Toroidal permanent magnet

[Circlotron]

Suppose we get a toroid shaped piece of magnetically hard material and thread wire through it in the usual way, crank up the current and magnetise it. The lines of force run around in a circle, not toward the outside world. Now we have a permanent magnet but nothing sticks to it. Is there any use for such a magnet?

Also, suppose that the material resists demagnetisation. If we now apply a dc current the opposite direction to what magnetised it originally, as the magnetism gets pushed down to zero and then toward the opposite direction, during this time would the current rise in a controlled manner like it would in a gapped inductor? And when the winding is opened circuited would there be much energy released in the same way as a gapped inductor makes an arc.

What I am trying to find out is, could you use a magnet as part of an inductor core to store energy e.g. for a dc filter choke instead of having an air gap, therefore getting more inductance for the same amount of turns?

[T3sl4co1l]

Depends on the exact properties of the core, but the hardest ones exhibit very low permeability below the coercive limit (basically, the self-magnetization is so strong that they remain saturated), so yes you would have an inductor just like ever, except it would be air cored.  So you'll need a lot of turns.  But you'll need a lot of amp-turns to push such a hard material into a different magnetization state anyway, so that's fine.

The trouble with magnets as cores is, they don't do anything until you exceed the coercive limit, and only then do they start storing flux.  Until it saturates in the opposite direction (which is how you know it's fully magnetized in that direction), then it does nothing for a while again.  This is a huge hysteresis loop, and it only saps power.  So, it's utterly useless for power control applications; the energy density also means just one cycle heats up the material significantly, so you can't use it at AC anyway.

But you're probably on the right track.  Saturation itself can be useful: if you can control the flux into a core, you can control when it saturates, and use that to transmit (or not) power.  Such materials are used for magnetic amplifiers and saturable reactors.  They are defined by a small hysteresis loop (low losses), high permeability, sharp saturation.  The loop has high remenance (it remains permanently magnetized very near saturation), but low coercivity (so it can be very easily pushed away from saturation).

Examples include square ferrite (a few special kinds: not something you're going to find in generic ferrite beads or whatever), amorphous/nanocrystalline (most types), and permalloy and related (traditional rolled strip) alloys.

Such materials are generally made to exhibit as high permeability as possible (in the linear range), so that energy storage is intentionally minimized.  For energy storage, the ideal inductor has a permeability of 1, i.e., no core at all (a core is normally used only because we don't have anything more conductive than copper).

Tim

[bobcat]

You are describing the basic principle behind magnetic core memory.

[ali80]

the magnetic energy density stored in any given volume is e=0.5*B^2/(2*u)
which "B" is magnetic field density and "u" is permeability of the volume
since magnetic materials have very high "u" in range of 1000 time of "u" of air, it is almost impossible to store energy in the core itself, so we use gapped cores and most of the energy is stored in that tiny gap.

[Circlotron]

Okay then.
The seed for my idea was that some modern day car ignition coils use a biasing magnet as part of the magnetic circuit (did not know if they also had an air gap but suspected they had) and the idea is that when they charge up they can swing the full width(?) of the BH curve instead of only from the centre to the edge.

[T3sl4co1l]

Ah.  While that's a thing that happens... you're probably better off buying a bigger core (cheap material) rather than sticking expensive magnets on it.  Or use an air-core coil altogether, as I think some/most CDI coils are made for?

Tim

[NiHaoMike]

In old CRT TVs and monitors, it is very common to find an inductor with a magnet glued to it, in order to oppose the DC bias and let them use a smaller and cheaper core.

[T3sl4co1l]

No, the magnetized cores are for horizontal linearity correction, and as I understand it, are intended to actually reduce the flux before saturation.  They'll also saturate more gradually, as the magnet is attached at one end of the core, whereas the coil is full length.  Which should have important implications for how soft/sharp the resulting linearity correction is, and would have to be matched to the needs of the deflection system.

This is understandably hard to photograph (and even harder to screenshot... ), but this is sort of what my Trinitron looks like with a 1 pixel alternating vertical line pattern:
http://seventransistorlabs.com/Images/Trinitron_Fringe.jpg
The video bandwidth is just enough to convey the on-off pattern, but it doesn't quite align with the phosphors on the screen.  (You can attempt to operate at a higher resolution -- this thing goes up to 2048 x 1536 x 60Hz -- but it looks like muddy lines of a very different pitch.  Good for screen area and working with soft graphics, but useless for fine per-pixel detail.)  The interference pattern changes over the screen, so that near the middle, it's a little off (skipping a phosphor band every 50 pixels or so), but it's worse near the edges (top and bottom, indicative of pincushion type distortion; left and right, indicative of sweep nonlinearity).

This is one problem that LCDs will never have problems with, though (at native resolution, anyway).

Tim

[Alex Eisenhut]

You are describing the basic principle behind magnetic core memory.

As I was reading the post I was thinking he just invented the bit...