Debunking SL's „Stored Energy“

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
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.

Underdamping not only increases the amplitude of oscillation at a given frequency, it also increases the bandwidth of oscillations that are multiples of the fundamental resonance (harmonics).

Your quote above demonstrates your lack of experience and understanding on the subject of active versus passive signal level control.

"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."

Active and passive systems behave essentially the same when energy is being subtracted from the oscillating system. When energy is being selectively added via signal level eq and subsequent amplification, a fundamental difference is occurring. Effectively, you are introducing a negative impedance. A close examination of the physical q of the driver throughout the affected frequency range will reveal a spike at the frequency in which boost is actively applied. In this case, there is no passive equivalent. Please read the entire post link provided above on the subject of Q thoroughly before casting further judgments on other's understanding of the subject of energy storage in active and passive circuits.
 
Claiming support from the silence of gurus is even sillier than claiming support from the statements of gurus, as some try to do. This is a new rhetorical technique. I will have to try it myself one day.

Look, on the one hand, you suggest that you have a lot of experience in this area. On the other hand, when you make statements like:

"Transient reponse and damping are certainly connected, but they are linear phenomena which have little to do with distortion."


red flags go up. I try to avoid arguments with folks that raise red flags in my mind. So perhaps we can just agree to disagree.
;)
 
But transient response and damping are linear phenomena! Distortion comes from non-linearity. These are simply mathematical facts. Now I suppose it is conceivable that in the world of loudspeaker design these words mean something different from what they mean in other branches of electrical engineering, in which case I apologise for wasting your time.

In the world of loudspeaker design, the concepts and terminology mean the same thing as they do in the field of electrical engineering. Non linear distortion is defined as the production of different frequency spectra from that of the applied signal. Poorly damped (high Q) damped oscillators (what a loudspeaker actually is) produce harmonics. By definition, this is non linear distortion. When examined from the frame of reference of energy versus time - we can see that linear distortion is also present. The terminology or semantics with respect to claimed linear and non linear behavior is really irrelevant in the end. The fact is that the Q of the driver/amplifier system can selectively be substantially increased at a particular frequency. This cannot be done with passive networks. It can be approximated by damping all frequencies except the band in question but because you are not adding energy - the response behavior is usually very different from an active implementation versus passive when a loudspeaker is operating in a region of high energy storage or resonance. This is because you are not actually increasing physical q at the selected frequency - you're decreasing Q everywhere else. I know this because I've conducted experiments that clearly establish the phenomenon of increased non linear distortion with applied eq boost at or near a driver's fundamental resonance.

Nothing you are likely to tell me will dissuade me from that conclusion. If you can produce actual test data, I might be inclined to look at it but given physics, it is highly unlikely that you will be able to produce data that contradicts what I've gathered.
 
It seems that you two are just arguing semantics. :rolleyes:

"Distortion" is commonly assumed in conversation to mean non-linear distortion unless otherwise stated, eg harmonic and intermod distortion which introduces new frequency products.

However more correctly "Distortion" refers to any deviation in waveform between the input and output of a device, and this includes linear distortion like amplitude and phase variations where existing frequencies are altered in amplitude or phase but no new frequencies are created.

I'm generally careful to state "Linear distortion" or "Non-linear distortion" to avoid getting into exactly these type of arguments ;)
 
It seems that you two are just arguing semantics. :rolleyes:

"Distortion" is commonly assumed in conversation to mean non-linear distortion unless otherwise stated, eg harmonic and intermod distortion which introduces new frequency products.

However more correctly "Distortion" refers to any deviation in waveform between the input and output of a device, and this includes linear distortion like amplitude and phase variations where existing frequencies are altered in amplitude or phase but no new frequencies are created.

I'm generally careful to state "Linear distortion" or "Non-linear distortion" to avoid getting into exactly these type of arguments ;)

I'm being very careful to sidestep the semantic argument by focusing on system Q as it relates to active and passive systems. There is most definitely a valuable lesson to be learned here by a significant percentage of DIY'ers who overlook important distinctions between active and passive systems.

I've read over and over again on different forums the opinion that "A Transfer Function is A Transfer Function, Is a Transfer Function". This overly simplistic view completely misses the boat as the end result very much depends on the net energy applied to system and the energy storage characteristics of the driver being driven.
 
Last edited:
"Distortion" is commonly assumed in conversation to mean non-linear distortion unless otherwise stated, eg harmonic and intermod distortion which introduces new frequency products.

fntn specifically stated that high Q oscillators produce increased nonlinear distortion. If the displacement is held constant, this is just simply incorrect; it's easy to see with a basic harmonic oscillator equation with a small second order term in the spring constant. fntn has produced no mathematical analysis or data to support his contention.
 
fntn specifically stated that high Q oscillators produce increased nonlinear distortion. If the displacement is held constant, this is just simply incorrect; it's easy to see with a basic harmonic oscillator equation with a small second order term in the spring constant. fntn has produced no mathematical analysis or data to support his contention.

Your statement

"If displacement is held constant"

speaks volumes about your perspective.
 
fntn specifically stated that high Q oscillators produce increased nonlinear distortion. If the displacement is held constant, this is just simply incorrect; it's easy to see with a basic harmonic oscillator equation with a small second order term in the spring constant. fntn has produced no mathematical analysis or data to support his contention.

Sorry, I should have quoted which comment I was responding to, as by the time I finished my post two more had been posted, losing the context of my reply. I was responding to:
fntn said:
Look, on the one hand, you suggest that you have a lot of experience in this area. On the other hand, when you make statements like:

"Transient reponse and damping are certainly connected, but they are linear phenomena which have little to do with distortion."


red flags go up. I try to avoid arguments with folks that raise red flags in my mind. So perhaps we can just agree to disagree.

Yes it will be interesting to see fntn's analysis. I decided to keep quiet on this point and see how that aspect of the discussion played out :)
 
Sorry, I should have quoted which comment I was responding to, as by the time I finished my post two more had been posted, losing the context of my reply. I was responding to:


Yes it will be interesting to see fntn's analysis. I decided to keep quiet on this point and see how that aspect of the discussion played out :)

A sample of the data I have collected over time was posted on the ZDL thread. It shows a substantial increase in non linear distortion when a 4 db boost was added to an active system at 150hz or at the fourth harmonic of the driver's fundamental resonance. What was notable about the data was that the overall signal level was reduced for the case in which high distortion was measured versus the case in which no selective frequency boost was applied. In other words, the net increase at the "boosted" frequency actually amounted to a net decrease in signal drive by about 6 db since the overall drive level for the non "boosted" case was 10db higher than the case in which a selected 150hz 4 db boost was applied. Excuses were put forth that the driver was inferior and that the increase in distortion was due only to amplitude increases. Clearly, the overall amplitude of the applied signal didn't go up - but the harmonics sure did. ;)
 
So did the displacement. Duh.

Relative to the "unboosted" case, it did not. Relative to the remainder of the spectral output at the reduced drive level (-10db case) it did. The absolute level of excursion or displacement did not - otherwise the microphone would have picked up a net increase in acoustic output of the "boosted " case over the "unboosted" case. The graph of response shows that the net result was for the "boosted" signal to have an acoustic ouput 10 db lower than the "unboosted" signal - meaning that while the 150hz boost increased output - it's absolute output was still 6db lower than that of the "unboosted" case at the frequency in question - 150 hz.

Edit:
You are apparently trying to assert that reduced damping only increases amplitude. It also increases the number of harmonics - both occur and this should be obvious to anyone - even someone who's famous for making plastic corks.


Duh.
 
Last edited:
No. It is obvious to me that reduced damping reduces harmonics of an oscillator at a given amplitude and given non-linearity. This is because the higher Q helps filter out the harmonics due to narrower bandwidth. Have you ever read any oscillator theory?

If a system has a peculiar non-linear response then it is possible to get higher non-linear distortion (as a percentage) from a smaller signal. In effect, a lot of high order terms conspire together to cancel out the effects of lower order terms but this is very level-dependent. I can imagine stiction in a mechanical system doing this. Voice-coil rubbing, maybe?

If I do an experiment, and measure certain things, then the things I measure are true (provided I know what I am doing). My deductions from those measurements may or may not be true, depending on whether they are based on sound science. A statement that higher Q leads to more harmonics may be a possible deduction from an experiment, but it is in general a false statement, so is likely to be false in the context of that experiment too.

Could you please review your understanding of the solutions of second-order linear differential equations, and specifically the damped driven oscillator? Without this you cannot understand loudspeakers.

BTW you can increase Q by passive means - just add some energy storage.
 
Do you mean this "measurement"?

http://www.diyaudio.com/forums/multi-way/178350-zdl-36.html#post2477519

Your only response to the criticisms of what you presented (including from, in your words, gurus) was, oh, I did this months ago for a different reason and didn't expect people to actually analyze it. I think some better taken and better controlled data (or a few minutes with some basic differential equations as Dave suggested) would be a more appropriate way of supporting a remarkable assertion. At constant displacement, the distortion will not increase.
 
According to fntn, the truth of a belief is established by it being disputed by no more than one person. As his ideas on loudspeakers now seem to have been disputed by at least three or four people (quite independently in two different threads - I have only just seen the other one) would he now agree that, on his own criterion, his view must be false?

Please note that I am not saying that he must be wrong because people disagree with him. I don't believe that, but he does.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.