is this explaination of a microwave circulator correct?

[coppercone2]

So this guy says its like a hall effect device. The wikipedia page says something very confusing and does not mention the hall effect. He says that the ferrite itself is just a magnetic field spreader, while wikipedia claims that the ferrite is a specific type of ferrite. Is the video explanation just over simplified?

https://en.wikipedia.org/wiki/Circulator

Also, what is a circularly magnetized material? How do they make that? And like for explainations, if you gutted a circulator and stuck it into a uniform helmholtz coil, would it work, without the ferrite? It seems that the ferrite is critical according to one explanation but not the other.

I think it would be interesting to have a
1) how its made (practically), I guess its simple machining other then the ferrite
2) how its designed (say what it you introduce a air gap between the ferrite and the copper by making the ferrite disk thinner?)

[paul@yahrprobert.com]

That guy on youtube is pretty much just making stuff up.  And the wikipedia article, with its "circularly magnetized" nonsense is a little to brief and misleading.  But you're right, its all in the ferrite.  Its a special material, known for its non-reciprocal (anisotropic) properties when magnetized.  If you decompose the RF field in the ferrite into a part rotating left around the magnetic axis and a part rotating to the right, those fields will have different propagation properties depending on the frequency and the magnetic field strength.
  There are some expensive texts on Microwave devices that will go into great detail, but maybe some of the references in that wikipedia page would give you a good exposition.

[coppercone2]

Ok I have a little trouble imagining the decomposition but from the sound of that behavior it sounds like there will be something like fringing happening.. is it like super imposing two magnetic paths in one location or something like that? (very confused).

I guess you have to solve it to some number of points to get an idea of how the wave is reconstructed or superimposed later or however its called for a practical explaination?

It would help to know about the ferrite disk manufacture. Is it like just a special mix where the crystalline structure (possibly with special cooling procedure/initial grain structure/pressing pressure requirements?) does the magic, or do they actually have some weird way of magnetizing it after the part is made in the kiln? When I hear magnetize I imagine something like how a magnetic part is manufactured, on a special high powered machine. I thought it might have a magnetic pattern on it like a hard drive platter the way that is phrased, but it also says it does not have a field of its own so I don't know.

What comes to mind is like a tesla-valve (the one for fluids)  that does not recirculate the water back into the stream, but bleeds it off in a different direction (like if you put a buncha tiny pipes hooked up to a typical tesla water valve thing that have more flow directed to the outlets depending on flow direction. I guess dropping a big concrete slab that is at some particular angle to a flow of water in a river built on a giant tilting table would do the same thing, its easy to imagine it might work for some kind of material like grain in some kind of grain chute that tilts back and forth, or some kind of 'marble machine'.

It also makes me think of lens for some reason (the one that looks like a ferrite post inside of a wave guide Y)


Might it be easier to understand how the circulator functions in a wave guide rather then a stripline for learning the mechanics? The picture I see of a ferrite pillar in the middle of a Y waveguide piece makes some what more sense to me then a strip line based one.

[T3sl4co1l]

Don't think about it in terms of point fields alone.

The bias is just that, bias.  Ignore it for AC purposes.  Its only effect is to bias the material, which causes Faraday rotation of propagating (EM) fields through it.

Then apply the EM field.  Though there is a conductor carrying EM into the system, a substantial amount of the EM field propagates through the space around that conductor -- in particular, into the ferrite.  The velocity factor within the ferrite is vastly reduced compared to air, so not much material is needed to achieve a 1/4 wave section, and sufficient rotation (Faraday effect) occurs in that span to reinforce one direction while interfering (canceling out) the opposite direction.  And, through design tweaks and whatever, the bandwidth can be actually quite good, despite the dependency on wave interference.

The extra width of the conductor can also probably be explained as a means of keeping the impedance matched.  Zo = sqrt(Lo/Co) and Lo is greatly increased (Co is increased modestly, as ferrites have notable dielectric constant too, just not to the extreme that permeability has), so a wider conductor can compensate for that.

The ferrite material itself may be NiZn or similar types, or YIG (yttrium iron garnet) which is even less magnetic (lower mu) than power (ZnMn) or radio (NiZn) ferrites, but much lower loss still, and I think the physical properties (paramagnetism and etc.) are more important (but I don't recall offhand if that's because EPR is tuned to the bandpass frequency in these, or if it's involved at all).  So, they're quite efficient despite the high frequency.

Or at least, that's one possible explanation.  It's been some years since I read about circulators.  Information is not hard to find.  Can be a bit technical though, being on the overlap of material physics and EM.

Tim

[paul@yahrprobert.com]

I found this on the web: https://www.rf-ci.com/wp-content/uploads/2015/09/KB-001_Operating-Principle.pdf,  The text on the first page talks about the counter-rotating fields.

You're right the actual device doesn't contain just pure left and right plane waves, but boundary conditions and fringing all occur and make the solution very messy.  I'll bet they made these devices work by trial and error long before they could solve for the exact fields.

A similar device in the optical world employs faraday rotation.  If you put a magnetic field along the axis of a long glass rod, then the left circularly polarized wave goes at a slightly different speed than the right polarized wave. If you put polarizers at either end of the rod and make the rotation difference be 90 degrees you can make an isolator.

[coppercone2]

Ok I guess the problem is I don't know about polarization when it gets to the 1/4 wave plate. I get the first one to make direction ally polarized light (this one can be explained with a shoe lace and a shopping cart) but I think I need to understand that plate first. I think a GIF of the output of a polarized antenna with the phase shifter would help me too, I will try to look for that. I know that if you take orthogonal antenna pair and supply one of them with a delayed signal you get a polarized RF signal but I don't know how that looks like (I avoided this because it seemed that its only useful for satcom). Since its two antenna and a phase shift it also probobly has something to do with cancellation like a phased array.


it almost looks like it deforms it, like 45 degree and 135 degree angle adjustment

[paul@yahrprobert.com]

The way to think about these things is to first understand that you can decompose a certain wave into a sum of other types of waves.  For instance, a vertically polarized wave (E field in the Y direction, propagating in the Z direction) can be added to a horizontally polarized wave (E in the X direction) and if the two are in phase the result will be a wave polarized at 45 degrees. If you delay one of the waves by a quarter wavelength you'll end up with a circularly polarized wave either left or right depending on which one gets delayed.  Conversely you can add two circularly polarized waves, left and right, to get a vertically or horizontally polarized wave depending on the phases.  The quarter wave plate shows how this idea pays off: what the plate does is delay the vertical by 90 degrees with respect to the horizontal.  So decompose your 45 degree incoming wave into a horizontal part and a vertical part, delay one by 90 degrees, and the result is a circularly polarized wave when you add the two waves together.

[coppercone2]

Well this book says something interesting (AL LANCE) that a linearly polarized wave can be considered to be composed of two circularly polarized waves equal in magnitude but rotating in opposite directions. These two circularly polarized waves can at all times be combined the yield the linearly polarized wave.

Then it says that if the ferrite does not resonate strongly with the wave, the two circularly polarized (which means a linearly polarized wave that is decomposed for the explanation?) can pass through it, but the two components experience difference phase shifts since one wave aids precession of the spinning electron while the other wave opposes the precission**. The phase rotation is equal to one half the difference between the phase shifts of the circularly polaried waves. The direction of the resultant phase rotation is determined by the direction of the magnetic field and not by the direction of propagation of the microwave energy. This phenomenon is applied to produce a isolator. *

I need to understand the part before it better, but it sounds like what you are talking about.

But he calls all of his faraday rotators, so isolator is just the marketing/engineering name.. if anyone wants to find more physics level information about this, I suggest you study faraday rotators.


*So if you managed to take a isolator apart, and swap the magnets position (poles), would it change direction from CW to CCW?

**if the decomposed waves are equal in magnitude and are both causing the preccision to change, wont the net effect be 0?

And the hard part is that its compared to gyroscopes, which are not the most easy thing to understand.


If it seems that the decomposed wave has parts that are both doing something to precession ,why does it not cancel out? I feel like I am missing a very simple piece of information. Does this have something to do with vectors? Like if they both have the same strenght to effect it but one of them effects it head on while other one has only 45 degree relationship or something like that since its a 90 degree phase shift?

[profdc9]

Magnetically susceptible materials (for example ferrite and YIG - Yttrium Iron Garnet) are unlike electrically susceptible materials in that they break time reversal symmetry, and so a wave traveling one direction is affected differently than the wave passing in the opposite direction.

In an electric medium, approximately speaking, the positive and negative charges of an electric dipole are separated (polarized) by an electric field.  The dipoles can not respond any differently to a wave or its time reversed version because the dipole itself has no "state" that depends on time.  Even a helical electrically polarizable molecule, such as dextrose (right-handed glucose) can not break time symmetry, because while the shape of the molecule allows charges to move helically along its length, which produces a different in retardance between circular polarizations, it still responds based on the chirality of the molecule which does not change with time or direction. 

On the other hand, a magnetic dipole breaks time symmetry.  If you imaging a magnetic dipole, it's a current loop.  That current loop (for example the spin of an electron or its motion around a nucleus) has a definite direction the current flows around the loop.  If you reverse time, that direction reverses as well.  So the magnetic dipole responds different to a circular polarization which turns in the same direction that the current is flow in the loop, vs. the polarization which turns in the opposite direction, tending to be magnetically polarized more by when they rotate in the same direction, resulting in increase retardance of that circular polarization as compared to the other.

Also, if you imagine the magnetic dipole being viewed from one side or the other, the direction the current appears to flow reverses, so that a linear polarized wave incident on the magnetic dipole for one side or the other, which is a superposition of left and right circular polarization, is going to have its two constituent polarizations retarded differently depending on direction. The linear polarizers on an isolator can be set up to allow the wave through if it travels in one direction through the magnetic medium and block it if it travels back the other way.

Basically, a material like ferrite or YIG is a bunch of electron spins that have been aligned due to paramagnetism or ferromagnetism, and each electron spin acts like a tiny current loop with a direction of current flow given by the right hand rule.  You can imagine classically that each electron is like a little spinning ball of charge, however, quantum spin is simply an inherent property of the electron and has no physical manifestation except given by the effects of symmetry (Pauli exclusion principle) and electromagnetism. 

In classical electromagnetism, the loss of time reversal symmetry is because the magnetic field depends on the first derivative of time, so that if time is reversed, the derivative is multiplied by -1.  Or in other words, if you reverse time, the direction of a current reverses and it becomes the opposite current.  This loss of time reversal symmetry because the medium has a net magnetic polarization is why it is used in an isolator/circulator.

[coppercone2]

So is the dextrose molecule you are thinking about like a twist drill kind of? Its hard to tell from the diagrams, I assume it looks like it has a... twist/saddle/bend in it that makes its behavior different then something like a flat aromatic molecule?, and the idea is that the mass particles? I guess are responsible for the distribution of electron cloud, and if you reversed time the electron cloud would still be the same?

I think I clearly understand the effect of reversing flow of time on the dipole, I think, which is that the orbital electron velocity direction parameter changes sign, right? so that is like changing the polarity in a coil. So its like north and south poles change if time were to run backwards..?

[coppercone2]

the book had me confused because the drawing is bad, i thought someone left a pizza on a basket ball or something, its actually like a mechanical drawing but I thought it was showing something like small signal gain plot the way its drawn that its like explaining a concept that occurs at a point and not a path (its not actually abstract shit).

so I think I understand the structure of it now and how the time symmetry principal works and from different points of view that the directions of things are changed. I am not sure if I should be thinking about phase shift in terms of distance though, like one of the decomposition products being rerouted through a some how longer squiggly path? or does it go slower? the way you put it makes me think of a partial boat 4th dimensional boat going up river. I just start thinking about delay lines here and I am not sure if that's productive ( hard not to when everything seems like some kind of mechanical structure with these explanations).