Hi,
In conventional 3 ways or more loudspeakers, best soundstages often result from a treble to medium being near a 1/2 wave length center to center spacing. Because the 1/4 wave length rule is impossible at those frequencies.
Now, Most of music is under 10K hz, there will be always few exceptions and instruments that climb high as piano or some instruments having very high harmonics that we still can perceive in the flow, but again all the story is often below 10 k Hz.
We often wan to cut-off outside the said area of ears sensibility as exposd in the Fletcher&Mundson curve. It translates by several approachs like hrned tweeter crossed below 1500 Hz, Full range from 4" to 2" sizes, conventional dome tweeters of 1.25" size, some ribbons and AMT of big size that handle for the hugest XO below 1k/1k5 Hz and of course all the horned compression drivers with possible low XO in home situation if the wife supports horn and have not fulled your suitcases yet because of that !
Now what happen with some design with very high cut-offs and more or less stiff slopes ? Say above 10K hz . For illustration like some design of ZU loudspeakers , some english style design like old Spendor or newer Harbeth with two differently sized tweeters and so on ???
I'm asking myself if I could not XO my ScanSpeak 2.5" 10F FR driver around 10 K hz or more with a first and second order slope 😕 Why not with some excelent tiny 3/4" Peereless, Monacor and numerous not so bad if excelent cars tweeters (Audax made some crazy good sounding ones with Fs around 6 k Hz for instance).
For the FR - full range- enthusiasts, a <300 hz to > 10 K hz standalone driver has certainly some appeal ! A logic behind it could be a size that makes the driver good enough for the task : not slam to perform below 300 hz (XO dependent of course) and not laser beaming treble as well (same XO dependent of course)
Trade offs discussion of course !🙂
In conventional 3 ways or more loudspeakers, best soundstages often result from a treble to medium being near a 1/2 wave length center to center spacing. Because the 1/4 wave length rule is impossible at those frequencies.
Now, Most of music is under 10K hz, there will be always few exceptions and instruments that climb high as piano or some instruments having very high harmonics that we still can perceive in the flow, but again all the story is often below 10 k Hz.
We often wan to cut-off outside the said area of ears sensibility as exposd in the Fletcher&Mundson curve. It translates by several approachs like hrned tweeter crossed below 1500 Hz, Full range from 4" to 2" sizes, conventional dome tweeters of 1.25" size, some ribbons and AMT of big size that handle for the hugest XO below 1k/1k5 Hz and of course all the horned compression drivers with possible low XO in home situation if the wife supports horn and have not fulled your suitcases yet because of that !
Now what happen with some design with very high cut-offs and more or less stiff slopes ? Say above 10K hz . For illustration like some design of ZU loudspeakers , some english style design like old Spendor or newer Harbeth with two differently sized tweeters and so on ???
I'm asking myself if I could not XO my ScanSpeak 2.5" 10F FR driver around 10 K hz or more with a first and second order slope 😕 Why not with some excelent tiny 3/4" Peereless, Monacor and numerous not so bad if excelent cars tweeters (Audax made some crazy good sounding ones with Fs around 6 k Hz for instance).
For the FR - full range- enthusiasts, a <300 hz to > 10 K hz standalone driver has certainly some appeal ! A logic behind it could be a size that makes the driver good enough for the task : not slam to perform below 300 hz (XO dependent of course) and not laser beaming treble as well (same XO dependent of course)
Trade offs discussion of course !🙂
Last edited:
Look at Toels' site where he analyses and modifies a Spendor 3way.
Celestion tweeter and supertweeter, crossed around 13k. So it can work.
With modern small tweeters, top octave should be even better.
Celestion tweeter and supertweeter, crossed around 13k. So it can work.
With modern small tweeters, top octave should be even better.
Dali do this in their Epicon and Rubicon with a ribbon tweeter cutting in at 10khz
DALI EPICON 6 - Incredible sound and design
And Fostex do their “bullet” tweeters to fill in top end of their 208EZ full ranger (T900A but cuts in a bit lower).
https://www.fostexinternational.com/docs/speaker_components/pdf/recom_enclose/208ez_enclrev.pdf
DALI EPICON 6 - Incredible sound and design
And Fostex do their “bullet” tweeters to fill in top end of their 208EZ full ranger (T900A but cuts in a bit lower).
https://www.fostexinternational.com/docs/speaker_components/pdf/recom_enclose/208ez_enclrev.pdf
Last edited:
Hi,
Such high frequencies have really short wavelenght, 1/4 at 10kHz is about 8mm. Not sure if there is any benefit trying to optimize driver position and delays too accurately unless your listening distance and height is always the same.
Super tweeters don't make much sense in this regard, they feel like slap on patch and marketing trickery. I mean, all tweeters play past 10kHz so where is the catch? Maybe put the supertweet backwards playing to the wall or something, for air?
I haven't heard any supertweeters, not sure if I even could, so take my post as a thinking excercise 🙂
Such high frequencies have really short wavelenght, 1/4 at 10kHz is about 8mm. Not sure if there is any benefit trying to optimize driver position and delays too accurately unless your listening distance and height is always the same.
Super tweeters don't make much sense in this regard, they feel like slap on patch and marketing trickery. I mean, all tweeters play past 10kHz so where is the catch? Maybe put the supertweet backwards playing to the wall or something, for air?
I haven't heard any supertweeters, not sure if I even could, so take my post as a thinking excercise 🙂
The SS 10F needs to be crossed over not higher than 4 KHz, if not you will create a hole in the power response. For that frequency you won't have a problem in finding a good performance inexpensive tweeter. Look here for an example: Zaph|Audio - ZDT3.5
Ralf
Asking myself, hence the topic. Maybe not having crossover where the ears are said so much sensitive ? Maybe not having phase offset due to a filter there ? I really dunno ?! Some are talking about some designs, there seems not to have well established opinion though. Are the designers that do that are just crazy ? What's the trade offs and why they do it ?
I have several measurments that show it is flat till 8k/10k up to 30° and a part till 60 hz.(the 4 k hz ou're talking about).
My problem is not to find a tweeter. Btw I had some very good advices as yours, Casull445 advised me the little 3/4 bulett tweeter with no front plate which is 2.5 cm diameter 🙂... if bigger I will go to the NE19V something from Peerless as well.
My experience with listening the 10F due to his transparency, is it brings well higher frequencies like high saxo, bells, cymballs bells... more better than many FR advised here most of the times like the Fostex, Faitals, etc of the same size... question of ears, I dunno !
@ X : it's your fault I bought the 10F/8424g00 from your reccords 🙂 ! Remember the guy who find blind test it was a scanspeak due t othe sound signature 😉 ! I was not disapointed after buying it though never was able alas to make the foam horn that was the trigger to buy them ! (itchy hands, thanks to God I am not sirgeon)
Is a XO around 10K still be named super tweeter ? I mean we are not above 18/20K and not if all of us have still 20 yo 😀.
Why Zu Loudspeaker with such compression crossed at 12K are liked by reviewers at Steroidphil ??? Why the guy at Harbeth made a loudspeakers with two tweeters with such a trick ? Just good because the ads ?
I'm ok with you with the theory you expose, but if the trade offs will be worthing it ? 😕 ... brands who do so risk their money !
Yes readed it already, fascinating read, hard to balance without having heard them and judge for the crossover tweaks...
c-c = 1/2 wave length at XO frequency is the worst case scenario for directivity index and power response assuming quite flat on-axis. In addition, it's not the best also for early reflections RMS so possibility to get the most broken sound balance especially with 2-way is increased.
c-c = 1.0-1.4 x wave length at XO f provides the most balanced power and ER. Closer to 1.0 if vertical directivity of radiators is so high that vertical early reflections are significantly reduced i.e. DI & power rules c-c dimensioning.
Hi diyiggy, yeah these few commercial products you've mentioned seem to be purely business motivated engineering decisions. With quick look the Zu speakers marketing people seem to do double duty as engineers. Harbeth does the shoebox thing and has broadened their catalog, I mean how many different price point shoebox 2-way models one can sell? Gotta add another tweeter to justify price and to have something to offer for the deep pocket customer, while still staying in the niche.
Sorry for blatant words. There is nothing wrong with what they do and it is completely understandable from the business perspective. I'm not able to think any reasons to add tweeter that high for audio quality reasons so this is why I think it is only for the business.
If you look at speaker design purely from the audio quality perspective and forget all the technologies: how many ways and what size and brand woofer is best, or tweeter, you probably end up to a design that doesn't need a supertweeter unless you have super hearing 🙂 If you take the science: frequency response is king, and to get in room frequency response right you gotta have directivity under control. Diffraction, distortion and ability to play comfortably loud are some of the other things you want to pay attention to in the design. Room acoustics, positioning, looks, cost are some other issues that are in play. It is only the looks that might benefit from a additional tweeter, everything else is hurt. These are my thoughs, maybe there is more to it, maybe not? 🙂
Add a small tweeter to the 10f, check out with measurements and simulation best crossover point to it, should be about 4-5kHz possibly, what others have mentioned already.
Sorry for blatant words. There is nothing wrong with what they do and it is completely understandable from the business perspective. I'm not able to think any reasons to add tweeter that high for audio quality reasons so this is why I think it is only for the business.
If you look at speaker design purely from the audio quality perspective and forget all the technologies: how many ways and what size and brand woofer is best, or tweeter, you probably end up to a design that doesn't need a supertweeter unless you have super hearing 🙂 If you take the science: frequency response is king, and to get in room frequency response right you gotta have directivity under control. Diffraction, distortion and ability to play comfortably loud are some of the other things you want to pay attention to in the design. Room acoustics, positioning, looks, cost are some other issues that are in play. It is only the looks that might benefit from a additional tweeter, everything else is hurt. These are my thoughs, maybe there is more to it, maybe not? 🙂
Add a small tweeter to the 10f, check out with measurements and simulation best crossover point to it, should be about 4-5kHz possibly, what others have mentioned already.
Last edited:
Frankly I would trust SS own measurement, which show the expected dispersion from such a driver: https://www.scan-speak.dk/datasheet/pdf/10f-8424g00.pdf
From 6KHz and up the difference between on-axis and 30 deg off-axis is more than 5dB, and to my knowledge this dictates a crossover point below that for optimum performance.
Troels Gravesen used the 10F in a design (ScanSpeak-3W-Discovery), and this is what he said about the driver:
I would go to a 3.5-4 KHz crossover.
Ralf
Damn... I believed 1/2 WL was ok till 90° off-axis where a null occurs 😱 and so it was good to avoid ceilling and floor reflexion a lot...
Thanks Kimmosto for the formula. I will follow that 🙂.
It's certainly less important cause not so talked here but I should be ok between the bass and the mid which is in the 1/4 center to center . Anyway I made the back and front removable à la old english style to change easily the design...
hello Tmuikku, thanks for your point. Ah it increases a lot the tweeter choice and make it more confortable. I planned 3k hz to stay in the 1/2 WL, but I wanted more a 4K hz from the beginning. Maybe I should go for a little wave guide from AugerPro member with for instance the Boutique Haut-parleurs & Audio DIY advised by Casull445 member from Canada to me some times ago.
I find this one appealing to if no Wave Guide needed : Boutique Haut-parleurs & Audio DIY
But as I recessed the 10F 1.1 mm behind the front bafle surface , I certainly need some recess to with the tweeter for a better impulse response matching...
This is a learning project on the picture... my first non avorted diy loudspeakers project 😀. Bass is the SB23NBAC (aluminium planned LR12 electrical at 500 to 600 hz). I wanted an Harbeth look as well, so I called mine the Hard-bet !
So full passive LR12 planned or pehaps a close transcient perfect asymetric JMLC and derivates style. - 18 dB -
Giralfino : yep, I'm not sure to know till where the off axis matters in the overlapp areas between the two drivers ? 45° off axis ? 60° ?
For the thread documentation and others, I join some 10F/8424G00 I collected. I have the one from Mark Audio as well but the pdf capture from his long measurement with 2 others drivers is too much big. One can also find Zapha Audio measurement in the blog pages : Zaph|Audio
Thanks Kimmosto for the formula. I will follow that 🙂.
It's certainly less important cause not so talked here but I should be ok between the bass and the mid which is in the 1/4 center to center . Anyway I made the back and front removable à la old english style to change easily the design...
hello Tmuikku, thanks for your point. Ah it increases a lot the tweeter choice and make it more confortable. I planned 3k hz to stay in the 1/2 WL, but I wanted more a 4K hz from the beginning. Maybe I should go for a little wave guide from AugerPro member with for instance the Boutique Haut-parleurs & Audio DIY advised by Casull445 member from Canada to me some times ago.
I find this one appealing to if no Wave Guide needed : Boutique Haut-parleurs & Audio DIY
But as I recessed the 10F 1.1 mm behind the front bafle surface , I certainly need some recess to with the tweeter for a better impulse response matching...
This is a learning project on the picture... my first non avorted diy loudspeakers project 😀. Bass is the SB23NBAC (aluminium planned LR12 electrical at 500 to 600 hz). I wanted an Harbeth look as well, so I called mine the Hard-bet !
So full passive LR12 planned or pehaps a close transcient perfect asymetric JMLC and derivates style. - 18 dB -
Giralfino : yep, I'm not sure to know till where the off axis matters in the overlapp areas between the two drivers ? 45° off axis ? 60° ?
For the thread documentation and others, I join some 10F/8424G00 I collected. I have the one from Mark Audio as well but the pdf capture from his long measurement with 2 others drivers is too much big. One can also find Zapha Audio measurement in the blog pages : Zaph|Audio
Last edited:
Be careful when attempting to justify your design with off-the-shelf consumer speakers!
Yeah, they tend to be very marketing focused because that is what sells and they are not a charity.
My favorite marketing driven speaker of the last 10 years was done by JBL. You'd think they would know better but it fit all the buzzwords for people that don't know.
It was a horizontal "3-way" center channel because 3-way centers are better than 2-way centers. Sure, but...only when done right. What JBL did was take a flawed 2-way MTM design and to make it "3-way" they added a super tweeter above the regular dome tweeter and crossed it at 20 KHz! Now you get "3-way" and no matter how bad the super tweeter is, humans can't hear it anyway for the win! 🙂
So it hit all the buzzwords, 3-way center channel, it went to 30 or 40 KHz (more KHz is gooood!) it had more pieces parts on the bezel so it looked cool and maybe added $15 to the price per speaker for fat profits.
This abomination of marketing was done by JBL...not Zu or other fashion brands, JBL did this so be careful who you pick as audio heroes.
As far as using super tweeters go, they do have a use and a function as long as they fix a problem. For example, say you have a wide band large compression driver that rolls off over 10KHz-say it runs from 600 Hz to 10KHz or somthing then dies. I can see adding in a "super tweeter" to pick up the final octave if you like.
Some people like "full range" speakers so adding a Fostex at 8KHz or so will give the treble extension one demands that allows the "full range" to be larger, dig deeper and be more efficient.
Since human hearing gets very inaccurate over 10KHz--you have more slop to play around with so can get away with more acoustic mess than you can at 1 or 4KHz. This also holds true on the bottom octave of sub bass so use the poor hearing of humans at the extremes to get improvement in sound.
Generally speaking I don't strive to use super tweeters, I have a pair of bullet super tweeters that cross at around 12KHz that I use with line arrays when having a party outside. I basicly point them about 10 meters back to add a little sizzle for BBQs and such. I'd get kicked out of the audiophile club for that but it sounds better with BBQ, beer and a buzz for large areas. To me it does but I had a few already so not scientific testing by any means. I call them "sparkle boxes".
At the end of the day, this is DIY and by all means have at it! Nothing wrong with drawing outside the lines, screwing around with drivers and experimenting--all part of the hobby.
Yeah, they tend to be very marketing focused because that is what sells and they are not a charity.
My favorite marketing driven speaker of the last 10 years was done by JBL. You'd think they would know better but it fit all the buzzwords for people that don't know.
It was a horizontal "3-way" center channel because 3-way centers are better than 2-way centers. Sure, but...only when done right. What JBL did was take a flawed 2-way MTM design and to make it "3-way" they added a super tweeter above the regular dome tweeter and crossed it at 20 KHz! Now you get "3-way" and no matter how bad the super tweeter is, humans can't hear it anyway for the win! 🙂
So it hit all the buzzwords, 3-way center channel, it went to 30 or 40 KHz (more KHz is gooood!) it had more pieces parts on the bezel so it looked cool and maybe added $15 to the price per speaker for fat profits.
This abomination of marketing was done by JBL...not Zu or other fashion brands, JBL did this so be careful who you pick as audio heroes.
As far as using super tweeters go, they do have a use and a function as long as they fix a problem. For example, say you have a wide band large compression driver that rolls off over 10KHz-say it runs from 600 Hz to 10KHz or somthing then dies. I can see adding in a "super tweeter" to pick up the final octave if you like.
Some people like "full range" speakers so adding a Fostex at 8KHz or so will give the treble extension one demands that allows the "full range" to be larger, dig deeper and be more efficient.
Since human hearing gets very inaccurate over 10KHz--you have more slop to play around with so can get away with more acoustic mess than you can at 1 or 4KHz. This also holds true on the bottom octave of sub bass so use the poor hearing of humans at the extremes to get improvement in sound.
Generally speaking I don't strive to use super tweeters, I have a pair of bullet super tweeters that cross at around 12KHz that I use with line arrays when having a party outside. I basicly point them about 10 meters back to add a little sizzle for BBQs and such. I'd get kicked out of the audiophile club for that but it sounds better with BBQ, beer and a buzz for large areas. To me it does but I had a few already so not scientific testing by any means. I call them "sparkle boxes".
At the end of the day, this is DIY and by all means have at it! Nothing wrong with drawing outside the lines, screwing around with drivers and experimenting--all part of the hobby.
There are few exceptions. A bit longer c-c than 1/2 wave length could be good option in some special cases. For example small mid-range driver installed to open baffle significantly wider than the driver. That could be very close to omni-polar at XO frequency so higher total directivity at XO range achieved with c-c=0.6-0.7 x wave length and phase matching slopes could be valuable for compensating too wide dispersion in horizontal plane.