A study on directivity... ribbons, domes, WG, horns, baffled or naked

[tktran]

As we all (should) know, a single on-axis measurement doesn't really characterize how a speaker sounds. You really have to listen to the speaker.
As I design / observe more speakers with spinorama data, which incorporate measurements in the horizontal plane and vertical plane in 360 degrees:


and presented in this manner:


...it started to dawn on me that although it answers some questions, it leaves out other things. You still have to listen to the speaker.

To quote the author/designer of the VituixCAD, things like "non-linear distortion, dynamics, stability of sound balance, timing, level of directivity, diffraction, location of radiators, half space and corner concepts are ignored..."
Reference:
https://kimmosaunisto.net/Software/V...ference_rating

Some of these are unexplored, or at least unpublished in the public domain. Here's what Sean Olive wrote iabout the research that was conducted some 20 years ago:


Reference:
Author's own post on distortion:

https://www.diyaudio.com/community/threads/who-makes-the-lowest-distortion-speaker-drivers.294787/post-7094796

I started my own study into non-linear distortion in 2022, but in 2023 I also started a small study on directivity.

To this aim, I've been collecting measurements of high frequency devices. This includes dome tweeters, soft and hard, ribbon tweeters, compression drivers, with and without waveguides or horns, with and without baffles. The goal is to see if measurements can determine the stability of the tonal balance as one moves around the listening area. Or perhaps how it may relate to spatial characteristics, broadly imaging, as one listens from the main listening position. How can measurements help tell us things like:

Reference:
https://extranet.itu.int/brdocsearch...2017-PDF-E.pdf


To help with some definitions, I will start off with a theoretical 3/4" tweeter. The graph below represents a circular driver, with a flat diaphragm of 3/4" diameter, flush mounted on a slim 5 1/8" wide x 12" H cabinet. The unit is centered on the baffle, with large 1.5" round-overs.



Grey- Reference: On axis response- directly in front of the tweeter, at 1 meter.
Green- Listening window: The average of the on axis, horizontal ±10, horizontal ±20, horizontal ±30, vertical ± 10 responses.


Here is the Directivity chart.


In the past it's has been called dispersion.
lower directivity ~ wider dispersion, and higher directivity ~ narrower dispersion.
constant directivity ~ constant dispersion - as one moves from directly in front to the sides; the sound levels drops evenly for all frequencies.
controlled directivity ~ controlled dispersion - by manipulating how much sound is radiated forwards, sideways, or backwards, we seek to "control the direction" of the sound.

This panel can be converted to normalized line chart, which shows how the frequency responses would behave at off axis angles, when compared to the normal (on-axis) response. Each colored line represents a frequency response measured in angular steps of 10 degrees.

You can also view this as:
A waterfall chart:

Or a surface chart:

Or a polar map:

Some critics will say that these are all representations of the same data, and they would be correct.
But as I will show, different graphs may be more useful than others to view/understand different things.
For instance, after you’ve seen a few of these big red stripes, you will start to be able to determine beamwidth.
And you will see that it's constant directivity from 500Hz to 5KHz, as it starts to become slightly narrower from 5-20 KHz, it starts to become more directive (higher directivity)

Next up is the Power (response) and Directivity Index chart:


Sound Power Directivity Index (DI) For the purposes of this standard the Sound Power Directivity Index is defined as the difference between the listening window curve and the sound power curve. An SPDI of 0 dB indicates omnidirectional radiation. The larger the SPDI, the more directional the loudspeaker is in the direction of the reference axis

Early Reflections (ER)The early reflections curve is an estimate of all single-bounce, first-reflections, in a typical listening room. • Floor Bounce: 20º, 30º, 40º down • Ceiling Bounce: 40º, 50º, 60º up • Front Wall Bounce: 0º, ± 10º, ± 20º, ± 30º horizontal • Side Wall Bounces: ± 40º, ± 50º, ± 60º, ± 70º, ± 80º horizontal • Rear Wall Bounces: 180º, ± 90º horizonta

Early Reflections Directivity Index (ERDI Total) The Early Reflections Directivity Index is defined as the difference between the listening window curve and the early reflections curve.

(Estimated) In-room response ... a Predicted In-Room (PIR) amplitude response is obtained by a weighted average consisting of 12 % listening window, 44 % early reflections and 44 % sound power.

Power Response: The (sound) power response is the weighted rms average of all 70 measurements, with individual measurements weighted according to the portion of the spherical surface that they represent. Calculation of the sound power curve begins with a conversion from SPL to pressure, a scalar magnitude. The individual measures of sound pressure are then weighted according to the values shown in Appendix C and an energy average (rms) is calculated using the weighted values. The final average is converted to SPL


Reference:
These definitions are covered in:
https://shop.cta.tech/products/standard-method-of-measurement-for-in-home-loudspeaker

[tktran]

Now, onto real devices:

First up:

Scan-Speak D2604/8330-00 26mm soft dome tweeter.
Cabinet/baffle: 9.5" W x 12" H x 8" D cabinet
Location: 60mm from top edge, offset 1/2" from center line
Diffraction management: nil (rectangular box)





Looks messy!

Let's remove some traces to simplify:




But when we compare to the Scan-Speak's measurement, this is what the manufacturer shows:


Reference:
Scan-Speak Datasheet

What? Why are they different?
Well it's the baffle-

The orange line is what the response SHOULD look like if the baffle was infinitely large. But due to finite dimensions of the baffle, and how sound interacts with the edges of the cabinet, it causes interference that results in peaks and dips in the frequency response.

Shall we straighten out the response, the best we can?
We can use as many passive crossover parts as we can imagine, or use a fine tuned signal processor (via PC or DSP etc)

Let's do that with VituixCAD2 to see what happens.
Below, the on-axis response is flat as a pancake from 500Hz to 20KHz. It's the gray line: +/- 0.0 dB!

Now it's flat on axis, but errr... there's a peak just below 3KHz, off axis between 20-50 degrees. Hmmm...



Going back, what did the Power and DI chart show?


Here, the on axis looks ragged. The other curves aren't flat, but seem to look smoother than the on-axis curve.


What happens when the on-axis is flattened?


Now the on-axis is flat, but now the early reflection, in-room response and power response looks ragged.
NB. Directivity Indices didn't change. They are half-baked in (more on this later)

Did we make it worse? Or better?
Should be accepting the tweeter's response as is, or flatten it?

If you haven't decided yet, let's take a look at the polars charts:

Left as is-



On-axis flattened:




What are those patches?

At under 2KHz and under 3KHz respectively, they represent excess energy.
Excessive 2-3KHz energy from across a 30 degree span, from 20-50 degrees. How does this sound?

Too much energy around 2-3KHz has "tinny/crunchy" character.

Reference:
Interactive Frequency Chart from Independent Recording Network (alexiy.nl)

[tktran]

Let's add some high pass filters.

First, let's consider a low crossover point with a shallow slope, eg. LR2 @ 1.8KHz




And let's flatten the on-axis, so we have a perfectly textbook acoustic LR2 @1.8KHz




These Line Charts can be represented as Polar Maps:


Here is the Polar Map of LR2@1.8KHz:



Polar Map of LR2 1.8KHz with perfectly flattened on axis:

Polar Map of LR4 1.8KHz:

Polar Map of LR4 1.8K with perfectly flattened on axis:

In the same fashion, let's look at the Polar Map for with a higher crossover point.


LR2 @ 3K:




Perfectly flattened LR2 @ 3K:



LR4 @ 3K:



Perfectly flattened LR4 @ 3K:


Comparisons:

LR2@1.8K:



LR4@1.8K:



LR2@3K



LR4@3K









Whether you use passive or active or DSP to completely flatten the response, in all cases completely smoothing out the on-axis affects ALL the other response, and can affect them in undesirable ways.

LESSON #1:If you measure/design with the on axis only, (*and don't have diffraction control eg. large round overs) don't completely flatten the on-axis response!

[tktran]

NEXT UP:

The Ring Radiator cousin, the
Scan-Speak R2604/8330-00

Exact same cabinet and measurement conditions. (9.5" W x 12" H x 8" D cabinet Location: 60mm from top edge, offset 1/2" from center lin Diffraction management: nil (rectangular box))






Here is a measurement from 1K - 20K

and the Line Chart of the R2604:


let's flatten the on-axis of the R2604:

What do you do when you pop a different tweeter into your speaker? Typically you adjust the levels so that the SPL are matched.

So lets match the SPL levels of the D2604 based on the on flattened axis:


apply a LR2 XO @3KHz:


Here's the Polar Chart:
With an LR2 @ 3KHz:


This Polar Map looks quite different from the D2604.

The Polar Chart when the on-axis is flattened:



Recall the Line Chart of the D2604...


Flattened, and then SPL matched:


Insert an LR2 @3K crossover filter:


The Polar Map:


The Polar Map when flattened:



Finally, Power and I charts:
First, R2604/8330 with the perfect acoustic LR2 @3KHz

The D2604/8330 with the perfect acoustic LR2 @3KHz



And when we didn't try to flatten the on-axis measurement.
For the R2604:

D2604/8330 not flattened-

[JonMarsh]

Kudos for taking the time to make this effort- I'm in the middle of a bunch of stuff, but have some thoughts to share, some in reference to specific designs (commercial and otherwise) and the attempt to address some of these issues.

[tktran]

Next up-
Let’s have a break from all those messy graphs.

Now, instead of the simulation from Post 1, a real 19mm soft dome tweeter on a relatively narrow baffle.


SB Acoustics SB19ST-C000-04 19mm soft dome tweeter.
Cabinet/baffle: 5 3/16" W x 10 3/8" H x 6 11/16" D cabinet
Location: 58mm from top edge, centerd in horizontal plane

First off, a cabinet with sharp edges ie.
1. No diffraction management: (rectangular box)


Frequency response measured from 0-90 degrees horizontally:


On axis now flattened:




As a Polar Map:




2. Diffraction management: 1" round-over on side panels

Frequency response measured from 0-90 degrees horizontally:




On axis completely flattened:

Flattened as a Polar Map:



Let's add crossovers:

LR2 @4Khz (no round-overs)


As a polar map:




As a Polar Map:


A LR2 @4Khz to the cabinet with the 1" round-over:



As a Polar Map:

On axis flattened with LR2@4KHz without round-over:


As a Polar Map:




On axis flattened with LR2@4KHz to cabinet with 1" round-over:


As a Polar Map:

Power Response and Directivity graph:

[Reet]

Thanks for posting your analysis Thanh. Images in post 3 are missing however

[tktran]

JonMarsh or theSven

hmm I can see them in my end? Can you guys see the pics on post #3?

[theSven]

The photos never uploaded to the post. All the images show as temp when I look at the source of the post. Can you try and edit the post and re-upload the photos please :-).

[tktran]

Are pictures up now?

[theSven]

Yes I'm now seeing the images in post 3.

[Jonasz]

No, no pics in post #3...

Interesting thread!

[tktran]

Next up...

A naked tweeter:

A Bowers and Wilkins diamond dome tweeter in a Nautilus tapered tube... (baffleless)

But first, a ribbon tweeter first on a narrow baffle, then on a wide baffle...

TBC...

[JonMarsh]

One of the things the newer updates on forum software does is flag a post when there are a lot of edits- send me a message, but I check in every day when possible, and will fix it, like I just did today.

[tktran]

Next up is a Fountek ribbon tweeter, model CD3.5H
Measurements are courtesy of www.audioexcite.com

"Baffle size (WxH): 24×35cm with 18mm round-over.
Driver position: No flush-mount, mounted on center-line with driver unit center 8cm from the top of the baffle.
Mic position: 1m distance, on tweeter-axis.
Smoothing: 1/24 octave smoothing applied.
Protection: A 47uF protection cap is used."


All measurements are extracted, converted and imported in VituixCAD2 for visualization:

First the line chart:



Polar Map:



On axis completely flattened from 1-20 KHz:





The practical limit for the Fountek NeoCD 3 is a higher crossover point eg. LR4 3Khz or LR2 4KHz. Here I insert a LR2 filter @4 Khz:


Polar Map:


Flattened and filtered LR2 @4Khz:





Power and DI chart, standard vs flattened


With a crossover LR2 @4Khz:



Reference:
www.audioexcite.com » Fountek NeoCD3.5H

[tktran]

Next ribbon comes courtesy of member Reet

This is his baffle:
8.75" W x 16 5/16 H Roundover 10mm
Tweeter centred, 3.75" from top edge
Diffraction pattern as below:



This is the Neo CD3.0 ribbon tweeter. Unlike the 3.5H, it’s not horn loaded.

Here’s the line chart:



Polar Map:



Flattened 1-20Khz:


Polar Map:



LR2 @ 4 Khz crossover:


Polar map:


Flattened and LR2 @ 4Khz:


Polar Map:


Power & DI:




The Fountek NeoCD 3.0, without an waveguide or horn, exhibits different directivity/dispersion pattern to dome tweeters. Long chastised for having higher harmonic distortion than dome tweeters, does this directivity pattern the reason for its favourable sonic attributes?

[tktran]

Now the Fountek NeoCD3.5H nude! ie. Not mounted onto a cabinet.

Courtesy Juhazi

Please ignore everything below 2.5KHz (VituixCAD2's highest low frequency limit is 1KHz)


Polar Map:



Flattened:


Polar Map:



LR2 @ 4Khz crossover:


Polar Map:


Flattened with LR2 @ 4Khz:

Polar Map:

[tktran]

Next up:
A hard dome tweeter in a minimal baffle arrangement-



This tweeter is the B&W HF00963
For people unfamiliar with this tweeter- it's a 1" tweeter with a diamond dome and small neo motor, mounted in a tapered tube:



Tweeter only:


Measurements courtesy of Petr Kocourek, who maintains a website of measurements of drivers and projects here:
PKAudio: DIY loudspeaker projects (webnode.cz)

Measurement conditions:
Measured @1m
Time window: ~4ms
Note: drive level is not 2.83V

Tweeter, measured as part of 802D2 speaker, 4cm below middle of dome, no crossover:


Polar map:



Completely flattened, reference level at 10KHz:




Polar Map with 1dB divisions:


Note how the Polar Map demonstrates nearly 180 degree covered below 3.5KHz, but above 4KHz a controlled directivity with a -6dB beamwidth of about +/-50 degrees:


B&W use a single capacitor on their tweeter! A 4.7uF (Mundorf brand 1200VDC Silver/Gold oil capacitor).
Let's attach one:



B&W seems to have a habit of crossing their tweeters around 4KHz in their 800 series. Perhaps this is why...



Power and DI Chart with the 4.7uF cap for crossover:

[tktran]

Coming up...

SB Acoustics SB29SDAC centered on a 8” wide baffle (Jeff Bagby Revolution mini):
Tweeter of KEF Reference/Blade Meta coaxial (2022 model)

[tktran]

Next up:

Tymphany DFM-2535R00-08 in the SB Audience H225 horn
(free standing, no cabinet)

Measured on-axis:

HORIZONTAL off-axis measurements:

Vertical off axis measurements:



Flattened- horizontal responses:

Flattened, vertical responses:


POLAR MAP: horizontal



POLAR MAP, vertical:




A nice combination of compression driver and horn with constant directivity. Just off the shelf, bolts on easily:


BONUS data !!!

Before you think I'm cherry picking results... here's a horn that DIDN'T work so well with this compression driver...

DFM2335 in the SB Audience H250 horn:


flattened on-axis response, POLAR MAP, horizontal reponses:



flattened on-axis response, POLAR MAP, vertical responses:



Just like drivers, some horns are more suitable than others...

[JonMarsh]

Let's add some high pass filters.

First, let's consider a low crossover point with a shallow slope, eg. LR2 @ 1.8KHz




And let's flatten the on-axis, so we have a perfectly textbook acoustic LR2 @1.8KHz




These Line Charts can be represented as Polar Maps:


Here is the Polar Map of LR2@1.8KHz:



Polar Map of LR2 1.8KHz with perfectly flattened on axis:

Polar Map of LR4 1.8KHz:

Polar Map of LR4 1.8K with perfectly flattened on axis:

In the same fashion, let's look at the Polar Map for with a higher crossover point.


LR2 @ 3K:




Perfectly flattened LR2 @ 3K:



LR4 @ 3K:



Perfectly flattened LR4 @ 3K:


Comparisons:

LR2@1.8K:



LR4@1.8K:



LR2@3K



LR4@3K









Whether you use passive or active or DSP to completely flatten the response, in all cases completely smoothing out the on-axis affects ALL the other response, and can affect them in undesirable ways.

LESSON #1:If you measure/design with the on axis only, (*and don't have diffraction control eg. large round overs) don't completely flatten the on-axis response!

This is very true, and one reason that an approach I find very useful with RøDETest Fuzzmeasure is to take the RMS average of a range of polar angles, and use that for an initial POC design. This simplifies the data processing in design, but with the ability to quickly shift between compare, average, and RMS average, it grants insights with overlaying these graphs that are pretty useful.

[Reet]

With VituixCAD I can almost turn off on-axis response completely. I find the listening window average to be a better metric for forward facing response. It can be configured to include any angles, by default it follows CTA-2034A, +/-30 deg horizontal, +/- 10 deg vertical. Diffraction problems will be clear in DI.

[Reet]

tktran Can I assume you've come across the study of cabinet diffraction at Heissmann?

https://heissmann-acoustics.de/en/kantendiffraktion-sekundaerschallquellen-treiberanordnun/

https://heissmann-acoustics.de/en/schraege-fasen/

[tktran]

Yes, Alexander’s small studies are very well documented and explain how to minimize diffraction and get a smoother response. I like them very much.

The DI line is useful. Sometimes I gets only a cursory glance. Perhaps I should pay more attention to it. I'm yet to clarify, in my mind at least, why I prefer Polar Maps.

[tktran]

...moved to bass directivity thread:

Directivity study- woofers and subwoofers. - HTGuide Forum

[Reet]

Yes, bass is not truly Omnidirectional as people say, but rather Omni-ish, and distance of sound travel to the listener will affect what the listener hears, as well as how the bass interacts in the room.

[tktran]

@Perhaps this was why Zaph asserted that front facing ports were superior. I'm not certain.

He was going to document a study showing why, but never got around to it. (Feb 2011 of his blog- Zaph|Audio (zaphaudio.com))

[tktran]

I've listened to the LS60, as well as the Blade 2 (non-Meta). When I do auditions, I listen to the same 3 most familiar recordings (about 15 minutes) and I didn't care for it. I've always wondered why.

kimmosto thoughts?

[tktran]

theSven JonMarsh please approved above posts (too many edits)

[theSven]

I approved them.

[Reet]

@Perhaps this was why Zaph asserted that front facing ports were superior. I'm not certain.

I don’t agree with Zaph on that. I prefer my midrange leakage to come out the back side. Also I think ports are mostly ugly, so put that where I can’t see it.

[tktran]

What he wrote was:

This goes against everything that has been said on the subject of port placement, but here goes: I believe front firing ports outperform rear firing ports. At least, that's just what my ears have told me over the years. I think I may have mentioned this before but now I expect to get empirical evidence from a carefully constructed experiment. This experiment will isolate some of the issues which limit testing that is intended to show midrange leakage from a port.
To tell the truth, I've never seen anyone do a front port measurement that takes all the steps required to remove the output of the front side of the woofer. This combined woofer output causes some measurement artifacts that aren't really there, and sometimes these artifacts are misdiagnosed as organ pipe resonance or midrange leakage. The simple fact is that you can't accurately measure and isolate a port's output if there is a radiating woofer anywhere within several feet. I'll get into the details of why this is the case when I complete the experiment and publish the results. I'll also go into some methods that will allow you to at least get more accurate port measurements. This is something that is seriously lacking in DIY world.
The experiment will use the Peerless 830946 SLS 6.5" Woofer. This is an excellent woofer that is well designed on all fronts. It is closer to a subwoofer than a midwoofer, with some unique properties that make it perfect for this test and the final application I had in mind.

Bold & italicized commented by me, for emphasis.

Anyway, I thought it was interesting.

Do you have any thoughts on this Reet or
kimmosto

[Reet]

Same comment as above. Regardless of how easy the port resonant modes are measured, they are real, and often at frequencies above baffle step, so placing the port on the back can have a reduction of up to 6dB for that alone. That comment may apply more so for 2 -way speakers, possibly a 3-way can help with a low pass crossover below port resonant modes.

If there's any small loss efficiency from the port location on the back, I'll live, like I said I generally prefer the portless look, so I hide them on the back.

[tktran]

<deleted<

[technodanvan]

That 3" mid looks interesting! At first I read 'volt' and thought of an old cheap (maybe car audio)? brand that might have been yellow and black. This....does not look like that!

Edit: That whole design is interesting, thanks for posting it.

Edit 2: Just for grins, I have the last pair Solen had. They are...impressive.

[tktran]

From 300Hz to 3KHz it's a good one. I wonder if yours also has that dip at 2.5KHz? Curiously when Vance measured it a few years ago it didn't have that dip:

Test Bench: Volt Loudspeakers VM752 3” Midrange Dome Driver | audioXpress

Please let us know how you go.

best regards,
Thanh

[tktran]

NEXT UP:
Fountek NeoCD3.0 8mm x 60mm ribbon tweeter, as part of the Statements II - A musical evolution! - HTGuide Forum, introduced to me by Jim Holtz

(This is a beautiful speaker in looks and sound IMHO... but you'll have to follow that thread to see my own variant finished in 2015)

Jim's speaker:



Data traced from Curt C off axis measurements, from the full design here: https://www.speakerdesignworks.com/copy-of-anthologys

Here's a brief of the baffle dimensions
Cabinet/baffle: 292mm (11.5") wide x 55" tall
Tweeter location: centered, 22" from top edge
Diffraction management: 3/4" round-over

Curt's original measurements are horizontal from 200Hz to 20KHz at 0/15/30/45/60/75/90 degrees
Here they are imported into VituixCAD2 with the original passive filter:

ON-AXIS:




Let's zoom into region of interest- on and off axis.
LINE CHART:





POLAR MAP:




Notice something interesting?

POWER and DI chart:


Notice any interesting about the DI line?

Now, if we used a complex filter to completely flatten the on-axis response, this is what we'd get:

ON AXIS:



Well, it doesn't get any better than that.
But what about the off axis measurements?

LINE CHART:



POLAR MAP:



POWER and DI:







LESSON #2:

A ribbon tweeter WITHOUT waveguide can give constant directivity.
#2.1 When the baffle dimensions is chosen appropriately
#2.3 The complex crossover that gives perfect on axis response gives little (if any) improvement.

Well done Jim Holtz and Curt C

[Reet]

Does traced data include both horizontal and vertical response?

I notice most data shown here includes 90 degrees of information, however some is 180 degrees. For DI calculation using only 90 degrees of data, consider checking "half space" in options to shift up DI to be more comparable with 180 degree response DI. Alternatively, use only 90 degree data consistently for consistent comparison.

[tktran]

I extract as much data as available. So far I've focused on the horizontal because the vertical is often not available.

I do note that you’ve got vertical data for your NeoCD 3. I need to go back to it and add some bonus data.

Thank you for reminding me of clicking on the right option to shift the DI when measurements are limited to only the front hemisphere.

[Reet]

For non round drivers, waveguides or horns, both horizontal and vertical is required to make any conclusions on power response characteristics.

[Reet]

Here, you can see that power response and DI curve drop a fair bit when full horizontal and vertical response is used, due to the limited vertical directivity. Dashed lines include only horizontal response, solid includes vertical.

[Jim Holtz]

NEXT UP:
Fountek NeoCD3.0 8mm x 60mm ribbon tweeter, as part of the Statements II - A musical evolution! - HTGuide Forum, introduced to me by Jim Holtz

(This is a beautiful speaker in looks and sound IMHO... but you'll have to follow that thread to see my own variant finished in 2015)

Jim's speaker:



Data traced from Curt C off axis measurements, from the full design here: https://www.speakerdesignworks.com/copy-of-anthologys

Here's a brief of the baffle dimensions
Cabinet/baffle: 292mm (11.5") wide x 55" tall
Tweeter location: centered, 22" from top edge
Diffraction management: 3/4" round-over

Curt's original measurements are horizontal from 200Hz to 20KHz at 0/15/30/45/60/75/90 degrees
Here they are imported into VituixCAD2 with the original passive filter:

ON-AXIS:




Let's zoom into region of interest- on and off axis.
LINE CHART:





POLAR MAP:




Notice something interesting?

POWER and DI chart:


Notice any interesting about the DI line?

Now, if we used a complex filter to completely flatten the on-axis response, this is what we'd get:

ON AXIS:



Well, it doesn't get any better than that.
But what about the off axis measurements?

LINE CHART:



POLAR MAP:



POWER and DI:







LESSON #2:

A ribbon tweeter WITHOUT waveguide can give constant directivity.
#2.1 When the baffle dimensions is chosen appropriately
#2.3 The complex crossover that gives perfect on axis response gives little (if any) improvement.

Well done Jim Holtz and Curt C

I've got to jump in and give Curt all the credit for the exceptional performance. He is a crossover "wizard" and always gets the most out of a speaker its capable of achieving.

I'm really sad the mids have gone NLA. Fountek is back and Madisound has them in stock. Driver availabilty has become our speaker design nemesis.

Enjoy!

Jim

[duvixan]

I'm really sad the mids have gone NLA.

MR13P are probably just as good, if not better

[Jim Holtz]

Curt and I are both SB fans. I can't argue with you on MR13's. However at the time the NE123 went NLA, Fountek also disappeared so the decision was made to focus on the Anthology II's. We're also big fans of the SB CAC series drivers. I don't regret the decision.

Jim

[tktran]

ONWARDS and OUTWARDS:

Next up the SICA 5.5" coaxial. Product info here: https://sica.it/prodotto/5-5-c-15-cp/

This is a 5.5" nominal midrange with 1.5" diameter voice coil, to which Sica put 1" soft dome tweeter:



Data comes courtesy of vineethkumar01

He installed it in an frame with a 18cm (7") outer diameter and 12mm (1/2") roundovers.

Again we focus on the tweeter performance, for now:

Here's the on-axis performance:


Off axis LINE CHART:


FLATTENNED, but we ignore trying to correct for the dips, just like we would with a passive crossover:





Here's the POLAR MAP:



Add a LR2 @2KHz HP shall we?



POLAR MAP:


Looks fairly controlled from a directivity point of view, but can we do anything about that dip around 9.5K and 12.5 K Hz?

Supposed we try to flatten everything from 1KHz to 20KHz. In the previous post that was a bad idea. Is this still the case?

LINE CHART:


POLAR MAP:



LESSON #3
The directivity control of a tweeter is a coaxial is controlled by midrange cone itself. The cone acts as a "waveguide"
#3.1 Don't try and correct for any dips in the on-axis response. Let it go!

[duvixan]

the SICA 5.5" coaxial

It's currently used in $$$ commercial speakers by Sigberg Audio and Antelope Audio

[tktran]

Am aware. And it’s not even the “best” coaxial!

more to come. 😎
5th element

[tktran]

By now we've had a few examples of how these POLAR MAPs are created.
For future posts, I will speed things up a bit, and skip all the explanatory parts for the equalization and insertion of crossovers.

I will post POLAR (heat) MAPs in the horizontal (and vertical plan, if available), as well as Power / DI chart to show on/off axis frequency responses and thus directivity (dispersion) ONLY.
Assume sensitivity to be 2.83V and measurements are at 1m, unless otherwise stipulated, and dimensions of the cabinet to which the driver is installed, as well as diffraction management eg. large round-overs/facets.

Of course, I will update the posts when questions are asks and new information is available.

Please review earlier previous posts if you have challenges understanding, or ask away.

NEXT UP:
This driver is a prototype coaxial by HT Guide's 5th element .

This is the 3rd coaxial driver he's developed using off-the-shelf parts, and the refinement is evident.

If I'm not mistaken, this tweeter is optimized to match the directivity of the midrange, which is to say directivity control is from 2.5KHz, where the midrange cone tends to beam. Also, the performance is optimized for 15 degrees off axis.

So without further ado, here's the SB17NBACoaxial tweeter performance, flattened and with acoustic LR4 @2.5KHz.
Data courtesy of Matt Storey / UK:





Research findings have been communicated with SB Acoustics.

You can look forward to a genuine SB Acoustics (non-Satori) SB17-based coaxial, manufactured by SB Acoustics, sold at your favorite DIY distributor in future!


Reference:
SB17Coax (5een.co.uk)

PS.
(Pssst 5th element - what is the dimension of the baffle to which this driver was installed, and what diffraction management did you employ? How large were your round-overs or facets?)

[duvixan]

You can look forward to a genuine SB Acoustics (non-Satori) SB17-based coaxial, manufactured by SB Acoustics, sold at your favorite DIY distributor in future!

Is this a hope or a hint?