Ironless Motor Assemblies and STEALLUS

[Heinrich]

STEALLUS is an acronym ...

There are two older bass drivers that show an up to date performance: JBL 2206 and JBL 2020. Especially the latter is of interest - the Xmax as given by JBL is unexpected, but real. The virtually non-existant distortion is real also. So much on roughly 40y old designs.

18sound implemented AIC, active impedance control to full success, but is was a commercial fail.

Now Purify seems to address the motor non-linearity again. But they consider the surround also, which at some point is unavoidable because it becomes a major issue compared to the optimized motor. Purify dismisses the caveat with Doppler distortion, but to me it appears as not well justified.

My question is, how far this should be driven (sic!). One may ask, if it wasn't more effective in many aspects to just double up the cone area, taking a bigger driver, or two of the original size and equalize the output at will in the digital domain.

Is there anything in advantage of the new center pole?

 

[AOR]

Publishing the STEALLUS concept in 2007 seemed to influence B&W's 800D woofer design.
https://pearl-hifi.com/06_Lit_Archi...s/Mowry_Steve/B&W_Dual_Magnet_vs_Steallus.pdf

View attachment 439271
https://www.linkedin.com/pulse/bw-800d-series-motor-toplogy-review-steve-mowry/?trackingId=E2lcBnaiSi2ZaiAYHRyHYQ==

It also seemed to influence the KEF Blade Woofer design somewhat but we could ask Jack.

View attachment 439274
https://www.linkedin.com/pulse/kef-blade-wofer-review-steve-mowry/?trackingId=E2lcBnaiSi2ZaiAYHRyHYQ==

The Blade LF has a motor geometry for slightly different reasons. The woofer uses nodal drive to drive the diaphragm such that the first breakup mode isn't excited. This means that the voice coil has to be really large. Therefore we had to use a pot motor rather than a ring motor. In order to get the required B, we needed some extra ferrite discs. Other than that, it's a fairly standard overhung motor with dual shorting rings.

 

[Chris Brunhaver]

There are two older bass drivers that show an up to date performance: JBL 2206 and JBL 2020. Especially the latter is of interest - the Xmax as given by JBL is unexpected, but real. The virtually non-existant distortion is real also. So much on roughly 40y old designs.

18sound implemented AIC, active impedance control to full success, but is was a commercial fail.

Now Purify seems to address the motor non-linearity again. But they consider the surround also, which at some point is unavoidable because it becomes a major issue compared to the optimized motor. Purify dismisses the caveat with Doppler distortion, but to me it appears as not well justified.

My question is, how far this should be driven (sic!). One may ask, if it wasn't more effective in many aspects to just double up the cone area, taking a bigger driver, or two of the original size and equalize the output at will in the digital domain.

Is there anything in advantage of the new center pole?

Well, purifi is doing something very similar to what steve proposed by using a pre-charged NEO magnet in place of the steel in the pole piece but, rather than no return path or "air" (like steve) they are using a ceramic magnet. In short, they are similar and trying to achieve the same thing.

Purifi "balances" the tendency to DC rectify towards wherever there is the most steel by adding an extra little piece of steel on the top of the pole. They are also doing the "neutral" surround, both of which are pretty novel.

These jbl drivers from the 70's are good but not as good as modern stuff with these kind of techniques involved (i.e. there is no reason why you couldn't incorporate this stuff into larger drivers. Larger drivers, while higher in sensivity require proportionately larger enclosures and have lower high frequency "beaming" limits and we now have very high power class-D amps and people may want smaller boxes etc. so designs tend to favor smaller footprints and these technologies become more important. However, they benefit larger drivers too.

 

[Heinrich]

Well, purifi is doing something very similar to what steve proposed by using a pre-charged NEO magnet in place of the pole piece but, rather than no return path or "air" (like steve) they are using a ceramic magnet. In short, they are similar and trying to achieve the same thing.

Purifi "balances" the tendency to DC rectify towards wherever there is the most steel by adding an extra little piece of steel on the top of the pole. They are also doing the "neutral" surround, both of which are pretty novel.

These jbl drivers from the 70's are good but not as good as modern stuff with these kind of techniques involved (i.e. there is no reason why you couldn't incorporate this stuff into larger drivers. Larger drivers, while higher in sensivity require proportionately larger enclosures and have lower high frequency "beaming" limits and we now have very high power class-D amps and people may want smaller boxes etc. so designs tend to favor smaller footprints and these technologies become more important. However, they benefit larger drivers too.

Thanks for the explanations. To mention Purify was meant as an example of, to me, exaggerated effort. They cannot be used with ease because of foremost Doppler, then enclosure size. Neither a port nor passive is viable for the intended form factor. 2-way with a waveguide spoils the vertical directivity. Going 3-way would further either increase cost, or would blame the fine woofer, it is, for its quite high cost. In a 3-way it doesn't need to be that perfect in its own right. On and on it goes.

Given the limited capabilities of the human hearing, and considered cheap conventional drivers used in multiples w/ equalization, what would a STEA be made for?

 

[smowry]

A core issue with high displacement transducers (woofers) is the asymmetric topologies that typically lead to asymmetric large signal parameters. So yes one must consider the motor and the suspension (spider(s) and surround). If you look at my sketch of the KEF blade woofer below, I ask the question why are the spiders not mounted in a complementary top[ology? I would be amazed if the Kms(x) curve of the Blade woofer is symmetrical. Any asymmetry in one spider will be x 2, rather complementary spiders almost guarantee symmetrical Kspider(x). The annular radius of the spiders is small and this will constrain Xmax. Whereas the simple surround geometry can be made symmetrical by using non-linear reaction force simulations. However, look closely and this surround has different boundary conditions on the ID and OD.

Additionally, because the top magnet is without steel on one side it seems too thin and this will result in a low operating point on the magnet's BH curve. The H within the magnet will be low and significantly lower than the other 2 magnets within the assembly. Magnets without steel need to be thick. This topology is similar to what was called a bucking magnet that was add to assemblies to increase BETA and reduce stray magnetic fields.

I never liked the Blade Woofer design and I was critical in my review in 2015. https://www.linkedin.com/pulse/kef-blade-wofer-review-steve-mowry/?trackingId=E2lcBnaiSi2ZaiAYHRyHYQ== My comment was "disappointing". I don't see any shorting rings, do you?

 

[smowry]

Purifi is new to me but seems to be on another level. They are doing 3D non-linear FEA on their surround design iterations, that most likely take hours to converge on a solution. Their motor assemblies, although traditional topology seem well designed with close attention to detail and clearly have been optimized. When I started working with Ilpo at GENELEC, Carsten was already supporting GENELEC as a freelance transducer engineer. I was curious so I asked Jussi how does Carsten designed his spiders. Jussi replied that he had a formula. Well that's Johnson (linear analysis). In effect no analysis at all! One cannot design a spider with a formula; one must use Newton's Method. I was using ABAQUS at that time but only axi-symmetric 2D models that would solve in minutes but have obvious geometric limitations. This was a powerful tool but just for axi-symmetric models. Most if not all spiders are axi-symmetric. Most surrounds are also axi-symmetric but the Purifi surround is on another level and appears to be a very effective implementation but Carsten did not design that with formulae. This was really not surprising because when I first visited GENELEC in Finland their engineers claimed that spiders made no difference. They claimed to have test results to prove it. Sure if displacement is 10^-4 (m) then measurements of various spiders may not show significant difference when mounted in a transducer but that's for small signal only. When they told me that spiders don't make a difference, I felt like when I was a kid and found out there was no Santa Claus.

FAQ: Why did GENELEC bring in Carsten and I? Well, their engineers were ineffective in new transducer development at that time. GENELEC had one COMSOL license but the mechanic running it was inexperienced.

As a transducer engineer for 25+ years, I can says that transducer engineering in one word (acronym) is FEA.

 

[AOR]

Congratulations on retiring! I'm not sure if I'll get to retire in a tropical paradise. Let me know if there are any loudspeaker design gigs in Thailand that I should be aware of

I don't know of any patents of theirs on their motor but the former international sales guy from my work (Travis Townes) now works there, so I might ask him. He is a sales guy and very non-technical so I'm not sure how much info I'll get.

I found a European patent EP3634013B1. I believe the higher end børresen speakers use the geometry described there.

 

[smowry]

Outstanding, AOR!
https://patents.google.com/patent/EP3634013B1/en?oq=EP3634013B1

Well there are four Neo rings but the voice coil is underhung and the spider is tiny. This one may have low small signal distortion but Xmax will be very small. It looks like the assembly can be pin gauged from the bottom. This is more like a midrange driver rather than a woofer. Borresen seems to use midranges for their woofers but they have the correct power cords and interconnectors.

This motor assembly looks like my tweeter motor from 2008 - 2009 just with a taller gap. Michael Borresen has the balls to claim an invention. Has he no shame?

https://pearl-hifi.com/06_Lit_Archi.../Mowry_Steve/Air_Core_Tweeter_Magnet_Assy.pdf

 

[Heinrich]

Outstanding, AOR!
https://patents.google.com/patent/EP3634013B1/en?oq=EP3634013B1

I'm neither an engineer nor a technician. From that perspective the effort presented here is crazy. Very much like the German combustion engine, beyond complicated, fragile and so by chance pretty shortlived, overdone for no good reason. Saves fuel (hopefully) but all effort is spoiled by the car's hybris.

The JBL 2020, a 12" mid/bass is virtually unaffected by harmonics or intermodulation. The design is like 35 years old, stoneage.

The pole-less design can't focus the magnetic field as good, hence the effectiveness is less than optimal. Cost is going to be (very?) high, reminding me of ribbon tweeters or 'AMTs'. Compare the design from the end-user's perspective to multiple conventional drivers in the same enclosure plus digital equalization. Compare to 18sound's AIC, which costs just a few meters of extra wire on the pole piece but was still unsuccessful commercially.

All this fancy designs may address the high end market for wealthy customers. As with German cars, wrong direction, me thinks. As an afterthought, KEF does is right with their pedestrian series. Tone down complications once a stellar optimum was achieved ;-)

ps: radial magnetization, really? Wow!

 

[smowry]

Purifi is on another level.

The "PermanentCore" appears to be a Ferrite Core similar what is used in transformer cores. It is not a ferrite magnet. It has low Hc but high permeability, B/H.

"Ferrites that are used in transformer or electromagnetic cores contain iron oxides combined with nickel, zinc, and/or manganese compounds. They have a low coercivity and are called "soft ferrites" to distinguish them from "hard ferrites", which have a high coercivity and are used to make ferrite magnets."

Purifi claims high DC permeability (B/H) and low AC permeability (B/H). Clearly we see that in the curves above. Additionally, the resistivity is very high ~10^4 ohm-m. This is not an induction heater and eddy current will be low. Then compare these curves to the BH curve of low carbon steel.

The ferrite conducts the flux but not as well as the steel. Note that the Purifi pole is not vented and is solid. Purifi has a serious transducer engineering team but no Patents. This is how respect is earned. Keep an eye on these guys but using ferrite as a core for a coil is really no invention. In fact it's almost trivial but in the lame world of moving coil audio transducers it is clearly innovative. Purifi is "Rock and Roll".

 

[bmc0]

The "PermanentCore" appears to be a Ferrite Core

It's a neodymium slug (source)—

In terms of assembly the only headache is that the pole (a neodymium magnet) needs to be installed after both it and the ferrite ring have been magnetized.

 

[smowry]

Heinrich,


Then all Doug Button's work at JBL was all for naught. Is that your claim?

https://patentimages.storage.googleapis.com/78/a8/36/94f940556815e5/US5748760.pdf

https://patentimages.storage.googleapis.com/9e/b5/ed/fa92c848a170bc/US5042072.pdf

https://patentimages.storage.googleapis.com/6c/71/93/4f1ca7895fd806/US8014555.pdf

https://patentimages.storage.googleapis.com/49/87/c1/e648af0d094386/EP2871856B1.pdf

 

[smowry]

bmc0,

Well their description sure fooled me. Here's the BH curve of a NEO magnet.

The curve is in a different quadrant of axis. Note that H is negative. Certainly the pole can be a Neo magnet but the resistivity is much lower (higher eddy current) and the permeability is that of air (low) u0. However, the following does not describe a Neo magnet.

From Purifi website

PURIFI PermanentCore​

A practical solution that realizes the advantages of "iron-free" designs by replacing the pole piece with a material that has low AC permeability while still allowing for proper conductance of the DC field lines. This is crucial in avoiding Force Factor Modulation (FFM) and producing a grain-free backdrop for music.

The Neo magnet is actually a poor conductor of flux. The Neo magnet is a source not a conductor. A Neo magnet does have low AC permeability but also low DC Permeability. The Permanent designation could mean permanent magnet. But anyway, I will send a message to Carsten and ask him and I will get back to you ASAP.

Sent message!

 

[smowry]

If one were to not charge the Neo slug, it would act as a huge air gap. The Neo magnet is a source and not a conductor. This is not the first time I got lost in the technical rhetoric. At best, Purifi's description is misleading.

 

[Heinrich]

Heinrich,
Then all Doug Button's work at JBL was all for naught. Is that your claim?
https://patentimages.storage.googleapis.com/78/a8/36/94f940556815e5/US5748760.pdf
...

No, but! I've seen not that convincing implementations, e/g the mid/bass-driver of JBL's famous "m2". Alas, I didn't have it at hand, as the non-engineer and non-tech I am. Seems the mechanical probs of surround, spider, cone overweigh the motor's today. That was my aim in pointing to the ol' 2020. Its motor is as good as it gets, judged by the result (a) and (b) human perception.

Once there is only that much you can do, the other day there is only that much you have to do. Hope this is not Dunning/Kruger on my side. Anyway, when using contemporary drivers in DIY speakers I quickly came to limits that were posed by the box itself, radiation pattern and so on. You driver designers are well ahead of the SOTA in that respect. My impression that is.

 

[smowry]

AOR,

Please don't be offended if I am critical of KEF woofers. Every commercial system in my home is KEF. I have the 5 x eggs with sub home theater. I have 2 pairs of Q100's and a pair of Q350's that I removed the drivers for a DIY project. I also purchased a pair of Q3 Meta's and removed the drivers for another DIY project. I am presently awaiting a pair of LS50 Meta drivers (red) to arrive from Hong Kong for yet another DIY project. I love KEF and KEF's engineers are great folks especially Jack. However, we disagree on how a woofer should be designed. Your coaxials are something special to me.

 

[smowry]

Here's an alternative transducer topology that got me excited when it was introduced.

https://pearl-hifi.com/06_Lit_Archive/14_Books_Tech_Papers/Mowry_Steve/Bucks_Induction_Driver.pdf

 

[AOR]

Outstanding, AOR!
https://patents.google.com/patent/EP3634013B1/en?oq=EP3634013B1

View attachment 439412

Well there are four Neo rings but the voice coil is underhung and the spider is tiny. This one may have low small signal distortion but Xmax will be very small. It looks like the assembly can be pin gauged from the bottom. This is more like a midrange driver rather than a woofer. Borresen seems to use midranges for their woofers but they have the correct power cords and interconnectors.

This motor assembly looks like my tweeter motor from 2008 - 2009 just with a taller gap. Michael Borresen has the balls to claim an invention. Has he no shame?

View attachment 439415
https://pearl-hifi.com/06_Lit_Archi.../Mowry_Steve/Air_Core_Tweeter_Magnet_Assy.pdf

I think the Xmax (xBL) is not too bad. I went down a bit of a rabbit hole of simulating this approach, and the resulting B(x) profile is quite similar to the Adire XBL2, with a double hump. This means that with a correctly sized voice coil, it can have appreciable excursion capability.
The problem that I found was magnet efficiency. In order to get a usable BL, the quantity of Neo was enormous. Maybe this explains why these motors are included in Børresen's >€100,000 speakers...

 

[Scrivs]

I seriously thought I'd stumbled upon a troll-- designs named "steal-us," and Phallus. Great.

 

[smowry]

STEALLUS required ~10x the volume of NdFeB of a typical motor assembly in FEA simulations.