Tekton style like tweeter array good idea?

[Bryguy]

I suspect that some people like beaming as it can introduce perceptual effects that give an impression of "soundstage".

It is generally agreed that narrow dispersion speakers are more "focused" while wide dispersion speakers create a bigger perceived soundstage width. Anyways, the problem isn't with beaming, as there are many great speakers with narrow and wide dispersion. The issue arises when a speaker beams only in small chunks of the frequency response, which can result in a clunky power response. If you already know all this just let me know and I'll shut up

 

[Penelinfi]

It is generally agreed that narrow dispersion speakers are more "focused" while wide dispersion speakers create a bigger perceived soundstage width. Anyways, the problem isn't with beaming, as there are many great speakers with narrow and wide dispersion. The issue arises when a speaker beams only in small chunks of the frequency response, which can result in a clunky power response. If you already know all this just let me know and I'll shut up

Maybe said chunky beaming also gives "interesting" soundstage or imaging qualities. For me due to the wide variety of music I listen to, perhaps such a speaker will get "found out" over time. But for some with a more limited genre set, it could work great.

 

[Rick Sykora]

Tekton uses SBAcoustic dome tweeter which provides several different tweeter with 10cm² Sd - see the link in the previous post or here and here.
In the Tekton speaker models with "circular arrays" are 6-14 tweeter used. So dome area isn't a real problem as long as the crossover frequency is chosen sensibly.


I agree with your statement for stiff materials like Al or Be domes, sorry but for slik domes it might be wrong. Easy to show, using the measurements of the Bliesma midrange drivers (identical except for dome material):
View attachment 363346
Source: HifiCompass

The Al and Be midrange domes run practically parallel to each other up to 6kHz, the difference in SPL is due to different dome weights.
The silk dome is completely different, above 2kHz there is probably no more piston-like behavior.

As mentioned in the post you quoted, there is another resonance in the usable range of the midrange domes (and even more so of all midrange woofers) - all kinds of surround resonances.
Of course the Bliesma midrange drivers are SOTA, but even there you can see the fundamental resonance (half-wave)*** at 700Hz and the multiple at 1.4kHz in the diagram.

Don't get me wrong, I don't say that this is clearly audible, but one can find arguments for the use of a midrange array with 2-2.5'' midrange driver.


*** midrange driver Sd is 50cm², so radius is about 4cm and thus the circumference at 25cm. This means that the half-wave for the first surround resonance is around 700Hz.

The Monacor SPX-20M looks almost ideal for this array. Since it is used on the CBT24, am sitting on enough Dayton ND64-16s to try them. Juggling a few too many projects currently though. Maybe next year…

 

[musicforcities]

If I’m going line array, i am going either with:
A) a small company that employs and publishes rigorous testing and measurements following established principles
B) an essentially diy based on such principles.
C) a company with a proven track record for engineering such things at a professional scale/market and transferring such knowledge to consumer markets. Aka, JBL

And one better understand how such a thing will work in the sort of space one has at home. lol.

 

[CedarX]

For what is worth, Tekton use SB Acoustics drivers on some (all?) of their speakers. Here is a picture from their website:

If you zoom it:

There is a SB17xxx (can’t read the exact reference).

 

[cavedriver]

I put the post below in the M Lore thread but I think I probably should have put it here so I'm just copying it over. Like Stereophile's review of the Impact Monitor, and specifically addressing the 6-tweeter midrange array, I'll say that in this review of the Moab we again see some fairly poor vertical axis smoothness, although the horizontal is "ok". In this Moab speaker the center-to-center distance of the two virtual midranges is also quite large, and with the relatively high crossover to the center tweeter (3400 Hz), I suppose it shouldn't be surprising that the vertical axis looks like there's some lobing? There's a dip just below the crossover frequency (772 Hz) that I assume has more to do with the crossover design than the use of 6 tweeters. Since it's Stereophile we don't have distortion measurements or understanding of what the speaker measures like at high volumes so the question of Sd, Xmax, and dynamic range is unanswered, but from this review it's still not looking good for the tweeter array concept.
***
should have looked a little harder earlier. Stereophile just published another Tekton review just 10 days ago. Model is the "Moab Be", so this literally is the one where someone took a speaker with a base price of ~$5,000 and added a $12,000 upcharge to replace all the tweeters with Satori Be tweeters. From JA's measurements I'm not getting much of a feel for this speaker. Too hot by about 5 dB from 33 to 200 Hz, then ~3 dB dip from 600 to 700 Hz (just below the very precisely called out "772" Hz crossover), one of the hardest loads John says he's ever measured, and some strange off-axis behavior again, especially in the vertical, as we could see with Stereophile's older Impact Monitor review (but not as bad this time?). To be fair, quite a few speakers have a bass bump somewhere in the 50~200 Hz range, including recently the Wharfedale Dovedales at $7k a pair. John very kindly struggles to say something positive about the speaker at the end of the measurements. Personally, I just can't see spending $12000 on a tweeter upgrade, nor wasting half the available beryllium diaphragm material in the western hemisphere on one mediocre speaker.

Image per Stereophile courtesy of JA (Thank you!):

And the vertical axis response measurements - note the big holes at -5 and +10 degrees at ~3kHz:

and a link to the review:

Tekton Moab Be loudspeaker Measurements​

Sidebar 4: Measurements

www.stereophile.com

 

[Randolf]

For what is worth, Tekton use SB Acoustics drivers on some (all?) of their speakers. Here is a picture from their website:


If you zoom it:


There is a SB17xxx (can’t read the exact reference).

I think Tekton at least uses/used drivers from Scan Speak, SB Acoustics and Audax, depending on the model and production time. All 3 well known highly rated brands, nothing to complain here. Maybe these brands even produce OEM versions for Tekton. Many manufactures do not produce all the drivers they use on their own. Nothing wrong with that since you have the extra freedom to pick those drivers which best serve your needs from the market and switch if something better comes up.

 

[Randolf]

from 33 to 200 Hz, then ~3 dB dip from 600 to 700 Hz (just below the very precisely called out "772" Hz crossover), one of the hardest loads John says he's ever measured, and some strange off-

Thanks for pointing us to this recent review. Since the usage of the SB Acoustics Satori Berylium tweeters and the 772 Hz crossover point are explicitly mentioned, I would personally rate this design as problematic. Satori specs are:

Satori - Sbacoustics - Page 3

- Page 3

sbacoustics.com

Satori - Sbacoustics - Page 4

- Page 4

sbacoustics.com

SATORI TW29B / SATORI TW29B-B: fs: 600Hz
SATORI TW29BN / SATORI TW29BN-B: fs: 700Hz
SATORI TW29BN-8 / SATORI TW29BN-B-8: fs: 750 Hz

crossover point is extremely close or even at resonance frequency.

 

[ctrl]

Model is the "Moab Be", .... Too hot by about 5 dB from 33 to 200 Hz

All Stereophile measurements of the overall frequency response are incorrect below 200-300Hz.

When combining gated and nearfield measurements, the half-space (2pi) to full-space (4pi) transition of the nearfield measurement is not taken into account.

You are probably the thousandth person to misinterpret the faulty measurements of by Stereophile. It is a complete mystery to me why Stereophile continues to use this incorrect representation of the overall frequency response of a loudspeaker or at least not puts a huge warning on the diagram explaining the incorrect presentation.

Depending on the baffle dimensions, the SPL must be reduced by 6dB below 200-300Hz in the Stereophile diagram.




It is also a pity that Stereophile does not question Tekton's marketing story at all, but reproduces it one-to-one:
Stereophile:

You see, years earlier, one of Alexander's most eye-popping and ear-pleasing ideas had been to cluster seven small high-frequency drivers in a tight circle and set the crossover to feed them midrange and low-treble only (in the case of the Moabs, between 772Hz and 3.41kHz). The cluster acts as a midrange driver with unusually low mass and thus exceptionally quick response. This helps especially with delicate overtones and harmonics, whose formation can be smothered by slower, traditional midrange cones.

As long as the driver used exhibits piston-like behavior, the size of the driver and the mass of the driver doesn't matter - e.g. cone mass can be counteracted with "stronger" driver motor to reach the same SPL.

A very nice real world example of the relationship between dome mass and SPL can be seen in the midrange domes from Bliesma. As long as the driver behaves ideally (piston like behavior) the mass only influences the SPL of the driver (the driver with heavier Al-dome is as "quick" as the lighter Be-dome ):

Source: HifiCompass

The driver with the lowest Mms, Be dome, shows the highest SPL and the heavier Al dome runs perfectly parallel (till about 5kHz) but with lower SPL with identical drive motor.
With a dome mass in between, the silk dome shows as expected a SPL in between, but due to the lower stiffness, it no longer shows ideal behavior much earlier.


But of course, in reality resonances inevitably occur. As the cone or dome area of a driver increases, resonances such as eigenmodes or surround resonances...., are shifted to lower frequencies.
The occurrence of resonances naturally also depends on the dome or cone material and surround material used. A pure paper cone or silk dome shows no more piston-like behavior at lower frequencies than when stiffer materials are used (see Blisma midrange example above).

One could now start a "high-end" argument that every resonance requires a certain settling and decay time and therefore affects the sound. And that out-of-phase resonances cause a wavy frequency response.

This is not wrong in principle and every woofer, for example, shows a more or less pronounced out-of-phase surround resonance.
Here is an example of an 8'' fullrange paper driver with severe out-of-phase surround resonance at 480 Hz:

If one follows this line of reasoning, then one should therefore avoid the resonance frequency of a driver in the used range as well, as the driver usually exhibits particularly pronounced settling and decay behavior there. So at all cost avoid to use a tweeter in the range of its resonance frequency - Tekton does the opposite with the tweeter in the "circular midrange array". So this "high-end" argumentation cannot be applied to Tekton.

Since it's Stereophile we don't have distortion measurements or understanding of what the speaker measures like at high volumes so the question of Sd, Xmax, and dynamic range is unanswered, but from this review it's still not looking good for the tweeter array concept.

Since the usage of the SB Acoustics Satori Berylium tweeters and the 772 Hz crossover point are explicitly mentioned, I would personally rate this design as problematic.
SATORI TW29B / SATORI TW29B-B: fs: 600Hz
SATORI TW29BN / SATORI TW29BN-B: fs: 700Hz
SATORI TW29BN-8 / SATORI TW29BN-B-8: fs: 750 Hz

crossover point is extremely close or even at resonance frequency.

The Moab uses 2x7 tweeters in the midrange array. This means that the excursion hardly plays a role even at the resonance frequency of the driver. Especially as the SPL at the alleged crossover frequency of 770Hz is already -3dB or -6dB, depending on the filter used.

Even without any filter (i.e. only the high pass behavior of the tweeter), 14 tweeters should have no problem with excursion at normal SPL <110dB.

 

[cavedriver]

All stereophile measurements of the overall frequency response are incorrect below 200-300Hz.

When combining gated and nearfield measurements, the half-space (2pi) to full-space (4pi) transition of the nearfield measurement is not taken into account.

You are probably the thousandth person to misinterpret the faulty measurements of stereophiles. It is a complete mystery to me why Stereophile continues to use this incorrect representation of the overall frequency response of a loudspeaker or at least not puts a huge warning on the diagram explaining the incorrect presentation.

---
The Moab uses 2x7 tweeters in the midrange array. This means that the excursion hardly plays a role even at the resonance frequency of the driver. Especially as the SPL at the alleged crossover frequency of 770Hz is already -3dB or -6dB, depending on the filter used.

Even without any filter (i.e. only the high pass behavior of the tweeter), 14 tweeters should have no problem with excursion at normal SPL <110dB.

Yes, I had read that years ago and in the back of my mind I knew that Stereophile's measurements below 200 Hz did not reliably mate up with their higher frequency measurements, although it's still interesting how John tries to do the summing and I suppose I had hoped after all these years that he had at least adopted some standard methods to make the process somewhat consistent. Perhaps I'm giving too much credit to the idea that this can be done in the absence of an anechoic measurement (NRC, etc.) or single mic measurement with background subtraction (Klippel). Fair enough that my words above do not convey this understanding but no, I am not unaware of this issue.

Regarding the Sd/Xmax comment, I didn't bother calculating it and only mentioned in passing because yes, it was likely not an issue given the high crossover. Since it wasn't taken for granted I did calculate it and I'm getting a limit of ~121 dB at 800 Hz for all 14 drivers. Of course this is just based on the Xmax using the formula below and not the usable driver displacement based on motor strength and power inputs, which I'm not prepared to hazard a guess at. Note that this puts the limit on the Tekton Impact Monitor at ~115 dB if the center tweeter is carried with the other 6 down to 800 Hz. So yes, at least by that evaluation there appear to be "enough" tweeters.

SPL (in dB) = 20 log [ f^2*VD /3.7] where VD is volume in cm3

 

[ctrl]

I had hoped after all these years that he had at least adopted some standard methods to make the process somewhat consistent. Perhaps I'm giving too much credit to the idea that this can be done in the absence of an anechoic measurement

The measuring process by Stereophile is consistent, but about 6dB wrong below 300Hz.

If done correctly, you can determine the low frequency range of a speaker to around +-1.5dB (my estimate, often deviation is less) via near-field measurement and subsequent merging of near-field measurement (with baffle correction) and gated far-field measurement. Depends on the complexity of the speaker (sealed vs. BR, multiple BR ports,...)

Here is an example of two merged near-far field measurements (red, orange - independently created as part of a round robin test in a German DIY forum without knowledge of the anechoic measurement) compared to a measurement in an anechoic chamber of the same speaker driver (blue - please note, the AC has a measurement accuracy of +-1dB down to 70Hz. Below this, the reflections can be greater).


So there is no excuse for not doing this properly (one don't need a Klippel NFS for reasonably correct measurements) or at least putting a warning about the inaccuracy on the measurements.

I did calculate it and I'm getting a limit of ~121 dB at 800 Hz for all 14 drivers.

Sounds right to me, but I didn't check myself.

 

[Randolf]

SPL (in dB) = 20 log [ f^2*VD /3.7] where VD is volume in cm3

Interesting idea to look at the effect of adding tweeters.

So it comes down to

SPL = 20 log (f^2 * Sd * Xmax * N / 3.7)

SPL: Maximum SPL in dB
f: Frequency
Sd: Effective piston area
Xmax: Maximum linear excursion
N: Number of tweeters

For the SATORI TW29B-B we have:

Sd= 9.6cm2
Xmax= 0.05cm

If the goal is SPL=120db we get

120 = 20 log(f^2 * 9.6 * 0.05 * N / 3.7) or
f = SQRT(7708333/N)

So we can plot the lowest possible frequency at 120dB with respect to the number of tweeters:

 

[Salt]

View attachment 363932

That's not anechoic room ... but You know, for sure ...

 

[Duke]

Yes, I had read that years ago and in the back of my mind I knew that Stereophile's measurements below 200 Hz did not reliably mate up with their higher frequency measurements, although it's still interesting how John tries to do the summing and I suppose I had hoped after all these years that he had at least adopted some standard methods to make the process somewhat consistent.

My understanding is that @John Atkinson kept using the same measurement technique largely for the sake of consistency, so that measurements made recently can be directly compared with measurements made long ago.

Correcting on-axis frequency response measurements below the baffle-step frequency is complicated when you don't have an anechoic chamber or a Klippel near-field scanner. Not only would the correction to be applied depend on the baffle width, it would also depend on the baffle height, the woofer diameter(s), and the number and locations of the woofer(s) on the baffle, not to mention what happens when the woofers are not on the front baffle. That's a lot of extra work for a frequency region that is going to be dominated by room interaction anyway. The informed reader can eyeball the speaker and the curve and draw his or her own ballpark conclusions about the net in-room results south of the baffle step.

John includes off-axis curves that go down well into the midrange; full-range in-room measurements; and often a comparison with a similar speaker's in-room measurements in the same room; plus other measurements... so if we look at the big picture he's providing a great deal of information even if it's not in a CEA 2034 format.

 

[Salt]

Why does a Stereophile not include SOTA measurements (that hobbyists do and dare to be sued for it)?
Just wondering.

 

[Bryguy]

Why does a Stereophile not include SOTA measurements (that hobbyists do and dare to be sued for it)?
Just wondering.

"My understanding is that @John Atkinson kept using the same measurement technique largely for the sake of consistency, so that measurements made recently can be directly compared with measurements made long ago." -Duke

 

[kemmler3D]

Why does a Stereophile not include SOTA measurements (that hobbyists do and dare to be sued for it)?
Just wondering.

Upgrading to a Klippel NFS would help them do better measurements, but maximum measurement quality may not be their objective.

They give a balanced take between subjective and objective, and a lot of their readers are only sort of interested in measurements in the first place.

 

[Duke]

Upgrading to a Klippel NFS would help them do better measurements...

Two limiting factors of the Klippel NFS system and loudspeaker WEIGHT and SIZE. The speaker needs to be light enough that it can be safely lifted onto the platform, and its dimensions need to fit within the range that the Klippel's arm can accommodate.

Arm extensions may be available, assuming the ceiling in the room is high enough, but there are no easy ways to lift a speaker than weighs hundreds of pounds up onto the platform, and I'm not sure who would be comfortable bearing the financial risk of doing so.

 

[NTK]

How Magico tested their M9 (weighs 1000 lb) with their NFS.

 

[Doodski]

How Magico tested their M9 (weighs 1000 lb) with their NFS.

That is a very serious speaker test! Nice video. Thanks for posting that.