Connect with us

Magnetic Flux Density Depending on Current

Discussion in 'Electronic Design' started by D from BC, Mar 25, 2007.

Scroll to continue with content
  1. D from BC

    D from BC Guest

    I don't know what it is about magnetics..it just never sinks in.. :(

    Here's what's on my inductor:
    Voltage: 240Vpp 40% duty square wave @ 100khz
    Current: +/-100mA triangle wave hovering around 1.5amps.
    L=1.3mH
    u=125

    Problem:
    1) What's Bmax
    2) What's delta B

    I thought about solving this is terms of voltage using
    Bmax=(Vrms*10E8)/4.44fAN
    But the Vrms is supposed to be from a sine wave...I have squareish.

    However, it looks easier to solve in terms of current.
    Neglecting the ripple current.. (low impact of accuracy)
    and just calling it 1.5Amps of DC current and using:

    magnetizing force H
    expressed by:
    H=(0.4*pie*NI)/l
    N=turns
    I=peak current Amps
    l=magnetic path length cm

    along with u = B/H
    (u=125)

    B can be solved. (Gauss)
    Hopefully to be less than the datasheet core sat.

    Delta B would be from the +/-100mA or 200mApp.
    200mA is put into the B=uH equation.
    I'd use that to look up the core loss on the datasheet.
    (The more Bchange the more core loss.)

    Then I found this eqn:
    Bdc=(L*Idc*10E8)/(N*Ac)
    L=inductance
    Idc=DC amps
    N=turns
    Ac=cross section cm^2
    Which should give the same results.

    Have I goofed up somewhere?
    D from BC
     
  2. John  Larkin

    John Larkin Guest

    Magnetics datasheets are a pain. For things like pot cores and
    toroids, for a given material, why don't they give us gauss per
    ampere-turn? That, plus Al which they usually *do* supply, would be a
    big help.

    John
     
  3. Fred Bloggs

    Fred Bloggs Guest


    Pick one terminal of inductor as reference so that for flux balance you
    must have VH*0.4+VL*0.6=0 and VH-VL=240 making VH=240*0.6=144V and
    VL=144-240=-96V. Then you have 144V=N*d/dt(phi)=N*d/dt(Aeff*B) so that
    dB=144*dt/(N*Aeff)=144*0.4/(N*Aeff*100E3). Then Bmax=Bdc+dB where
    Bdc=N*Idc*u(1.5A) where u(1.5A) denotes permeability of core at 1.5A dc
    bias.
     
  4. Fred Bloggs

    Fred Bloggs Guest

    Actually that's a bit conservative. Since flux balance requires that
    +dB=|-dB|, or equivalently, Bdc=(Bmax+Bmin)/2, then since Bmax=Bmin+db
    and vice versa, you have Bmax=(Bmax+Bmin)/2+(Bmax-Bmin)/2=Bdc+dB/2.
     
  5. legg

    legg Guest

    B units of Tesla, area units in square meters, inductance units of
    henries - knocks out the 10E8 multiplier from this formula, (if it is
    the correct multiplier with the original units).

    DC bias has no frequency relation, so the earlier sinusoidal formula
    is misapplied in the average static case, regardless of its potential
    suitability elsewhere..

    B= n . I . ue / le in teslas amps and meters.

    For gapped structures, out of saturation:

    B = n . I . uo / lg

    l in amps
    uo = 4 . pi . 10E-7
    lg = gap length meters

    AC flux, as always, is

    deltaB = E . t / ( n . a )

    volts, seconds, meters and Teslas,
    Half the delta is arithmetically additive/subtractive to the DC flux
    value.

    RL
     
  6. D from BC

    D from BC Guest

    In this case, just using the nearly DC low ripple current waveform
    through the inductor, I could simply use the relation:

    B=(n.I.ue)/le
    ue is that the permeability?
    le is the the magnetic path length?

    I guess this would be useful too if I ever needed to make an
    electromagnet powered by a DC current source.

    D from BC
     
  7. D from BC

    D from BC Guest


    Oops my OP is a little fuzzy...

    There's no current bias through the inductor. The inductor current
    waveform is the result of being in a continuous mode converter. The
    converter doesn't let the inductor current drop to zero.
    Using the inductor voltage waveform in a B calculation and then adding
    on B from the inductor current waveform doesn't seem right.
    But ok for dc bias as described.

    Bdc=N*Idc*u
    Is this a lean equation?..
    No other inductor parameters like Across and lmag path length??
    But then, I'm ok with B error up to 20%.

    Note: I'm watching the u and aware it can be current dependent.
    D from BC
     
  8. legg

    legg Guest

    yes - the permeability of the medium used to store the energy.
    In gapped media, the gap will tend to dominate, hence dominance of the
    gap length.

    In homogenous media the full path length carries the permeability
    of the material used. This may degrade with direct current bias.
    An electromagnet develops a useful external field in the (varying)
    gap.The field does not have to be DC; is often more effective with AC
    due to the potential reduction in residual magnetism.

    RL
     
  9. Tim Williams

    Tim Williams Guest

    I wish I knew. Just what *IS* the purpose of quoting Bmax for ferrite,
    anyway? That doesn't help me a damn when I want to measure the fucker.
    Geometry is fixed, give me amp-turns, vs. gap if need be!

    Tim
     
  10. D from BC

    D from BC Guest

    I'm thinking of the electromagnet used by the auto wreckers to pick up
    cars.. AC or DC?
    So an electromagnets core might stay a little magnetized after a shot
    of DC.
    Wouldn't an AC powered electromagnet just be kinda jiggy when
    attracting say some iron..
    Reminds me of those buzzing solenoids for doors.
    D from BC
     
  11. Fred Bloggs

    Fred Bloggs Guest

    Not exactly, I was thinking of the case where the manu plots B vs H at
    DC for you.
     
  12. legg

    legg Guest

    The function is the same for a shaped field, and the average forces
    applied average out.
    There are many ways to configure electronic noise makers, not all of
    them intentional. No doubt German-made models would be configured to
    produce a more 'noble' sound, if noise was actually considered to be
    neccessary to do the job of unlatching an electronic door.

    RL
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-