Ironless Motor Assemblies and STEALLUS

[smowry]

Hi Lars,

Where can we find a typical Bl(x) curve of a Purifi woofer?

I was specific that asymmetry was related to Xdc. I agree that linearity relates to the shape of the Bl(x) curve. Dr. Klippel uses the term "Symmetrical Nonlinearity". However, both examples are related to distortion in the transducer output. I even referenced my post with a link to the Klippel poster. Furthermore, when a Bl(x) curve was posted without showing the displacement limits, my claim was extraordinarily linear without knowing if Xdc was present. My comment was it looked like a line. I don't see any misconception(s) within my post. Did I miss something?

So we can see the shape of the Bl(x) curve on the top plot is symmetrical but highly nonlinear; however, in the lower Bl(x) plot the curve is more linear but asymmetrical and with Xdc. I agree that DC Offset and Bl(x) nonlinearity are different animals; however, the primary cause of Bl(x) asymmetry is a motor's asymmetric geometry. It could also be related to the placement of the voice coil's zero position during assembly.

"If we use the definition of Xmax as Bl(Xmax) = 0.7Bl(0.0) @ 10% THD, this would be at least moderately nonlinear without any significant flat plateau about x = 0. What I consider equally important is the symmetry of the even Bl(x) function about x = 0. Any offset about x = 0 is what Dr. Klippel calls Xdc, (DC Offset)."

 

[Lars Risbo]

I said common misconception and I was not meaning your posts

what Klippel calls Xdc is the mean excursion (ie DC value ) when you drive sinusoidal voltage on the driver in free air. Xdc is frequency dependent and zero if Bl(x) is symmetric. Xdc is of course problematic since is eats up the linear Bl and suspension compliance range. the suspension and box air can also cause Xdc.

 

[smowry]

Patrick used his DA2 to acquire the Bl(x) parameter of the Purifi 10 in woofer. Below is what was published by Brother Vance including his comments.

Really nice Bl(x) curve but there is still ~1.0mm of DC Offset as indicated in the lower symmetry plot.

 

[thewas]

Where can we find a typical Bl(x) curve of a Purifi woofer?

Test Bench: The Purifi Audio PTT8.0X04-NAB-02 8” Aluminum Cone Woofer

The PTT8.0X04-NAB-02 is the fifth Purifi woofer to be examined in Test Bench. The new 8" aluminum cone woofer includes the same constant Sd surround, cone edge stiffening ring, and low magnetic hysteresis motor design as previous models. The addition of the unique Purifi NBR surround, make the...

audioxpress.com

Test Bench: Purifi Audio PTT10.0X04-NAB-01 10” Midbass Woofer

The eagerly awaited Purifi Audio PTT10.0X04-NAB-01 10" Midbass Woofer is the sixth Purifi driver to be examined in Test Bench. This 4Ω woofer features extensive proprietary technology that includes the constant Sd surround, cone edge stiffening ring, and low magnetic hysteresis motor design. The...

audioxpress.com

Purifi Audio PTT4.0W04-01A 4 Inch Midwoofer Review

Purifi Audio PTT4.0W04-01A 4 Inch Midwoofer Review

www.erinsaudiocorner.com

Purifi Audio PTT6.5W04-01A 6.5 Inch Woofer Review

Purifi Audio PTT6.5W04-01A 6.5 Inch Woofer Review

www.erinsaudiocorner.com

Possibly even more can be found than from my 30 seconds googling.

 

[smowry]

Well Lars, the only person to mention the word symmetry in this thread was I. You could have made your point without the generalization of common misconception. I feel that you implied that I had misconception whether meaning to do so or not.

 

[smowry]

Then after all is said and done, the primary cause of Xdc is the suspension and can be observed in Kms(x) plots. Without FEA nonlinear simulation is almost impossible to make the suspension reaction force symmetrical and yes there is hysteresis of the reaction force. Anyway there's another source of DC Offset, Dr. Klippel's poster presents this as Le(i). That is what I used to call Bl(i) or induced AC Bl; however the effect of Bl(i) on DC Offset is small. In my VF OPERA AC simulations, I could create several cases with different voice coil positions at different frequencies. I could then acquire the AC flux linkage to the voice coil. I had written a command file that applied a weighted average to the Bl(i) acquisitions. Then conveniently, using those AC models, I could subsequently run the thermal solver. I purchased my perpetual VF OPERA license in 2002 after I left P.Audio and decided to go it alone!

A way to eliminate DC Offset from Kms(x) is to utilize complimentary toroidal surrounds and 2 x complimentary spiders. Actually, one could possibly end up with symmetrical suspension without any FEA due to symmetrical suspension geometry. I presented this concept to Jack at KEFF more than a decade ago. He said he liked it but I never saw any indication in KEF products. KEF has 2 x spider in their Blade woofer but they are not complementary. The only reason for this that I can identify, is that its easier that way.

 

[Lars Risbo]

Patrick used his DA2 to acquire the Bl(x) parameter of the Purifi 10 in woofer. Below is what was published by Brother Vance including his comments.

View attachment 440339
Really nice Bl(x) curve but there is still ~1.0mm of DC Offset as indicated in the lower symmetry plot.

we also have a Klippel LSI and it does not measure Bl exactly and particular challenged when Bl is very flat. LSI does unlike our analyser not use the laser during the LSI. we use our own analyser and a first principles setup that moves the coil at constant speed whilst recording induced coil voltage to verify the Bl curve.

all our data sheets show Bl graphs.

note that Voice Coil mixed up some Bl graphs so one of them is not for our driver

 

[smowry]

Lars,

I have been exceptionally complimentary of Purifi within this thread. Kindly let me quote myself.

"..... 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....."

Please note that my post received a "Like" from Amir and that has significance to me. Amir operates at the highest standards and I respect him for that.

 

[Lars Risbo]

Lars,

I have been exceptionally complimentary of Purifi within this thread. Kindly let me quote myself.

"..... 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....."

Please note that my post received a "Like" from Amir and that has significance to me. Amir operates at the highest standards and I respect him for that.

thank you Steve. All designs are compromises based on priorities on what to fix first etc. Feel free to point out the flip sides of what we do. This is how science operates.

 

[Anton S]

Firstly, the curve is extraordinarily linear; it looks like a line. However, something seems amiss here. The acquisition is hosed because the displacement limits were improperly set by the individual operating the Klippel DA. You indicated Xmax = 6.0mm when in fact this curve does not identify Xmax. Frankly, I have never seen a Bl(x) curve like this one, either from experience or from anything published. Having said that I must question the validity of the acquisition and believe that the curve above is controversial!

Suppose you post some information on the transducer that was measured. The curve is atypical and this implies that the transducer is atypical.

The driver is a proprietary design used exclusively in a single speaker model and unavailable separately on the retail market. Looking back at the manufacturer's white paper, I see that I had incorrectly cited an Xmax of 6mm for this driver. The manufacturer claims a 6mm Xmax for the older driver (black trace), but an 8mm Xmax for the new driver (the red trace in the graph above). The woofer is a 10"x6" racetrack design with a polymer/carbon composite cone and cast aluminum frame. I have seen no other details regarding its materials or construction. However, the manufacturer claims extensive use of the Klippel R&D system, and that its engineers received their training in its use at Klippel's headquarters in Germany.

The same white paper containing the graph above also included a graph of the driver's harmonic distortion (below). As indicated on the chart, the magenta trace was the measured THD, with the black trace being the contribution of the 2nd harmonic and the blue trace being that of the 3rd. Independent testing by the Canadian NRC of the complete speaker system appears to support this level of performance.

 

[smowry]

Anton S,

1. Who is the manufacturer?
2. What is the model?
3. What was the input for the above THD acquisition? Was it 1.0 W? Was it 10^-3 W?
4. Can you please share the "white paper"?
5. Can you post a picture of the transducer?
6. Can you post a link to the manufacturer's website.

 

[Anton S]

Anton S,

1. Who is the manufacturer?
2. What is the model?
3. What was the input for the above THD acquisition? Was it 1.0 W? Was it 10^-3 W?
4. Can you please share the "white paper"?
5. Can you post a picture of the transducer?
6. Can you post a link to the manufacturer's website.

1 & 2. Irrelevant with respect to an opinion regarding the quality of the driver's Bl (X) performance, which was my question.
3. 1.0 W.
4. Same as 1 & 2.
5. Same as 1 & 2.
6. Same as 1 & 2.

All I wanted was an honest assessment of the Bl (X) curve that I posted. I wasn't expecting the Spanish Inquisition. To my eye, the Bl (X) curve that I posted appears nearly identical to the one above for the Purifi driver, so it is certainly not unique. Since the Purifi driver is highly regarded, I'll consider it good. Thanks.

 

[bmc0]

1. Who is the manufacturer?
2. What is the model?

Based on the description given, I'm thinking that the driver in question might be the subwoofer used in the Definitive Technology Mythos ST-L.

 

[olieb]

All I wanted was an honest assessment of the Bl (X) curve that I posted.

Well, with all that cryptic non-information, this one can say:
The curve is quite flat, end of story, but this has been said already.

Whether that is a good woofer depends on how the curve continues outside of what is shown and on other stuff.
If Xmax is actually 8mm, the suspension seems not to keep up with the motor. And if 1W (about 1A) produces 4% distortion at 40Hz, distortion is not really that good (for a 10").
Whether that is a good driver for mids depends on other stuff that is not shown.

 

[NTK]

@Anton S, if you are looking for some baseline Bl(x) curves as references, here are 3 curves from Klippel's paper:

https://www.klippel.de/fileadmin/klippel/Files/Know_How/Literature/Papers/Loudspeaker%20Nonlinearities_Causes%2CParameters%2CSymptoms_06.pdf

 

[Anton S]

@NTK

Thank you very much.

 

[smowry]

As I indicated, I have not practiced in more than 10 years. So I did misstate the guidelines for determining Xmax from the large signal parameters,

Bl(Xmax) = 0.82Bl(max)

Cms(Xmax) = 0.75Cms(max) where Cms = 1/Kms

There are also displacement limits for Doppler and Le(x,i) but it's unlikely that one would design a transducer whose displacement was limited by Doppler and/or Le before Bl or Cms gave up.

So in theory, one could have a flat Bl(x) curve, Bl(Xmax) ~ Bl(0.0) and Xmax would be limited by Cms(x) = 1/Kms(x). I just have never seen this nor do I see it in any of Dr. Klippel's examples.

 

[smowry]

Michael Børresen was on LinkedIn puffing and complaining about patent offices.
https://www.linkedin.com/feed/update/urn:li:activity:7313594732541050881/

Then I added my comment below his post.

Respect must be earned. This guy is a joke! I can add this to my lists of stories.
https://pearl-hifi.com/06_Lit_Archi.../Mowry_Steve/Air_Core_Tweeter_Magnet_Assy.pdf

Voice Coil is an industry wide publication and it's free. What is the probability that he reads Voice Coil? Please wait while I integrate.

Bob Carver said, make the magnets 30% thinner.

 

[olieb]

So in theory, one could have a flat Bl(x) curve, Bl(Xmax) ~ Bl(0.0) and Xmax would be limited by Cms(x) = 1/Kms(x). I just have never seen this nor do I see it in any of Dr. Klippel's examples.

Maybe I do not understand you fully and I am lost in translation, but it is not uncommon that CMS is the limiting factor for Xmax. I just checked and this are the very first two examples of Klippel analysis I could find. In both cases CMS is the limiting factor.

Test Bench: The E180HE-44 Full-Range Subwoofer from the Dayton Audio Epique Line

In this article, Vance Dickason characterizes the E180HE-44 woofer, from the Dayton Audio Epique lineup. As with the E150HE from the same range, the E180HE-44 uses the patented Multiple Magnet Air Gap (MMAG) technology, invented by Enrique Stiles and patented by Step Technologies. The E180HE is...

audioxpress.com

Test Bench-Purifi Audio’s PTT5.25X04-NFA-01 Midbass Woofer

For this Test Bench, Vance Dickason focused his full attention on another innovative driver from Purifi Audio, this time the PTT5.25X04-NFA-01 5.25” Midbass Woofer. At the core, this Purifi midbass woofer is pretty much identical to the PTT6.5X and PTT4.0X, and includes the same constant Sd...

audioxpress.com

 

[smowry]

I strongly agree. At one time Dr. Klippel was of the opinion that the surround should be the displacement limiting component. I disagreed then and I disagree now. In a robust low frequency transducer design the spider(s) should be the limiting component. The surround should not be allowed to pull tight. This is primarily related to reliability. Surround lift is a common failure mode.

Cms(x)/Kms(x) vary with position but they do not vary directly with current, i (A). Motor force varies with both position and current, F(x) = iBl(x) N. So in the limit, as
iBl >> xKms(x) N and x > Xmax, then we are looking at an unstable system. The red curve below is effectively constant! Perhaps I should have said Cms(x) better limit displacement or else! Without additional information, one must consider such a system potentially unstable.

Was Anton S attempting to bait me with his cryptic question? So is that a "good" Bl(x) curve?