LTC1066-1
10
10661fa
APPLICATIONS INFORMATION
WUU
U
DC PERFORMANCE
The DC performance of the LTC1066-1 is dictated by the
DC characteristics of the input precision op amp.
1. DC input voltages in the vicinity of the filter’s half of the
total power supply are processed with exactly 0dB (or
1V/V) of gain.
2. The typical DC input voltage ranges are equal to:
V
IN
= ±5.8V, V
S
= ±7.5V
V
IN
= ±3.6V, V
S
= ±5V
V
IN
= ±1.4V, V
S
= ±2.5V
With an input DC voltage range of V
IN
= ±5V, (V
S
=
±7.5V), the measured CMRR was 100dB. Figure 1
shows the DC gain linearity of the filter exceeding the
requirements of a 14-bit, 10V full scale system.
3. The filter output DC offset V
OS(OUT)
is measured with the
input grounded and with dual power supplies. The
V
OS(OUT)
is typically ±0.1mV and it is optimized for the
filter connection shown in the test circuit figure. The
filter output offset is equal to:
V
OS(OUT)
= V
OS
(op amp A) –I
BIAS
× R
F
= 0.1mV (Typ)
4. The V
OS(OUT)
temperature drift is typically 7µV/°C
(T
A
> 25°C), and –7µV/°C (T
A
< 25°C).
5. The V
OS(OUT)
temperature drift can be improved by
using an input resistor R
IN
equal to the feedback resis-
tor R
F
, however, the absolute value of V
OS(OUT)
will
increase. For instance, if a 20k resistor is added in series
with pin 3 (see Test Circuit), the output V
OS
drift will be
improved by 2µV/°C to 3µV/°C, however, the V
OS(OUT)
may increase by 1mV
(MAX)
.
6. The filter DC output offset voltage V
OS(OUT)
is indepen-
dent from the filter clock frequency (f
CLK
250kHz).
Figures 2 and 3 show the V
OS(OUT)
variation for three
different power supplies and for clock frequencies up to
5MHz. Both figures were traced with the LTC1066-1
soldered into the PC board. Power supply decoupling is
very important, especially with ±7.5V supplies. If nec-
essary connect a small resistor (20) between pins 5
and 18, and between pins 10 and 4, to isolate the
precision op amp supply pin from the switched
capacitor network supply (see the Test Circuit).
INPUT VOLTAGE (VDC)
–6 –5 –3 –1 1 3 5
V
IN
– V
OUT
(µV)
75
50
25
0
–25
–50
–75
100
125
2
1066-1 F01
–4
–2
0
46
V
S
= ±7.5V
T
A
= 25°C
f
CLK
= 1MHz
f
C
= 20kHz
Figure 1. DC Gain Linearity
CLOCK FREQUENCY (MHz)
0 0.5 1.5 2.5 3.5 4.5
FILTER OUTPUT OFFSET VOLTAGE CHANGE (mV)
0.2
0
0.2
0.4
0.6
0.8
1.0
1.2
4.0
1066-1 F03
1.0
2.0
3.0
5.0
V
S
= ±2.5V
V
S
= ±5V
V
S
= ±7.5V
T
A
= 25°C
f
CLK
/f
C
= 50:1
CLOCK FREQUENCY (MHz)
0 0.5 1.5 2.5 3.5 4.5
FILTER OUTPUT OFFSET VOLTAGE CHANGE (mV)
0.2
0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
4.0
1066-1 F02
1.0
2.0
3.0
5.0
V
S
= ±2.5V
V
S
= ±5V
V
S
= ±7.5V
LINEAR PHASE
T
A
= 25°C
f
CLK
/f
C
= 100:1
Figure 2. Output Offset Change vs Clock
(Relative to Offset for f
CLK
= 250kHz)
Figure 3. Output Offset Change vs Clock
(Relative to Offset for f
CLK
= 250kHz)
LTC1066-1
11
10661fa
APPLICATIONS INFORMATION
WUU
U
AC PERFORMANCE
AC (Passband) Gain
The passband gain of the LTC1066-1 is equal to the
passband gain of the internal switched-capacitor lowpass
filter, and it is measured at f = 0.25f
CUTOFF
. Unlike conven-
tional monolithic filters, the LTC1066-1 starts with an
absolutely perfect 0dB DC gain and phases into an “imper-
fect” AC passband gain, typically ± 0.1dB.
The filter’s low passband ripple, typically 0.05dB, is mea-
sured with respect to the AC passband gain.
The LTC1066-1 DC stabilizing loop slightly warps the
filter’s passband performance if the – 3dB frequency of the
feedback passive elements (1/2πR
F
C
F
) is more than the
Figure 4. Passband Behavior
FREQUENCY (Hz)
10
GAIN (dB)
100 1k 10k 20k
1066-1 F04
1.00
0.75
0.50
0.25
0
0.25
0.50
0.75
1.00
T
A
= 25°C
f
CLK
/f
C
= 50:1
R
F
= 20k,
C
F
= 1µF
CURVE D: f
CUTOFF
= 20kHz = 2500 ×
2πR
F
C
F
1
CURVE C: f
CUTOFF
= 5kHz = 625 ×
2πR
F
C
F
1
CURVE B: f
CUTOFF
= 2kHz = 250 ×
2πR
F
C
F
1
CURVE A: f
CUTOFF
= 1kHz = 125 ×
2πR
F
C
F
1
A B CD
cutoff frequency of the internal switched-capacitor filter
divided by 250. The LTC1066-1 clock tunability directly
relates to the above constraint. Figure 4 illustrates the
passband behavior of the LTC1066-1 and it demonstrates
the clock tunability of the device. A typical LTC1066-1
device was used to trace all four curves of Figure 4. Curve
D, for instance, has nearly zero ripple and 0.04dB passband
gain. Curve D’s 20kHz cutoff is much higher than the 8Hz
cutoff frequency of the R
F
C
F
feedback network, so its
passband is free from any additional error due to R
F
C
F
feedback elements. Curve B illustrates the passband error
when the 1MHz clock of curve D is lowered to 100kHz. A
0.1dB error is added to the filter’s original AC gain of
0.04dB.
LTC1066-1
12
10661fa
INPUT
90%
50%
10%
OUTPUT
t
r
t
d
t
s
1066-1 F05
RISE TIME (t
r
)
SETTLING TIME (t
s
)
DELAY TIME (t
d
)
50:1 ELLIPTIC
0.43
f
CUTOFF
3.4
f
CUTOFF
0.709
f
CUTOFF
±5%
±5%
±5%
0.43
f
CUTOFF
2.05
f
CUTOFF
0.556
f
CUTOFF
±5%
±5%
±5%
100:1 LINEAR PHASE
APPLICATIONS INFORMATION
WUU
U
Transient Response and Settling Time
The LTC1066-1 exhibits two different transient behaviors.
First, during power-up the DC correcting loop will settle
after the voltage offset of the internal switched-capacitor
network is stored across the feedback capacitor C
F
(see
Block Diagram). It takes approximately five time constants
(5R
F
C
F
) for settling to 1%. Second, following DC loop
settling, the filter reaches steady state. The filter transient
response is then defined by the frequency characteristics
of the internal switched-capacitor lowpass filter. Figure 5
shows details.
DC loop settling is also observed if, at steady state, the DC
offset of the internal switched-capacitor network suddenly
changes. A sudden change may occur if the clock fre-
quency is instantaneously stepped to a value above 1MHz.
and on the value of the power supplies. With proper layout
techniques the values of the clock feedthrough are shown
on Table 7.
Table 7. Clock Feedthrough
POWER SUPPLY 50:1 100:1
Single 5V 70µV
RMS
90µV
RMS
±5V 100µV
RMS
200µV
RMS
±7.5V 160µV
RMS
650µV
RMS
Wideband Noise
The wideband noise of the filter is the total RMS value of
the device’s noise spectral density and is used to deter-
mine the operating signal-to-noise ratio. Most of its fre-
quency contents lie within the filter passband and cannot
be reduced with post filtering. For instance, the LTC1066-
1 wideband noise at ±5V supply is 100µV
RMS
, 95µV
RMS
of
which have frequency contents from DC up to the filter’s
cutoff frequency. The total wideband noise (µV
RMS
) is
nearly independent of the value of the clock. The clock
feedthrough specifications are not part of the wideband
noise. Table 8 lists the typical wideband noise for each
supply.
Table 8. Wideband Noise
POWER SUPPLY 50:1 100:1 (Pin 8 to GND)
Single 5V 90µV
RMS
80µV
RMS
±5V 100µV
RMS
85µV
RMS
±7.5V 106µV
RMS
90µV
RMS
Speed Limitations
To avoid op amp slew rate limiting at maximum clock
frequencies, the signal amplitude should be kept below a
specified level as shown in Table 9.
Table 9. Maximum V
IN
INPUT FREQUENCY MAXIMUM V
IN
250kHz 0.50V
RMS
700kHz 0.25V
RMS
Clock Feedthrough
Clock feedthrough is defined as the RMS value of the clock
frequency and its harmonics that are present at the filter’s
output pin (9). The clock feedthrough is tested with the
input pin (2) grounded and depends on PC board layout
Figure 5. Transient Response

LTC1066-1CSW#TRPBF

Mfr. #:
Manufacturer:
Analog Devices Inc.
Description:
Active Filter 14-Bit DC Acc. 8th Order LP Filter
Lifecycle:
New from this manufacturer.
Delivery:
DHL FedEx Ups TNT EMS
Payment:
T/T Paypal Visa MoneyGram Western Union

Products related to this Datasheet