below

FM-Frequency Modulation bandwidth

and the proof of the BESSEL's Series
Leonardo Rubino
February 2023

Abstract: here is a proof of the Bessel’s Series used to figure out the bandwidth in the Frequency
Modulation-FM.

Contents:

GENERAL VIEW____________________________________________________________Pag. 1

Appendix 1 - EULER GAMMA FUNCTION ______________________________________Pag. 3

Appendix 2 – BESSEL’S DIFFERENTIAL EQUATION ____________________________Pag. 3

Appendix 3 - REFLECTION FORMULA _________________________________________Pag. 5

Appendix 4 - BESSEL’S SERIES _______________________________________________Pag. 6

GENERAL VIEW

FM-Frequency Modulation:

Carrier: v(t )  VP cos   VP cos(t   ) , Modulating signal: vm (t )  Vm cos  m t

The modulated pulsation is:    P  K f Vm cos  m t , so:
Kf
f  fP  Vm cosmt  f P  f cos mt ; f is the frequency deviation. In general:
2
d
  P  K f vm (t ) , so:
dt
f
   dt  2  fdt  2  ( f P  f cos mt )dt  2 [ f Pt  sin 2f mt ] , and:
2f m
f
v(t )  VP cos  VP cos[2f Pt  sin 2f mt ] . (1.1)
fm
f
mf  is the Modulation Index.
fm
We develop the above (1.1) in Bessel’s Series (see (4.3) and (4.4) in Appendix 4, here reported):

cos( x sin  )  J 0 ( x)  2 J 2 ( x) cos 2  2 J 4 ( x) cos 4  ......

sin( x sin  )  2 J1 ( x) sin   2 J 3 ( x) sin 3  2 J 5 ( x) sin 5  ......

( cos(a  b)  cos a cos b  sin a sin b )

Well, v(t )  VP cos(Pt  m f sin mt )  VP [cos Pt cos(m f sin mt )  sin Pt sin(m f sin mt )] 
 VP cos Pt[ J 0 (m f )  2 J 2 (m f ) cos 2mt  2 J 4 (m f ) cos 4mt  ....] 
 VP sin Pt[2 J1 (m f ) sin mt  2 J 3 (m f ) sin 3mt  .....]
1 1 1 1
Now, as: cos a  cos b  cos(a  b)  cos(a  b) and sin a  sin b  cos(a  b)  cos(a  b) ,
2 2 2 2
it follows that:
v(t )  VP [ J 0 (m f ) cos Pt  J1 (m f ) cos(P  m )t  J1 (m f ) cos(P  m )t  J 2 (m f ) cos(P  2m )t 
 J 2 (m f ) cos(P  2m )t  J 3 (m f ) cos(P  3m )t  J 3 (m f ) cos(P  3m )t  .....] (1.2)
This equation (1.2) gives us the spectrum of the modulated signal; it has the carrier and the side
components, spaced by m and whose amplitudes are given by the Bessel functions which depend
on m f . As usually small m f (<0,5) give small values of the Bessel functions, then only the side
rows at freq. f P  f m and f P  f m are considered. As J 0 (m f )  0 when m f  2,40 , there the
carrier disappears.
On the basis of all that, the bandwidth is approximately B  2(f  f m ) .

The expressions for the Bessel Functions J n (x) are given by the following (see (2.2) below):

(1) k ( x 2) n  2 k
J n ( x)   and are represented in the above image. For instance:
k  0 ( n  k  1) k!
 x2 x4 x6
J 0 ( x)  1     ....... . In order to figure out the value of such a J 0 ( x) by
2 2 2 4 (2!) 2 26 (3!) 2
adding terms of a series, do not worry, because as long as k gets high enough, the terms will start
being negligible.

Appendix 1: EULER GAMMA FUNCTION


( z )   e  t t z 1dt
0
Of course (1)  1 . Then, (n  1)  n(n) ; in fact, after an integration by parts:

tz t 1  t z 1
( z )  e  e t dt  0  ( z  1) and by iterating the (n  1)  n(n) , we get:
z 0
z 0 z
(n  1)  n! .

Stirling’s Formula:
 
(n  1)   x n e  x dx   e n ln x  x dx and the function n ln x  x has a maximum in x=n, as you
0 0
can see by putting to zero its derivative. Now, after saying x=n+y, we have:
  n ln n  n ln(1 y )  y  n ln(1 y )  y
(n  1)  e  n  e n ln( n  y )  y dy  e  n  e n
dy  n n e  n  e n
dy
n n n

x 2 x3
Now, according to Taylor, we have that: ln(1  x)  x    .... ; we put first x=y/n and later
2 3
2 3
 n ln(1 y )  y  y y  ......
n n n n n 2n 3n 2
y  nv , so: ( n  1)  n e  e dy  n e  e dy 
n n
2 3
 v v  ......
n n 2
n e n e 3 n
dv , and if n is large enough: (n  1)  2n n ne  n (Stirling’s Formula),
 n

as the last Gauss integral was 2 .

Appendix 2: BESSEL’S DIFFERENTIAL EQUATION

When you face a differential equation like this:

x 2 y ' ' xy' ( x 2  n 2 ) y  0 (2.1)
you have a Bessel Differential Equation.
If we look for a solution as a series J n (x) , then the differential equation becomes:
(1) k ( x 2) n  2 k 
x 2 J ' 'n  xJ 'n  ( x 2  n 2 ) J n  0 and the solution series is: J n ( x)   , (2.2)
k  0 ( n  k  1) k!
where  is the Gamma Euler Function. In fact, if you insert the (2.2) into the (2.1), you get 0=0:

(1) k ( x) n  2 k 
(1) k (n  2k )( x) n  2 k 1 
(1) k (n  2k )(n  2k  1)( x) n  2 k  2
y   n 2k , y'   , y ' '  
k  0 ( 2) (n  k )!k! k 0 (2) n  2 k (n  k )!k! k 0 (2) n  2 k (n  k )!k!

(1) k ( x)n  2 k  2 
(1) k n 2 ( x) n  2 k 
(1) k (n  2k )( x) n  2 k
( x 2  n2 ) y   n2k
  , xy '   ,
k  0 ( 2) (n  k )!k! k  0 (2) n  2 k (n  k )!k! k 0 (2) n  2 k (n  k )!k!
2

(1) k (n  2k )(n  2k  1)( x) n  2 k
x y' '  
k 0 (2) n  2 k (n  k )!k!
 2 2 2 (1) k (n  2k )(n  2k  1)( x) n  2 k  (1) k (n  2k )( x) n  2 k

x y ' ' xy '( x  n ) y    
k 0 (2) n  2 k ( n  k )!k! k 0 (2) n  2 k ( n  k )!k!

(1) k ( x) n  2 k  2 
(1) k n 2 ( x) n  2 k 
(1) k ( x) n  2 k  2 
(1) k [4k (k  n)]( x) n  2 k
 n  2k
      (2)n  2k (n  k )!k! 
k  0 ( 2) (n  k )!k! k  0 (2) n  2 k (n  k )!k! k  0 (2) n  2 k (n  k )!k! k  0

(1) k 1 ( x) n  2 k 
(1) k 4( x) n  2 k 
(1) k 4( x) n  2 k
 n  2k  2
     n  2k

k 1 ( 2) (n  k  1)!(k  1)! k  0 (2) n  2 k (n  k )! k! k 1 ( 2) (n  k  1)!(k  1)!
(n  k ) k

(1) k 4( x) n  2 k
  n2k , so:
k  0 ( 2) (n  k  1)!(k  1)!

(1) k 4( x) n  2 k 
(1) k 4( x) n  2 k
x 2 y ' ' xy '( x 2  n 2 ) y   n  2k
  , but we can also
k 1 (2) (n  k  1)!(k  1)! k  0 (2) n  2 k (n  k  1)!(k  1)!

(1) k 4( x) n  2 k 
(1) k 4( x) n  2 k
say that: x 2 y ' ' xy '( x 2  n 2 ) y   n  2k
   0,
k 1 ( 2) (n  k  1)!(k  1)! k 1 (2) n  2 k (n  k  1)!(k  1)!
where also the latter summation can start from k=1, as its term in k=0 has in its denominator:
(k  1)! (k )  (1)! (0)   .
--------------------------------------
n
We see that: J  n ( x)  (1) J n ( x) ; in fact:

(1) r ( x 2)  n  2 r n 1 (1) r ( x 2)  n  2 r  (1) r ( x 2)  n  2 r
J  n ( x)    
r  0 r! ( n  r  1) r  0 r!(  n  r  1) r  n r! (  n  r  1)
but as (n  r  1)   for every r = 0,1,…..,n-1, the first summation is zero. Moreover, in the
second summation we put r = n+k , so:

(1) n  k ( x 2) n  2 k 
(1) k ( x 2) n  2 k
  (1) n   (1) n J n ( x) , qed.
k  0 ( n  k )! ( k  1) k  0  ( n  k  1) k!
The denominators in both summations are the same after considering that: (n  1)  n(n) .
2
We also have that: J 'n ( x) J  n ( x)  J ' n ( x) J n ( x)  (2.3)
x(n)(1  n)
In fact, as both J n (x) and J  n (x) satisfy the Bessel Equation, then:
x 2 J ' 'n  xJ 'n  ( x 2  n 2 ) J n  0 and x 2 J ' ' n  xJ ' n ( x 2  n 2 ) J  n  0 and by multiplying the
former by J  n and the latter by J n and finally after subtracting side from side, we get:
x 2 ( J ' 'n J  n  J ' ' n J n )  x( J 'n J  n  J ' n J n )  0 , or:
d d
x ( J 'n J  n  J ' n J n )  ( J 'n J  n  J ' n J n )  0 , or again: [ x( J 'n J  n  J ' n J n )]  0 
dx dx
c
J ' n J  n  J ' n J n  . (2.4)
x
xn x n 1
Now, in order to calculate c, we remind that: J n   ..... , J ' n   ..... ,
2n (n  1) 2 n  (n)
x  n 1 xn
J ' n   ..... , J n   ..... , and all these in (2.4) yield:
2  n  (  n) 2  n (n  1)
(((.......(n  1)  n(n).......)))
1 1 1 1 1 1
c     
(n)(1  n) (n  1)(n) (n)(1  n) [n(n)]( n) (n)(1  n) (n)[n(n)]
 1 1 1 1 2
     , so:
(n)(1  n) (n)[ n( n)] (n)(1  n) (n)(1  n) (n)(1  n)

2 2
c . So: J 'n J  n  J ' n J n  .
(n)(1  n) x(n)(1  n)

Appendix 3: REFLECTION FORMULA

2 2 sin n
Here it is:  . (3.1)
(n)(1  n) 
sin x
In fact, we know that is zero in   , ,2 ,2 ,...., n , n , so:
x
sin x x x x x x2 x2 
x
 (1  )(1  )(1  )(1  )..........  (1  2 )(1  2 ).....  [1  ( ) 2 ] , and so:
x   2 2  4 k 1 k

x x
 [1  ( ) 2 ]1 (3.2)
sin x k 1 k
Moreover, we can see a factorial like this:
k! k n (k  1)(k  2)...(k  n)
n! lim  lim ; in fact, the denominator of the first limit is
k   ( n  1)...( n  k ) k   kn
(n  k )!
(n  1)...(n  k )  , while from both numerators we can collect: k!(k  1)...(k  n)  (k  n)! ,
n!
(k  1)(k  2)...(k  n)
so just n! is left. Furthermore, we see that the second limit is 1: lim  1 , as it
k  kn
kn k!k n
gets of the form lim n . Therefore: n! lim . As ( x  1)  x! , we have:
k  k k   (n  1)...(n  k )

k!k x
( x  1)  x! lim and after dividing both numerator and denominator by k!:
k   ( x  1)...( x  k )

kx
( x  1)  lim . (3.3)
k   (1  x )(1  x 2) ...(1  x k )

1 1 1
Now, let’s introduce the constant  k  1    ...   ln k , and let’s say:   lim  k ; on the
2 3 k k 
x x
 x ex e 2 e k
basis of that, the (3.3) becomes: ( x  1)  e lim .... . (3.4)
k   (1  x ) (1  x 2) (1  x k )
x x
In fact, all the e x , e 2 ,…, e k cancel with terms in e x and the term k x  e(ln k ) x still comes from
e x . Moreover, as ( x  1)  x( x) , then the (3.4) becomes:

1 (1  x) (1  x 2) (1  x k ) x x
 xex lim x x
.... x
 xe x
 [1  ( )]e k (3.5)
( x ) k  e e 2 e k k 1 k
1
Now, as ( y  1)  y( y ) , then ( y )  ( y  1) and after replacing y by –x:
y
1
( x)   (1  x) and so: (1  x)   x( x) . Let’s figure out ( x)(1  x) :
x
 
x x x x 1  x
( x)(1  x)  [ x 1e  x  (1  ) 1e k ]  [ex  (1  ) 1e k ]  [ A]  [ B]   [1  ( ) 2 ]1 , (3.6)
k 1 k k 1 k x k 1 k
where A is an inverted (3.5) and B is the (3.4) with x changed into –x. So, (3.6) is:
1  x x 
x
( x)(1  x)   [1  ( ) 2 ]1 , and for the (3.2) ((  [1  ( ) 2 ]1 )) we have:
x k 1 k sin x k 1 k
2 2 sin n
 which is the (3.1) indeed.
(n)(1  n) 

Appendix 4: BESSEL’S SERIES

2
Now, as according to (2.3): J 'n ( x) J  n ( x)  J ' n ( x) J n ( x)  and using in it the (3.1),
x(n)(1  n)
2 sin n
we get: J 'n ( x) J n ( x)  J 'n ( x) J n ( x)  . (4.1)
x
x 1 
(t  )
The following equality holds: e 2 t
 J n ( x)t n ; (4.2)
n  
x 1 xt x
(t  )  
( xt 2) r  ( x 2t )k  
(1) k ( x 2)( r  k ) t ( r  k )
in fact: e
2 t
e e 2 2t
 [ ][ ] 
r 0 r! k 0 k! r 0 k 0 r! k!
and defining n=r-k, it follows that n will go from   to   , so:
x 1
(t  )  
(1) k ( x 2)( n  2 k ) t n  
(1) k ( x 2)( n  2 k ) n 
e 2 t
     [ ]t   J n ( x)t n qed.
n   k 0 (n  k )! k! n   k  0 k!(n  k )! n  

----------------------------------------

The following equalities (Bessel’s Series) hold:

cos( x sin  )  J 0 ( x)  2 J 2 ( x) cos 2  2 J 4 ( x) cos 4  ...... (4.3)

sin( x sin  )  2 J1 ( x) sin   2 J 3 ( x) sin 3  2 J 5 ( x) sin 5  ...... (4.4)

In fact, if we put t  ei into the (4.2), we get:

1 
x ( e i  e  i )
e 2
e ix sin x
  J n ( x)ein 


  J n ( x)[cos n  i sin n ]  {J 0 ( x)  [ J 1 ( x)  J1 ( x)] cos   [ J  2 ( x)  J 2 ( x)]cos2  ....} 


 i{[ J1 ( x)  J 1 ( x)]sin  [ J 2 ( x)  J 2 ( x)]sin2  ....} 

 {J 0 ( x)  2 J 2 ( x)cos2  ....}  i{2 J1 ( x) sin   2 J 3 ( x)sin3  ....} , and making equal real and

imaginary parts just obtained with those from eix sin x , we have (4.3) and (4.4) indeed.
 FM-Modulazione di Frequenza-larghezza di banda
e dimostrazione della Serie di BESSEL
Leonardo Rubino
Febbraio 2023

Abstract: ecco una dimostrazione della Serie di Bessel utilizzata per valutare la larghezza di banda
nella Modulazione di Frequenza-FM.

Indice:

CONCETTI GENERALI______________________________________________________Pag. 7

Appendce 1 – FUNZIONE GAMMA DI EULERO__________________________________Pag. 9

Appendce 2 – EQUAZIONE DIFFERENZIALE DI BESSEL_________________________Pag. 9

Appendce 3 – FORMULA DI RIFLESSIONE_____________________________________Pag. 11

Appendce 4 – SERIE DI BESSEL ______________________________________________Pag. 12

CONCETTI GENERALI

FM-Modulazione di Frequenza:

Portante: v(t )  VP cos   VP cos(t   ) , Segnale modulante: vm (t )  Vm cos  m t

La pulsazione modulata è:    P  K f Vm cos  m t , dunque:
Kf
f  fP  Vm cosmt  f P  f cos mt ; f è la deviazione di frequenza. In generale:
2
d
  P  K f vm (t ) , così:
dt
f
   dt  2  fdt  2  ( f P  f cos mt )dt  2 [ f Pt  sin 2f mt ] , e:
2f m
f
v(t )  VP cos  VP cos[2f Pt  sin 2f mt ] . (1.1)
fm
f
mf  è l’Indice di Modulazione.
fm
Sviluppiamo la (1.1) qui sopra in Serie di Bessel (vedi le (4.3) e (4.4) in Appendce 4, qui riportate):

cos( x sin  )  J 0 ( x)  2 J 2 ( x) cos 2  2 J 4 ( x) cos 4  ......

sin( x sin  )  2 J1 ( x) sin   2 J 3 ( x) sin 3  2 J 5 ( x) sin 5  ......

( cos(a  b)  cos a cos b  sin a sin b )

Bene, v(t )  VP cos(Pt  m f sin mt )  VP [cos Pt cos(m f sin mt )  sin Pt sin(m f sin mt )] 
 VP cos Pt[ J 0 (m f )  2 J 2 (m f ) cos 2mt  2 J 4 (m f ) cos 4mt  ....] 
 VP sin Pt[2 J1 (m f ) sin mt  2 J 3 (m f ) sin 3mt  .....]
1 1 1 1
Ora, siccome: cos a  cos b  cos(a  b)  cos(a  b) e sin a  sin b  cos(a  b)  cos(a  b) ,
2 2 2 2
segue che:
v(t )  VP [ J 0 (m f ) cos Pt  J1 (m f ) cos(P  m )t  J1 (m f ) cos(P  m )t  J 2 (m f ) cos(P  2m )t 
 J 2 (m f ) cos(P  2m )t  J 3 (m f ) cos(P  3m )t  J 3 (m f ) cos(P  3m )t  .....] (1.2)
Questa equazione (1.2) ci fornisce lo spettro del segnale modulato; essa contiene la portante e le
component laterali, spaziate di m e le ampiezze delle quali sono date dalle funzioni di Bessel, che
dipendono da m f . Siccome, di solito, piccoli m f (<0,5) danno piccolo valori delle funzioni di
Bessel, allora solo le righe laterali a frequenze f P  f m e f P  fm sono considerate. Siccome
J 0 (m f )  0 quando m f  2,40 , lì la portante scompare.
Sulla base di tutto ciò, la larghezza di banda è approssimativamente B  2(f  f m ) .

Le espressioni per le Funzioni di Bessel J n (x) sono date dalla seguente (vedi la (2.2) più sotto):

(1) k ( x 2) n  2 k
J n ( x)   e sono rappresentate nella figura qui sopra. Per esempio:
k  0 ( n  k  1) k!
 x2 x4 x6
J 0 ( x)  1     ....... . Nel calcolare il valore di tale J 0 ( x) sommando termini
2 2 2 4 (2!) 2 26 (3!) 2
di una serie, non preoccupatevi, poichè appena k cresce un po’, i termini iniziano ad essere
trascurabili.

Appendice 1: FUNZIONE GAMMA DI EULERO


( z )   e  t t z 1dt
0
Naturalmente, (1)  1 . Poi, (n  1)  n(n) ; infatti, dopo una integrazione per parti:

tz t 1  t z 1
( z )  e  e t dt  0  ( z  1) ed iterando la (n  1)  n(n) , otteniamo:
z 0
z 0 z
(n  1)  n! .

Formula di Stirling:
 
(n  1)   x n e  x dx   e n ln x  x dx e la funzione n ln x  x ha un massimo in x=n, come si può
0 0
vedere ponendo a zero la sua derivata. Ora, ponendo x=n+y, abbiamo:
  n ln n  n ln(1 y )  y  n ln(1 y )  y
(n  1)  e  n  e n ln( n  y )  y dy  e  n  e n
dy  n n e  n  e n
dy
n n n

x 2 x3
Ora, in accordo con Taylor, abbiamo che: ln(1  x)  x    .... ; poniamo dapprima x=y/n e
2 3
2 3
 n ln(1 y )  y  y y  ......
n n n n n 2n 3n 2
poi y  nv , così: ( n  1)  n e  e dy  n e  e dy 
n n
2 3
 v v  ......
n n 2
n e n e 3 n
dv ,e se n è grande abbastanza: (n  1)  2n n ne  n (Formula di
 n

Stirling), siccome l’integrale di Gauss valeva 2 .

Appendice 2: EQUAZIONE DIFFERENZIALE DI BESSEL

Quando vi imbattete in un’equazione differenziale come questa:

x 2 y ' ' xy' ( x 2  n 2 ) y  0 (2.1)
avete una Equazione Differenziale di Bessel.
Se cerchiamo una soluzione in forma di serie J n (x) , allora l’equazione differenziale diventa:
(1) k ( x 2) n  2 k 
x 2 J ' 'n  xJ 'n  ( x 2  n 2 ) J n  0 e la soluzione in forma di serie è: J n ( x)   , (2.2)
k  0 ( n  k  1) k!
dove  è la Funzione Gama di Eulero. Infatti, se inserite la (2.2) nella (2.1), otterrete 0=0:

(1) k ( x) n  2 k 
(1) k (n  2k )( x) n  2 k 1 
(1) k (n  2k )(n  2k  1)( x) n  2 k  2
y   n 2k , y'   , y ' '  
k  0 ( 2) (n  k )!k! k 0 (2) n  2 k (n  k )!k! k 0 (2) n  2 k (n  k )!k!

(1) k ( x)n  2 k  2 
(1) k n 2 ( x) n  2 k 
(1) k (n  2k )( x) n  2 k
( x 2  n2 ) y   n2k
  , xy '   ,
k  0 ( 2) (n  k )!k! k  0 (2) n  2 k (n  k )!k! k 0 (2) n  2 k (n  k )!k!
2

(1) k (n  2k )(n  2k  1)( x) n  2 k
x y' '  
k 0 (2) n  2 k (n  k )!k!
 2 2 2 (1) k (n  2k )(n  2k  1)( x) n  2 k  (1) k (n  2k )( x) n  2 k

x y ' ' xy '( x  n ) y    
k 0 (2) n  2 k ( n  k )!k! k 0 (2) n  2 k ( n  k )!k!

(1) k ( x) n  2 k  2 
(1) k n 2 ( x) n  2 k 
(1) k ( x) n  2 k  2 
(1) k [4k (k  n)]( x) n  2 k
 n  2k
      (2)n  2k (n  k )!k! 
k  0 ( 2) (n  k )!k! k  0 (2) n  2 k (n  k )!k! k  0 (2) n  2 k (n  k )!k! k  0

(1) k 1 ( x) n  2 k 
(1) k 4( x) n  2 k 
(1) k 4( x) n  2 k
 n  2k  2
     n  2k

k 1 ( 2) (n  k  1)!(k  1)! k  0 (2) n  2 k (n  k )! k! k 1 ( 2) (n  k  1)!(k  1)!
(n  k ) k

(1) k 4( x) n  2 k
  n2k , così:
k  0 ( 2) (n  k  1)!(k  1)!

(1) k 4( x) n  2 k 
(1) k 4( x) n  2 k
x 2 y ' ' xy '( x 2  n 2 ) y   n  2k
  , ma possiamo
k 1 (2) (n  k  1)!(k  1)! k  0 (2) n  2 k (n  k  1)!(k  1)!
anche dire che:

(1) k 4( x) n  2 k 
(1) k 4( x) n  2 k
x 2 y ' ' xy '( x 2  n 2 ) y   n  2k
   0 , dove anche
k 1 ( 2) (n  k  1)!(k  1)! k 1 (2) n  2 k (n  k  1)!(k  1)!
l’ultima sommatoria può iniziare da k=1, visto che il suo termine in k=0 ha nel suo denominatore:
(k  1)! (k )  (1)! (0)   .
--------------------------------------
n
Vediamo che: J  n ( x)  (1) J n ( x) ; infatti:

(1) r ( x 2)  n  2 r n 1 (1) r ( x 2)  n  2 r  (1) r ( x 2)  n  2 r
J  n ( x)    
r  0 r! ( n  r  1) r  0 r!(  n  r  1) r  n r! (  n  r  1)
ma essendo che (n  r  1)   per ogni r = 0,1,…..,n-1, la prima sommatoria è zero. Inoltre,
nella seconda sommatoria poniamo r = n+k , sicchè:

(1) n  k ( x 2) n  2 k n

(1) k ( x 2) n  2 k
  ( 1 )   (1) n J n ( x) , cvd.
k 0 ( n  k )!  ( k  1 ) k 0  ( n  k  1) k !
I denominatori in entrambe le sommatorie sono uguali, visto che: (n  1)  n(n) .
2
Abbiamo anche che: J 'n ( x) J  n ( x)  J ' n ( x) J n ( x)  (2.3)
x(n)(1  n)
Infatti, siccome entrambi J n (x) e J  n (x) soddisfano l’Equazione di Bessel, allora:
x 2 J ' 'n  xJ 'n  ( x 2  n 2 ) J n  0 e x 2 J ' ' n  xJ ' n ( x 2  n 2 ) J  n  0 e moltiplicando la prima per
J  n e la seconda per J n e, infine, dopo aver sottratto membro a membro, otteniamo:
x 2 ( J ' 'n J  n  J ' ' n J n )  x( J 'n J  n  J ' n J n )  0 , o:
d d
x ( J 'n J  n  J ' n J n )  ( J 'n J  n  J ' n J n )  0 , o ancora: [ x( J 'n J  n  J ' n J n )]  0 
dx dx
c
J ' n J  n  J ' n J n  . (2.4)
x
xn x n 1
Ora, nel calcolare c, ricordiamo che: J n  n  ..... , J 'n  n  ..... ,
2 (n  1) 2  (n)
x  n 1 xn
J ' n   ..... , J n   ..... , e tutte queste dentro la (2.4) danno:
2  n  (  n) 2  n (n  1)
(((.......(n  1)  n(n).......)))
 1 1 1 1 1 1
c     
(n)(1  n) (n  1)(n) (n)(1  n) [n(n)]( n) (n)(1  n) (n)[n(n)]
1 1 1 1 2
     , dunque:
(n)(1  n) (n)[ n( n)] (n)(1  n) (n)(1  n) (n)(1  n)

2 2
c . Così: J 'n J  n  J ' n J n  .
(n)(1  n) x(n)(1  n)

Appendice 3: FORMULA DI RIFLESSIONE

2 2 sin n
Eccola:  . (3.1)
(n)(1  n) 
sin x
Infatti, sappiamo che vale zero in   , ,2 ,2 ,...., n , n , sicchè:
x
sin x x x x x x2 x2 
x
 (1  )(1  )(1  )(1  )..........  (1  2 )(1  2 ).....  [1  ( ) 2 ] , e dunque:
x   2 2  4 k 1 k

x x
 [1  ( ) 2 ]1 (3.2)
sin x k 1 k
Inoltre, un fattoriale possiamo anche vederlo così:
k! k n (k  1)(k  2)...(k  n)
n! lim  lim ; infatti, il denominatore del primo limite è:
k   ( n  1)...( n  k ) k   kn
(n  k )!
(n  1)...(n  k )  , mentre da entrambi i numeratori possiamo raccogliere:
n!
k!(k  1)...(k  n)  (k  n)! , così resta solo n! . Poi, vediamo che il secondo limite vale 1:
(k  1)(k  2)...(k  n) kn k!k n
lim  1 , poichè si riduce alla forma lim . Perciò: n!  lim .
k  kn k  k n k   (n  1)...(n  k )

k! k x
Siccome ( x  1)  x! , si ha: ( x  1)  x! lim e dopo aver diviso sia numeratore
k   ( x  1)...( x  k )

che denominatore per k!:

kx
( x  1)  lim . (3.3)
k   (1  x )(1  x 2) ...(1  x k )

1 1 1
Ora, introduciamo la costante  k  1    ...   ln k , e poniamo:   lim  k ; sulla base di
2 3 k k 
x x
 x ex e 2 e k
ciò, la (3.3) diventa: ( x  1)  e lim .... . (3.4)
k   (1  x ) (1  x 2) (1  x k )
x x
Infatti, tutti gli e x , e 2 ,…, e k si elidono con i termini in e x ed il termine k x  e(ln k ) x pure
proviene da e x . Inoltre, siccome ( x  1)  x( x) , la (3.4) diventa:

1 (1  x) (1  x 2) (1  x k ) x x
 xex lim x x
.... x
 xe x
 [1  ( )]e k (3.5)
( x ) k  e e 2 e k k 1 k
1
Ora, poichè ( y  1)  y( y ) , segue che ( y )  ( y  1) e dopo aver sostituito y con –x:
y
1
( x)   (1  x) e dunque: (1  x)   x( x) . Valutiamo ( x)(1  x) :
x
  
x x x x 1  x
( x)(1  x)  [ x 1e x  (1  ) 1e k ]  [ex  (1  ) 1e k ]  [ A]  [ B]   [1  ( ) 2 ]1 , (3.6)
k 1 k k 1 k x k 1 k
dove A è la (3.5) invertita e B è la (3.4) con x scambiato con –x. Perciò, la (3.6) è:
1  x x 
x
( x)(1  x)   [1  ( ) 2 ]1 , e per la (3.2) ((  [1  ( ) 2 ]1 )) abbiamo:
x k 1 k sin x k 1 k
2 2 sin n
 che è proprio la (3.1).
(n)(1  n) 

Appendice 4: SERIE DI BESSEL

2
Adesso, siccome, in accordo con la (2.3): J 'n ( x) J  n ( x)  J ' n ( x) J n ( x)  , usando
x(n)(1  n)
2 sin n
in essa la (3.1), otteniamo: J 'n ( x) J n ( x)  J 'n ( x) J n ( x)  . (4.1)
x
x 1 
(t  )
Vale la seguente eguaglianza: e 2 t
 J n ( x)t n ; (4.2)
n  
x 1 xt x
(t  )  
( xt 2) r  ( x 2t )k  
(1) k ( x 2)( r  k ) t ( r  k )
infatti: e
2 t
e e 2 2t
 [ ][ ] 
r 0 r! k 0 k! r 0 k 0 r! k!
e definendo n=r-k, segue che n andrà da   a   , così:
x 1
(t  )  
(1) k ( x 2)( n  2 k ) t n  
(1) k ( x 2)( n  2 k ) n 
e 2 t
     [  ]t   J n ( x)t n cvd.
n   k 0 (n  k )! k! n   k  0 k!(n  k )! n  

----------------------------------------

Valgono le seguenti eguaglianze (Serie di Bessel):

cos( x sin  )  J 0 ( x)  2 J 2 ( x) cos 2  2 J 4 ( x) cos 4  ...... (4.3)

sin( x sin  )  2 J1 ( x) sin   2 J 3 ( x) sin 3  2 J 5 ( x) sin 5  ...... (4.4)

Infatti, se poniamo t  ei nella (4.2), otteniamo:

1 
x ( e i  e  i )
e 2
 eix sin x   J n ( x)ein 


  J n ( x)[cos n  i sin n ]  {J 0 ( x)  [ J 1 ( x)  J1 ( x)] cos   [ J  2 ( x)  J 2 ( x)]cos2  ....} 


 i{[ J1 ( x)  J 1 ( x)]sin  [ J 2 ( x)  J 2 ( x)]sin2  ....} 

 {J 0 ( x)  2 J 2 ( x)cos2  ....}  i{2 J1 ( x) sin   2 J 3 ( x)sin3  ....} , ed eguagliando parte reale e

parte immaginaria appena ottenute con quelle scaturenti da eix sin x , otteniamo appunto le (4.3) e
(4.4) .