Jump to content



Sign up for a free account to remove the pop-up ads

Signing up for an account is fast and free. As a member you can join in the conversation, enter contests and remove the pop-up ads that guests get. Click here to create your free account.

Photo
- - - - -

Reverse engineering XBL2


This topic has been archived. This means that you cannot reply to this topic.
33 replies to this topic

#1 of 34 OFFLINE   Bill Fagal

Bill Fagal

    Stunt Coordinator



  • 166 posts
  • Join Date: Jul 08 2002

Posted April 23 2003 - 02:32 AM

Board seems kind of dead, so I thought I'd have some fun with a little speculation.

Anyone besides me still intrigued with Adire's XBL2 technology? Now that XBL2 drivers have been shipping for some time, they won't mind if we indulge in a little reverse engineering, will they? Posted Image

So for the sake of public edification, anyone willing to take their Brahma/Tumult down to their local metalshop and bandsaw the motor in half so we can see a cross section? Didn't think so.

Well, short of that, here's a sketch of my guess at the innards of a Tumult XBL2 motor. The brilliance of this topology is how the voice coil traverses between two gaps, with special attention paid to keeping the fringing fields symmetrical and the VC winding hight optimized for a constant BL over the transition.

http://image1ex.vill....sp?id_=2297319

Any thoughts?

If I have some time over lunch, I'll try to do a quick finite element model of the field density distribution and post it.

Bill

(Edit: Switched to a link to save bandwidth.)

#2 of 34 OFFLINE   Seth_L

Seth_L

    Screenwriter



  • 1,553 posts
  • Join Date: Apr 05 2002

Posted April 23 2003 - 03:17 AM

Bill,

I think you have too much free time at your job. Posted Image

Seth

#3 of 34 OFFLINE   Bill Fagal

Bill Fagal

    Stunt Coordinator



  • 166 posts
  • Join Date: Jul 08 2002

Posted April 23 2003 - 03:31 AM

Shhhhh... I think my boss might have heard you. :wink:

#4 of 34 OFFLINE   Bill Fagal

Bill Fagal

    Stunt Coordinator



  • 166 posts
  • Join Date: Jul 08 2002

Posted April 23 2003 - 05:41 AM

Alright, here's a finite element flux density plot. I roughly modeled the outer dimensions based on the info on Adire's website. I specd 1006 steel pole/top-plate and grade-8 ceramic magnets. (I don't know if they use grade-8 or 5, so I gave them the benefit of the doubt--certainly enough to saturate the pole.)

I'm just guessing at the top-plate geometry, but a really scooped-out channel between the gaps seems to model best, and saves weight to boot.

Take the actual values with a grain of salt since there are so many unknown variables (like, for example, whether or not this is actually XBL2 at all. Posted Image)

http://image1ex.vill....sp?id_=2299011

Notice the symmetry of the fringing fields and negligible flux between the gaps. I believe these are crucial to the design to keep BL constant from inward VC excursion, across the gap crossover zone, and through outward excursion.

These twin gaps allow the VC to completely traverse a magnetic gap with each half cycle, shortening the winding hight necessary to achieve a given excursion. Brilliant!

I don't know if anyone else out there has any interest in this stuff, but driver design is what I do for fun, so I find this line of conjecture facinating.

Bill

Edit: Switched to a link--bandwidth, ya know..

#5 of 34 OFFLINE   Steven Kephart

Steven Kephart

    Agent



  • 33 posts
  • Join Date: Oct 01 2002

Posted April 23 2003 - 05:53 AM

Actually I've been involved in a speculation discussion similar to this that spanned two forums. Maybe our thoughts might give you more information. Here's the links:
http://www.teamamp.c....t=5344&hl=xbl2
We continued the discussion on Acoustic Visions's forum in the car audio section. Unfortunately it's not there anymore, so it must have been erased. Maybe we were closer than we thought.

#6 of 34 OFFLINE   Bill Fagal

Bill Fagal

    Stunt Coordinator



  • 166 posts
  • Join Date: Jul 08 2002

Posted April 23 2003 - 07:00 AM

Interesting link, Steven--Thanks!

The link back to this forum was especially enlightening. I hadn't seen that discussion before.

Quote:
We continued the discussion on Acoustic Visions's forum in the car audio section. Unfortunately it's not there anymore, so it must have been erased. Maybe we were closer than we thought.


Posted Image Perhaps so. Well, I'm not out to steal Adire's lunch. I first arrived at my abovementioned conclusions almost a year ago while the Tumult specs were getting hashed out on this forum (my first visits here). I enjoyed the challenge of trying to turn the scattered clues into a topology that matched published dimensional and performance specs, and I believe I succeeded.

I've been sitting on this idea since then, not wanting to be a thorn in Adire's side. Of couse, by now I expect that all the industry types who have reason to be curious have already thoroughly disected XBL2 motors, so I feel my speculations won't cause harm.

Of course, if this thread disappears, I guess I'll have some idea of why!

Bill

#7 of 34 OFFLINE   Kyle Richardson

Kyle Richardson

    Screenwriter



  • 1,076 posts
  • Join Date: Jan 01 1998

Posted April 23 2003 - 07:18 AM

It wasnt deleted! I think this was the discussion you were thinking of Steven.
Kyle Richardson
Acoustic Visions

AOL Instant Messenger Name: kyler70
MSN IM: acousticvisions@hotmail.com

#8 of 34 OFFLINE   Steven Kephart

Steven Kephart

    Agent



  • 33 posts
  • Join Date: Oct 01 2002

Posted April 23 2003 - 08:26 AM

Thanks Kyle, that was exactly it. I looked throught that section several times and couldn't find it. How odd. Thanks again for the help. I really appreciate it.

#9 of 34 OFFLINE   Ryan Schnacke

Ryan Schnacke

    Supporting Actor



  • 877 posts
  • Join Date: Feb 05 2001

Posted April 23 2003 - 04:13 PM

I don't think Adire will consider this kind of discussion as threatening. They know that once the patent is approved then it will be made available for anyone and everyone to read. And their rights are already fairly protected since the patent is pending.

#10 of 34 OFFLINE   Bill Fagal

Bill Fagal

    Stunt Coordinator



  • 166 posts
  • Join Date: Jul 08 2002

Posted April 24 2003 - 07:22 AM

Since I think XBL2 relies on twin gaps, each used for 1/2 a cycle, The design makes less efficient use of the combined field strength than a garden-variety overhung motor (wasting the flux lost across whatever gap doesn't have VC windings in it at a given excursion point). Of course, this is a necessary compromise to extend Xmag with a relatively short coil.

This would seem to mean that, to keep efficiency up, they'd want to spec a big magnet.

In my FEA output, notice that the polepiece seems to be the flux bottleneck, reaching saturation first, (unless the gaps are shorter/narrower than I specd, which they very well could be).

Looking into my crystal ball, I'm thinking the next XBL2 extreme sub may feature a 4" VC. Fattening the pole piece from 3" to 4" seems an obvious step if you want to boost the max total flux in the motor to allow for even more Xmag at a given efficiency.

Another nice thing about XBL2's short coil is that they can probably move from a 3" to a 4" VC without getting giant Le numbers.

One downside of a new 4" XBL2 super-excursion motor: It'll probably weigh more than your little brother!

Chubbiness is pretty much an unavoidable fact if you want to carry a lot of flux a long way through a magnetic return circuit. I'm working on a concept driver with a 70mm Xmag, and just the steel in the magnetic circuit calculates to about 80lbs--and that's largely flux-saturated!

Anyway, It'll be fun to see where Adire takes XBL2 from here. But remember--you heard the 4" VC prediction here first! Posted Image

Bill

#11 of 34 OFFLINE   eric

eric

    Agent



  • 42 posts
  • Join Date: Apr 16 2002

Posted May 09 2003 - 11:05 AM

i had a look at what the xbl2 motor looks like from one that got destroyed and was taken apart (spider, former, and cone removed). i dont clearly recall what it looks like tho--and if i did, i wouldnt be able to disclose it anyhow...i dont know about how flux flows or any of that stuff, and i would have loved to have seen a 1/2 brahma, but it didnt happen.

anyhow, my idea was a sort of dual gap design (this was before seeing the brahma gutted), tho it was contoured instead of being squared off as you suggest. the biggest problem from how i pictured it was that it would requite magnets above and below the tp (did i mention i dont understand flux Posted Image ).

eric
nismo240sx on soundillusions.net
nismo240sx on maxima.org
nismo240sx @ EVERYWHERE ELSE...but they wont let me here

#12 of 34 OFFLINE   Bill Fagal

Bill Fagal

    Stunt Coordinator



  • 166 posts
  • Join Date: Jul 08 2002

Posted May 13 2003 - 12:35 AM

Hi Eric,

Quote:
i dont clearly recall what it looks like tho--and if i did, i wouldnt be able to disclose it anyhow...


Did you sign an NDA or something? Spill! Spill! Posted Image

Quote:
the biggest problem from how i pictured it was that it would requite magnets above and below the tp (did i mention i dont understand flux ).


Yeah, I played around with the idea of multiple magnets/tps until I modeled it in my FEA program and got no joy.

#13 of 34 OFFLINE   Steven Kephart

Steven Kephart

    Agent



  • 33 posts
  • Join Date: Oct 01 2002

Posted May 13 2003 - 05:23 AM

Quote:
I first arrived at my abovementioned conclusions almost a year ago while the Tumult specs were getting hashed out on this forum (my first visits here). I enjoyed the challenge of trying to turn the scattered clues into a topology that matched published dimensional and performance specs, and I believe I succeeded.

Dan has recently told me how it works, so I can't really comment on this anymore. But I will say that you will be surprised at it's simplicity and logic.

Anyway, for more info, I thought maybe another tech thread would help you piece more together. At least, Mark Seaton was able to figure it out early on in this thread. Here's the link:
http://www.carsound.....;f=12;t=007641

I'll warn you though, it's a LONG read. Especially at 40 posts a page and 16 pages. Plus you know how long winded Dan can be in tech threads when he argues with someone. Posted Image

#14 of 34 OFFLINE   Bill Fagal

Bill Fagal

    Stunt Coordinator



  • 166 posts
  • Join Date: Jul 08 2002

Posted May 19 2003 - 02:32 AM

All right folks,

Not to toot my own horn or nuthin' but I now have it on *very* good authority that my above guesses about XBL2 are correct...in principle if not in detail.

For the next chapter of this discussion, I thought I'd share some more FEA results, this time BL over excursion. I spent waaay too much time on this over the weekend, so I hope you guys appreciate it! Posted Image

I began by trying to optimize the coil geometry in an effort to replicate the unique XBL2 BL curve.

All plots run from the bottom of the stroke, through dead center, to the top of the stroke. They're a tad rough, containing only 40 data points (hey, it was the weekend!). X axis is tenths of an inch, Y axis is BL (derived from my Tumult motor sim which does not necessarily reflect an actual Tumult's field strength, etc.)

This was one of my first attempts. Ugly, eh? The coil was too short, ergo the BL dip in the segment corresponding to the rest position where the coil is straddling the two gaps.

My next shot landed a lot closer. It's surprising what a difference 4/100" of coil height makes in XBL2. The coil height is now essentially optimized. Any longer and the curve starts to dome up. However, there's still a small problem. On the inward excursion over the lower gap, the BL peaks a little higher than on the outward excursion as the VC cuts through a little more flux from stray lines choosing the shortest path back home.

I tried shaving off a couple hundredths from the lower plate thickness, subtly shortening the gap height. Bingo. Now the BL is essentially equal front-to-back.

I didn't ever see an exact replication of the distinctive 3-hump BL curve Adire publishes in their XBL2 paper. I suspect it's because my model has a deeper cut in the top plate than Adire uses, therefor less fringe flux between the gaps.

Bill

#15 of 34 OFFLINE   Steven Kephart

Steven Kephart

    Agent



  • 33 posts
  • Join Date: Oct 01 2002

Posted May 19 2003 - 06:19 AM

Those curves look good. I am trying to reply without breaking my oath to Adire about not telling how it works. So I think I can comment on two problems I can see without there being a problem.

First of all, I think Adire has a little more control over the curve than you do in your design. Maybe you could draw a picture of you design so I can make sure I am picturing it correctly?

And second, you forgot the corn pup. Posted Image

#16 of 34 OFFLINE   Bill Fagal

Bill Fagal

    Stunt Coordinator



  • 166 posts
  • Join Date: Jul 08 2002

Posted May 19 2003 - 06:44 AM

Hi Steven,

For a rough sketch, see my first post.

In my second post, you can see a snapshot of the model I drew up. That's the model I used to generate the BL curves. I just added a voice coil straddling the gaps and moved it through 40 1/10" steps of excursion from bottom to top.

I guess I don't understand what you mean by "control over the curve." Can you be more specific within the bounds of integrity? My curve has less ripple in it than the one Adire publishes, IIRC. Of course, there is quite a delta of time/money/R&D between me and Adire. I mean, they're Adire Audio, fer cryin' out loud, and I'm just a DIY audio goofball who had a Sunday morning to kill (though a keen-minded, visionary, and ruggedly handsome goofball, I hasten to add.) Posted Image

Of course, I'm sure they chose their curve with a certain performance/pricepoint objective in mind and with an eye to how it would interface with the suspension compliance curve to contribute to the performance of the total system. Now, maybe if I'd give *all* Sunday to it...ya think? :b

Oh, and a cornpup. (You guys at the Acoustic Visions board are *out there*, man!)

PS. I apologize in advance for the fact that my (free) host bandwidth for the day is already half gone. Ya'll will soon be seeing some sort of "bandwidth exceeded" message instead of pics.

#17 of 34 OFFLINE   Steven Kephart

Steven Kephart

    Agent



  • 33 posts
  • Join Date: Oct 01 2002

Posted May 19 2003 - 10:41 AM

Quote:
In my second post, you can see a snapshot of the model I drew up. That's the model I used to generate the BL curves. I just added a voice coil straddling the gaps and moved it through 40 1/10" steps of excursion from bottom to top.

Thanks. I just wanted to verify that I was picturing what you are doing correctly. I was.

[quote]
I guess I don't understand what you mean by "control over the curve." Can you be more specific within the bounds of integrity? My curve has less ripple in it than the one Adire publishes, IIRC.
[quote]

I really can't explain further. Once you see what they are doing, you will understand. And yes, it does look like your curve has less ripple. Are you keeping all parameters the same with your changes? Maybe they added more ripple as a compromise for something? I know one reason is for more excursion. But you have them beat with your model. So it will be interesting to see what the difference is. Maybe it's because your using a higher BL since you are roughly using the Tumult's numbers, and the DUMAX report is for the Brahma.


Quote:
Of course, there is quite a delta of time/money/R&D between me and Adire. I mean, they're Adire Audio, fer cryin' out loud, and I'm just a DIY audio goofball who had a Sunday morning to kill (though a keen-minded, visionary, and ruggedly handsome goofball, I hasten to add.)

hehe. I agree with most of your above statement. I can't comment on the looks though. Posted Image You definitely have impressed me with your work thus far.

Quote:
Oh, and a cornpup. (You guys at the Acoustic Visions board are *out there*, man!)

I'm not sure how that started. But I will say that I once had a customer tell me that he gained huge SPL with his subs when he took the "SPL limiters" out of his subs. At least the cornpup is an actual item. Posted Image

#18 of 34 OFFLINE   Bill Fagal

Bill Fagal

    Stunt Coordinator



  • 166 posts
  • Join Date: Jul 08 2002

Posted May 20 2003 - 03:24 AM

Quote:
And yes, it does look like your curve has less ripple.


As I mentioned, I suspect the difference may have to do with the depth of the channel in the top plate. If the cut was shallower than my model, more flux would cross beween the gaps and you'd have to shorten the coil to keep the curve flat. I suspect you'd begin to see a center hump flanked by two side humps that correspond to the coil centering on the upper and lower gaps. I also suspect that one of the compromises of a shallower channel is more ripple in the curve.

Quote:
Are you keeping all parameters the same with your changes?

The only parameter I'm changing is coil height. This, of course, affects average BL.

Quote:
Maybe they added more ripple as a compromise for something? I know one reason is for more excursion. But you have them beat with your model.

I optimized my model strictly for BL flatness. I don't know what kind of compliance curve Adire had to superimpose on their BL curve. The result may end up looking more linear than their raw BL curve.

Quote:
Maybe it's because your using a higher BL since you are roughly using the Tumult's numbers, and the DUMAX report is for the Brahma.


Since all I had to go on for my model was the external motor dimensions, many of the quantities in my model were wild guesses. Looking at the numbers I got, I agree that I'm probably using more magnet than an actual Tumult has.

Quote:
I'm not sure how that started. But I will say that I once had a customer tell me that he gained huge SPL with his subs when he took the "SPL limiters" out of his subs. At least the cornpup is an actual item.


Don'tcha just love all the imaginary voodoo people dream up about audio? By comparison, some of it makes a cornpup seem like a reasonable tweak. I'll add one to my next model, if I can track down its pertinent properties. Anybody know the B-H curve, coercivity, or electrical conductivity of a cornpup? Posted Image

Bill

#19 of 34 OFFLINE   Steven Kephart

Steven Kephart

    Agent



  • 33 posts
  • Join Date: Oct 01 2002

Posted May 20 2003 - 07:39 AM

Quote:
I'll add one to my next model, if I can track down its pertinent properties. Anybody know the B-H curve, coercivity, or electrical conductivity of a cornpup?

If my calculations are correct, it is 2.

But I think if you can make one without MSG, then that number will improve.

#20 of 34 OFFLINE   Bill Fagal

Bill Fagal

    Stunt Coordinator



  • 166 posts
  • Join Date: Jul 08 2002

Posted May 20 2003 - 09:25 AM

Posted Image


Back to Members' Theaters, HT Projects, Remotes, Seating, Accessories & Lighting



Forum Nav Content I Follow