Revel F206 by ASR https://www.audiosciencereview.com/forum/index.php?threads/revel-f206-tower-speaker-review.53691/
F228 https://www.audiosciencereview.com/forum/index.php?threads/revel-f228be-review-speaker.23659/
F328Be
.. I would not say that 2,7kHz harmonic peak is a problem, it is still lower than both above
F228 https://www.audiosciencereview.com/forum/index.php?threads/revel-f228be-review-speaker.23659/
F328Be
.. I would not say that 2,7kHz harmonic peak is a problem, it is still lower than both above
This had nothing to do with the cost being reflective of a tweeter of a particular type but just to do with if you're going to destruction test a bunch of tweeters you don't want them to cost $500 each.
Besides the cost has nothing to do with it you see tweeters designed for, somewhat, different applications, at all price points.
The point is that you keep harping on about tweeter reliability and robustness when used in certain applications. Such as the Wilsons. Complaining that you don't want the tweeter to die on you when along comes a cymbal crash. You keep going on about how tweeters won't tolerate more than 10-15 watts continuously. No one is disputing this fact. But when it's mentioned that music is transient in nature and with very little energy in the top octaves you dig your heels in and keep repeating the same things without any evidence to back it up.
People come along with technical reasons for why this is going to be much less of a problem than you're trying to make it out to be and you then reject their reasons without anything other than your opinion.
What I am saying is that if you're going to be this belligerent and dogmatic about your position on tweeters then you need hard evidence to back it up. 10-15 watts continuously? Great. Go and create models and simulate the average power dissipation for a tweeter with variables for acoustic crossover slope/peak SPL and the ability to load in a music stimulus of your choice. Either that or actually destruction test some tweeters instead using music.
I expected this type of reaction from you. My opinions are in the end my own. I’ve accumulated enough knowledge over the course of my lifetime working with speaker design and most of this is practical in nature. I don't need to quote engineering literature or theory just to make a fairly objective, technical based statement regarding fundamental operating principles of speaker components. I've worked in driver manufacturing and understand all the basic aspects of the subject. If you demand accreditation of my knowledge and skillset, that's a massive waste of time and I'm not going to feed you the line you expect.
That said, i can show you endless examples of 25 - 28mm dome tweeters which ALL handle about 10 - 15 watts continously before blowing up. Its simple physics. What exactly they can tolerate for some specific transient is hard to determine, even with expensive destructive testing. There are too many tolerancesat play. My whole issue of starting this thread lies in the claim that you can supposedly have linear behavior of an underrated HF transducer used in conjunction with other much more thermally robust drivers, which can deal with proportionately larger transients without failing (sooner or later) or showing signs of audible distress.
I agree with your theory of harmonic down modulated peaks in metal cones being a possible cause for the observed 3rd order harmonic peak in that Revell F226. Why are you just trying to bash me about all these little technical details that don't lead to anything other than coming across as a pi$$ing contest? Do you just need to feel like you have the upper hand at something? Maybe its my outward disapproval of your own force fed info and/or from those other "experts" I disagree with? Is it because I disagree with their hard "science" or theories they put forward, claiming certain things to be inaudible or imperceptible despite my findings to the contrary? You can disagree, but don't rag on me about it.
Claiming I'm belligerent is pretty unreasonable. I just don't like big companies like Wilson using smoke and mirrors when its obvious reality is far from stated performanc.
That said, i can show you endless examples of 25 - 28mm dome tweeters which ALL handle about 10 - 15 watts continously before blowing up. Its simple physics. What exactly they can tolerate for some specific transient is hard to determine, even with expensive destructive testing. There are too many tolerancesat play. My whole issue of starting this thread lies in the claim that you can supposedly have linear behavior of an underrated HF transducer used in conjunction with other much more thermally robust drivers, which can deal with proportionately larger transients without failing (sooner or later) or showing signs of audible distress.
I agree with your theory of harmonic down modulated peaks in metal cones being a possible cause for the observed 3rd order harmonic peak in that Revell F226. Why are you just trying to bash me about all these little technical details that don't lead to anything other than coming across as a pi$$ing contest? Do you just need to feel like you have the upper hand at something? Maybe its my outward disapproval of your own force fed info and/or from those other "experts" I disagree with? Is it because I disagree with their hard "science" or theories they put forward, claiming certain things to be inaudible or imperceptible despite my findings to the contrary? You can disagree, but don't rag on me about it.
Claiming I'm belligerent is pretty unreasonable. I just don't like big companies like Wilson using smoke and mirrors when its obvious reality is far from stated performanc.
Last edited:
Why? I'm just trying understand why they use such a small tweeter crossed so low (in relation to how much bandwidth it actually has to deal with). Thats an honest question. So far I've come to the conclusion its a considerable mismatch in spectral power distribution. Trying to get a small dome tweeter with a HP set at 1.4k (at any slope) to keep up with so much more thermally and acoustically capable drivers leaves a big hole in output capability, creating an imbalance in the mids.
Here's the fundamental distribution of power all conventional multi way speakers have to deal with (taken from my old German speaker design book I've owned since the mid 80s)
You can see the fundamental power distribution for specific crossover frequencies and their spectral distribution of power across all the drivers. Yes, music has changed, but not by that much. With large orchestral transients, the tweeter on these Wilson's will be overwhelmed. It won't matter how steep the crossover slope is at below 1.4 - 1.5k
I've owned this book for a long time. It's what got me into building speakers back in 1984 when I grew up in (West) Germany. Lots of great specifc info in this book.
if you're not able to translate from German -
Rosa Rauschen = pink noise
The top curves are referenced to different types of music.
A typical system with a tweeter HP crossover point of 1.4k will require 10 - 25 % power distribution to the tweeter.
If system power handling is 100 watts while playing "music", the tweeter will see an average of 10 - 25 watts.
Even at very conservative power distribution, the tweeter on the Alexandria will need to deal with at least 10 watts. This is considering an average tweeter sensitivity to be close to the same as the other driver chassis. You can fudge it a little with more sensitive tweeters to increase power handling through voltage deviding networks (resistors).
This is all pretty fundamental speaker design knowledge. Just posting it in case there are some others who are doubtful.
Rosa Rauschen = pink noise
The top curves are referenced to different types of music.
A typical system with a tweeter HP crossover point of 1.4k will require 10 - 25 % power distribution to the tweeter.
If system power handling is 100 watts while playing "music", the tweeter will see an average of 10 - 25 watts.
Even at very conservative power distribution, the tweeter on the Alexandria will need to deal with at least 10 watts. This is considering an average tweeter sensitivity to be close to the same as the other driver chassis. You can fudge it a little with more sensitive tweeters to increase power handling through voltage deviding networks (resistors).
This is all pretty fundamental speaker design knowledge. Just posting it in case there are some others who are doubtful.
Last edited:
The curve from your book, is it about acoustic or electric power? More recent scientific work seems to point out otherwise and even rock music doesn’t take more than an average 16% acoustic energy in the range 1k-4K. Above 4K the energy drops rapidly to a mere 2%.
Plus you’d have to take into account that a lot of speakers have tweeters that are (far) more efficient than the mids and woofers, especially when baffle step is part of the deal. Imho (low) mid frequency drivers often are more limiting than tweeters.
Funny you have Berndt Stark lying around. I must have it somewhere too. Did you read the theoretical pieces of the Lautsprecherhandbuch from Michael Gaedtke too? Really underrated work from the ‘80s to me.
@markbakk I have his book as well buried somewhere. I sometimes brush through the old stuff to reacquaint myself with how things used to be done with charts, scientific calculators and even slide rules. I think I still have my old ported box tuning chart to figure out crude box tunings. 4th order bandpass just started to come out. We licensed it from Procus for the ACR Isostatik.
Anyways, I think the tweeter issue I have beef with is sort of self answering. Tweeter efficiency levels have risen a bit over time, but I don't see much evidence of this with speakers that utilize wide bandwidth tweeters.
The Alexandria essentially appears to be a large 2 way with an extended range subwoofer underneath it. It looks and sounds alot like the equivalent of an old 60s Camaro with a 396 big block and Muncie M22 under the hood.
Anyways, I think the tweeter issue I have beef with is sort of self answering. Tweeter efficiency levels have risen a bit over time, but I don't see much evidence of this with speakers that utilize wide bandwidth tweeters.
The Alexandria essentially appears to be a large 2 way with an extended range subwoofer underneath it. It looks and sounds alot like the equivalent of an old 60s Camaro with a 396 big block and Muncie M22 under the hood.
I agree the Alexandria's tweeters (there are two per cabinet) are unlikely to be able to provide realistic "front row" orchestral transient levels, even with THD exceeding 10%.
Seems your old German speaker design book is assuming the crest factor to be the same across the audio spectrum, like pink noise.
Meyer Sound did considerable research into the spectral content and dynamic range of music, in 2019 they came up with a new test signal to better represent it, "M-Noise" (Music Noise).
Although the peak levels are similar to pink noise (~12dB crest factor), musical crest factors rise with frequency above ~500Hz.
The resultant average power decreases by ~-3dB by 1400Hz, -10dB by 10kHz.
That would cut your tweeter average power estimate by 50% or more, reducing the 25% to under 12%.
Art
These 29mm soft domes with the wide surround were amongst the most capable tweeters back in 2011. They are not tiny IMHO.
We really should be moving forward with FSAF testing with -real music- at SPLs we like to listen at.
Further reading:
https://www.roomeqwizard.com/betahelp/help_en-GB/html/fsafmeasurement.html
I'm not holding my breath. The log chirp or exponential Sine Sweep was introduced a quarter of a century ago, yet still some people are still using MLS, introduced almost half a century ago...
We really should be moving forward with FSAF testing with -real music- at SPLs we like to listen at.
Further reading:
https://www.roomeqwizard.com/betahelp/help_en-GB/html/fsafmeasurement.html
I'm not holding my breath. The log chirp or exponential Sine Sweep was introduced a quarter of a century ago, yet still some people are still using MLS, introduced almost half a century ago...
FSAF testing came to my mind with regards to this too as the music stimulus will prevent tweeters from attempting to spontaneously combust.
It is absolutely true that a tweeters power handling is lower than that of other drivers within a typical system and if you're going to do high level sine sweeps of any duration, stepped sine testing, or subject them to any steady state dual tone sine stimulus then they will quickly go up in smoke. It's also a fairly gross misrepresentation of a speakers capability too. If sine testing causes a tweeter to fail, or show excessively high distortion, but FSAF does not with 99.9% of music, then you can conclude that a speaker is fine with music signals specifically because of it's typical spectral power distribution.
But we can do power analysis using software that doesn't require looking at books with decades old power distribution charts making assumptions about what the tweeter is being delivered in terms of power. We can analyse music.
So I loaded up a piece of heavily compressed modern jazz/rock/pop. Lots strings, brass, distortion guitar, drums, bass, so it's spectrally dense.
Average level -11.6dB relative to a maximum peak level of 0dB.
Now lets apply a 4th order highpass at 1kHz average level -18.5dB.
2kHz -22dB.
3khz -25dB.
Lets say you were listening full-range at a level that required 100 watts when a peak of 0dB came along. Full range the speaker would be dissipating an average of, wait for it, 7 watts.
How about with the 1kHz high pass? 1.4 watts.
2kHz? 0.6 watts.
3kHz? 0.3 watts.
Scale this up by a factor of 10, so that the peaks require one thousand watts, and the tweeter is dissipating 14 watts on average.
I don't think Wilson have anything to be concerned about when it comes to thermal issues and music playback with their low crossover point. Clipping amplifiers on the other hand? Now that'd smoke the tweeter way faster than if you didn't have a thousand watt amplifier and were clipping your 100 watt.
This helps to explain why small neo tweeters, with their slug magnets, don't go up in smoke. They can probably only handle 2-3 watts continuously before dying but they are used everywhere. It's not that big of a problem if all you're listening to is music with unclipped amplifiers.
Xmax and linear excursion, on the other hand, are a different matter entirely.
Of course it's possible that some music signal could come along that would throw an unusually high amount of high frequency energy, for an unusually long amount of time, at your tweeter and it would fry, but the same could be said for some insane bass content in one specific movie and whoops there goes a subwoofer. But at this point the tweeter will be exceeding xmax by quite a bit and sound awful so hopefully you'd turn it down.
It is absolutely true that a tweeters power handling is lower than that of other drivers within a typical system and if you're going to do high level sine sweeps of any duration, stepped sine testing, or subject them to any steady state dual tone sine stimulus then they will quickly go up in smoke. It's also a fairly gross misrepresentation of a speakers capability too. If sine testing causes a tweeter to fail, or show excessively high distortion, but FSAF does not with 99.9% of music, then you can conclude that a speaker is fine with music signals specifically because of it's typical spectral power distribution.
But we can do power analysis using software that doesn't require looking at books with decades old power distribution charts making assumptions about what the tweeter is being delivered in terms of power. We can analyse music.
So I loaded up a piece of heavily compressed modern jazz/rock/pop. Lots strings, brass, distortion guitar, drums, bass, so it's spectrally dense.
Average level -11.6dB relative to a maximum peak level of 0dB.
Now lets apply a 4th order highpass at 1kHz average level -18.5dB.
2kHz -22dB.
3khz -25dB.
Lets say you were listening full-range at a level that required 100 watts when a peak of 0dB came along. Full range the speaker would be dissipating an average of, wait for it, 7 watts.
How about with the 1kHz high pass? 1.4 watts.
2kHz? 0.6 watts.
3kHz? 0.3 watts.
Scale this up by a factor of 10, so that the peaks require one thousand watts, and the tweeter is dissipating 14 watts on average.
I don't think Wilson have anything to be concerned about when it comes to thermal issues and music playback with their low crossover point. Clipping amplifiers on the other hand? Now that'd smoke the tweeter way faster than if you didn't have a thousand watt amplifier and were clipping your 100 watt.
This helps to explain why small neo tweeters, with their slug magnets, don't go up in smoke. They can probably only handle 2-3 watts continuously before dying but they are used everywhere. It's not that big of a problem if all you're listening to is music with unclipped amplifiers.
Xmax and linear excursion, on the other hand, are a different matter entirely.
Of course it's possible that some music signal could come along that would throw an unusually high amount of high frequency energy, for an unusually long amount of time, at your tweeter and it would fry, but the same could be said for some insane bass content in one specific movie and whoops there goes a subwoofer. But at this point the tweeter will be exceeding xmax by quite a bit and sound awful so hopefully you'd turn it down.
I guess the low crossover point was chosen by Wilson to somewhat mitigate the disadvantages of the MTM design.
I myself don't feel that comfortable either when the smallest of the drivers in a BIG system has to carry the largest relative part of the spectrum (almost 4 octaves).
Regards
Charles
I myself don't feel that comfortable either when the smallest of the drivers in a BIG system has to carry the largest relative part of the spectrum (almost 4 octaves).
Regards
Charles
Is there a record of just how many of the units needed servicing?
I assume a company owner would count on future income as well.
I assume a company owner would count on future income as well.
Three highest octaves are mostly just harmonics in music, and low spl. Ie. up from 4kHz
Crossover region 1.5-4kHz on the other hand is very important, response, directivity and phase shift should be very smooth
Crossover region 1.5-4kHz on the other hand is very important, response, directivity and phase shift should be very smooth
Yes exactly. What would fry a tweeter is if a prominent fundamental came along that happened to be within its passband. The lower you take the tweeter the more chance there is of this happening.
A soprano high C is just over 1kHz. That has the potential to be a threat to a low crossed tweeter luckily forté high C's, without the orchestra going forté too, tend to be few and far between.
A soprano high C is just over 1kHz. That has the potential to be a threat to a low crossed tweeter luckily forté high C's, without the orchestra going forté too, tend to be few and far between.
https://www.diyaudio.com/community/threads/spectrum-of-musical-genres.310900/post-5152404
Reasonable, but as this thread shows some music has fairly predictable falling power with frequency.
Reasonable, but as this thread shows some music has fairly predictable falling power with frequency.
OTOH what would be interesting is how dynamics relate to that spectral distribution. Haven't found something on that yet.
Considering how many designs use a 1" dome tweeters if there were issues with power handling it would be obvious and the industry as a whole would have moved on long ago.
Rob 🙂
Rob 🙂
Oh yes, I completely missed that. Assuming M-noise is a good representation of music, high frequency (>2,5k) peak signals are only 3 to 4dB lower than the midband peak signals. For the region 1-2,5k it is equal to midband.
There are a few tracks in my music collection with a 20dB peak to average ratio. Obviously short transients and most people wont have these records but still within THX specs if you listen at an average of 85dB at the listening chair. I tend to listen a bit lower than that.