hi Ben
i have heard the Fostex tweeters. I like them, but nothing of the sort of the h100. The only other tweeters at the h100 level are Goto's, YL's, ALE's etc. The very expensive japanese high-end stuff. My Beymas TPL150 which i listen to now has the advantage of sounding amazingly natural, and the treble does not attract any attention to it. But it has not the dynamics, sharpness, precision and ease of the Coral. Its not king of awe inspiring or spectacular. Its just different. A work horse with its excellence of its own.
Also bought these from Serbia with amorphous core, silver-gold primary wire and unpainted stainless front plate. There is a proper high end feel to them. Haven't listened them yet 😱.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Does that mean the primary wires are made from a gold/silver alloy?
If so what would be the point of using an alloy as conductor?
I think it's more or less the same stuff as Mundorfs silver-gold wire, quite established high end hook up wire. I have read gold is supposed to make the silver conductor's sound more mellow and prefereble to some tastes compared to pure silver. Raal's Aleksandar also told me this was the best sounding step-down transformer they could make, amorphous core with silver-gold primary wire.
Alloys don't behave as a a conductor as one would expect.
Conductivity is usually referenced to copper (IACS system) which thus rates 100%.
Pure silver manages 105% and gold 70% but if you alloy gold with 10% silver the conductivity of the resulting alloy drops to around 30% or about the same as brass.
The actual numbers I dug out of the interwebs some time ago lead one to assume that the conductivity of the alloyed gold keeps dropping as the amount of silver is increased and suggest that the worst possible situation is 50% gold 50% silver or thereabouts.
Hi all!
Just wanted to comment on a few points in comparing distortion in loudspeakers.
First of all, the method we measure and present distortion data is borrowed from measuring amplification stages.
While it is OK to use it in electronics, it's pretty much invalid for speakers. Speakers have stored energy, while amplifier doesn't.
The distortion measurement for a speaker is meaningless if it is not multiplied by decay time.
for example, most hard cone bass-mids will have a peak at the end of bandwidth. Say, a 8" magnesium cone rings 10dB at 4.5k.
It will have 0.5% of 3rd harmonic there. That 3rd harmonic will come from a 1.5k signal that was applied to motor.
So, you apply the signal, it generates a 3rd harm, then stop the 1.5k signal and your 3rd harmonic gets to ride on top of response 4.5k peak and last 2ms after the signal stopped. Now, what a 0.5% distortion actually means in that case? Pretty much nothing, as that will sound like nothing we can relate to. It's not a "timbre" any more, it's a horror story.
Now, that is an extreme example, nevertheless, all moving systems will do this to some extent.
It is of utmost importance to reduce energy storage in pass-band and if it must exist, decay must be smooth, with the least amount of "ridges" appearing in CSD.
That will dismiss the impact of distortion lasting than the signal that caused it. That is the primary objective of my tweeter design.
As for the amount of distortion, yes, a tight flat ribbon will initially show greater distortion than a corrugated foil ribbon.
Provided that those two ribbons, have the same decay (they don't, corrugations have breakup), the flat tight one will be worse if you listen to music at 85dB level with no dynamics (Bob Rock?).
However, if you do listen to music that has dynamics, more important part is whether the tweeter will change the "nature" of it's distortion with increasing level. Tight flat ribbons (when properly suspended in homogenous magnetic field) do not appreciably change the distortion spectrum, while corrugated ones tend to "explode" into distortion after a certain level is reached, and it is reached all the time in normal listening.
So, a speaker wit initially higher distortion will sound clean if the distortion rises linearly with level and it's spectrum doesn't change.
That's because the distortion in ears change in a similar way.
Only at that point, I can dare to speak of whether a tweeter has colorations, but if it does what I described, it will sound clean, nevertheless. With a "timbre" perhaps, but still exceptionally clean.
This was the secondary objective of the tweeter design that I do.
All other things that we value in tweeters come naturally if you sort out these two things, and you can play with variables later (response shape, dispersion, bandwidth...), as they are in a far third place.
Here is the THD vs. SPL of RAAL 140-15D, without foam pads, crossed with 4th order at 2k5:
Note the THD spectrum and amount vs. SPL. That's the point when we talk audible distortion, not whether we're having -40, -50 at 85-90dB SPL.
So, a full picture can be grasped from (CSD x (THD vs SPL)) and distortion in speakers takes a whole different meaning and interpretation.
Discussions of it becomes pointless with looking only at data fragment gotten by measuring methods borrowed from electronics.
Cheers!
Just wanted to comment on a few points in comparing distortion in loudspeakers.
First of all, the method we measure and present distortion data is borrowed from measuring amplification stages.
While it is OK to use it in electronics, it's pretty much invalid for speakers. Speakers have stored energy, while amplifier doesn't.
The distortion measurement for a speaker is meaningless if it is not multiplied by decay time.
for example, most hard cone bass-mids will have a peak at the end of bandwidth. Say, a 8" magnesium cone rings 10dB at 4.5k.
It will have 0.5% of 3rd harmonic there. That 3rd harmonic will come from a 1.5k signal that was applied to motor.
So, you apply the signal, it generates a 3rd harm, then stop the 1.5k signal and your 3rd harmonic gets to ride on top of response 4.5k peak and last 2ms after the signal stopped. Now, what a 0.5% distortion actually means in that case? Pretty much nothing, as that will sound like nothing we can relate to. It's not a "timbre" any more, it's a horror story.
Now, that is an extreme example, nevertheless, all moving systems will do this to some extent.
It is of utmost importance to reduce energy storage in pass-band and if it must exist, decay must be smooth, with the least amount of "ridges" appearing in CSD.
That will dismiss the impact of distortion lasting than the signal that caused it. That is the primary objective of my tweeter design.
As for the amount of distortion, yes, a tight flat ribbon will initially show greater distortion than a corrugated foil ribbon.
Provided that those two ribbons, have the same decay (they don't, corrugations have breakup), the flat tight one will be worse if you listen to music at 85dB level with no dynamics (Bob Rock?).
However, if you do listen to music that has dynamics, more important part is whether the tweeter will change the "nature" of it's distortion with increasing level. Tight flat ribbons (when properly suspended in homogenous magnetic field) do not appreciably change the distortion spectrum, while corrugated ones tend to "explode" into distortion after a certain level is reached, and it is reached all the time in normal listening.
So, a speaker wit initially higher distortion will sound clean if the distortion rises linearly with level and it's spectrum doesn't change.
That's because the distortion in ears change in a similar way.
Only at that point, I can dare to speak of whether a tweeter has colorations, but if it does what I described, it will sound clean, nevertheless. With a "timbre" perhaps, but still exceptionally clean.
This was the secondary objective of the tweeter design that I do.
All other things that we value in tweeters come naturally if you sort out these two things, and you can play with variables later (response shape, dispersion, bandwidth...), as they are in a far third place.
Here is the THD vs. SPL of RAAL 140-15D, without foam pads, crossed with 4th order at 2k5:
Note the THD spectrum and amount vs. SPL. That's the point when we talk audible distortion, not whether we're having -40, -50 at 85-90dB SPL.
So, a full picture can be grasped from (CSD x (THD vs SPL)) and distortion in speakers takes a whole different meaning and interpretation.
Discussions of it becomes pointless with looking only at data fragment gotten by measuring methods borrowed from electronics.
Cheers!
Attachments
Last edited:
RAAL:
Nice posts!
I have to agree that for my own aesthetics, lack of stored energy is a bigger deal than a flat or smooth frequency response, and the way I normally look for this is at the high-resolution waterfall plots.
I'll have to spend a few days reading this post again. 🙂 Wonder if I can do this type of testing with what I have.
Best,
Erik
Nice posts!
I have to agree that for my own aesthetics, lack of stored energy is a bigger deal than a flat or smooth frequency response, and the way I normally look for this is at the high-resolution waterfall plots.
I'll have to spend a few days reading this post again. 🙂 Wonder if I can do this type of testing with what I have.
Best,
Erik
Would this correlate with compression? That is, that a tweeter with little or no compression would not show this exploding distortion, but a tweeter with appreciable compression (adding +5 db input causes less than 5 db increase in output) ?
Just for my own education, what is an example of a corrugated ribbon I might have seen at the Madisound shop?
Best,
Erik
I'm listening to the Mundorf AMT 25CM1.1Rs right now, but I'm afraid I can't compare to the Raal's. If you'd like me to compare to Raal's please send me a pair in a box with an Automatica Clio Pocket Personal Measurement system and I'll get them back to you when I'm done..... 😀😀😀
I will say the Mundorfs are pretty fabulous. Very little stored energy to my ears, no compression I can hear, and measurably flat frequency and impedance making them very easy to integrate. The tweeter also is one of the few that sounds as good close up as 10' away. Most tweeters of any quality tend to loose detail as you get further from the speaker. This of course is probably due to the dimensions of the radiating surface.
Compared to the Beryllium tweeters in the older Focal Utopias, the Mundorfs have the same clarity without the colorations. Sorry I haven't heard the ScanSpeak or other versions of them except for the custom Be/Diamond one's used by Magico.
As some one else on this thread has mentioned, I can't find a commercial speaker I've listened to lately I'd trade for them, this includes the latest Wilson Sasha's, the latest Magico S1/Mk II's, etc. If a fire hits my building it's the cats, myself and the speakers that I'm saving. 🙂
Last edited:
Without going in depth into thermal, mechanical (compliance linearity, diaphragm breakup) and magnetic field (gap depth, field homogeneity and linear excursion) reasons for compression and their different rates of compression, you're right, you could say that this would be an indicator.
Differences are small, though, to make valid evaluations. It takes very little compression expressed dB to get 5-10% of 3rd harmonic...same thing as with amplifiers, they don't get appreciably compressed audibly when driven into clipping.
Well, for some time now, all Founteks are with flat foils. I just don't fancy that reinforced sandwich they use. Too stiff with a breakup at 10k, but they had to do it instead of corrugations to mitigate the problem with stronger field they have at ribbon edges. In such magnetic fields, the ribbon gets 5 times more driving force on edges than in the middle. You can't have the edges flapping like wings, so you use stiff sandwich, and then you provoke it's breakup at 10k...
Basically, they should handle THD vs SPL very good, but 10k breakup will stop you cranking the volume up very soon, so not much of a gain in clarity of presentation...
In contrast, I use the softest tempered Al there is, but in fairly or completely homogeneous field. That way, you don't need stiffness. All points of conductor are immersed in same field, force acting equally at all points, so no breakups, no tears. It would act as ideal piston if it weren't clamped at the ends and no stiffness means no breakups and no "kitchen foil" noises.
LCY is the only corrugated ribbon at Madisound. The corrugated foil is cut in the middle, so the "wing flapping" wont tear it at the middle on its own after 5 minutes...nice workaround, but nothing is genuinely solved.
All of the solutions you choose to make in a simple thing like a ribbon tweeter, will influence 3-5 variables at the same time, so one must be very careful in what choices to make and what are the pros and cons.
I think I made good ones, I guess, and I can apply them all without thinking of whether they are applicable in "mass" produced product, as I don't have that problem.
The latest addition is elastic suspension/dampener at the ribbon ends, so distortion gets a tad better, but reliability is much better. High excursion overloads are handled without damage or with very little damage.
I don't want to turn this into winding my own windmill 😀
I sure ain't the smartest guy around by any stretch, so I'll stop here and maybe chime in on general discussion on validity of test data at some point...
I usually get smashed by signal theory guys that apply too general assumptions about the inner life of speakers and how they are tested, so I better shut up right now! 😉
Best,
Alex
Last edited:
So to answer the question about Mundorf AMT measurements, I've created a new thread, please see here:
http://www.diyaudio.com/forums/multi-way/284527-mundorf-amt-ss-7-compression-distortion.html
http://www.diyaudio.com/forums/multi-way/284527-mundorf-amt-ss-7-compression-distortion.html
I recently heard a small floor standing speaker with ribbon tweeters and the treble sounded bad to me. The detail was all garbled and seemed to blend into itself - everything sounded lispy. I was unimpressed. The dome tweeter sounded better on a similar system right next to the ribbon sounded better to me.
a slotted pipe and decent one inch format compression driver is good and self-equalizing in the sense that no contour is needed after the highpass filter for "flat" response
video of Karlson slotted pipe aka "K-tube" wavguide on top of a little K-coupler
https://www.youtube.com/watch?v=T71pRwTGfuc
that product's ad - you can make diy versions from plastic or metal pipe, rolled paper, laminated veneer
video of Karlson slotted pipe aka "K-tube" wavguide on top of a little K-coupler
https://www.youtube.com/watch?v=T71pRwTGfuc
that product's ad - you can make diy versions from plastic or metal pipe, rolled paper, laminated veneer
Sounds like the ribbon tweeter was crossed below 5kHz.
They all seem to struggle there, make decent super tweeters though.
There are a few ribbons that are fine crossed lower. But that new Beyma AMT looks awesome in that regard
That's my experience too. I tried using a Raven R1 from 3kHz and it sounded woeful. OTOH, The larger RAAL's are sublime right down to 3kHz as long as a 4th order HP is used.
With ribbons, size matters.