I've not had time to dive deep into the FSAF thing, but it seems like it might be a (more sophisticated) version of the "Distortion Isolation" process I was working with in Praxis back in the '00s? https://www.libinst.com/distortion_isolation_in_the_time.htm (Or, maybe not?)
I wonder whether anyone has done a distortion residual test like FSAF on just the voice coil current signal in a cone driver? That is, drive the voice coil with a music (or sweep) voltage signal and use the driver's pre-determined small-signal voice coil current impulse response to remove the Linear Time Invariant part., and listen to what's left. I may have to drag out my old stuff and see if I can do it here. It might be interesting to compare similar drivers with and without demodulation rings or pole caps, and drivers using ferrite vs neo magnets.
Or maybe that's already been done in this thread somewhere -- sorry, I kind of feel like the kid asking questions not having done his homework before.
I am not aware of such detailed study. We used two years developing a model and system identification such that we can decompose all the distortion contributions into separate sources of origin. One outcome was the Wolfgang changed his own model 😀
If a single frequency can dsplace the cone by 1mm, then that should be factored in at all frequencies of interest.
This is probably correct, but without phase information we don't really know how the various harmonics stack together, or where the distortion occurs along the length of a sine wave.
A good old-fashioned harmonic spectrum could also be useful, along with the 'shape' of the error on a time scale.
Are you sure it's not just 1/frequency? With a given force, the cone movement will be the inverse of frequency due to inertia. I don't see where the extra frequency term comes in unless your assumption is the cone has already entered breakup. If you take the inductance of the coil as the default, then the directivity must increase for a flat response, so in terms of power transfer the inductance is a parasitic element.
For a given force input, ideal cone movement will be 1/f and SPL will be flat with frequency below breakup or diffraction frequencies.
@bwaslo: I made music audio recordings (stupidly with a microphone). Listening to this recordings was very promising. It is easier to hear something with headphones, compared to a blasting paper cone.
Then i used "DeltaWave Audio Null Comparator". This was very promising too.
Then my main SSD...
With all new hardware i made something to look at:
https://www.diyaudio.com/community/threads/drive-current-distortion-measurement.402566/
But the icing (with cherries) on that subject are the FSAF audios from @mikets42 . He put very successful effort in his microphony. And i learned here that one can attach zipped audio files to a posting.
A sketch of my current probe is in my thread first page.😉
Thank you for sharing your idea. We are two. 🙂
Best regards
Bernd
Then i used "DeltaWave Audio Null Comparator". This was very promising too.
Then my main SSD...
With all new hardware i made something to look at:
https://www.diyaudio.com/community/threads/drive-current-distortion-measurement.402566/
But the icing (with cherries) on that subject are the FSAF audios from @mikets42 . He put very successful effort in his microphony. And i learned here that one can attach zipped audio files to a posting.
A sketch of my current probe is in my thread first page.😉
Thank you for sharing your idea. We are two. 🙂
Best regards
Bernd
above the bass resonance we have constant acceleration (Newtons’s second law). the velocity is the integral of the acceleration and the displacement is the integral of the velocity. Hence, the position is the electrical signal integrated twice. this means that the displacement is proportional to 1/f^2. The sound pressure is proportional to the cone acceleration, ie for a constant voltage we get constant SPL above the resonance frequency.
Many years ago I did use Audio Diffmaker to measure the difference between resistors etc. I wasn’t aware of this work Bill. My math is entirely undergrad level, so I’ll have to pass to others who can compare and contrast your work to what @mikets42 has done.
Hello All,
I am trying to parse out what is being said here about about "Bark..." noise or distortion and "Hysteresis distortion". Both are said to be due to discontinuities in the magnetic field and both are reported to sound like not so random crackling and popping.
Might they be the same thing or are they similar and overlapping?
Thanks DT
I am trying to parse out what is being said here about about "Bark..." noise or distortion and "Hysteresis distortion". Both are said to be due to discontinuities in the magnetic field and both are reported to sound like not so random crackling and popping.
Might they be the same thing or are they similar and overlapping?
Thanks DT
Steel is not continuously magnetic, and the magnetic sections in steel do not all magnetize at the same level or at the same time. The various magnetic domains change polarity suddenly at a certain threshold, and also respond to other magnetic domains near them which influence the field. This threshold behavior creates bifurcations in the magnetization curve, tipping points where the outcome is erratic and unpredictable because it depends on the entire prior history of the material up to that point. For the same reason, bifurcations play a central role in most systems that exhibit chaos or randomness.
The reason we do not always notice is because there are so many of these domains that they normally blend together, the same way that when we measure signals in the countable electrons or photons the data is noisy, but in most circuits there are so many electrons acting at once that the "grain" is an irrelevant fraction of the actual signal.
Hysteresis is the action of sticky magnetic domains en masse, because when they have all been polarized by a positive signal, they will all tend to repolarize at a similar threshold for the negative signal. So when we test with periodic/repeating waveforms, hysteresis emerges as the result because we have created the ideal conditions for numerous magnetic domains to flip at the same time. With non-repeating signals the outcome can be very different. This is shown in Purifi's excellent demonstration of signal history that remains in a ferrite core after a non repeating waveform:
https://purifi-audio.com/blog/tech-notes-1/this-thing-we-have-about-hysteresis-distortion-3
There is also noise from the magnetic field itself which at the lowest level has discrete steps, just like how at the lowest level signals have discrete steps in the form of single electrons. I think this is different from magnetic domains in steel and represents the lower limit of the resolution allowed by the laws of physics.
The reason we do not always notice is because there are so many of these domains that they normally blend together, the same way that when we measure signals in the countable electrons or photons the data is noisy, but in most circuits there are so many electrons acting at once that the "grain" is an irrelevant fraction of the actual signal.
Hysteresis is the action of sticky magnetic domains en masse, because when they have all been polarized by a positive signal, they will all tend to repolarize at a similar threshold for the negative signal. So when we test with periodic/repeating waveforms, hysteresis emerges as the result because we have created the ideal conditions for numerous magnetic domains to flip at the same time. With non-repeating signals the outcome can be very different. This is shown in Purifi's excellent demonstration of signal history that remains in a ferrite core after a non repeating waveform:
https://purifi-audio.com/blog/tech-notes-1/this-thing-we-have-about-hysteresis-distortion-3
There is also noise from the magnetic field itself which at the lowest level has discrete steps, just like how at the lowest level signals have discrete steps in the form of single electrons. I think this is different from magnetic domains in steel and represents the lower limit of the resolution allowed by the laws of physics.
I only know observational data.
@5th element told me:
“the slowly rising HD3 from the lower mids/upper bass until the drivers natural response falls off. Obviously the HD3 falls off at 1/3 the frequency where the driver does… is the typical HD3 pattern of old SEAS drivers that did have copper rings, above and below the pole piece, but none that went through it…The same is true for the shape in drivers without any copper actually.
Drivers without shorting had the highest levels of that HD3 profile. Add in some aluminium rings, or shorting, and the levels would fall, like with the RS series from Dayton. Add in some copper, a la old Seas Excel, and the levels fall even further.
Use copper caps/sleeves, however, such as the Scan-Speak SD motor, SB Acoustics motor, B&W motors, new SEAS Excel, Tymphany TC/TG9 and full-rangers, and the rise is significantly reduced.”
We sure are glad you didn’t skimp on the copper, Lars.
There always is, but here on the ground I prefer that the level of improvement at least be consummate with the level of cost.
That for me also is the million dollar question. Furthermore, in this (and other threads) regular off the shelf moving coil driver motors are criticized for not being current drive optimized by design. Now what would the motor of a current drive optimized driver look like?
https://audiogroupdenmark.com/articles/article-optimizing-the-loudspeakers-magnet-system/
The sophisticated drive concept from Borresen might be a consequent approach to avoid the elaborated effects by leaving out all iron materials. Just Neodymium magnets and a coil. Shorting ring made of silver to shield the coil strayfield and so improve drive counterfield stability. Kyrogenic treating is "a bit" overengineering....
https://static.dali-speakers.com/en/sound-academy/tech/patented-soft-magnetic-composite-smc/
Another approach used by Dali is to make the pole piece from sintered powder that has optimized B-H curve with minimized hysteresis. Purify has similar approach by replacing most of the pole piece with a neodymium magnet.
P.S.: I don't see that much difference between optimizing drivers for current or voltage drive.
As shown in the measurements before, special care might be taken on resonances in the passband that affect the impedance where current drive reacts more sensitive to amplify these. It was also shown that current drive is not the preferred operating method for bass drivers on their main resonance as the driver is then just damped by the highly nonlinear mechanical damping. So a bass driver optimized for current drive will include an acceleration sensor for overall position feedback and have minimized dynamic coil offset as this cannot be compensated by acceleration sensor (2x integration to calculate the position --> no zero pos reference).
P.P.S.: An optimized driver for current drive might be designed with care for all nonlinearities like for voltage drive, but leave out optimizations for nonlinearities that appear in the current as analyzed by Klippel - these get compensated by the current drive:
The sophisticated drive concept from Borresen might be a consequent approach to avoid the elaborated effects by leaving out all iron materials. Just Neodymium magnets and a coil. Shorting ring made of silver to shield the coil strayfield and so improve drive counterfield stability. Kyrogenic treating is "a bit" overengineering....
https://static.dali-speakers.com/en/sound-academy/tech/patented-soft-magnetic-composite-smc/
Another approach used by Dali is to make the pole piece from sintered powder that has optimized B-H curve with minimized hysteresis. Purify has similar approach by replacing most of the pole piece with a neodymium magnet.
P.S.: I don't see that much difference between optimizing drivers for current or voltage drive.
As shown in the measurements before, special care might be taken on resonances in the passband that affect the impedance where current drive reacts more sensitive to amplify these. It was also shown that current drive is not the preferred operating method for bass drivers on their main resonance as the driver is then just damped by the highly nonlinear mechanical damping. So a bass driver optimized for current drive will include an acceleration sensor for overall position feedback and have minimized dynamic coil offset as this cannot be compensated by acceleration sensor (2x integration to calculate the position --> no zero pos reference).
P.P.S.: An optimized driver for current drive might be designed with care for all nonlinearities like for voltage drive, but leave out optimizations for nonlinearities that appear in the current as analyzed by Klippel - these get compensated by the current drive:
Last edited:
Leaving the iron out of the center polepiece has already been discussed here, albeit not in depth:
Hello All,
Both Barkhausen Noise and Hysteresis Distortion are caused by magnetic domains flipping orientation causing the same not so random crackling and popping sound.
If we are doing Experiments I purchased a not too many dollars dual voice coil woofer from Parts Express.
https://www.parts-express.com/CES-VW-8820D-8-DVC-Mid-Woofer-299-4156?quantity=1
The test amplifier connects to Voice Coil #1 and Voice Coil #2 connects to the audio analyzer input. Think transformer output from Voice Coil #2 goes to the analyzer input. Filter out the test sine wave and what remains is Barkhausen Noise or Hysteresis Distortion, what ever you wish to call it, not so random crackling and popping sound.
Thanks DT
On the subject of hysteresis caused by iron in the motor magnetic circuit, I am aware of three drivers built with only magnets and no iron.
Two are from Aurasound, the 1" Cougar and the 2" Whisper drivers, readily available from Madisound. Parts Express now has a GRS driver that looks the same, but I have no experience with that one. I used 20 of the 1" drivers on a 6" diameter spherical speaker with good results. The 2" driver is used in the Pluto kit. As links tend to stop working after a few years here is a picture.
https://www.madisoundspeakerstore.com/index.php?p=catalog&mode=search&search_str=aurasound
The third driver is not individually available, but is used in the very unique and expensive LeedH Acoustic Beauty speakers. A pair of these speakers sells for around 16,000 euro. These are very interesting as they have a ferro-fluid, essentially liquid O - ring, suspension and linear travel of around 1" peak to peak for what is a ~ 2" diameter driver. It would be an ambitious DIY project to try to build a driver like this.
https://www.acoustical-beauty.com/index-uk.php
" ... LEEDH designers have entirely invented a new type of electrodynamic speaker, the HPAB (Acoustical Beauty Speaker). The HPAB does not contain a soft iron pole piece in the motor, a foam edge, or a spider in the suspension, because these three components produce the majority of defects in traditional speakers.
Two breakthrough technologies, the ironless motor and suspension with a Ferro fluid seal, internationally patented with the University of Maine, have replaced the traditional motor and suspension. "
Two are from Aurasound, the 1" Cougar and the 2" Whisper drivers, readily available from Madisound. Parts Express now has a GRS driver that looks the same, but I have no experience with that one. I used 20 of the 1" drivers on a 6" diameter spherical speaker with good results. The 2" driver is used in the Pluto kit. As links tend to stop working after a few years here is a picture.
https://www.madisoundspeakerstore.com/index.php?p=catalog&mode=search&search_str=aurasound
The third driver is not individually available, but is used in the very unique and expensive LeedH Acoustic Beauty speakers. A pair of these speakers sells for around 16,000 euro. These are very interesting as they have a ferro-fluid, essentially liquid O - ring, suspension and linear travel of around 1" peak to peak for what is a ~ 2" diameter driver. It would be an ambitious DIY project to try to build a driver like this.
https://www.acoustical-beauty.com/index-uk.php
" ... LEEDH designers have entirely invented a new type of electrodynamic speaker, the HPAB (Acoustical Beauty Speaker). The HPAB does not contain a soft iron pole piece in the motor, a foam edge, or a spider in the suspension, because these three components produce the majority of defects in traditional speakers.
Two breakthrough technologies, the ironless motor and suspension with a Ferro fluid seal, internationally patented with the University of Maine, have replaced the traditional motor and suspension. "