STEORN DEMO LIVE & STREAM in Dublin, December 15th, 10 AM

[lumen]

@OC,

I seem to always think of remanence as a bad thing. I can see no way it could play in as a benefit.
The last condition the core would be left in would be a circular path that was generated to ignore all external fields.
The next approaching rotor magnet will have no attraction until some core domains start flipping so the remanence at this point could be of no real help.
By the time the rotor reaches TDC, any remanence of the previous cycle will be totally gone so it will not exist before the coil is again energized.

Unless I'm missing something in there?

[0c]

I seem to always think of remanence as a bad thing. I can see no way it could play in as a benefit.

If it was an AC application or if here were alternating magnetic fields, you would be correct.

The last condition the core would be left in would be a circular path that was generated to ignore all external fields.

Not true. The last state of the core is when the coil is deenergized. When that happens, the core returns to its remanent state which, if you look at one of the permeability vs. saturation references I provided, you will see is actually close to maximum permeability and maximum attractive force. In remanence, the domains in the core are already mostly aligned with each other and will rotate easily, reorienting to present an external field to attract the magnet even better as the magnet approaches.

The next approaching rotor magnet will have no attraction until some core domains start flipping so the remanence at this point could be of no real help.

The domains don't "flip", they just "rotate". Get a document on Barkhausen Noise. The noise occurs during domain growth and alignment (flipping), the steep vertical portion of the BH curve. This is the hardest part of the magnetization process. Subsequent rotation (without much noise) is easy in comparison. The saturation flux density (B field) is not easily increased above the knee, but the polarity can easily be rotated up to about 90 degrees or so, possibly more.

By the time the rotor reaches TDC, any remanence of the previous cycle will be totally gone so it will not exist before the coil is again energized.

Not true. The domain alignment remains and all that's required is rotating prealigned domains by about 90 degrees.

EXPERIMENT:

Materials:

1) Four axially magnetized rod magnets, say about 1/4" dia x 1/2" long
2) One larger magnet, say 1" dia x 1" long
3) Lid from a pickle or mayonnaise jar
4) Large round plastic or cardboard container (Quaker Oats container works good)

Take the 4 small magnets and arrange them on the lid near one edge so they form a square with N end of each connected to S pole of the next (similar to the natural domain orientation of unmagnetized material).

Set the jar lid with magnets in the bottom of the oats container and using the large magnet try to get all the magnets to rotate so they are oriented in parallel and have the same pole facing the large external magnet.

Now take the jar lid and rearrange the rod magnets so they are all as parallel as you can get them with same poles facing the same direction. Stick the lid and magnets back into the oats container and use the large external magnet to get the jar lid and magnets to all align with the same pole facing the external magnet. Now move the external magnet around the outside of the container and see how easy it is for the rod magnets to follow it and maintain their relationship with each other and the external magnet.

I know it's a pretty crude experiment, but hopefully it'll give you a better feel for what I'm saying.

Have fun!

[0c]

Another way to see what I said above is to just take the 1st quadrant of the BH curve for the material in question. Draw a minor loop from zero B out to Bsat. Draw another minor loop from Br out to Bsat. Which loop is smaller? Which one requires the most work?

[gravityblock]

Quote from: gravityblock on Today at 08:37:58 AM
    This is the reason for the fast rise time in current for the Orbo.

Quote from OC in response:
I disagree. This is the primary reason for the initial delay before the sharp rise in current for the Orbo. Some of the factors I have been mentioning do not seem to be incorporated in Orbo, remanence is one. That inductive delay at the beginning of the pulse indicates to me that Orbo's cores have very little remanence and require more effort to realign the domains with each other at the beginning of each cycle.

It is possible that Steorn has deliberately sacrificed some of the efficiency remanence might contribute in order to leverage some other material property (like magnetic viscosity). It's also possible they have simply overlooked it.

My Response:  The reason why it appears to have very little remanence is because the core material is returned to its virginal state.  There is an easy axis and a hard axis.  It requires less energy to align the domains along the easy axis than the hard axis.  From M_r to M_s, "B" and "H" are related to the voltage and current by an integral.  This means, the deeper you get into the flat saturation portions, the harder it is to continue aligning the domains, which means your voltage and current requirements go up by an integral the deeper into the saturation curve you go.  This is the hard axis.

On the linear portions of the curve, "B" and "H" are proportional to the voltage and current by a constant.  This means the voltage and the current remains constant and doesn't require additional energy the deeper into the linear portions you go.  This is the easy axis.

Quote from: gravityblock on Today at 08:37:58 AM
    ... "B" is proportional to the voltage related by a constant on the linear portions?

Quote from OC in response:
Huh???

My Response:
The 4th and 5th paragraphs below the second image of the page, http://www.beigebag.com/case_xfrmer_4.htm , you will find how "B" is related to the voltage by a constant and "H" is proportional to the current when you're on the linear portions of the curve.  The flat portions representing the saturation area is not linear or related by a constant but is related by an integral.

Quote from: gravityblock on Today at 08:37:58 AM
    Between these two regions is a transition region. In this portion of the curve, which can be relatively large dependent on the material, the number of unaligned magnetic regions is becoming smaller, and the B-H curve flattens.  As you can see, the steep linear regions do lead to saturation according to the material. This region, dependent on the core material, can be rather small or even quite large.

Quote from OC in response
Not "wrong" but doesn't really sound quite right either. Let's call that "transition region" the "knee". Below the knee, the primary effect is domain alignment. Above the knee, the primary effect is rotation. The transition region itself is varying degrees of both.

My response:
The 4th paragraph below the first image you will find my words are almost identical to this page, http://www.beigebag.com/case_xfrmer_4.htm  The author of this page is claiming to be an expert on Cores and Transformers.

Quote from: gravityblock on Today at 08:37:58 AM
    The Orbo is sitting at it's highest point of this steep linear curve.

Quote from OC in response:
I'm afraid I don't see it this way. I think Orbo is straddling the permeability curve, leaning a bit towards the saturated side. Remember their first measurement of inductance? The change was only from about 960 to 990 mH. That's not a massive change like they got in their supplementary demo with the A and large magnet in direct contact. I think greater gains in torque and inductive return can be had if better use can be made of the saturated side of that permeability curve.

My Response:  There is no need to fully saturate the core when there is a large gap between the magnets and toroid.  This is just wasted energy.

Quote from: gravityblock on Today at 08:37:58 AM
    [Edit:]  The "flat" portions representing the fully saturated regions is nonlinear in respect to B, which means it is related by an integral and not related by a constant.  This means going from Mr to Ms, "B" will be related by an integral and not by a constant, which I think means the current won't have a fast rise time and the voltage/current won't remain steady under load.

Quote from OC in response
I'm afraid the math is out of my reach. I'll leave that to those more capable than me. I just have "visions".

My Response:  No need to do the math.  Just knowing that it is related by an integral should tell you that it's going to require more energy the deeper you go into full saturation.  On the linear portions, the energy requirements don't increase the deeper you go into the linear regions, which will lead right up to the "knee" or transitional area.  At this position, the magnets won't be attracted to the toroid because full saturation isn't needed due to the gap between the toroid and magnets.

GB

[gravityblock]

Another way to see what I said above is to just take the 1st quadrant of the BH curve for the material in question. Draw a minor loop from zero B out to Bsat. Draw another minor loop from Br out to Bsat. Which loop is smaller? Which one requires the most work?

Smaller doesn't necessarily mean less work, especially when the energy requirements are increasing by an integral the deeper you go into it.

Now, I may have the easy and hard axis backwards.  If I do, then please correct me.

GB