The graph on page 230 (fig 5.15) doesn't fit the definition of µ
given on the same page. The authors emphasise that its value is B/H,
not dB/dH. The peak value should therefore occur at the knee of the
BH curve, from where its tangent passes through the origin. RDH has
got the peak in the right place, above the knee rather than above
the steepest part of the BH curve, which is where it would be were µ
to be dB/dH.

But the right hand side of the curve is wrong by miles. µ is shown
to plummet towards zero as the gradient of the BH curve flattens
towards saturation. This would be true if µ were dB/dH, but if it is
B/H then it should drop more gradually with an asymptote along the B
axis.

The shape of the curve shown on page 216 (fig 5.13D) fits their
definition better, but that is for AC, which is another can of
worms.

Checking several sites and books, there is no agreement on whether µ
is B/H or dB/dH.

For example, Menno van der Veen says that µ reflects the mobility of
the magnetic domains. This observation seems to favour the dynamic
definition, because at saturation there is no mobility even though
the value of µ may still be high according to the static RDH
definition. He shows a graph of inductance versus secondary voltage,
in which the inductance plummets to near zero at saturation, which
is what I would expect. He also correctly points out that µ is the
only variable in the equation for inductance for a given inductor.
Hence where the inductance falls to near-zero, so must µ.

Further, µ0 is used for the permeability of free space, and also for
initial permeability, as in the table on page 208. The unit for µ
and µ0 is also generally omitted. It does have a footnote that says
"strictly" the unit is gauss/oersted. It also states that µ0 = 1.
Actually its unit depends on what system of units you are using, as
does its value. In a table using both gauss and lines per inch, the
unit of µ should be stated.

The confusion is everywhere. It has infected various spice core
models that don't work properly. Much data is barely intelligible
because units and definitions are not clear.

Incidentally, Van der Veen, of Plitron fame, gets the formula for
inductance wrong in his book, transposing core area and magnetic
path length. I assume this is a misprint, otherwise Plitron
transformers would be very long and thin.

cheers, Ian

Ian Iveson wrote:

The graph on page 230 (fig 5.15) doesn't fit the definition of µ
given on the same page. The authors emphasise that its value is B/H,
not dB/dH. The peak value should therefore occur at the knee of the
BH curve, from where its tangent passes through the origin. RDH has
got the peak in the right place, above the knee rather than above
the steepest part of the BH curve, which is where it would be were µ
to be dB/dH.

But the right hand side of the curve is wrong by miles. µ is shown
to plummet towards zero as the gradient of the BH curve flattens
towards saturation. This would be true if µ were dB/dH, but if it is
B/H then it should drop more gradually with an asymptote along the B
axis.

The shape of the curve shown on page 216 (fig 5.13D) fits their
definition better, but that is for AC, which is another can of
worms.

Checking several sites and books, there is no agreement on whether µ
is B/H or dB/dH.

For example, Menno van der Veen says that µ reflects the mobility of
the magnetic domains. This observation seems to favour the dynamic
definition, because at saturation there is no mobility even though
the value of µ may still be high according to the static RDH
definition. He shows a graph of inductance versus secondary voltage,
in which the inductance plummets to near zero at saturation, which
is what I would expect. He also correctly points out that µ is the
only variable in the equation for inductance for a given inductor.
Hence where the inductance falls to near-zero, so must µ.

Further, µ0 is used for the permeability of free space, and also for
initial permeability, as in the table on page 208. The unit for µ
and µ0 is also generally omitted. It does have a footnote that says
"strictly" the unit is gauss/oersted. It also states that µ0 = 1.
Actually its unit depends on what system of units you are using, as
does its value. In a table using both gauss and lines per inch, the
unit of µ should be stated.

The confusion is everywhere. It has infected various spice core
models that don't work properly. Much data is barely intelligible
because units and definitions are not clear.

Incidentally, Van der Veen, of Plitron fame, gets the formula for
inductance wrong in his book, transposing core area and magnetic
path length. I assume this is a misprint, otherwise Plitron
transformers would be very long and thin.

cheers, Ian

I agree its difficult to follow what mu is all about.
Parts of RDH4 are rather difficult to follow.

Its too late here to go into all the sordid details.

If RDH is confused about u,
I can sympathise with the author,
since he was only human.

Patrick Turner.

Ian Iveson wrote:

The graph on page 230 (fig 5.15) doesn't fit the definition of =B5
given on the same page. The authors emphasise that its value is B/H,
not dB/dH. The peak value should therefore occur at the knee of the
BH curve, from where its tangent passes through the origin. RDH has
got the peak in the right place, above the knee rather than above
the steepest part of the BH curve, which is where it would be were =B5
to be dB/dH.

But the right hand side of the curve is wrong by miles. =B5 is shown
to plummet towards zero as the gradient of the BH curve flattens
towards saturation.

At saturation the mu would be one, as it is in air or non-magnetic
materials.
OTOH, it should be shown somewhat higher than it is while crossing the
origin.

Looks like someone has just done a fast sketch of the variables.

In some magnetic materials such as used in magnetic core memory the mu
at the origin can be quite low. JLS

This would be true if =B5 were dB/dH, but if it is
B/H then it should drop more gradually with an asymptote along the B
axis.

The shape of the curve shown on page 216 (fig 5.13D) fits their
definition better, but that is for AC, which is another can of
worms.

Checking several sites and books, there is no agreement on whether =B5
is B/H or dB/dH.

For example, Menno van der Veen says that =B5 reflects the mobility of
the magnetic domains. This observation seems to favour the dynamic
definition, because at saturation there is no mobility even though
the value of =B5 may still be high according to the static RDH
definition. He shows a graph of inductance versus secondary voltage,
in which the inductance plummets to near zero at saturation, which
is what I would expect. He also correctly points out that =B5 is the
only variable in the equation for inductance for a given inductor.
Hence where the inductance falls to near-zero, so must =B5.

Further, =B50 is used for the permeability of free space, and also for
initial permeability, as in the table on page 208. The unit for =B5
and =B50 is also generally omitted. It does have a footnote that says
"strictly" the unit is gauss/oersted. It also states that =B50 =3D 1.
Actually its unit depends on what system of units you are using, as
does its value. In a table using both gauss and lines per inch, the
unit of =B5 should be stated.

The confusion is everywhere. It has infected various spice core
models that don't work properly. Much data is barely intelligible
because units and definitions are not clear.

Incidentally, Van der Veen, of Plitron fame, gets the formula for
inductance wrong in his book, transposing core area and magnetic
path length. I assume this is a misprint, otherwise Plitron
transformers would be very long and thin.

cheers, Ian

"John Stewart" wrote

At saturation the mu would be one, as it is in air or non-magnetic
materials.

Yes...or 4pi * 10e-7, or whatever, depending on units...

OTOH, it should be shown somewhat higher than it is while crossing
the
origin.

Maybe, but the right hand side is a real mess.

You would think that after making the point that the value of µ is
given by the angle of a straight line from the origin, and after
drawing a diagram to show an example line that crosses the BH curve
at two points, they would have made the value of µ equal at those
two points. Sigh...

Looks like someone has just done a fast sketch of the variables.

To demonstrate two features, one of which is totally wrong.

And how is this reconciled with the common equation for inductance
(L=µ*N^2*A/L)? If µ continues to have a high value beyond
saturation, then so does L. In which case the sudden increase in AC
current does not reflect an equally sudden fall in L with respect to
H, but a dynamic process in which increasing current results in an
increase in H, resulting in a small decrease in µ, resulting in more
current, etc. So the sudden fall in inductance only happens with
respect to steadily increasing driving *voltage*, not current. Also
µ can be negative, since the AC BH curve crosses all four quadrants.

Perhaps that's why µ gets confused with dB/dH

This is all your fault. I have made a model that includes hysteresis
and eddy current losses, but can't find reliable and sufficient
matching data to check the results. I have come across stuff on
variation of winding resistance in a magnetic field, but no useful
data. The preoccupation of transformer designers is with HF SMPS
with composite cores.

cheers, Ian

Ian Iveson wrote:
The graph on page 230 (fig 5.15) doesn't fit the definition of µ
given on the same page.....

cheers, Ian




I am the silly fellow who is (still) converting the RDH4 into
electronic-text.

I will occasionally look for posts of this nature (concerning mistakes
in RDH4) and save the opening post, but not all the rest. I found P.
Turner's page 523 one previous, for example. Some initial posts get too
many responses/off topic to store easily. As the individual chapters are
worked up, these matters can be revisited again. The corrections will go
onto the back of the book, and be linked both ways with the pages they
concern.

As for "the effort", I had not touched it for a few weeks there but
finished page 500 as of today.
~~~~~~~

DougC wrote:

Ian Iveson wrote:
The graph on page 230 (fig 5.15) doesn't fit the definition of µ
given on the same page.....

cheers, Ian




I am the silly fellow who is (still) converting the RDH4 into
electronic-text.

I will occasionally look for posts of this nature (concerning mistakes
in RDH4) and save the opening post, but not all the rest. I found P.
Turner's page 523 one previous, for example. Some initial posts get too
many responses/off topic to store easily. As the individual chapters are
worked up, these matters can be revisited again. The corrections will go
onto the back of the book, and be linked both ways with the pages they
concern.

As for "the effort", I had not touched it for a few weeks there but
finished page 500 as of today.

Only 1,100 pages to go.

Maybe you deserve a medal when you do the final saving "click".

Patrick Turner.


~~~~~~~