Having becoming aware that many, many seem to have pretty „personalized“/ weird ideas about energy storage in real world speakers, it may be a nice idea to go slightly „more in depth“ into this topic than usually happens.
The monumental Babylonian confusion being documented over hundreds of posts, at
http://www.diyaudio.com/forums/everything-else/173283-cmp-framing-what-you-mean.html
do show us that even „big“ boys (and SL himself of course) seem to have no clear idea about „Stored Energy“, hence rather relaying on superficial wording and some „easy on the eye“ plots instead of having investigated the very details.
So, lets see what possibly can be put back into perspective by some simple „monkeys button pressing“ plus giving the whole topic an additional thought or two.
What will be covered in order to correct some wide spread mis-beliefs:
- „Energy storage“ simply is *a must* of any real world loudspeaker
- „Energy storage“ is *not* (!) an „ill behaviour to be avoided“ - as its seen by many and implicitly was coined that way by SL (for sure its not the case in a general sense)
- loosing or gaining *control over cone motion* has nothing to do with „stored energy“ term, as - implicitly - is said by SL to be the case (and is the believe of many)
- the mass of any speakers diaphragm is directly linked to the „Energy storage“ in that speaker (as is the stiffness of the spider) – but - has *nothing whatsoever* to do with any sonic quality of that very speaker – besides sensitivity and band width (if this is thought of being a „sonic“ quality).
in addition it will be focused at some root causes for current Babylonian confusion / misinterpretations
- SL's „stored energy“ term actually does not tell us anything useful - neither anything new nor anything more pinpoint
- „stored energy“ term - as coined by SL's - is fogging the technical meaning regarding „energy storage“, as it is valid and perceived all over the world except in many's „audio speak“
- The really interesting thing that „could“ be seen in some measurements presented by SL sadly haven't been pin pointed ( - by him
) to the real roots
and finally a suggestion is made regarding how to possibly proceed from ancient TEF thinking towards more appropriate measurement methods if time domain behaviour has to be revealed
- The measurement method used from SL to show – what he thinks is - „stored energy“, I suggest to better be substituted by more elegant and less time consuming methods
( wavelet analysis for example)
Michael
„Disclaimer“:
Sorry for not being able to „go for the ball instead of going for the person“ here.
The topic at hand simply can not be kept apart from the person IMO – same case as the „minimum diffraction waveguide“ bogus could not be kept apart from just another „big“ guy.
The monumental Babylonian confusion being documented over hundreds of posts, at
http://www.diyaudio.com/forums/everything-else/173283-cmp-framing-what-you-mean.html
do show us that even „big“ boys (and SL himself of course) seem to have no clear idea about „Stored Energy“, hence rather relaying on superficial wording and some „easy on the eye“ plots instead of having investigated the very details.
So, lets see what possibly can be put back into perspective by some simple „monkeys button pressing“ plus giving the whole topic an additional thought or two.
What will be covered in order to correct some wide spread mis-beliefs:
- „Energy storage“ simply is *a must* of any real world loudspeaker
- „Energy storage“ is *not* (!) an „ill behaviour to be avoided“ - as its seen by many and implicitly was coined that way by SL (for sure its not the case in a general sense)
- loosing or gaining *control over cone motion* has nothing to do with „stored energy“ term, as - implicitly - is said by SL to be the case (and is the believe of many)
- the mass of any speakers diaphragm is directly linked to the „Energy storage“ in that speaker (as is the stiffness of the spider) – but - has *nothing whatsoever* to do with any sonic quality of that very speaker – besides sensitivity and band width (if this is thought of being a „sonic“ quality).
in addition it will be focused at some root causes for current Babylonian confusion / misinterpretations
- SL's „stored energy“ term actually does not tell us anything useful - neither anything new nor anything more pinpoint
- „stored energy“ term - as coined by SL's - is fogging the technical meaning regarding „energy storage“, as it is valid and perceived all over the world except in many's „audio speak“
- The really interesting thing that „could“ be seen in some measurements presented by SL sadly haven't been pin pointed ( - by him
) to the real rootsand finally a suggestion is made regarding how to possibly proceed from ancient TEF thinking towards more appropriate measurement methods if time domain behaviour has to be revealed
- The measurement method used from SL to show – what he thinks is - „stored energy“, I suggest to better be substituted by more elegant and less time consuming methods
( wavelet analysis for example)
Michael
„Disclaimer“:
Sorry for not being able to „go for the ball instead of going for the person“ here.
The topic at hand simply can not be kept apart from the person IMO – same case as the „minimum diffraction waveguide“ bogus could not be kept apart from just another „big“ guy.
Starting out and to summarize with what SL has shown under „Investigation of energy storage“ look at the plots / pix below:
Issues in speaker design - 2
referring to plots like:
we see that SL is tagging his „Stored Energy“ lable to the areas below „any extra tail“ after stimulating a speaker with a specific burst signal.
The stimulus signal he uses is a 4 cycle Blackman windowed sine of a specific frequency.
SL provides a 1kHz sample for download under „Shaped toneburst generator“ (to animate some playing aroung, I guess):
System Test
http://www.linkwitzlab.com/images/sound/1kblkman4.wav
The shape of that signal is close to what can be seen here:
as is the processed (peak rectified + log) outcome quite similar as seen in above pix (as we will see later).
This „transfer/ transform processing“ SL is performing by using a analog circuit like shown here:
redrawn/ rearranged by Rod Elliott / Ray Hernan as shown here:
Linkwitz Cosine Burst Generator
The smoth osciloscope plots finally shown by SL are envelope functions of whats coming out of above circuit.
Michael
Issues in speaker design - 2
referring to plots like:
we see that SL is tagging his „Stored Energy“ lable to the areas below „any extra tail“ after stimulating a speaker with a specific burst signal.
The stimulus signal he uses is a 4 cycle Blackman windowed sine of a specific frequency.
SL provides a 1kHz sample for download under „Shaped toneburst generator“ (to animate some playing aroung, I guess):
System Test
http://www.linkwitzlab.com/images/sound/1kblkman4.wav
The shape of that signal is close to what can be seen here:
as is the processed (peak rectified + log) outcome quite similar as seen in above pix (as we will see later).
This „transfer/ transform processing“ SL is performing by using a analog circuit like shown here:
An externally hosted image should be here but it was not working when we last tested it.
redrawn/ rearranged by Rod Elliott / Ray Hernan as shown here:
Linkwitz Cosine Burst Generator
An externally hosted image should be here but it was not working when we last tested it.
The smoth osciloscope plots finally shown by SL are envelope functions of whats coming out of above circuit.
Michael
Above transfer/ transform circuit of SL can be substituted by a simple „LTC1968 - Precision Wide Bandwidth, RMS-to-DC“ (or anything like) plus any log amp circuit.
In fact this its what I did first to get a handle on what SL was actually doing / measuring.
Being not only a monkey that loves to press buttons - but being a lazy monky at first hand – I diceded not to built that circuit but rather run a SPICE simu instead:
Its basically the same circuit as SL is using and this is whats it's output:
Green trace is the input signals as used by SL .
Blue trace is the output of SL's circuit
As said before – the plots SL is showning at his page are actually the envelope function of the circuits output.
SPICE is a very handy tool to perform all kinds of time domain or frequency domain analyisis.
Michael
In fact this its what I did first to get a handle on what SL was actually doing / measuring.
Being not only a monkey that loves to press buttons - but being a lazy monky at first hand
– I diceded not to built that circuit but rather run a SPICE simu instead:
Its basically the same circuit as SL is using and this is whats it's output:
Green trace is the input signals as used by SL .
Blue trace is the output of SL's circuit
As said before – the plots SL is showning at his page are actually the envelope function of the circuits output.
SPICE is a very handy tool to perform all kinds of time domain or frequency domain analyisis.
Michael
Most audio folks are more familiar with some dedicated audio measurement software though.
This is what ARTA displays in analysis „Energy Decay“ fed by SL's stimulus:
Same analysis type than SL's dedicated circuit - as we easily can see...
Again - as said before – the plots SL is showning at his page are actually the envelope function of these plots.
Michael
This is what ARTA displays in analysis „Energy Decay“ fed by SL's stimulus:
Same analysis type than SL's dedicated circuit - as we easily can see...
Again - as said before – the plots SL is showning at his page are actually the envelope function of these plots.
Michael
Now having all the toys together, I guess its time to have some fun to play with ...
Ah – one thing I forgot …
We either need to insert a real speaker into the measurements – or – we do that in a simplified form by adding some circuits into the SPICE simu – or – we do that by inserting the EQing of some hardware.
In the following measurements the Behring 24/96 was used to add some well defined humps and bumps to otherwise perfectly flat frequency response.
Starting out with what SL would call serious „stored energy“:
Above plot simply shows the behaviour of a boost of 6dB at 1000Hz with a Q of 2 – nothing more.
To even exaggerate this lets look at this one:
Above plot simply shows the behaviour of a boost of 6dB at 1000Hz with a Q of 10 – nothing more.
To even exaggerate this one lets look here:
Above plot simply shows the behaviour of a boost of 12dB at 1000Hz with a Q of 10 – nothing more.
Michael
Ah – one thing I forgot …
We either need to insert a real speaker into the measurements – or – we do that in a simplified form by adding some circuits into the SPICE simu – or – we do that by inserting the EQing of some hardware.
In the following measurements the Behring 24/96 was used to add some well defined humps and bumps to otherwise perfectly flat frequency response.
Starting out with what SL would call serious „stored energy“:
Above plot simply shows the behaviour of a boost of 6dB at 1000Hz with a Q of 2 – nothing more.
To even exaggerate this lets look at this one:
Above plot simply shows the behaviour of a boost of 6dB at 1000Hz with a Q of 10 – nothing more.
To even exaggerate this one lets look here:
Above plot simply shows the behaviour of a boost of 12dB at 1000Hz with a Q of 10 – nothing more.
Michael
Of course not only humps in the frequency response do show sub optimal time domain behaviour:
Above is a simple LR 48dB high pass at 1kHz – showning considerable „stored energy“ in SL's terms – no ?
Or this one:
Actually being only a notch of -6dB at 1kHz with Q=2
Or this one:
Actually being only a notch of -6dB at 1kHz with Q=10
Michael
Above is a simple LR 48dB high pass at 1kHz – showning considerable „stored energy“ in SL's terms – no ?
Or this one:
Actually being only a notch of -6dB at 1kHz with Q=2
Or this one:
Actually being only a notch of -6dB at 1kHz with Q=10
Michael
So having seen a lot of – what according to SL has to be called „stored energy“ - with measurements that do not even use a speaker, we come to the easy conclusion *that SL's „stored energy“ term does not tell us anything specifically related to ill behaviour of speakers*.
All the „sub optimal time behaviour“ shown above is a simple and mere function of frequency response deviations from plain flat.
At this point I'd like to stess that those „sub optimal time behaviour“ can be perfectly corrected by equalizing FR back to flat.
Well – sometimes even the most obvious has to be stated !
No big deal – but not the whole story either ...
Michael
All the „sub optimal time behaviour“ shown above is a simple and mere function of frequency response deviations from plain flat.
At this point I'd like to stess that those „sub optimal time behaviour“ can be perfectly corrected by equalizing FR back to flat.
Well – sometimes even the most obvious has to be stated !
No big deal – but not the whole story either ...
Michael
Last edited:
Now that we have seen that SL's „Stored Enerergy“ is not related to the mass of a loudspeaker diaphragm only – we possibly should have a closer look at the „real“ energy storage in a mechanical system – which a loudspeaker actually represents.
The most simplified mechanical equivalent of a loudspeaker can be embraced by the stiff mass of the diaphragm plus the ideal spring of the spider.
At any time of operating a speaker there is – either : movement of the diaphragm – or : displacemet of the diapragm happening.
So - from basic laws of physics: *there is energy stored in a speaker at any time of operation* – no matter what.
At this point I'd like to stess that *the energy that is stored in the speaker – more precisely in the mass' movement or in the spring's displacement - has no impact on the sonic pattern of that speaker whatsoever*.
Well – sometimes even the most obvious has to be stated !
No big deal – but not the whole story either …
Michael
The most simplified mechanical equivalent of a loudspeaker can be embraced by the stiff mass of the diaphragm plus the ideal spring of the spider.
At any time of operating a speaker there is – either : movement of the diaphragm – or : displacemet of the diapragm happening.
So - from basic laws of physics: *there is energy stored in a speaker at any time of operation* – no matter what.
At this point I'd like to stess that *the energy that is stored in the speaker – more precisely in the mass' movement or in the spring's displacement - has no impact on the sonic pattern of that speaker whatsoever*.
Well – sometimes even the most obvious has to be stated !
No big deal – but not the whole story either …
Michael
Now - having looked into analyzing behaviour of filters and into the mechanic system a loudspeaker represents, we safely can conclude that all that energy storage effects presented above are not a „speakers behaviour“ but a mere filter behaviour / pattern.
At this point I'd like to stress that *even if the root cause with respect to energy storage is in the mechanics of a speaker (springs and masses) it does not make any difference if its in caps and coils instead : same behaviour *.
As a subsequent conclusion i'd like to stress at this point that *„control“ over the speaker has nothing to do with energy that is stored in the system*.
To being able to actually engrave to our liking quite any specific filter behaviour into the system (speaker + EQing) actually is the kind of 100% control we have over such systems.
Well – sometimes even the most obvious has to be stated !
No big deal – but not the whole story either …
Michael
At this point I'd like to stress that *even if the root cause with respect to energy storage is in the mechanics of a speaker (springs and masses) it does not make any difference if its in caps and coils instead : same behaviour *.
As a subsequent conclusion i'd like to stress at this point that *„control“ over the speaker has nothing to do with energy that is stored in the system*.
To being able to actually engrave to our liking quite any specific filter behaviour into the system (speaker + EQing) actually is the kind of 100% control we have over such systems.
Well – sometimes even the most obvious has to be stated !
No big deal – but not the whole story either …
Michael
Now that we have seen that :
- SL's „stored energy“ term actually does not tell us anything useful - neither anything new nor anything more pinpoint
and:
- „stored energy“ term - as coined by SL's - is just fogging the technical meaning regarding „energy storage“, as it is valid and perceived all over the world except in some's „audio speak“
we come to the interesting part that was missed until now.
The conclusion out of whats been outlined above, is that the presented systems can be EQed (upstreams) to our liking = impulse response shaping at will.
This is a save bet at least for any „stringent“ MP behavior system.
For real world loudspeakers - being to different extent „Consecutive Min Phase“ systems (in RF speak known as multi path issue) this does not hold - CMP systems being by far more complicated animal.
Speakers / chassis that show pronounced CMP behaviour face some additional limitations
In particular :
- for CMP systems the concept of „frequency response“ is void
- CMP systems show a specific sonic pattern not seen elsewhere
- CMP systems face limitations with respect to equalization beyond MP systems
Of course CMP effects would show up in any analysis – even if I would do with above
- but thats quite another story ...
Part of that story is told here:
http://www.diyaudio.com/forums/multi-way/161627-horn-honk-wanted-17.html#post2492102
which brings us to "prove" for the last statement ( 8.) ) that was claimed at the beginning:
1.) „Energy storage“ simply is *a must* of any real world loudspeaker
2.) „Energy storage“ is *not* (!) an „ill behaviour to be avoided“ - as its seen by many and implicitly was coined that way by SL (for sure its not the case in a general sense)
3.) loosing or gaining *control over cone motion* has nothing to do with „stored energy“ term, as - implicitly - is said by SL to be the case (and is the believe of many)
4.) the mass of any speakers diaphragm is directly linked to the „Energy storage“ in that speaker (as is the stiffness of the spider) – but - has *nothing whatsoever* to do with any sonic quality of that very speaker – besides sensitivity (if this is thought of being a „sonic“ quality).
5.) SL's „stored energy“ term actually does not tell us anything useful - neither anything new nor anything more pinpoint
6.) „stored energy“ term - as coined by SL's - is fogging the technical meaning regarding „energy storage“, as it is valid and perceived all over the world except in many's „audio speak“
7.) The really interesting thing that „could“ be seen in some measurements presented by SL sadly haven't been pin pointed ( - by him ) to the real roots
8.) The measurement method used from SL to show – what he thinks is - „stored energy“, I suggest to better be substituted by more elegant and less time consuming methods to reveal time domain behaviour
- "wavelet analysis" that is
Michael
- SL's „stored energy“ term actually does not tell us anything useful - neither anything new nor anything more pinpoint
and:
- „stored energy“ term - as coined by SL's - is just fogging the technical meaning regarding „energy storage“, as it is valid and perceived all over the world except in some's „audio speak“
we come to the interesting part that was missed until now.
The conclusion out of whats been outlined above, is that the presented systems can be EQed (upstreams) to our liking = impulse response shaping at will.
This is a save bet at least for any „stringent“ MP behavior system.
For real world loudspeakers - being to different extent „Consecutive Min Phase“ systems (in RF speak known as multi path issue) this does not hold - CMP systems being by far more complicated animal.
Speakers / chassis that show pronounced CMP behaviour face some additional limitations
In particular :
- for CMP systems the concept of „frequency response“ is void
- CMP systems show a specific sonic pattern not seen elsewhere
- CMP systems face limitations with respect to equalization beyond MP systems
Of course CMP effects would show up in any analysis – even if I would do with above
- but thats quite another story ...
Part of that story is told here:
http://www.diyaudio.com/forums/multi-way/161627-horn-honk-wanted-17.html#post2492102
which brings us to "prove" for the last statement ( 8.) ) that was claimed at the beginning:
1.) „Energy storage“ simply is *a must* of any real world loudspeaker
2.) „Energy storage“ is *not* (!) an „ill behaviour to be avoided“ - as its seen by many and implicitly was coined that way by SL (for sure its not the case in a general sense)
3.) loosing or gaining *control over cone motion* has nothing to do with „stored energy“ term, as - implicitly - is said by SL to be the case (and is the believe of many)
4.) the mass of any speakers diaphragm is directly linked to the „Energy storage“ in that speaker (as is the stiffness of the spider) – but - has *nothing whatsoever* to do with any sonic quality of that very speaker – besides sensitivity (if this is thought of being a „sonic“ quality).
5.) SL's „stored energy“ term actually does not tell us anything useful - neither anything new nor anything more pinpoint
6.) „stored energy“ term - as coined by SL's - is fogging the technical meaning regarding „energy storage“, as it is valid and perceived all over the world except in many's „audio speak“
7.) The really interesting thing that „could“ be seen in some measurements presented by SL sadly haven't been pin pointed ( - by him
) to the real roots8.) The measurement method used from SL to show – what he thinks is - „stored energy“, I suggest to better be substituted by more elegant and less time consuming methods to reveal time domain behaviour
- "wavelet analysis" that is
Michael
Last edited:
If I understand you correctly, you are saying that:
- any system resonance involves stored energy
- an electrical resonance is the same as a mechanical resonance
- the resultant system frequency response can be adjusted to any other causal frequency response by the use of appropriate compensation
This is basic physics plus filter theory. If that is what you are saying then it is obviously true. However, at times it feels like you may be tilting at windmills. People who don't understand physics are not impressed by arguments that go over their heads; they merely continue to assert their intuitive misunderstanding.
- any system resonance involves stored energy
- an electrical resonance is the same as a mechanical resonance
- the resultant system frequency response can be adjusted to any other causal frequency response by the use of appropriate compensation
This is basic physics plus filter theory. If that is what you are saying then it is obviously true. However, at times it feels like you may be tilting at windmills. People who don't understand physics are not impressed by arguments that go over their heads; they merely continue to assert their intuitive misunderstanding.
If two crossovers have the same frequency response, and the same terminal impedances, then they should have the same effect on the total system response. One may use passive components, which may store energy; the other may use active circuits, which draw energy from a PSU. So fixating on energy storage can lead people astray.
Your title "Debunking SL's Stored Energy" implies that that the issue of stored energy is some inapt obsession of SL's. In the relevant section of Linkwitzlab.com SL addresses 29 individual topics which he identifies as important to optimizing speaker performance. Stored energy is topic "M" at 13 of 29 and gets about 1/2 of a page out of a total of 7 pages. As he does with other technical aspects of speaker design, SL explores the theoretical and practical implications of stored energy in speaker design. Anyone familiar with the basics of dynamic electro-mechanical systems will recognize that stored energy in the system, whether stored mechanically in the drivers, electrically in the crossover, or mechanically in the structure of the loudspeaker will necessarily decouple the corresponding electrical input into and sonic output out of the speaker.
In the opening paragraph of the section SL puts the topic in its proper context and addresses a number of the points you have raised so...what is it your 10 serial posts are debunking exactly???
In the opening paragraph of the section SL puts the topic in its proper context and addresses a number of the points you have raised so...what is it your 10 serial posts are debunking exactly???
Last edited:
Very simply put, the level of stored energy in any moving mass is a measure of the quality factor of that mass's motion. Critically damped oscillators possess the best distortion and transient characteristics. Those that are overdamped or underdamped exhibit qualities that are less desireable - containing higher levels of distortion. One can try to overcomplicate or confuse the subject with intensely theoretical discussions about minimum phase but the effects of stored energy are neither theoretical nor are they very difficult to document and observe. Linkwitz' transform circuit underscores the simple nature of the problem. He places zeros at the fundamental mechanical resonance of the driver to counter the effects of peak energy storage and poles at lower frequencies to boost response in opposition to the driver's natural rolloff. Nothing in the OP's "debunking" exercise has demonstrated any error, oversight, or misunderstanding on Linkwitz' part.
Unfortunately, threads like this suggest that there's a good mix of ego massaging in audio internet forums along with the more valuable sharing of real, useful, and actionable information.
http://en.wikipedia.org/wiki/Q_factor
Unfortunately, threads like this suggest that there's a good mix of ego massaging in audio internet forums along with the more valuable sharing of real, useful, and actionable information.
http://en.wikipedia.org/wiki/Q_factor
You may be confusing distortion with transient response - or are you using a generalised definition of distortion to mean 'anything I don't want to happen'?
If you are saying that the overall system response is best when any oscillators are critically damped then I guess you are right, as this is the best compromise between ringing and sluggishness.
If you are saying that the overall system response is best when any oscillators are critically damped then I guess you are right, as this is the best compromise between ringing and sluggishness.
Oscillations naturally produce harmonics. The concepts of transient response and damping are where the entities of non linear distortion and linear distortion meet. When the effects of damping are minimized (neither insufficient nor excessive), the applied force (input signal) and its effects are permitted to dominate. When stored energy or damping force dominates, the result is distortion. Anyone who has bothered to look at an ETC or non linear plot for a midwoofer at resonance has seen the obvious evidence of this.
I would have thought that most of the energy stored within passive xovers would be stored by the inductors (at least it is the series inductors which in my experience make passive xovers so vastly inferior, could well be just the way they store and release energy rather than the actual amount) which do not exist in active ones.
Either way they are not connected to the driver directly which may also reduce any negative effects.
No. Nonlinearity produces harmonics. Poorly damped oscillation may increase amplitude, which makes the effects of nonlinearity worse.
Transient reponse and damping are certainly connected, but they are linear phenomena which have little to do with distortion.
Capacitors store energy too. The difference is that typical inductors are more lossy and less linear than typical capacitors. It is not the energy storage which is the problem, but the non-ideal nature of the component. Your concerns might disappear if perfect inductors were available.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.