# Speed of magnetic flux propagation?

• I
• bhall
bhall
TL;DR Summary
What's the speed of flux? How long does it take to interact with other flux sources?
I have a case where I'm combining a N42 permanent magnet with two coils (20 AWG, 56 turns each, wired in parallel) to create flux transfer. Although model tools such as FEMM or SimScale model a static state of flux interaction, it seems clear based on experimentation that the flux interaction between the steady-state PM and the pulsed coils takes a significant amount of time to interact.

Here I show the models with 30A supplied to the coils in one direction, and then the reverse. As you see it causes the flux to either concentrate in the "back iron" at the top, or mostly at the load at the bottom. But that is steady state. If I instead apply a 10Khz (100uS) pulse that ultimately climbs to 80A before it is cut off, the same effect does not occur. I am currently adding more capacitance to power a longer pulse, but the question becomes - is there a "speed of flux", or the stabilization of the interaction of two sources? If so, how long does it take, and how much amperage is required for a pulse to have the same effect as steady-state? Also, are there any good (and not astronomically priced) software that can help model and answer this question? Or am I missing some important concept here?

Bryan

berkeman
Welcome to PF.
bhall said:
Although model tools such as FEMM or SimScale model a static state of flux interaction, it seems clear based on experimentation that the flux interaction between the steady-state PM and the pulsed coils takes a significant amount of time to interact.
The propagation of a magnetic field into a conductor is very slow because a back emf is induced, that reflects the incident field. Diffusion of the field into the material can be as slow as walking pace.

You need to calculate the skin depth for all your materials.

Laminations can be used to give the magnetic field faster access to the body of the magnetic material. The thickness of laminations should be of the order of the skin depth at the frequency of operation. To reduce eddy currents, lamination orientation should be edge on to the flux.

The flux enters the laminations at close to the speed of light, through the thin insulation gap between lamina.

Last edited:
Klystron, bhall, berkeman and 1 other person

This at least answers my question as to whether to treat this more as a DC problem or AC. Although the math shows that it is still not full on AC at least for the coil resistance. At 110v/80A = 1.375 ohms, with the amperage still climbing versus the 2.355 AC ohms measured at the same 10 KHz with repeating cycles.

Skin depth for the 20 AWG coil at 10 KHz would be 0.65 mm, which is more than the radius of the 0.40 mm wire of the coil, so at least at that frequency it should be fine.

I suspect eddy currents in the 0.25" steel sections, 4 each side, are then a large part of the issue. I can quickly get the same material cut in 0.030", which is close to 0.025", and try again by the end of next week.

Regards,
Bryan

bhall said:
I suspect eddy currents in the 0.25" steel sections, 4 each side, are then a large part of the issue.
Calculate the skin depth in the magnetic path and see what you get. Audio transformers always have thinner laminations than power transformers. Consider using an iron powder, or a ferrite, for the magnetic circuit.

Also take a look at the skin depth of the changing flux in the permanent magnet material.

• Electrical Engineering
Replies
12
Views
1K
• Electromagnetism
Replies
1
Views
2K
• Electrical Engineering
Replies
4
Views
3K
• Electromagnetism
Replies
26
Views
3K
• Electromagnetism
Replies
9
Views
4K
• Classical Physics
Replies
2
Views
3K