Slimelite BBS

Inductance of solonoid

Back to AMIGA_VIDEO

From Ron Freimuth · 1996-04-02 12:20 · #76594 in AMIGA_VIDEO

    Well, I said I was gonna try to figure the inductance per unit
 length of an infinity long solenoid.  The result should be a good
 approximation for any solenoid where L >> a.

    I know L should vary directly with volume (per length).

    I will try to figure the total magnetic energy stored per unit
 length at one ampere in the solenoid.  Then, I can equate W = 1/2 L
 I^2.

    People familiar with the field around a solenoid will be able to
 fill in the missing flux arrows below (hard to draw an arrow at 45
 deg).
                    ->         ->
            ->                             ->
    ->                                              ->

    ->  ->  ->   ->  ->  ->   ->  ->  ->   ->  ->  ->   ->
       xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx I
    <-  <-  <-   <-  <-  <-   <-  <-  <-   <-  <-  <-   <-      | Area
H<- <- -------------------------------------------------  A is  | of
    <-  <-  <-   <-  <-  <-   <-  <-  <-   <-  <-  <-   <-      | circle
      ...............................................a          | Pi a^2
    ->  ->  ->   ->  ->  ->   ->  ->  ->   ->  ->  ->   ->

    ->                                              ->
           ->                               ->
                    ->          ->

    'x' represents current into the page, '.', out.

    H is measured in A/m. In this case it is to the left.

    H = I*p     ; p = pitch, turns per meter
    A, Area of solenoid is Pi a^2

    This is hardly rigorous.  However, we might expect H inside the
 coil to equal I*p.  Also, assume H is constant inside the coil.

    W = B dot H ;magnetic energy per cubic meter.
    B = mu0 H   ;for free space

    therefore

    W/m^3 = H^2 mu0 ; mu0 = 4 Pi e-7

    Thus, the energy stored per unit length is simply

    W/m^3 * A   ;A = a^2 Pi (m^2)
    W/l = mu0 H^2 Pi a^2
        = mu0 I^2 p^2 Pi a^2

    Doubling I doubles H, which quadruples energy stored per unit
 volume, thus unit length.  Doubling the turns pitch, p has the same
 effect, since twice as much H is developed per unit length.  The energy
 also varies with radius, a squared.

    Note that W(ork)/sec = RI^2 in a circuit.  Somewhat analogously,
    W/l = mu0 p^2 I^2 for the magnetic field energy.

    What about the energy stored outside the coil?  Well, I remember a
 CHEAT: since H can spread out and return through an infinite distance,
 it approaches zero intensity on long flux return lines.  W varies with
 H^2, it ends up NO Energy is stored outside the solenoid!

    Thus, we need only equate

    1/2 L I^2 = W =  mu0 p^2 I^2 Pi a^2
    1/2 L = mu0 p^2 Pi a^2
    L = 4Pi 1e-7 p^2 Pi a^2
       ________________________________
****** L = 4/10 Pi^2 p^2 a^2 micro Hy/m ************
****** ________________________________ ************

    Note that L varies with vol/l as I said at the start.  Due to the
 a^2 term.

    Lets figure the inductance for a 10 cm long section of an infinite
 solenoid.  With a diameter of 1 cm and pitch of 10 turns per cm.

    l = 0.1 m                   3.96"
    a = 0.005 m                 0.198"
    p = 10*100 = 1000 turns/m   N = 100 turns total

    L = 0.1*4/10 *9.87*1e6*25e-6* micro Hy
    L = 9.87 micro Hy    (0.0000987 Hy/m)

    Let me check against a simple formula (uses inches)

        a^2 N^2
    L = --------    ; takes into account 'end effects'
        9a + 10l

    L =  0.198"^2 * 100^2
        --------------------    =  9.47 micro Hy
         9*0.198" + 10*3.96"

    HOT DAMN!

    Using lots of hand waving I get a theoretical value for inductance
 per meter.  Then, I figure the inductance of 0.1 m length of a specific
 diameter an turns per meter.  Finally, I convert the MKS units to
 inches so I can substitute in a common inductance I remember.  Look at
 how well the results compare:

    9.87 micro Hy per 0.1 m     Exact for an infinite length solenoid
    9.47 micro Hy     Approximation, takes into account finite length.

    They should compare well, since I figured a long (l/d = 10) coil.
 Where the reduction of inductance due to flux spreading out at the ends
 of a finite length coil is minimal.

    Note the denominator of the empirical equation is 9a + 10l.  For
 small a this reduces to 10l.  Lets try figuring L in that case.

    L = 9.90 micro Hy.  Very close to 9.87, the value for negligible
 effect from radius.  Anyway, the empirical formula clearly rounds off
 the coefficients '9' and '10'.  BTW, it's supposed to be good to about
 5% for appropriate a and l.  I used it a lot in figuring inductance of
 small, air wound RF coils.

    BTW, I did this all cold.  No cheating by checking text books.  All
 I had to remember were a few tricks from Freshman Physics in 1962.

    Calculating the capacitance per unit area of two plates in a vacuum
 is simpler.  Anyone care to try?


    Another simple calculation is for the inductance per unit length of
 a coaxial line. One can use a similar approach, figuring H(r), then
 energy stored per unit length.  Capacitance per unit length is also
 straightforward.  I think both vary as Log Di/Do.

    One might then develop the wave equation for propagation on a
 distributed L/C transmission line.  It's fair to assume a solution and
 simply solve for the coefficients.  Using the results for L and C, one
 should be able to show that voltage and current waves propagate along
 the line at +-c.  For a cable with a vacuum dielectric.

    I more or less did this one afternoon working at WHA while attending
 UW.  Many, many years ago.

---
 * OLX 2.1 TD *  The fault lies not in our stars but in ourselves
--- FidoPCB v1.5 beta-'j'
 * Origin: Lunatic Fringe * Richardson,TX * 214-235-5288 * (1:124/2113)