Selectivity
Rockwell Collins builds filters from two poles to 12 poles; selectivity is a function of the number of poles
and the type of design. The adjacent figure shows the selectivity for a .01 dB Chebyshev design with 3, 5,
7, 9 and 11 poles. The 3 dB bandwidth of all the filters is 3 kHz.
3, 5, 7, 9 and 11 Pole Designs
The closer a design is to a Butterworth function the greater its stability but the lower the selectivity. For
optimum stability, Rockwell Collins designs all filters of more than three poles with designs ranging from
Butterworth to .05 dB Chebyshev.
Ripple
The filters are designed using theoretical models with ripple less than .05 dB. Actual ripple, due to process
variations, will be higher. A typical Low-cost filter has room temperature ripple of less than 1 dB. Ripple
over temperature is conservatively specified as less than 3 dB.
Ripple is defined in several ways. The definition we use most often is the difference between the minimum
loss and the lowest dip within the passband. Using this method a monotonic round top filter would have 0
dB ripple. A second method to define ripple is from the highest to lowest points within a specified
frequency range. A third method, which is never used by Rockwell Collins Filter Products, is to use the
largest adjacent peak to dip value. This method understates the true ripple.
Insertion loss
Insertion loss, for our torsional series, is typically two dB or less for filters with a bandwidth of greater
than 1000 Hertz. Filters of lesser bandwidth can have insertion losses that are several dB higher. Loss will
increase with temperature, nearly doubling at 85°C.
Delay
3, 5, 7, 9 and 11 Pole Designs
Delay is also a function of the number of poles and type of design. A Chebychev filter will have more delay
than a Butterworth filter. The delay for a 3 dB bandwidth .01 dB Chebychev is shown above for 3, 5, 7, 9
and 11 poles.
Delay compensation can be achieved by bridging across two resonators within a filter. This is shown in the
plot below.
Intermodulation distortion
Intermodulation distortion (IMD) in torsional filters is quite low. The figure below shows the inband 3rdorder IMD of a typical SSB filter. The reference voltage is the source voltage Vo in the filter test circuit.
The wider the bandwidth, the higher the intercept point. The slope of the third-order product is 2.6 and
not 3, as would be expected of a LC bandpass filter. Crystal, ceramic and all electromechanical filters tend
to have slopes that are below 3. The usefulness of knowing the 3rd order intercept point is reduced
because the value is not 3. When doing system calculations, the intermod slope and 3rd order intercept
point need to be considered. A typical 3rd order intercept point for our filters is +30dB.
Shock and vibration
Rockwell Collins torsional mechanical filters pass the following shock and vibration standards as a
minimum:
Vibration: MIL-STD-202, Method 201,10 G’s, 10 Hz to 55 Hz.
Shock: MIL-STD-202, Method 213B, 50 G’s, 11 ms, half-sine.
Rugged construction of our torsional mechanical filters results in low sensitivity to vibration and shock as
shown in the measured data below.
Filter
Type
Shock
Duration
CW filter
100 g, 11
msec
Center
Frequency
Change
(Hz)
Bandwidth
Change
(Hz)
-3
0
SSB
Filter
AM Filter
200 g, 9 msec
-5
0
100 g, 11
msec
-5
-2
200 g, 9 msec
-12
-1
100 g, 11
msec
-14
+1
200 g, 9 msec
-18
+3
Phase matching
Rockwell Collins Filter Products builds phase-matched sets of filters. These are used in range finders and
sonar systems. A typical match would be ± 3 degrees over 70 percent of the passband. Typical groups are
from 2 to 8 filters. If the system has the capability of subtracting out the phase slope difference then
lower cost and improved specification tolerances can be achieved.
Packages
Most popular case styles for our Torsional Mechanical Filters
PS and PL
FP and LP
Dimension
FP
mm
(inches)
Tolerance metal
LP
PS
PL
metal
plastic
plastic
A
( max )
40.64
(1.60)
53.34
(2.10)
32.3
(1.27)
44.0
(1.732)
B
( max )
18.54
(.730)
19.3
(.760)
12.5
(.492)
14.3
(.563)
C
( max )
8.64
(.340)
9.27
(.365)
8.0
(.315)
7.8
(.307)
E
± .51
(± .020)
6.35
(.250)
6.35
(.250)
3.3
(.13)
3.3
(.13)
F
± .25
(± .010)
5.08
(.200)
5.08
(.200)
3.3
(.13)
3.3
(.13)
G
± .25
(± .010)
29.21
43.18
25.4
37.6
(1.150) (1.700) (1.000) (1.480)
H
± .25
(± .010)
.76
(.030)
.76
(.030)
.78
(.031)
.78
(.031)