I wonder whether it is something superior with AlNiCo or whether it is just weak enough that some cone anomaly’s are not stimulated. Does anyone have a theoretical or measurable explanation for this?
The advantage of AlNiCo is its high electrical conductivity and therefore lower levels of flux modulation that result. Ferrite and 'rare earth' magnet motors normally employ copper rings so that induced eddys can flow in a low resistance and act to essentially cancel the modulaton from the coil current. AlNiCo might be considered "self-shielding", although this is of course sometimes outweighted by the major disadvantage of AlNiCo to be relatively easily demagnetized.
The move from AlNiCo to ferrites was not for technical reasons, but for economic necessity as cobalt became very hard to come by. I suspect much of the "love" for AlNiCo is also born of "a young life gone by". But the technical advantage is still real, measurable and audible when compared to many modern drive units.
There is also a diffence to highlight between flux modulation that produces essentially second order harmonics and the linked effect caused by eddy currents induced in the pole pieces that produces third order harmonics - and (IMO) the latter is far more detrimental to the quality of reproduction. Quite possibly too is that the second order distortion might lend itself to a "pleasant sounding distortion", but I am unaware of any tests that have been done to assess this.
And whilst rare earth types improve upon the characteristics of ferrites, as they are magnetically "stiffer" and therefore produce less flux modulation, copper is still prevalent to reduce the audible effects. One big and often over-looked advantage remains with ferrites, however, in that they generally require large, flat motor components that also provide a large area for heatsinking (rather dissipation): The audible reduction in thermal modulation in bass channels is something I referred to earlier.
Hmm.
Factors I would consider include:
1. Relative permeability: neo's can have close to free space value, which means the magnet volume is the dominant reluctance in the circuit, whereas a higher permeability will make the gap dominate.
2. Permeability curve..how tight is the flux control given audio frequency push.
3. VC hang, over vs under. Overhung adds an additional non linearity during high excursion, as the coupling to the actual magnet increases as the VC drives back into the magnetic structure.
I wonder how much the love of alinco is just romantic memories vs actual. I've never worried about it nor tried to examine the issue, so have no firsthand audible experience to say one way or the other. I do suspect that as speaker tech advanced, everything is pushed harder, from VC epoxy to flux density, to excursion velocities.. So wonder how much is due to physical changes vs magnetic material.
If one is worried about the electrical conductivity of the magnet, just use 8 or 10 magnets in parallel, each clad in it's own copper sleeve. That will restrict hf signals from even touching the magnets without the non linearity of the shorting sleeve impacting eddy losses and flux dragging.
jn
Factors I would consider include:
1. Relative permeability: neo's can have close to free space value, which means the magnet volume is the dominant reluctance in the circuit, whereas a higher permeability will make the gap dominate.
2. Permeability curve..how tight is the flux control given audio frequency push.
3. VC hang, over vs under. Overhung adds an additional non linearity during high excursion, as the coupling to the actual magnet increases as the VC drives back into the magnetic structure.
I wonder how much the love of alinco is just romantic memories vs actual. I've never worried about it nor tried to examine the issue, so have no firsthand audible experience to say one way or the other. I do suspect that as speaker tech advanced, everything is pushed harder, from VC epoxy to flux density, to excursion velocities.. So wonder how much is due to physical changes vs magnetic material.
If one is worried about the electrical conductivity of the magnet, just use 8 or 10 magnets in parallel, each clad in it's own copper sleeve. That will restrict hf signals from even touching the magnets without the non linearity of the shorting sleeve impacting eddy losses and flux dragging.
jn
Real and measurable, I would agree, but audible has never been shown to my knowledge.
I asked you before and I'll ask again, where is it theoretically shown that eddy currents create a nonlinearity? From Vanderkooy's work this is not at all clear.
Lidia and I did a study on the audibility of nonlinearity more than 15 years ago. You should catch up on the current thinking in this regard. These low order nonlinearities are not highly audible and do often come across as "prefered" - which was shown in our work.
And yet no reference was given. Was this just a personal test or was it actually done scientifically?
I agree here, but would add that no real loudspeaker system designer is going to use a driver that does not have some kind of shorting ring. These magnetic effects are all well known, just as their solutions are as well. People who are concerned about them just do not seem to understand where the state-of-the-art is today.
With magnetic modeling, as I am sure you are aware, one can model and correct virtually all of these magnetic effects. So yes, the love of ALNICO is most likely "romantic memories".
The shorting ring is there to try exclude time varying flux. from the magnetic circuit. I see some consider it "done" when the vc inductance is symmetric through excursion.
My approach would be to make the pole and front plate non conductive in the frequencies of interest. Laminated front plate with profiling in the gap as well to linearize the flux. Pole piece, I'd either wire edm, or a ferrite composite, perhaps a layer of it over turned down pole piece. In addition, the pole piece can be profiled to further linearize the flux. This is bog standard for making high accuracy accelerator magnets, so is not exactly new ground for me.
I still see many writeups that are inaccurate. Still, the thinking is heading in the right direction.
The DC field stuff, the modelling is very good. As frequency climbs, the modelling gets significantly more complex. The big (expensive) modelling packages are pretty good, I'm not familiar with SOTA for standard industry/home use so don't know if they've caught up to the packages needed for particle accelerator magnetic design.
john
not going to argue, but i said 'much like' (similar), not equivalent. if you have more than 1 person on the trampoline, each with different mass, the situation is not so linear.
We are not arguing, we are having a discussion.
Actually, it is still quite linear until the point where the suspension elements leave their linear range of operation. The fact that it is a 3 dimensional multi-body problem does not make it non linear until the bodies impact, assuming they impact non linearly.
For a simple ideal speaker, the device can be modeled as a simple linear system, and much good work can come of it. But as we become more discerning, the simple linear models begin to fall apart, continued use of simple linearity will not allow the model to give accurate results.
A simple example is that of a two tone stimulus to a woofer. a LF high excursion drive taking the vc to near xmax with a second signal at 500 hz also riding it, you have the added complexity of the 500 hz energized vc being dragged through a metal tube by the LF. As the vc moves in one direction, the eddies of the 500 hz will be non symmetrical, the dissipation loss of the 500 hz will be asymmetric. And that's only considering the copper shorting sleeve, never-mind additional proximity based vc wire current distribution changes due to the net fields the vc sees.
jn
The designers that I know understand that inductance symmetry is not the goal, it's a non-changing inductance.
I don't follow why you would want the "front plate" (not sure what that is, a loudspeaker has a "top plate".) to be non-conductive. I have done gap profiling to manipulate the flux, but I don't want it flat. I want it to fall off as gradual as possible. This minimizes the higher order nonlinearities at high excursions - think soft clipping in an amp versus hard clipping. We all know that hard clipping sounds awful. A soft clipped amp will actually sound louder and never really sounds bad. Do the same thing with the BL(x).
If by "I still see many writeups that are inaccurate." you mean around here, then I agree. There is a tremendous level of misunderstanding here at DIY, mostly because the professionals don't come here very often. DIY is not nice to them so they leave.
FEA modeling at audio frequencies is fine.
Hi Earl,
this is of course true, but in my understanding - for the reason, that the speaker driver has nonlinear impedance, then even if you feed the driver with considerably linear voltage signal, the current becomes nonlinear (linear voltage/nolinear impedance) and so does the driver acoustical output. I tried to measure something
Current drive of speakers and speaker distortion
We are saying the same thing. Inductance symmetry is non changing vs position. (edit: I did mean flat inductance vs position as being symmetric, although I see the word symmetric on my part was insufficient, sorry)
(I call the front plate the one with the gap opening, back plate is where the pole is attached. The cone is on the front of the speaker for me.) It's semantics really.
However, the use of a shorting ring structure to make it non changing is by definition, not symmetrical. When the vc is pulled towards the back plate (or bottom, depending on what you want to call it), the vc couples more to the magnetic circuit. When it is being pushed away from the back plate, the front of the vc will extend further out of the magnetic circuit. Using a shorting ring to reduce the inductance as the coil pulls in is lowering the inductance by exclusion, whereas at the front, it is by reluctance increase. While an inductance meter can see the same inductance (Ls) despite vc location, the Rs will be higher with the vc pushed in, lower when it is pushed out.
Honestly, the professionals get a lot of it incorrect or partially correct. I was not talking about a website forum.
edit: I also see no reason the pro's would want to brainstorm new tech on a website, as they have to earn a living on this, not give it away. But I also see your point, some posters do make life difficult here.
Me, I do not make speakers for a living. If anything I post is used by a designer of speakers, good for them.
I've not seen the level of capability necessary for what I've mentioned. For example, modelling the difference in Rs vs position and velocity for a vc being dragged by a LF with a mid freq component added on. What is the eddy loss vs time on both sides of center? Remember, the asymmetry itself is non linear, can the FEA package handle that in amongst the rest of the electrodynamic stuff going on.
As to field profile in the gap, that is a design decision. What you choose is on you. And I know you have been very successful in that regard.
Oh, btw..stained glass? Really? I made thousands of red, white, and blue sailboats back in 1976, light catchers..and fix the occasional window or lampshade..Lead, copper foil..it's fun to do. But the only "stock" I have left is the glass cutter..
jn
Last edited:
i would like for someone to prove to me that such an approach is 'better'. if i hear something 'better', i will say so. if it is simply 'different', the same.
i was too much off the top here. i apologize to any and all offended since this is audio and not politics.
when i was in college i found something interesting in the JAES -- the jensen transflex design and copied this article for a later date (future reference).
at this 'later date' i found that some AE had taken this design, improved it, and turned it into a or 'the' premiere SW of the day. imagine this!
so while most of the 'better' systems of 'the day' used the alnico woofers & mids (jbl, altec, maybe more) they didn't use SW systems.
at some point along the way, systems using small cubes and SW boxes poked into corners crept into the market. the large heavy speakers: Paragons, etc. fell along the wayside. WAF became a critical criteria for audio.
due to the cost of alnico, which seemed to be the advertised reason for dropping this material, it was easier to use something else for bandwidth limited systems. also electronics were introduced with a LF response ending around 40hz.
also since may american manufacturers were importing speakers, switching from alnico reduced freight costs. also speaker enclosures dropped particle board & used lighter materials.
then with surround sound introduced, the small box/cube speaker systems with the SW box in the corner gained acceptance in mass marketing.
so part of the move from alnico, part of the change in audio is due to simple cost. and freight is far more expensive today than 30 years ago.
now off the soapbox.
I am not sure that the idea of a nonlinear impedance has any meaning. In fact nonlinearity can only be defined in terms of a transfer function and an impedance is not such a variable.
Your test does show that the back EMF of a driver has HD because the transfer function of the voltage to the back EMF current is nonlinear. It goes as BL(x)^2 meaning that the back EMF current has the harmonics of the square of the BL nonlinearity. That makes it a not so good thing to look at when one is trying to ***** the nonlinearity of the driver system.
To mean the whole concept of "impedance" is linear. It comes from linear systems theory. If one studies nonlinear systems the concept of impedance is never used as it really doesn't have a definition in that problem.
At any rate, just looking at the impedance does not tell use if its linear or nonlinear. Looking at the transfer of voltage to SPL, for example, will show how nonlinear the system is.
I am not clear on why a non-changing Le with displacement is not symmetrical. At any rate what you are describing is why many design extend the pole piece up past the top plate to make the Le more symmetrical with excursion.
I saw an AES paper many years ago which did a dynamic FEA of the whole magnetic circuit. To my understanding they were capturing the whole array of what we are talking about.
Oh, btw..stained glass? Really? I made thousands of red, white, and blue sailboats back in 1976, light catchers..and fix the occasional window or lampshade..Lead, copper foil..it's fun to do. But the only "stock" I have left is the glass cutter..
There is stained glass and then there is Stained Glass. I don't do window hangings, I do pieces that are usually only found in museums and worth untold thousands of dollars. I also do some of the bronze work.
What does "SW" stand for?
It was most certainly the rarity of Cobalt that made the price of ALNICO go through the roof. I was involved with buying speakers at that point in time and I knew a major metals importer who acknowledged that Cobalt had skyrocketed. It was the revolutions in Africa that closed the mines which were the main source for this metal. It became very scarce for quite awhile. All available supply was earmarked for military.
When I said symmetrical, I meant flat. It the dawned on me that symmetrical could also be something like a Gaussian profile. I meant flat.
I am very confident that they did not analyze it to the extent I speak of. So far, I've not seen anything close to what I have mentioned, in either audio, or particle accelerators. However, I would love to read what you mention. Learning is a perk we all should embrace, that is my goal.
Agreed. Window hangings were 3 to 5 dollar things, I never approached the level of real antique pieces. I do however, teach clock repair, with museum pieces from 1650 to about 1800. So I can certainly appreciate the level of effort and care needed to retain value of the old stuff.
When you work on a piece from almost 400 years ago and see what was done then, it is quite humbling. Like seeing what was done in the exhibits at the forbidden city in Beijing, the brass engineering pieces.. Whoa.
Jn
Last edited:
Sub woofer.
I admit, it took me hours to figure that one out..
As to cost of materials, Those I work with do not have that restriction.
Have I mentioned that I really have fun where I work?? It is so really cool, they actually pay me to do this e/m physics stuff!!!
Jn
This is one of the errors I speak of w/r to professionals and their understandings (or lack thereof).
It doesn't matter how far you extend the pole piece beyond the top plate, the reluctance path is still through the air. No matter how you cut it, vc position towards the bottom plate couples to the magnetic circuit, while position in front of the top plate reduces the coupling.
There are many flaws in understanding out there, we can only attack one at a time..
Jn
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.