ADDITIONAL TABLES FOR DESIGN OF

OPTIMUM LADDER NETWORKS *

BY

LOUIS WEINBERG 1
PartI**

ABSTRACT

In a preceding paper tables were presented for the design of three large classes
of ladder networks. These networks had characteristics given by Butterworth,
Tschebyscheff, and Bessel polynomials. In the present paper a number of additional
tables are presented, the tables now being classified on the basis of the parameter r,
which is the input-to-output resistance or conductance ratio. The tables give the
element values of normalized low-pass ladders with one of the following character-
istics : maximally flat magnitude (Butterworth), equal-ripple magnitude (Tsche-
byscheff), and maximally flat time delay (Bessel polynomial). By means of frequency
transformations the networks given by the tabulated element values for the Butter-
worth and Tschebyscheff networks may be converted to give high-pass, band-pass,
and band-elimination filters. Thus the tables may be used as a handbook for the
design of these optimum networks.

INTRODUCTION

Tables for the design of three large classes of networks were pre-
sented in a preceding paper (I).” It was shown there that by use of
these tables the engineer who knows little about the theory of modern
synthesis can synthesize useful networks. These networks had char-
acteristics given by Butterworth, Tschebyscheff, and Bessel polynomials.
In this paper the same characteristics are considered. However,
many new tables are added. In addition, the basis for classifying the
tables has been changed to one that is believed to be more useful for
most applications.
The following extensions and changes have been made:

1. The tables are no longer classified in terms of the decrement

ratio D, as they were in the preceding paper. The basis for classification
in this paper is r, where r = R./R1 (or r = G’JG’J, the ratio of the
input to the output resistance (or input-to-output conductance).
These tables give a much larger range of input-to-output terminations
than was available in the previously published tables.
* This paper is based on the author’s report with the same title, Technical Memorandum
No. 434, Hughes Research Laboratories, Culver City, Calif.
1 Research Laboratories, Hughes Aircraft Co., Culver City, Calif.
** Part II will appear in this JOURNAL for August, 1957.
* The boldface numbers in parentheses refer to the references appended to Part I of this
paper.
7
8 LOUIS WEINBERG [J. F. I.

2. Tables for the Tschebyscheff characteristic with &-db and &db.

ripple have been added.
3. In the preceding paper the Bessel-polynomial networks were
given only for a resistance termination at one end. In this paper the
networks are given for the same wide range of resistance ratios as are
the Butterworth and Tschebyscheff networks.
It is felt that since these tables eliminate tedious computations,
they will be of great value to engineers familiar with synthesis theory.
But the paper is directed mainly at those practical engineers who know
little about synthesis and desire a final working formula and a set of
tabulated values. It was therefore decided to make the paper as self-
contained as possible, within the space restrictions, by giving an
analytical discussion of each type of network. Though the discussion
is necessarily brief, it is believed that it is sufficient to acquaint the
reader with the characteristics of the networks whose element values
are given in the tables.
As mentioned in Sec. V, a number of different networks is possible,
each of which realizes the identical transfer function (including the
constant multiplier). The networks differ because of the different
choices for the zeros of the reflection coefficient. The tables in this
paper are based on the choice that gives maximum gain-bandwidth
product for a specified value of shunt capacitance (2), that is, all the
zeros are chosen to lie in one half-plane.
The paper is divided into five main sections. How to use the tables
of element values is discussed in the first section. The next two treat
Butterworth and Tschebyscheff networks, respectively, while the
fourth section treats the maximally flat time-delay networks obtained
by the use of Bessel polynomials. In each of these three sections
tables of the element values of the normalized low-pass ladder network
are given. In the final section it is briefly shown : (a) how to transform
Butterworth and Tschebyscheff networks to serve high-pass, band-pass,
or band-elimination functions; (b) how to remove the normalization of
the element values-that is, how to change the pass band of the network
from w = 1 to the desired radian frequency, and how to raise the level
of the network; (c) how to use duality and reciprocity to obtain sets
of new networks; and (d) how to convert the symmetrical Butterworth
and Tschebyscheff networks to unsymmetrical ones with any desired
ratio of input-to-output resistance.
I. USE OF THE TABLES OF ELEMENT VALUES

The general form of the low-pass ladder network whose element

values are given in the tables is a lossless network terminated in resist-
ance. In all the tables and in the figures the eIement values are in
ohms. henrys, and farads. The tabulated element values are normal-
ized in that the pass band has a cutoff radian frequency W, equal to
July, 1957.1 OPTIMUM LADDER NETWORKS 9

unity and the network has a one-ohm resistance load, that is, in all
the tables R1 = 1. As shown in Sec. V and in the illustrative examples
the removal of these normalizations requires only simple multiplications.
Six different values of I = RJR1 (or r = G’./G’r) are included:
0, f, 4, $, 4, and 1. Since RI = 1 for all the tables, the value of R,
for each table is given by the 7 of that table and thus need not be
tabulated.
The primed and the unprimed values in the tables yield dual net-
works so that a transfer impedance or a transfer admittance can be
realized. The networks for the general ladder, with a resistance
termination at both ends and with a current-source input, are shown in
Fig. 1 for n odd and in Fig. 2 for n even ; n is the degree of the denomina-

I’ ,.,,“r*--*“,*,
~q-$?g--g$-~R Et

FIG. 1. General form of low-passladder network FIG. 2. General form of low-pass

with a current-source input and 1z odd. ladder network with a current-source
input and n even.

tor of the transfer function and is thus also equal to the number of
reactances in the ladder. The transfer function realized by these
networks is the transfer impedance ZZ1 = Ez/ll.
For a voltage source used as the input, the dual of the above networks
can be used. The networks are shown in Figs. 3 and 4 for n odd and

FIG. 3. General form of low-pass ladder FIG. 4. General form of low-pass ladder
network with a voltage-source input and n network with a voltage-source input and n
odd. even.

even, respectively. The transfer function realized by these networks

is the transfer admittance Yz, = Iz/El.
It is also pointed out that Thevenin’s or Norton’s theorem can be
used to effect a source conversion and thus yield new network con-
figurations. For example, Thevenin’s theorem applied to R, and the
current source in Fig. 1 yields a voltage source and a series resistance.
In this way a transfer admittance or transfer voltage ratio may be
realized with a shunt capacitance branch at both ends of the coupling
network.
The primed and unprimed elements shown in Figs. 14 are intended
to correspond to the primed and unprimed values given in the tables.
IO LOUIS WEINBERG [J. F. I.

Thus for a specified Zzl and n odd the unprimed tabulated values yield
the network of Fig. 1, whereas for n even the primed tabulated values
yield the network of Fig. 2. When a transfer admittance Ytl is re-
quired, the primed tabulated values yield the network shown in Fig. 3
for n odd, and for n even the unprimed values yield the network of Fig. 4.
The tables are divided as follows:
(a) Table I gives the element values for the Butterworth filter,
where (a) applies for r = 0, (b) for r = +, (c) for r = a, (d) for r = Q,
(e) for r = 3, and (f) for r = 1.
(b) Tables II-VII apply to the Tschebyscheff filter, with the sub-
divisions being necessary to provide for the different ripple factors.
For example, Tables I I and III give the element values for a +db
ripple and a a-db ripple, respectively. The alphabetical subdivisions
are the same as for the Butter-worth case.
(c) Tables VIII and IX apply to the Bessel-polynomial networks.
Table VIII gives the frequencies at which significant values of time
delay and loss occur. The variable u is the normalized frequency W/W,,
where o. = l/to and to is the desired time delay, that is, the time delay
occurring at zero frequency. Table IX gives the element values; the
alphabetical subdivisions are the same as for the Butter-worth and
Tschebyscheff cases.
As has been mentioned, the parameter r is equal either to the ratio
of the input to the output resistance or to the input-to-output con-
ductance ratio. A little reasoning always suffices to determine which
applies, but for convenience the practical rules are expressed explicitly
below.
(a) Except in the case of r = 0, for networks formed from the un-
primed values in the tables, r is equal to the resistance ratio R,/R1.
For networks formed from the primed values r is equal to the conductance
ratio G’./Gfl (which is of course equal to R’JR’,).
(b) For r = 0 the parameter r is equal to the resistance ratio for the
combinations: (1) unprimed values and n even, and (2) primed values
and n odd. It is equal to the conoktunce ratio for (1) unprimed values
and n odd, and (2) primed values and n even. This is merely a detailed
way of stating that for r = 0, R, becomes a short for networks with a
series +ut, that is, those in Figs. 3 and 4, and R’, becomes an open
circuit for networks with a shunt input, that is, those in Figs. 1 and 2.
Inspection of the tables for the Tschebyscheff networks shows that
for a number of tables element values are not given for n even. For
these cases the specified resistance (or conductance) ratio is too large
to be physically realizable. This occurs for:
(a) r = 1 for all ripples.
(b) r = 3 for l-db, 2-db, and 3-db ripples.
(c) r = + for 2-db and 3-db ripples.
(d) r = 2 for 2-db and 3-db ripples.
July, ws7.1 OPTIMUM LADDER NETWORKS II

With the above preliminary remarks the steps in the procedure for
using the tables follow:

1. Determine from the specifications of the problem whether a

Butterworth, Tschebyscheff, or Bessel-polynomial network is to be used.
2. Calculate the value of n that gives the required degree of the
denominator polynomial of the transfer function and consequently the
required complexity of the network. For the Tschebyscheff character-
istic it is first necessary to calculate the ripple factor c.
3. Using this value of n look up the element values in the ap-
propriate table.
4. Remove the normalizations as shown in Sec. IV. The bandwidth
is thus changed from wc = 1 to the desired cutoff value, and the load
resistance and the network level are changed to the required values.

FIG. 5. Sketches of the first three orders of the Butterworth

approximation to the low-pass filter.

5. If a high-pass, band-pass, or band-elimination network is desired,

convert the element values by means of the frequency transformations
of Sec. V.
In carrying out the first step, if we require a filter whose magnitude
characteristic is specified, then the Butterworth or Tschebyscheff
characteristic may be used. For the same value of n, the Tschebyscheff
filter gives a better coverage of the pass band and a faster drop-off
outside the band than any other possible transfer function that is also
a constant divided by a polynomial; its phase characteristic, however,
is more nonlinear than that of the Butterworth filter. Thus the choice
between the two will depend on the importance of the phase character-
istic. Neither of these filters gives a linear phase characteristic, that is,
a pure time delay, over a specified frequency range. For this purpose
12 LOUIS WEINBERG 1J. F. I.

the Bessel polynomials are used. The filters obtained by use of the
Bessel polynomials also have a low-pass magnitude characteristic so
that they may also be used in those problems where the magnitude
characteristic is specified.
The method for calculating the value of n and, in the case of the
Tschebyscheff filter, the value of E, is shown in the respective sections.

TABLE I.-Element Values (in ohms, henrys, farads) for a Normalized Butterworth Filter.

Value al ” c, or L,’ Lf or C*’ c, or Ls L, or C,’ CJ or Lg’ Ls or c, c, or L,’ L* or Cg’ coor Lp’ L. 01 CM

“1”=O 1.oooo
3” 0.7071 1.4142
0.5OOil 1.3333 1.5000
4 0.3827 1.0824 1.5772 1.5307
6” 0.2588
0.3090 0.8944 1.3820 1.6944 1.5451
0.7579 1.2016 1.5529 1.7593 1.5529
7 0.2225 0.6560 1.0550 1.3972 1.6588 1.7988 1.5576
a 0.1951 0.5776 0.9370 1.2588 1.5283 1.7287 1.8246 1.5607
0.5155 1.1408 1.4037 1.6202 1.7772 1.8424 1.5628
lo’ Z:E 0.4654 E&Z 1.0406 1.2921 1.5100 1.6869 1.8121 1.8552 1.5643

!’ r = l’8 9.oGQO
0.0939
i 11.9764
12.4442 0.1735 4.1674
i 12.6076
12.5685 0.2032 8.92% 0.0493
0.2169 11.3305 2.5343
6 12.6190 12.6794 “o:::!!
i 12.6199 :::5; 13.5040 0.1778 GE 1.8121
0.2314 14.0417 0.1940 10.2279 4.6929 0.0248
0.2333 14.4102 0.2053 11.3856 0.0653 1.4086
1: :2”::::
12.6064 0.2346 14.6730 0.2135 12.2305 EE 0.0965 3.7699 0.0198

2
r=1’5.0000
E’ 4 6.2741
3 6.3870 EE 2.1699
4 6.3840 0.4180 4.6024 0.1018
0.4435 5.8036 0.2350 1.2992
: Et 0.4567 6.4673 0.3130 3.1601 0.0675
0.4641 6.8671 0.3618 0.1700 0.9225
;: ::ii;: 0.4687 7.1244 0.3940 “;:E 0.2417 2.3838 0.0503
0.4716 1.2984 0.4162 5.7720 0.2932 3.4607
1: 6.2941
6.2825 0.4735 7.4209 0.4321 6.1916 0.3312 4.2683 :::;:: 0.0401

:’r=1’3 4.OOoO
0.2761
i 4.8284
4.8473
4 4.8105 i:::;: ::;::: 0.1386
0.5997 4.4239 0.3186 0.9912
2 4.7743
4.7446 0.6156 4.9155 0.4233 2.4042 0.0913
7 4.7206 0.6244 5.2085 0.4882 0.2294 0.7006
0.6295 0.5308 :;Fi 0.3258 1.8075
t i:::: 0.6326 55:Zi 0.5601 4:3702 0.3948 2.6209 0.5410
10 4.6120 0.6346 5.607 1 0.5809 4.6833 0.4454 3.2293 1.4445 0.0540

e) I = l/2

: YEi 0.4483
3 3.2612 0.7189 1.1811
0.8826 2.4524 0.2175
: 3.1868
3.1331 0.9237 3.0510 0.4955 0.6857
6 3.0938 0.9423 0.6542 1.6531
0.9513 EZ 0.7512 2.2726 :::::i 0.4799
i 3.0408
3.0640 0.9558 3.6678 0.8139 2.6863 0.5003 1.2341 0.1042
0.9579 3.74’(6 0.8565 2.9734 0.6046 1.7846 0.2735 0.3685
1: 3.0072
3.0223 0.9588 3.7924 0.8864 3.1795 0.6808 2.1943 0.4021 0.9818 0.0825

f) I = 1

: ::iE 1.4142
3 1.0000 1.0000
::Ei 1.8478 0.7654
4 FiEt 1.6180 2smO 1.6180 0.6180
2 0:5176 1.9319 1.9319 0.5176
i 0.3902
0.4450 ::;K 1.8019 ::iZ 0.4450
1.1111 1.6629 ::E 1.%16 ::Ei 1.1111 0.3902
l.COOO 1.5321 1.8794 2.0000 1.8794 1.5321 0.3473
1: 0.3129
0.3473 0.9080 1.4142 1.7820 1.9754 1.9754 1.7820 ::Ei 0.9080 0.3129
July, 1957.1 OPTIMUM LADDERNETWORKS I3
II. BUTTERWORTH
CHARACTERISTIC
(3,4,5)

The Butterworth function is used to approximate the squared

magnitude of a transfer function. For the transfer impedance it is
given by

(1)

This function gives an approximation to a low-pass filter characteristic;

sketches of the Butterworth approximation for the first three values of
n are shown in Fig. 5. The Butterworth function is said to have a
maximally fiat magnitude characteristic.
By use of Eq. 1 the complete transfer function is given as

H
Z,,(s) = __
B.(s)’
where H is a constant multiplier. The polynomials B, are called the
Butterworth polynomials ; these polynomials have a unity coefficient
for sn and their zeros are the nth roots of unity that lie in the left half-
plane.
The element values are given in Table I; the resulting networks
realize the transfer function within a constant multiplier. To obtain
the constant multiplier we let s = 0 in the network and in the transfer
function.
An example of the use of the tables to design a Butterworth filter
is presented below.
Example 2.1. We wish to design a low-pass filter that has a resist-
ance termination at the output only. The cutoff frequency is wc = 10,000
radians/set and the output resistance is to be 750 ohms. At a fre-
quency w = 3w, the magnitude response is to be down at least 50 db.
The input source is a cathode follower which approximates a true
voltage source.
First we determine the value of n.

1
= 10-S
1 + 02n w=3
(1 + c+) (w=3 = 105
32” &XIO5
5 1
n = 2 log 3
= 5.23.

The next larger integer n = 6 must be used.

I4 LOUIS WEINBERG [J. F. I.

Since no input resistance is required, the table for r = 0 is used,

namely, Table I (a). Since the input is a voltage source and n is even,
the network form of Fig. 4 (with R, omitted) is applicable; that is, the
unprimed element parameters are used.
Consulting the tables yields the element values

RI = 1 Lb = 1.553
c1 = 0.2588 cg = 1.759
Lz = 0.7579 Lg = 1.553
CB = 1.202

To obtain a load resistance of 750 ohms, we multiply RI and all L’s

and divide all C’s by 750. To change the cutoff frequency to 10,000
rad/sec every L and C must be divided by this value.
The final values are therefore

R = 750 RI = 750
&z--z RL 1.16 x 10-l
ml?
c, = 2 = 3.45 x 10-g

C, = = 2.34 X lo-’
Lo = 5 = 5.68 X 1O-2 2 E

C, = 2 = 1.60 x 10-r Lf = % = 1.16 X 10-l

c

and the network is shown in Fig. 6.

FIG. 6. Final network achieved for Example 2.1.

At s = 0 the network becomes a pure resistance and therefore

Z2lld = & =R = 750.

Since the constant term of every B, is unity, the constant multiplier H

is 750. The transfer voltage ratio E2/E1, since Ez = 750 r2, is given by

_=-
E2 1
E1 &J(S)’
July, 1957.1 OPTIMUM LADDER NETWORKS I5

III. TSCHEBYSCHEFF CEUUCTERISTIC (6)

The Tschebyscheff approximation to the magnitude characteristic

of a low-pass filter is given by
1
I-W+) 12 = 1 + [&-&J)]2’ (3)

TABLE II.-Element Values (in ohms, henrys, farads) for a Normalized Tschebyscheff
Filter with A-db Ripple (e = 0.1526, 9 = 0.0233).

Value of ” c, or L,’ L* o* q csor LJ’ L, or C,’ c5 or L$’ LI or Ce’ c, or 4

a) * = 0
1 0.1526
2 0.4215 0.7159
1.0864
4 0.5158
0.5544 1.1994 ::fE 1.2453
2 0.5841
0.5734 1.2490 1.5562 1.5924 1.3759
1.2752 1.5999 1.6749 1.7236 1.4035
1.2908 1.6236 1.7107 1.7987 1.7395 1.4745
: 0.5949
0.59% 1.3008 1.6380 1.7302 1.8302 1.8070 1.8163
1.3076 1.6476 1.7423 1.8473 1.8343 1.8814 i:%g 1.5182
1: EZ 1.3124 1.6542 1.7503 1.8579 1.8489 1.9068 1.8585 i.4964

b) r = l/8
1 1.3736
3” 8.9466
5.9892 0.0567
0.1403 4.3787
5” 10.0512 0.1866 9.8722 0.0733
11.112a 0.2008 12.7123 0.1602 4.8368
6 11.2235 0.2179 13.5071 0.2041 10.46m 0.0770
;: 11.8455 0.2179 14.5445 0.2153 13.2359 0.1652 4.9726
0.2287 14.4866 0.2312 13.9349 0.2087 0.0783
:EE 0.2248 0.2297 14.9251 0.2191 :EZ 0.1672 5.0298
It: 11:wo 0.2336 :::;:ZZ 0.2401 14.8149 0.23% 14.OR87 0.2106 10.7462 0.0790

c) r = l/4
: 3.0912
0.7631 0.1220
0.2886 2.3272
: 5.0046
4.5446 0.3815 5.06% 0.1559
0.4037 6.4880 0.3281 2.5577
2 5.5377
5.5547 0.4403 6.7916 0.4167 5.3580 0.1632
0.4353 7.3489 6.7472 0.3379 2.6256
: 5.7441
5.8904 0.4601 7.2417 GE! 7.0032 0.4258 5.4537 0.1659
0.4478 0.4594 7.5404 0.4404 0.3419
1: 5.8435
6.0374 0.4690 ES! 0.4839 7.4072 0.4742 E8” 0.4295 0.1672

d) r = l/3

:. 0.6105
2.3497 0.1712
: 3.7120
3.4253 0.3914 1.8216
0.5161 3.8671 0.2172
0.5389 4.9341 0.4442
.z 4.0823
4.1422 0.5906 5.1046 0.5639 0.2272
0.5785 5.5512 0.4572 2.0481
;: 4.2247
4.3777 0.6155 5.4212 ::5::: 0.5760 4.1536
0.5942 5.7535 0.6119 0.5885 5.1967 ZE
1: 4.2930
4.4804 0.6266 5.5455 0.6490 0.6374 5.3199 0:5r310 0.2325
e) r = l/2
1 0.4579
3” 2.2746
1.5715 0.2880
0.6035 1.3341
0.7973 2.6600 0.3626
: 2.3545
2.6921 0.8042 3.3682 0.6853 1.4372
6 2.5561 0.8962 3.3962 2.8071 0.3785
0.8560 3.7594 ::Z3: 3.5246 0.7050 1.4932
: ::E 3.5762 0.9619 0.8950 2.8547 0.3843
::8”::; 0.9121 Z:EZ 0.8836 3.5703 0.7127 1.5084
lo” ;:ZE 0.9429 ::KZ 0.9887 3.6707 0.9765 3.5472 0.9027 2.8761 0.3870
nr=1
1 0.3052

: i.0316 i.1474 1.0316

: i.1-68 ;.3712 i.9750 i.3712 1.1468

6
7 i.181; i.4228 z.096; i.5;3; 2.0%; i.4228 1.1812

: ;.195; i.4426 ;.1346 i.6i6Y -2.2054 -i&7 -2.1346 1.4426 1.1957

10 --- - _ _ - _ _ _ _ _ - - _ - - _ - - _ _ _ _ ___ --_
16 LOUIS WEINBERG [J. F. I.

Here the parameter E is the ripple factor and T,(w) is the Tschebyscheff
polynomial of order (and degree) n ; T,(w) is defined by cos (n cos-1 CO).
The role played by E and the equal-ripple quality of the Tschebyscheff
approximation are illustrated in Fig. 7, where n = 3 and a 1-db ripple
are used.

I i
I I
I
-I 0 I w
FIG. 7. Low-pass filter obtained by using the Tschebyscheff
approximation with n = 3 and a l-db ripple.

From the magnitude given by Eq. 3 the complete transfer function

can be found. It is given by

H
ZZl(S)= v,(s)9

where H again is a constant and V, is formed from the left half-plane

zeros of the denominator of Eq. 3 ; Vn is a polynomial of degree n with
the coefficient of sn equal to unity and with its zeros lying on an ellipse.
The element values of the ladder networks for values of E correspond-
ing to A-, a-, $-, l-, 2-, and 3-db ripple are presented in Tables II
through VII.
After the value of E has been calculated from a specified ripple
factor, it is necessary to determine the required value of n. Formulas
useful for this purpose are

(w + LJJZ
- 11%+ (w + 4zTp-n
T,(w) = (5)
2
or
_)n + (w - _)--n
T,(o) = (w - (6)
2
July, 1957.1 OPTIMUM LADDER NETWORKS 17

TARLE III.-Element Values (in ohms, henrys, farads) for a Normalized Tschebyscheff
Filter with i-db Ripple (C = 0.2434, c2 = 0.0593).

Value of n c, cl1 L,' L* or C*’ csor L1’ L, or Cp’ csor Lgl CTo* LI’ L* or C,’ coor Lp

a) r = 0
1 0.2434
2 0.5566 0.8499
3 0.6517 1.2198 1.2248
4 0.6891 1.3215 1.5979 I.3003
: 0.7173
0.6912 1.3538 1.6741 1.6371 1.4480
1.3868 1.7271 1.7144 1.8105 1.4193
1.3999 1.7475 1.7450 1.8816 1.7497 1.5323
;i 0.7274
0.7234 I.4083 1.7598 1.7612 1.9099 1.8124 1.8806 1.4647
1.4140 1.7678 1.7711 1.9248 1.8365 1.9439 1.7927 1.5648
1: 0.7322
0.7302 1.4180 1.7733 1.7776 1.9338 1.8490 1.9676 1.8505 1.9119 1.4864

b\ I = l/8

4 ~2.1908
7.0446 0.0755
3 10.0648 0.1562 5.5746
4 10.4126 0.2051 0.0922
0.2052 :!:86E 0.1726 5.8758
i 11.7227
11.2734 0.2298 13.7315 0.2211 11.3063 0.0958
7 12.3557 15.2643 0.2183 0.1780 6.1472
::::;I: 0.2423 :::tE 0.2250 11.4700
t 11.5987
12.5942 0.2228 ::::Z$ 0.2287 15.6054 0.2214 14.4391 KG 6.2011
10 11.7539 0.2414 14.7148 0.2489 14.7268 0.2452 14.2166 0.2267 11.5438 0.0976

cl I = l/4
1’ 1.2171
0.1647
i 3.5907
5.1234 2.9867
4 5.1282 GE 5.4989 0.1990
0.4097 7.0092 0.3509
2 5.5080
5.8822 0.4669 6.3423 0.4540 :.::3278 0.2062
0.4322 7.7470 0.4354 7:3060 0.3611 3.2774
i 5.6503
6.1786 7.1583 0.4930 0.4618 5.8228 0.2087
::‘E 7.9582 0.4543 ;:Ez 0.4415 7.3830 0.3643 3.3046
1: 6.2867
5.7179 0.4880 7.2798 0.5050 7.3032 0.4989 7.0810 0.4650 5.8583 0.2099

d) r = l/3
: ;:Z
3 3.8716 Ei 2.3508
4 3.7640 0.5729 4.1610 0.2812
2 4.0210
4.4034 0.5439 5.3427 0.4720 2.45%
0.6299 4.3438
0.5713 55:~~~ “0:::;: 5.5708 “o:::i 2.5735
i 4.1167
4.6162 0.6480 5.3173 0.6668 5.2361 0.6285 4.4023 0.2945
0.5822 6.0240 0.6017 6.0086 0.5864 5.6288 0.4893 2.5944
1: 4.1620
4.6917 0.6561 5.4002 0.6817 5.4294 0.6749 5.2808 0.6328 4.4285 0.2%2

e) r = l/2
1 0.7303
2 1.7288 0.4104
3 2.5965 0.6465 1.7402
4 2.2aa4 0.9039 2.7832
0.8007 3.6926 “0::::; 1.8152
2 2.4162
2.8983 0.9771 3.2941 0.9837 2.9094 0.5100
0.8341 4.0204 0.85% 3.8585 0.7340 1.9007
i 2.4631
3.0294 mOoo 3.4072 1.0463 3.3925 1.0015 2.9490
0.8478 4.1088 4.1199 0.8670 3.8989 1.9157
1: 3.0724
2.4852 1.0100 3.4501 ~:E~ 3.4927 1.0606 3.4235 2.9666 0.5190

f) I = 1
: 0.4868

3 i.3034 ;.I463 1.3034

4 _ _ - --_ _ - -
i.3824 1.3264 2.2091 1.3264 1.3824
f!t
7 i.4468 ;.3;6; 2.3476 ;.4&9 2.3476 i.356; 1.4468

i ---
1.4604 i.3704 2.3800 i.s-HIi 2.4414 i.5000 2.3800 Y.3704 1.4604
10 --- _ _ _ _ - _ _ - _ _ - - - - - - - - --- - _ _ --_
18 LOUIS WEINBERG [J. F. I.

TABLE IV.-Element Values (in ohms, henrys, farads) for a Normalized Tschebyscheff
Filter (i-db Ripple (E = 0.3493. e2 = O.fZZO).

value of n c, or LI’ c,or C,’ c: or r, L, or C,’ esor Ls’ c, or r, L* or c, coor L#’ LMor GO
a)?-0
1 0.3493
3” z;;:: !%f
1:3916
1.3465
1.7279 1.3138
: 0.8529
0:8352
1.4291 1.8142
1.4483 EE
: 0:8725
:E 1.45% :zE 1:7371 1.5982
1.4666 1:8750 1.9571
1.4714 !E x;
2 t8877::
. 1.4748 ::Zt 1:7645 EE 1:8119 1.4539

b) r = l/8
1’ 3.1438
2 7.6905 0.0965
0.1646 6.81%
3 KE 0.2234 11.2532 0.1135
: 12:5367 0.2039 14.9223 0.1801 7.3211

! :;:Z.t Z’: :tEfl EE :E; “d:z 7.4478

! 13.1608
11.2694 14:0704 0:2556 13:8641 0:2422 11.8216
:z 16.4112 0.2180 15.4833 SE 7.5006
10 11.3804 Ok24 :f:Z ::ZE 14.3247 0.2583 13.9670 0:2436 11.8832 0.1188

c) I - l/4
1 1.7466
3” 3.8432
4 ;z 0.2504
2 65:g347& 0.3613

7 GE 4.0061
2% 0:5272 0.2602
! 6.6323 0.4389 ?‘Z! 4.0331
10 5.4437 0.5364 6:8522 8:Z 5.9413 0.2614

d) r - l/3
: 2.8282
1.3972
3 EEl 2.9371
4 $6”G ok399 4.1985
5 4:78% 0.5293 5.8898 kEE .3.1130
: 3.7922 0.6851 4.8770 0.6852 4.3536 0.3722
0.5603 0.4901
$:E :El ::ZE 0.7235 ZE ZE
! 4.9901 0:5572 0.5770 6:4061 EE 6:1064 3.1841
10 3.8860 0.7051 ;:EZ 0.7348 5.1229 0:7314 5.0307 4.4237 0.3776

2.1903
2.4881
::::; 2.3197
t::: 1:2403 2.5976
413575 0.8132 E;G 2.3566
1.2864 z:9 1:2628
ZE 0.8341 414546 0.8235 ::tZ: 2.3719
2:8366 1.2999 2.8964 1.3054 2.8744 2.6456 0.8104

01-l
:: 0.6986
3 i.5963 1.5963
: ---
1.7058 i.5408 i.22y6 1.7058
6
: i.7373 i.6383 i.3443 i.6383 i.258; 1.7373

;.6678 i.367; 2.7239 i.3673 2.6678 i.2& 1.7504

1: i.lSG
_-_ _ - _ _ - - -_- -_- ___
July,1957.1 OPTIMUM LADDER NETWORKS I9

TABLE V.-Element Values (in ohms, henrys, farads) for a Normalized Tschebyscheff
Filter with I-db Rifiple (e = 0.5088. 3 = 0.2589).

Value of n c, or L, L, or C;’ c, or L,’ L, or C“ c, or Ls’ c, or L1 coor LB’

a1r-O
0.5088
:: 0.9110
1.0118 !:E 1.5088
: 1.0495 1.4126 1.9093 1.2817
1.0674 1.4441 1.9938 1.5908 1.6652
: 1.0773 1.4601 2.0270 1.6507 2.0491 1.3457
7 1.0832 1.4694 2.0437 1.6736 2.1192 1.6489 l.ill8
1.0872 1.4751 2.0537 1.6850 2.1453 1.7021 2.0922 1.3691
: 1.0899 1.4790 2.0601 1.6918 2.1583 1.7213 2.1574 1.6707 1.7317
10 1.0918 1.4817 2.0645 1.6961 2.1658 1.7306 2.1803 1.7215 2.1111 1.3801

b) r = l/8
1
::E 0.1286
2 12.5563 0.1657 8.8038
4 9.9024 0.2517 0.1447
13.7259 0.1943 :zz 9.2596
2 10.3304 0.2677 12.7878 ::;3: 11.5115 0.1503
14.0719 17.5013 0.2045 16.9MO 0.1819 9.3890
: 10.4856 ::E 13.1313 0.2802 13.0465 0.2701 11.6220 0.1516
14.2174 0.2033 17.7827 0.2097 0.2066 17.0949 0.1830 9.4427
1: 10.5585 0.2746 13.2602 0.2842 E:E 0.2828 13.1287 0.2715 11.6709 0.1522

c) I = l/4
2.5442
f 3.7779 0.3001
0.3264 4.7927
: :::g: 0.5428 5.3680
7.0522 8.6301 “o:z
: 4.7366 ::E! 6.0240 0.3486
7.2126 0.3888 9.0689 oO:Z 0.3577 5.0989
: 4.7966 0.5803 0.6005 5.5869 0.3515
7.2800 0.3930 “9::::: 0.4064 “o:E 8.9024 0.3598 5.1270
1: 4.8247 0.5841 6.2098 0.6076 0.6063 6.1890 0.5864 5.60% 0.3528

?r=1’3 2.0354
2.5721 0.4702
3” 4.9893 0.4286 3.8075
0.7929 3.7589 0.5347
5” EEi 0.4915 6.6673 0.4622 3.9944
3:1307 0.8287 4.1451 0.8467 3.8812 0.5475
! 0.5050 6.9839 0.5177 6.8280 0.4696 4.0473
EZ 0.8395 4.2237 0.8764 4.2404 0.8580
!z 5.5459 0.5101 7.0783 0.5288 7.1141 0.5232 ::EZ
10 3.1806 0.8442 4.2532 0.8851 4.3088 0.8857 4.2691 0.5541

e) r = l/2
1.5265
::
3 3.4774 0.6153 2.8540
_-_
: 3.721; 0.694; 4.7440 0.6650 2.9936
__-
! 3.7916 0.7118 4.9425 0.7348 &36 0.675; 3.0331
8 __-
3.8&d 0.718; S.OG 0.748; ;.OG ;.7;2; 4.9004 0.6797 3.0495
1: --_ _ _ _ ___ ___ -_- - _ _ __- _-_ -_- ___

P=l 1.0177
3” ---
2.0236 0.994; 2.0236
4
;.1349 Gni ;.cGG i.091; 2.1349
2
;.1-MS i.111; ;.o93a i.173; i.0936 i.111; 2.1666
8’
2.1797 i.119; 3.1214 i.iG 3.1746 i.lS97 3.1214 i.1192 2.1797
1; ___ -__ _ _ _ -__ ___ _-- --_ _-_ - _ - _--
20 LOUIS WEINBERG [J. F. I.

TABLE VI .-Element Values (in ohms, henrys, farads) for a Normalized Tschebyscheff
Filter with Z-db Ripple Ce = 0.7648, 3 = 0.5849).

CJ or Lx’ L, or C,’ csor Ls’ Loor C,’ c, or Li’ Ls 0, Cs’ ccor Lo’ LlO01 Go’
a) * - 0
i%i 0.9766
1:3553 1.2740 1.7717
1.3%2 1.3389 2.2169 1.1727
5” 1.3610 1.4468 1.9004
1.3765 ;:;:t! 1.4974 2.3304 1.2137
t 1.3836 2.3551 1.5159 2.4063 1.9379
1.3881 2.3645 1.5251 2.4332 ::E: 2.3646 1.2284
9” 1.3911 2.3707 2.4463 1.5495 2.4386 1.4959 1.9553
10 1.4416 1.3932 2.3748 f:E7 2.4538 1.5536 .2.46O7 1.5419 2.3794 1.2353

b) r - l/8

: 6.8830
7.2895 0.1875
0.1541
: tE 0.3093 ?Ez 0.2088
: lp& 0.1729 19h74 0.1646
0.3224 10.9256 0.3260
1;;&5;; 0.1769 20.2574 0.1811 12.5671
;I 0.3263 11.1205 0.3368 10.3007 0.2144
0.1784 20.4848 20.0271
1; 16.3765
8.8824 0.3280 11.1932 :::fZ 11.1858 8:Z 0.2151

c) r - l/4
: 3.8239

3 i.9106 0.295; 6.5423

t 8.3859 0.328; 10.3300 0.3156 6.8118

! 8.5220 0.3354 10.7~zQ9 0.3443 10.5472 0.3199 6.8877

t 8.518; 0.3380 1o.ao; 0.3499 1o.n9i 0.3475 10.6i4Z 0.321; 6.9191

10 --- _ _ - _ - - _ - _ _ _ - _ - _ - _ _ _ _ - --_

d) I - l/3
2 3.0591

3 6G7i 0.3816 5.2161

: a.a7; 0.4219 iG81 0.4076 5.4294

6
: 6.597; 0.4302 i.34li 0.4425 8.2389 0.413; 5.4893

0.4334 8.4220 0.4493 8.4834 0.4467 8.2913 0.415; 5.5141

1; _-_
6.6391 _ _ - _-- - - _ _ _ _ _- _ - _ - _ - - _ - _

0) * = l/2
1 2.2943

; ---
4.3975 ;.5i26 3.9184

: 4.626; 0.583; i.G.0; 0.5698 4.0790

‘: ;.a&; 0.594; 6.0293 0.61136 i.9780 0.5176 4.1242

t ;.7ie7 0.5-G 6.082; 0.61226 6.1370 0.6195 6.0168 0.5-w; 4.1429

10 --- - _ - _ _ _ ___ __- - - _ - - _ _ - - _--

!-)r-1
1.5296
:.
;.7loi 0.832’1 2.7107
:
5 i.8310 0.898; 3.782; 0.898; 2.8310
li --_
i 2.8650 i.9iZi 3.877; 0.953i 3.8774 0.9120 2.8650
_ - -
8
;.ai9i 0.917; i&J56 0.964; i.959; O.&i 3.9056 0.9171 2.87%
1; ___ __- _-_ ___ _-- - _ _ _ _ - - - - __-
July, I9v.l OPTIMUM LADDER NETWORKS 21

TABLE VII.-Element Values (in ohms, henrys, farads) for a Normuli~ed Tschebyschgff
Filter with 3-db Ripple (e = 0.9976, ~2 = 0.9953).

Value Of” C, or L,’ Lr01C,’ cso* 4’ L, or C“ csor L,’

a) I = 0

:: 0.9976
1.55% 0.9109
1.1739 2.0302
4
3 :~~~~ 1.2292 2.5272 1.0578
6” 1:7522
1:7409 1.2501 2.6227 1.3015 21491
1.2606 2.6678 1.3455 2.6309

; 1.7591
1.7638
1.2666
1.2701
2.6750
2.6852
1.3614
1.369il
2.7141
2.7436
::E2
1.3687
2.1827
1.0982
1.2726 1.3733 2.7577 1.3827 1E
1: 1.7670
1.7692 1.2744 ;:3:: 1.3761 2.7655 1.3693 2.7683 SE 1.1032

b) r = l/E

1 8.9787
6.1219 0.25%
: 17.4070 0.1392
0.3884 0.2861
Q.1530 0.1481
0.4011 o/ma7 ‘EE 0.2913
0.1559 0.1600 22.9195 0.1499
0.4048 0.4190 9.0452 0.4128 0.2931
0.1569 0.1623 0.1614 15.6097
0.4061 0.4219 “9:::: 0.4223 ?l::TZ a.5679 0.2939

c) r = l/4

1 4.9881

: 9.3059 0.262; a.1669

4
5 9.7676 0.2866 12.057; a.4724
6 - _ - _--
i 9.8986 0.291; 1;.4ilT 12.2946 0.2&G a.5581
II - _ -
J 9.9530 0.293; 1;.5is; 1&S&; ;.3024 lG69 0.2&9 a.5935
10 --- - _ _ _ _ _ _ _ _ _ _ _ _ _ - - - _ - - _ mm_

d) r = l/3
1 3.9905

t ;.2-w3 0.335i 6.5207

4
;.a37; 0.3652 Glos a.357; 6.7635
: ___
7.7352 0.3712 9.7463 0.3822 GG 0.361; 6.8315
;
G.7760 0.373; ;.a243 0.3870 9.9043 0.3856 9.7256 0.3632 6.8597
1; ___ _ _ _ - _._ _ _ _ _ _ - _ - - _ _ _ - - _ _ - _

e) r = l/2
1 2.9929

; ;.,1, 0.4618 4.8991

4
5 Ti.5259 0.4993 6.9460 0.491; 5.0821
6
;.5i2; 0.5069 $56 o&3; ;.&a; 0.4979 5.1335
;: __- _ - _
5.6197 0.5-G ;&a; 0.5296 <.2-z OS282 ;.1292 0.5-30; 5.1547
1; ___ _ - - _ _ _ _ - - _ - - - - r _ - _ - _ _ - - _

f)r==l
: 1.9953
_ - -
0.7117 3.3487
2 3.3-s
_ _ -
2 3.4813 0.7619 4.537; 0.7619 3.4813
_ - _
7 ;.5ia5 0.772; ;.6390 0.8038 4.6390 ;.7;22 3.5185
_ _ _ _ _ _ - - _
; 3.5339 0.7760 4.6691 0.8118 4.7270 0.8118 4.6691 0.7;60 3.5339
10 --- _ _ _ - - _ _ _ _ _ - _ _ _ _ _ _ _
22 LOUIS WEINBERG [J. F. I.

The use of the tables is illustrated in the example below.

Example 3.1. Determine a ladder network that has the following
characteristics :
1. Low-pass filter with a peak-to-peak ripple in the squared mag-
nitude characteristic not exceeding 15 per cent of the maximum value.
2. A cutoff radian frequency wc = 5000 (the bandwidth being
measured at the minimum value of the ripple).
3. Resistance terminations at both ends with the load and input
resistances equal to 1000 ohms and 500 ohms, respectively.
4. The response is to be down at least 50 db at o = 4w,.
5. The network is to be driven by a current source.
We first calculate the required value of e2. At a trough of the ripple
we have
1
= 1 - 0.15 = 0.85
1 + E2Tn2(1)

E2= 0.176.

Since this value lies between 4-db and 1-db ripple we must use Table IV.
Now we calculate rz. At w = 4
1
1 + H%(4) = 1o-6
1 + e2Pn(4) = 106
e2P,(4) g 105
T’,(4) = 753.

Now using Eq. 5, we have

(w + GzY)n + (w + &F-X)-.
= 753
2 w-4

(w + 21W2-l)+_4 s 1506
(7.88)n = 1506
n = 3.58.

Therefore n = 4 will be more than satisfactory.

Since the specification calls for r = +, we use Table IV(e). Since
the input is a current source the unprimed values are used. Removing
the normalization by multiplying all C’s by & = 5 x’ 1o6, all L’s by
c
July, 1957.1 OPTIMUM LADDER NETWORKS 23

R
-=- 1
and the resistances by R = 1000, we obtain the final element
4 5’
values

5 x’ 106 Cl = 0.363 x 10-G

3 Lz = 0.227

5 x
1106c 3 = 0.498 x 10-C

; Lq = 0.155

1000 R, = 500.

The network is shown in Fig. 8.

1, - 0.155 0.227
0

500
Ez
1000

0 h
”
FIG. 8. Network obtained in Example 3.1 (values in ohms, henrys, and farads).

(To be continued)

REFERENCES FOR PART I

(1) L. WEINBERG,“Network Design by Use of Modern Synthesis Techniques and Tables,”

Technical Memorandum No. 427, Hughes Research Laboratories, Culver City, Cali-
fornia; also published in Proc. Nat. Electronics Conf., Vol. 12, 1956, and in Electronic
Design (four parts), September 1.5,October 1, October 15, and November 1, 1956.
(2) H. W. BODE, “Network Analysis and Feedback Amplifier Design,” New York, D. Van
Nostrand Co., 1945, pp. 361-368.
(3) R. F. BAUM, “A Contribution to the Approximation Problem,” Proc. I.R.E., Vol. 36,
pp. 863-869 (1948).
(4) V. D. LANDON,“Cascade Amplifiers with Maximal Flatness,” R. C. A. Rev., pp. 347-362
(1941).
(5) G. E. VALLEYAND H. WALLMAN,“Vacuum Tube Amplifiers,” Radiation Laboratory
Series, Vol. 18, New York, McGraw-Hill Book Co., Inc., 1948.
(6) E. GREEN, “Amplitude-Frequency Characteristics of Ladder Networks,” Marconi’s
Wireless Telegraph Co., Essex, England, 1954.