I've abused LTspice to do stuff like multi-pole electric motor modeling, 2D and 3D FDEC, lathe bit regenerative chatter based on cut history and overlapping cuts, etc. I surprise myself with what SPICE can be used for.
What was even more shocking was looking at the equations in multiple complex speaker modeling writeups and realizing that they were NOT doing anything that SPICE can't do! In some cases FEMM was brought in but only to generate BL(x) and Le(x) functions which could easily be loaded into SPICE.
This is true for even the papers that successfully model the acoustic distortion of the speaker.
Personally I have just been working on making parametric functions that can model the curves I see in Klippel analyzer data. John Curl posted a page from Olson that showed a good basic function for Kms(x). Still can't find Olson in a Google search though.
What was even more shocking was looking at the equations in multiple complex speaker modeling writeups and realizing that they were NOT doing anything that SPICE can't do! In some cases FEMM was brought in but only to generate BL(x) and Le(x) functions which could easily be loaded into SPICE.
This is true for even the papers that successfully model the acoustic distortion of the speaker.
Personally I have just been working on making parametric functions that can model the curves I see in Klippel analyzer data. John Curl posted a page from Olson that showed a good basic function for Kms(x). Still can't find Olson in a Google search though.
I may be wrong, but may this be of help
http://diy-audio.narod.ru/litr/ThePowerofLoudspeakerModels.pdf
http://diy-audio.narod.ru/litr/ThePowerofLoudspeakerModels.pdf
FEM Finite Element Method ? Finite Element Model.
Indeed that is the core of SPICE.
As far as I no, there is no access to this, the user interface is EE oriented.
Unless there is some hidden access, we are left
with electric analogies.
Indeed they are locked and now show you are falsifying my sayings. I wrote in #28553:jneutron said:
"The ring current and its consequent image current in the primary coil are another matter, and these are really in antiphase so that their magnetic effects cancel out, leaving the original vc flux unchanged (like the magnetizing flux in a power transformer remains unchanged)."
which is entirely in line with Bonsai's comment and transformers' basic operation.
😕😕😕 That consuming of the magnetic flux is just what you have insisted to happen with your popularized version of the Lenz' law, and I'm the one who has disputed it.
You have all along failed to explain where the extra 90 degree lag in your extended version of Lenz' law comes from - the lag without which there can be no counteracting or excluding of the original flux. You have also failed to point out errors in my spoon-feed explanation of Faraday's law in post #28595, so Faraday still remains valid and not overridden by the weird form of Lenz.
I am only making a clear distinction between the true Lenz' law, which is merely the minus sign already included Faraday's law, and the common fallacious extension of Lenz' law, which is claimed to counteract and exclude the original flux and which is plain nonsense, no matter how many seem to propagate it. I also doubt that Mr. Lenz has ever presented the latter form, as that would be in clear contradiction with Faraday's law.
Not hidden, most of the useful inputs to this problem have electrical analogies. I mentioned FEM separately because it might be useful to extract a one dimensional property such as inductance from a 3D model of an actual device.
Very nice link, thanks..
I note with interest the (page 10) impedance measured at +/-4mm and zero.
At zero it spikes at resonance, but at +/-4mm there is no resonance shown.
Clearly they are using mechanical means to shift the driver, and that destroys any possibility of seeing a resonance. That is why I was thinking of using an 18 incher to force displacement of the 8 and see (almost) unfettered behaviour through the full range. (looks like they are doing that on page 25)
On page 13, note they mention optimization of the shorting ring (they called it short cut ring)
I did not find any discussion of the shorting ring increasing dissipation at hf..
jn
I specifically stated that YOUR statements are yours, and I provided exact copies of YOUR statements..
The antiphase is precisely what Lenz's law states, that the ring currents will buck the rate of change of flux.
No. It is a strawman argument YOU built trying to disprove Lenz's law.
So it's up to me to prove Lenz's law? I don't think so. Rather, you are telling everybody who will listen that it doesn't work. So it's up to you to prove such.
If you insist on stating Lenz law is incorrect, prove it..
As I said before, I will stick with the physicist guy from 1834 who derived it, rather than someone who keeps dancing around the physics..claiming without any proof that Lenz's law isn't..
Meanwhile, your analysis is slowly drifting towards reality. At least now you agree the two coils are 180 degrees out. That is progress. Next, you might just start using phasor analysis...Don't worry, we'll get you there. And in the long mix of time and floobydust, nobody will remember your anti-Lenz remarks...
jn
Last edited:
Would using COMSOL be useful?
Modeling Low-Frequency Electromagnetics with COMSOL(R)
-RM
Comsol.html
Installation of COMSOL Multiphysics on a Windows PC
Download Comsol 5.4
Connect to engineering VPN server
Extract the package. You will need 7zip or a similar utility.
Right Click on "setup" & run as "administrator"
Choose "New Install"
Accept the license
In same window, click on the drop-down menu for license type and select "<port number>@<host name>"
Enter "1718" for port number
Enter "comsol.engr.ucr.edu" for hostname
Continue to click "Next" when prompted until the install is finished
To run comsol, make sure you are connected to engineering VPN server and then run the program like you normally would.
NOTE: These installation instructions have been tested for Windows 8.1
You can download the previous version (Comsol 5.3a) if needed.
Modeling Low-Frequency Electromagnetics with COMSOL(R)
-RM
Comsol.html
Installation of COMSOL Multiphysics on a Windows PC
Download Comsol 5.4
Connect to engineering VPN server
Extract the package. You will need 7zip or a similar utility.
Right Click on "setup" & run as "administrator"
Choose "New Install"
Accept the license
In same window, click on the drop-down menu for license type and select "<port number>@<host name>"
Enter "1718" for port number
Enter "comsol.engr.ucr.edu" for hostname
Continue to click "Next" when prompted until the install is finished
To run comsol, make sure you are connected to engineering VPN server and then run the program like you normally would.
NOTE: These installation instructions have been tested for Windows 8.1
You can download the previous version (Comsol 5.3a) if needed.
Last edited:
As I said before, I will stick with the physicist guy from 1834 who derived it
Err, not exactly....
Err, not exactly....
The Project Ryu guy has FEMM files for his field coil speakers for download. I never got them to work, but that doesn't say much.
It's a small world, I recall using Suprem IV (semi technology simulator) for solving some thermal problems 😀. Diffusion equation is the same for both.
Whats the most direct way to minimize flux modulation (not necessarily the cheapest way).
THx-RNMarsh
THx-RNMarsh
Circa 1986 the latest bipolar process had min geometry devices pushing 2000C/W so I helped one of the CAD guys add T as a state variable to our SPICE and it was amazingly easy. I think we had one of the first simulators that had thermal nodes with thermal resistance and capacity as components. More surprising it had almost no hit on speed performance.
When you do the locked voicecoil measurements, can you measure the inductance at the lowest possible frequency? What I've found is that it just keeps rising as frequency goes down, and no one has bothered to see what it is at say 30Hz or around Fs. It's important to know it's "ultimate" value because that gives us the necessary parameters to model the reluctance force. The inductance at LF can be several times the 1KHz inductance. I would say it has something to do with the skin depth of eddy currents on thick steel plates but not sure if skin depth is the right formulation.
I always run Ls/Rs from 20hz up, and use the 2, 5, 10, 20...frequencies as they plot very nicely on log. Taking the coil and shorting ring to 100Khz was simple enough and the data didn't look too outrageous.
I may have to grab a different meter for the speaker reads, as I was unhappy with the 20 and 50 Hz rise in the inductance. However, I suspect it was just a loss of resolution as the inductor was 250 uH so reactance at 20 and 50 hz is just too small.
jn
Another no-copper method of reducing motor inductance.
JBL's Diffrential Drive speaker motor uses both poles of a NdFeB motor to create an N-gap which powers a right-wound voice coil and the P-gap which powers a left-wound voice coil, which summed together cancel out the generated inductance. A very musical solution for modest Mms, modest Xmax woofers.
"In the traditional design, magnetic flux density B crosses a gap in which a coil of wire of length l is placed. The coil has an electrical resistance, RE. These quantities establish the value of (Bl)2/RE, which is the electromechanical coupling factor of the driver. In a Differential Drive topology there are two magnetic gaps that have opposing flux fields, and each gap has a B field equal to that of the traditional design. Two reversely wound coils are used, each using wire with one-half the cross-section width as before. The coils are scaled so that they are one-half the length of the traditional coil. In addition, the masses of the two thinner coils will equal that of the single traditional coil. Each of these thinner coils will then have a coupling factor of 0.5(Bl)2/RE, or one-half that of the traditional driver. Since there are two such voice coils, both acting in the same direction, the total coupling factor for the new motor structure will be identical to that of the traditional design."
A shorting ring is added on the former between the two wound voice coils to provide Xmax braking.
Last edited:
We went through the that design what, 6 months ago? I remember the shorting ring used for Xmax braking as well. I do like the symmetry of it. There is always beauty in symmetry.
I guess even JBL has been hoodwinked into believing in Lenz exclusion.
Oh, the humanity of it...
We also went through the neo ring structures where the flux did a really compact donut thingy, some paper put up here. Needed interesting magnetization rings glued up. My concern was worker fingers.
Jn
I guess even JBL has been hoodwinked into believing in Lenz exclusion.
Oh, the humanity of it...
We also went through the neo ring structures where the flux did a really compact donut thingy, some paper put up here. Needed interesting magnetization rings glued up. My concern was worker fingers.
Jn
Last edited:
So
It Looks to me that the non-linear Z source is the magnetic material needed to make the speaker function (assuming uniform fields etal). Reducing the Z (L) to a smaller number, proportionately lowers the distortion.
Maybe we need to go back to the electro-magnet powered from the wall ac outlet?
🙂 Who volunteers to make one?
OR, use the current feedback/voltage feedback combo to reduce distortion (via sense resistor).
THx-RNMarsh
It Looks to me that the non-linear Z source is the magnetic material needed to make the speaker function (assuming uniform fields etal). Reducing the Z (L) to a smaller number, proportionately lowers the distortion.
Maybe we need to go back to the electro-magnet powered from the wall ac outlet?
🙂 Who volunteers to make one?
OR, use the current feedback/voltage feedback combo to reduce distortion (via sense resistor).
THx-RNMarsh
Last edited:
- Status
- Not open for further replies.