[sci.electronics] Ferrite core loss vs. frequency

rkarlqu@hpscdc.HP.COM (Rick Karlquist) (05/06/89)

I notice in ferrite core catalogs that core loss (for a fixed value of B)
is always approximately proportional to switching frequency 
squared.  Assuming that the dominant loss mechanism is hysteresis,
it seems to me that each time you go around the B-H hysteresis loop
you give up a certain amount of energy.  In fact, the number
of joules of energy is equal to the area of the B-H curve.
If you double the frequency, you go around the curve twice as
many times per second resulting in twice as much energy lost
per second.  Since power is energy per second, twice as much
core loss power should be expended (not four times as much).
Where is my thinking wrong here?:

a.  Hysteresis is *not* the dominant loss mechanism (?)
b.  The B-H loop gets bigger if you traverse it more quickly (?)
c.  Some other misconception

                             Rick Karlquist
                             HP Santa Clara