New Celestion "AxiPeriodic Driver"

[Ro808]

Back on topic.
There are some similarities between the Axidriver2050's diaphragm and those from the Thiel and Fostex drivers.

The Thiel Coaxial Coincident Driver 4.5" midrange driver is not a cone. Its an aluminum diaphragm that resembles a flattened, corrugated doughnut. The tweeter nests into the center, where the dustcap would go in a conventional design. The "corrugations" are radial lobular indentations in the aluminum.
It's a very thin diaphragm that operates pistonically—very close to ideally according to Thiel.
The woofer is another essentially flat driver with radially mounted strengthening lobes. Like the midrange driver, it also employs Thiel's long-gap, short-coil drive system, which keeps the coil within a uniform magnetic field throughout its entire range.

For the Fostex a hi-rigid and reduced inherent resonance HR diaphragm is employed by using a "radial papermaking technique"
and whizzy hybrid materials (whatever that means).
Obviously the claim is reduced distortion. Even the spider is "radialized" (image no. 5).

 

[Zvu]

I don't know what that cone in Fostex serves for, but it measures quite dreadfully.

 

[Ro808]

That's what I thought. On top of that they're hideously expensive.

 

[Pallas]

I don't know what that cone in Fostex serves for, but it measures quite dreadfully.

The Thiel, at least as implemented by them, is also very fatiguing sounding. That doesn’t necessarily mean anything about the midrange, except that it’s a horrid tweeter baffle/waveguide.

 

[phase_accurate]

I once heard these on exorbitantly priced electronics. Although the combination did many things right I also had the impression that I probably wouldn't like to listen to them for long.

Regards

Charles

 

[Ro808]

These comments sound familiar.

Whereas later Thiels were superior in many respects, they didn't grab me in the same way the CS3s did.
The SEAS 10" & 5" + the Dynaudio D28AF tweeter provided a relaxed listening experience, but later Thiels were much too bright/forward for my taste.

 

[Ro808]

I highlighted Fostex, because it's a Japanese company and... apart from the well known Laboratory Series drivers, they also manufactured some impressive monitors in the early 1980's.

Celestion made pioneering use of Laser Doppler Interferometry to analyse diaphragm behaviour in the 1980's.
Once again, in the seventies the Japanese lead the way in analysing diaphragm behaviour and experimenting with diaphragm materials.


One of the most remarkable research projects was published by ISAMI NOMOTO, MAKOTO IWAHARA, AND HIDEO ONOYE from Victor Company of Japan, Ltd., Tokyo in 1975.

"For a shaped pulse excitation, responses at 3000 points in a plane through the major axis of a loudspeaker are measured and stored. The data are then processed and converted to visual. representations at 10-µs intervals. The resulting display is photographed sequentially for observation of wave-front motion away from the loudspeaker."

In other words, these guys made wave fronts visible!

 

[Ro808]

However, during this period Celestion became "big in Japan" 😉

 

[Ro808]

The Axi2050's diaphragm is basically flat, but deviates from existing designs in some respects.
One is the location and size of the voicecoil, another is the shape of the features.

Thiel might have taken a pretty good look at the flat diaphragms developed by the Japanese in the late 1970's and early 1980's. Besides Sony, Technics has put a lot of R&D effort into flat diaphragms.
In 2015 they revived this technology with the launch of Technics SB-10 and SB-R1.

The diaphragms are based on those developed for the SB series from the 70's and 80's.
These vintage Technics speakers were never very popular outside of Asia, but highly regarded by many today.

An externally hosted image should be here but it was not working when we last tested it.

 

[Ro808]

Here's the AES paper "Loudspeaker with Honeycomb Disk Diaphragm" by
Sakamoto, Naraji; Satoh, Katsuaki; Satoh, Kazue; Atoji, Nobuhisa
(Acoustic Research Laboratory, Matsushita Electric Industrial Company, Ltd., Osaka, Japan, 1981)

Description:
New loudspeakers have been developed, using honeycomb disk diaphragms, which exhibit low distortion, wide frequency range, and uniform frequency response. These characteristics are realized by using a very rigid diaphragm. The rigidity is brought about by varying the cell size of the honeycomb core in the direction of the radius and by driving the diaphragm at its first nodal circle. The result is a very desirable density gradient and equivalent Young's modulus.

 

[kipman725]

In other words, these guys made wave fronts visible!

Is that the fabled HOM I see in the horn plot?

 

[Ro808]

Those would be the ripples in the plots (on the left) below.

 

[Ro808]

Here's the SB-10 Whitepaper.

 

[Ro808]

Finally, some images of Technics' SB drivers.

 

[Patrick Bateman]

Is that the fabled HOM I see in the horn plot?

Keep in mind, HOMs can be very easily visualized in the Hornresp wavefront simulator.

I find it interesting that such a simple simulator achieves results that are surprisingly similar to BEM tools like axidriver and ABEC, but the Hornresp results are *animated*, which really helps me figure out what the issue is. This is particularly handy for phase plugs.

I guess if we had thirty two core CPUs and gigabytes of ram, it might be practical to use software that does animated BEM sims. My laptop is a quad core Ryzen and it still takes about two minutes to run an Axidriver sim.

 

[nc535]

the latest ryzen does have 32 cores!
AMD's wild 32-core Ryzen Threadripper chip goes on sale

 

[Ro808]

Most laptops - with the exception of a few high end HP's - are insufficient to run professional FEM/BEM software.
A workstation with dual XEON 2695V4, a couple of Tesla GPUs + 192GB RAM will run most audio related sims pretty much real-time.

 

[Patrick Bateman]

But ABEC and Axidriver don't use GPUs 🙁

Which really sucks; this type of software is ideal for that.

It is kinda funny that the Celestion BEM simulations of this driver, featured in the Youtube video, aren't a whole lot different than Hornresp simulations like this:

from here : Cloning a $3200 Speaker for $400

Animations created with Hornresp wavefront simulator and Image Magick

 

[Ro808]

These wavefront sims can be misleading (see attachment).

I know GP Acoustics' subsidiaries utilise high-end software including Siemens PLM.

ABEC and Axidriver are very powerful tools. Hard to beat in terms of price/performance,
but obviously limited compared to Ansys, Comsol, Siemens etc.

 

[kipman725]

May be available this month:
Celestion Announces NAMM 2019 Debut of Axi2050 Axiperiodic Driver | Live Design

Also found a related PhD thesis:
https://eprints.soton.ac.uk/348798/1/Jack%20Oclee-Brown%20PhD%20Thesis.pdf