The secondary effect is clear; derailed discussions
Mm, or are they the primary effect? 🤔
Sorry, this is an artifact from an old comment that I didn't notice and forgot to delete. If you can, please edit my comment and erase this phrase.
These distortions are related to the voice coil rather than the cone displacement. Their nature is well undestood and we know how to deal with them optimizing the magnet structure.
It is a trivial truth: linear distortions aeffects the frequency response of a loudspeaker. However, If your hipotesis were right, we should obtain correlation between on-axis SPL changes due to linear effects and an increase of THD level in the corresponding frequency range. Such situation can be obtained in compression drivers, when the harmonics frequency coincides with the break-up mode so we can obtain THD peaks in THD frequency response. But this is not the same situation with LF drivers. For example, in picture from paper by Yamamoto & Tsukagoshi I cite above both LF drivers has quite comparable on-axis frequency response with a couple of dB variation. Whereas variation in THD level in the midrange is almost 10dB, which is hard to explain just by difference in frequency response due to linear break-up effect.
Similar situation can be obtained in modern LF loudspeakers with the identical magnet system and the cone/suspention, but with different cones.
https://hificompass.com/en/reviews/satori-mw16tx-4
For example I post the picture from HiFiCompass page for Satori MW16TX(textreme) vs. Satori MW16P(paper). As we can see, the paper cone has break-up problem above ~1kHz (apparently it is a typical structural mode appearing in the paper cone loudspeakers due to mismatch of mechanical impedances of the loudspeaker cone and suspention). On-axis SPL variation of the the both speaker lies in the range 2-4 dB, whereas variation in the second harmonic is in the range of 5-20dB (softer paper cone has higher distortions) in the midrange below 1.5kHz, and the THD level is drastically reduced in the paper cone above ~1.5 kHz. Like in the Yamamoto and Tsukagoshi case, it is hard to explain 5-20 dB larger THD level in the loudspeaker with softer paper cone just by linear break-up effect.
But so far, your hypothesis about linearity of the break-up does not quite fit into the experimental data for LF drivers.
Last edited:
To add on to this, Klippel specify that at higher frequencies where the cone is not pistonic, there exist mechanical nonlinearities that lead to distortion. See the image below (taken from Loudspeaker Nonlinearities - Causes, Parameters, Symptoms):
There are also further examples in the paper about measured drivers, where the distortion in current and sound pressure are compared, showing harmonic distortion due to nonlinear cone vibration at higher frequencies (see page 51 and page 58) as well as more details about the kinds of distortion you can expect from nonlinear cone vibration (page 38, 41, and 44). The paper is freely available on Klippel's site: https://www.klippel.de/fileadmin/_m...linearities–Causes_Parameters_Symptoms_01.pdf
There are also further examples in the paper about measured drivers, where the distortion in current and sound pressure are compared, showing harmonic distortion due to nonlinear cone vibration at higher frequencies (see page 51 and page 58) as well as more details about the kinds of distortion you can expect from nonlinear cone vibration (page 38, 41, and 44). The paper is freely available on Klippel's site: https://www.klippel.de/fileadmin/_m...linearities–Causes_Parameters_Symptoms_01.pdf
Yes I remember, but theres gotta be something, Crossing at 200hz?.... Excursion looks nice with a set of 4. The pic below is using ther force flat FR feature and SPl/1m matched at 117db. The lowest excursion is a 8 - AE18"s, highest is 4 - AE18"s and slightly above the lowest excursion example are simply 4 Dayton 21ers.... I'd have them facing forward, 2 front 2 rear.... Yes I have the 15" that can still be the bandain between the lowest driver... but... Mating this 316g mms to a light weight diaphragm on horn.... Why does it sound like it could have issues... Or would it have been issues except that 200hz is too low for those issue to come to perception. Might it be an issue of timbre, in regards to thd.... yet, with excursion so low.... I remember @gedlee reminding me that is all in the measurements of the actual woofer, but what does theory suggest in regards to mms vs frequency? @b_force I'll have to go back and read your thread focusing on the performance of a driver based around Low excursion and F, but I remember that as always, excursion is the heart of most issues. I envision that the negatives of mass are seen at large excursions. I picture decay lengths increasing with mass due to storage, but at 200hz and below? Is it just not an issue? The Dr. Geddes in me says that CSD of woofers are never an issue except in poor performing drivers... I wonder if there is already a thought that, one not want to use a sub for a reference monitor if the mass to motor force ratio isn't low enough? Otherwise, this Dayton appears to be a good candidate for 200hz and lower. Possibly a better candidate, but I am not confident.
AE18h+
PSS555
Last edited:
Actually they are related to cone displacement since the VC and cone are connected, but I get your point.
And you make some good points.
But, it is certainly wrong to assume that all cone modes are nonlinear, and I will agree that it is possible for them to be nonlinear. One usually assumes linearity until forced to acknowledge that nonlinearity occurs. Each case must then be looked at independently.
I will contend that in a well designed cone they are most likely linear and that nonlinear is the exception not the norm.
But all that is irrelevant in my mind as no test has ever shown THD to be correlated to perception. Hence, either way, it is uninteresting to me.
Far too much emphasis is placed on THD measurements in drivers. I've build competent loudspeaker system and never even considered THD. Experts like Toole and Olive are also in agreement with this point of view.
Last edited:
It's a resonance of the spider where the outside edges are clamped (although the inside on is not really clamped, in this frequency range it is.) This puts the resonance mostly out of phase from the cone and it bleads out cone displacement resulting in a FR dip. But then on a waterfall plot, it usually has a ringing as this mode is not well damped.
I searched the literature on the problem of nonlinear break-up modes in membranes. There is not much literature on this problem, but I managed to find something.
Modelling and Measurement of Nonlinear Intermodal Coupling in Loudspeaker Diaphragm Vibrations (please, send me PM the article if anyone has it )
Mechanical Resonances and Geometrical Nonlinearities in Electrodynamic Loudspeakers (I can share via PM not for public distribution)
Modal Analysis and Nonlinear Normal Modes (NNM) on Moving Assemblies of Loudspeakers (I can share via PM not for public distribution)
I will probably end the further discussion on this issue, since further debates will require a sufficient in-depth study of the specific academic literature.
Modelling and Measurement of Nonlinear Intermodal Coupling in Loudspeaker Diaphragm Vibrations (please, send me PM the article if anyone has it )
Mechanical Resonances and Geometrical Nonlinearities in Electrodynamic Loudspeakers (I can share via PM not for public distribution)
Modal Analysis and Nonlinear Normal Modes (NNM) on Moving Assemblies of Loudspeakers (I can share via PM not for public distribution)
I will probably end the further discussion on this issue, since further debates will require a sufficient in-depth study of the specific academic literature.
Last edited:
what do you consider most when building a loudspeaker systems ? what things briefly govern your driver and crossover choices ? for any given size loudspeaker do you feel it Is essential for instance to avoid the cone breakup as much as is possible ? do you sort of over engineer the speaker then everything is just cruising ?
thanks very much
James
I'm no Geddes, but if you start your playback system design from fundamentals like ability to output full bandwidth 20-20kHz considering effects of room, you'll anyway endup with a system thats quite big, and has enough SPL capability so that distortion is not an issue. It's just never a problem. If you want a small system for some reason, for aesthetics or trends you wanna follow, then you'd endup with all kinds of issues like not being able to reproduce 20-20kHz with sufficient SPL. Big systems also have issues, like difficulty of building, and bulk, a different set of compromises. It's just what you prioritize and why.
The most important thing is the free field frequency response, followed by the smoothness/flatness of the DI (Directivity index), then the level of the DI (higher is better.) LF driver choices are always dictated by system size limitations, but bigger is always better (TS values mean almost nothing to me since I always use closed box multiple subs with DSP EQ for this region.) It is important to avoid the cone breakup mode region, but if this cannot be avoided (as is often the case) then how well controlled it is becomes critical - this varies dramatically from driver to driver. I don't over-engineer deliberately, but it tends to work out that way since bigger drivers have higher DI and better low end response with lower excursion requirements, so they tend to seem over-engineered. Crossover point is dictated by the DI and the DI must match between drivers at the crossover.
The single biggest problem that I noted in my designs over the years was the cone breakup of the woofer. Even though this was usually above the crossover, the LF drivers response would often bleed through into the HF drivers polar response. With passive crossovers (which 80% of the systems that I built were) very high order LP filters were not viable. With DSP, of course, this was never a problem - you just notch it out in the EQ.
I virtually never did THD measurements of my designs and have never found this to be a shortcoming. Toole once commented to me that he covered THD in his book at about the same rate as the importance that he gave it - 1 paragraph in 300 pages. That's about right IMO.
But most of your systems are also done with high quality woofers that are 12 inch or bigger or so?
Could you elaborate on why, again? I think I want to ask, At what frequency does TS values begin to matter? Is it a matter of excursion? Nothing matters at 1-2mm of excursion?
Last edited by a moderator:
Basically yes, as well as if size doesn't matter anyway and the woofers are large enough.
Yeah, than you can basically correct everything afterward anyway, especially when you're far from being SPL limited.
Still you have to watch out for speakers with a small enough Qt en Vas
Definitely not your average setup and general approach.
Wouldn't recommend that for like a good 8 or 10 inch subwoofer.
Unless you like to blow up things quickly.