Excursion of loudspeaker model is drifting massivly

Hello everyone,

for a project that I am working on I need a nonlinear loudspeaker model( currently only the Force and Suspension) so I made one in Simulink. While Parameters like SPL and Distortion look fine, the cone excursion suffers from a massive offset drift in both the linear and nonlinear version of the Model:

While I expect some offset with it beeing nonlinear and all, 2.5 Meters seem quite excessive... here is the Schematic:

I am at a loss as to where this comes from. Has anybody an Idea what could cause this? I attached the .slx of the Speaker if annybody needs it. I already tried different solvers to no avail.

Have a nice day and thanks for anny pointers or Help.

Answer 1

[Mathieu NOE]

HELLO

I think I have an idea of where the problem lies...

I first had a look how the non linear BL and suspension force equation looks like when plotted

u = linspace(-5,5,100); % displacement (in m,  mm ? I believe mm )
F = 0.014761904761905*u.^2+0.05*u+4.952380952380952;
figure(1),plot(u,F)
BL = 1.212121212121212e-04*u.^4+3.980519423082485e-19*u.^3-0.070909090909091*u.^2+9.890342055685577e-17*u+11.826839826839825;
figure(2),plot(u,BL)

My first concern was if those equations where given for a displacement in mm (as I believe) and not meters (otherwise the output variation is quite neglectable)

so I added a gain block to convert the output from m to mm before the NL blocks ; please double check if this is true;

secondly, the "suspension force" function is not outputing a force (IMO, as for negative x we still have a positive force , so this is the root cause for the drift ) but the stiffness , so to get the force you have to multiply the output of this block with the actual displacement (in m)

combining those mods have solved the drift issue , but still you have to double check a few open points as listed above (mm / m NL blocks )

I send you back in version 2017 (I am running R2020b)

hope it helps

[Christian Noack]

Hello Mathieu,

sorry for the late reply and thanks for loocking in to it! The proposed changes worked like a charm!

[Mathieu NOE]

Glad it helped !