Question on Configuring Demagnetizing Field for a Specific Geometry

[Alexandra B.]

Dear Mumax Community,

I am currently working with Mumax+ and am having trouble configuring the demagnetizing field to behave as expected.

My goal is to model a ferromagnet where the in-plane (XY) dimensions are significantly larger than the thickness (Z), under the influence of elastic waves along the z-axis. For the demagnetizing field calculation, I require the geometry to behave as an infinite plate. For all other aspects of the simulation, however, I woud like the magnet to be treated as a finite-length rod along the z-axis (magnet = Ferromagnet(world, Grid((1, 1, 1000)))).

I have experimented with various mastergrid and pbc_conditions configurations setting the World but have not succeeded in achieving the desired demagnetizing tensor, where only the Nzz component is non-zero and equal to 1.

Could you please advise if such a configuration is possible in Mumax+? Any insight you could offer would be greatly appreciated.

Best regards,
Alexandra B.

[Ian Lateur]

Hi,

I guess you just need to take an enormous amount of pbc_repetitions? Although the demag kernel might be slow to calculate and not very accurate...

With very small variations and a nearly uniform magnetization, you could disable the demagnetization and add an effective external bias field to compensate? But that will not update throughout time.

Being able to set the demag kernel would be a cool feature for cases like this...

Ian

Op ma 17 nov 2025 om 11:34 schreef Alexandra B. <alexandr...@gmail.com>:

--
You received this message because you are subscribed to the Google Groups "mumax2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mumax2+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/mumax2/80cbe7bf-453a-472b-98c8-236d0f2d85ddn%40googlegroups.com.

[Alexandra B.]

Hi,

Thank you so much for your reply,  I really appreciate it! You've confirmed my suspicions.

I tried implementing the large PBC setup, but unfortunately, as you've mentioned, the performance is too slow.

Given this, I'll probably try to move on to modify the source code to implement a custom demagnetizing field. 

Best regards,
Alexandra B.