are extremely tiny and guarantees the fundamental is now the only thing that can exit the stage. The red shift / blue shift torque is held at a balanced point by a hair trigger which is the velocity detector. All of the correction as I mentioned is done on the horizontal axis. (time domain line). To do this the detector has to be rotated by 90 degrees so that it is literally seen by the music signal as a path that it must travel through to reach the output. (unlike monitoring the voltage or amplitude levels as in the vertical axis).
Trying to use negative feedback driven by amplitude measurements is no way to accurately correct anything. As you may know by the time you get this "sample" of output to use as a countermeasure its too late. That "piece" of music already left the circuit as distortion.
By using 90 degree phase detection capable of seeing a micro-degree of phase shift you have plenty of time to fix the problem in real time. Once this process takes over and we know the sensitivity of the detector we know the max deviation from the true fundamental. At this point it is right in the ballpark of the same phase shift you would expect sound to experience traveling through air. (virtually zero).
At this level of phase purity we have emulated the linear property of air.
The sound of the unit now seems more like a hole in the wall or portal through which sound waves are allowed to pass through unaltered.
The hardware is cloaked.
Hope this helps.