AD797
NOISE AND SOURCE IMPEDANCE CONSIDERATIONS
LOW FREQUENCY NOISE
The AD797’s ultralow voltage noise of 0.9 nV/√Hz is achieved
with special input transistors running at nearly 1 mA of collector
current. It is important then to consider the total input referred
noise (e
N
total),
which includes contributions from voltage noise
(e
N
), current noise (i
N
), and resistor noise (√4
kTr
S
).
e
N
total =
[e
N
2
+
4
kTr
S
+
4 (i
N
r
S
)
2
]
l/2
where
r
S
= total input source resistance.
This equation is plotted for the AD797 in Figure 30. Since opti-
mum dc performance is obtained with matched source resis-
tances, this case is considered even though it is clear from
Equation 1 that eliminating the balancing source resistance will
lower the total noise by reducing the total r
S
by a factor of two.
At very low source resistance (r
S
<50
Ω),
the amplifiers’ voltage
noise dominates. As source resistance increases the Johnson
noise of r
S
dominates until at higher resistances (r
S
>2 kΩ) the
current noise component is larger than the resistor noise.
100
Equation 1
Analog Devices specifies low frequency noise as a peak to peak
(p-p) quantity in a 0.1 Hz to 10 Hz bandwidth. Several tech-
niques can be used to make this measurement. The usual tech-
nique involves amplifying, filtering, and measuring the amplifiers
noise for a predetermined test time. The noise bandwidth of the
filter is corrected for and the test time is carefully controlled
since the measurement time acts as an additional low frequency
roll-off.
The plot in Figure 4 was made using a slightly different tech-
nique. Here an FFT based instrument (Figure 31) is used to
generate a 10 Hz “brickwall” filter. A low frequency pole at
0.1 Hz is generated with an external ac coupling capacitor, the
instrument being dc coupled.
Several precautions are necessary to get optimum low frequency
noise performance:
1. Care must be used to account for the effects of r
S
, even a
10
Ω
resistor has 0.4 nV/√Hz of noise (an error of 9% when
root sum squared with 0.9 nV/√Hz).
2. The test set up must be fully warmed up to prevent e
OS
drift
from erroneously contributing to input noise.
NOISE – nV/
Hz
10
TOTAL NOISE
1
RESISTOR
NOISE
ONLY
3. Circuitry must be shielded from air currents. Heat flow out
of the package through its leads creates the opportunity for a
thermoelectric potential at every junction of different metals.
Selective heating and cooling of these by random air currents
will appear as 1/f noise and obscure the true device noise.
4. The results must be interpreted using valid statistical
techniques.
100kΩ
0.1
10
100
1000
10000
+V
S
**
1Ω
2
7
1.5µF
6
**
V
OUT
HP 3465
DYNAMIC SIGNAL
ANALYZER
(10Hz)
SOURCE RESISTANCE –
Ω
Figure 30. Noise vs. Source Resistance
The AD797 is the optimum choice for low noise performance
provided the source resistance is kept <1 kΩ. At higher values of
source resistance, optimum performance with respect to noise
alone is obtained with other amplifiers from Analog Devices (see
Table I).
Table I. Recommended Amplifiers for Different Source
Impedances
AD797
3
4
–V
S
** USE POWER SUPPLY BYPASSING SHOWN IN FIGURE 32.
Figure 31. Test Setup for Measuring 0.1 Hz to 10 Hz Noise
WIDEBAND NOISE
r
S
, ohms
0 to <1 k
1 k to <10 k
10 k to <100 k
>100 k
Recommended Amplifier
AD797
AD707, AD743/AD745, OP27/OP37, OP07
AD705, AD743/AD745, OP07
AD548, AD549, AD645, AD711, AD743/
AD745
The AD797, due to its single stage design, has the property that
its noise is flat over frequencies from less than 10 Hz to beyond
1 MHz. This is not true of most dc precision amplifiers where
second stage noise contributes to input referred noise beyond
the audio frequency range. The AD797 offers new levels of per-
formance in wideband imaging applications. In sampled data
systems, where aliasing of out of band noise into the signal band
is a problem, the AD797 will out perform all previously avail-
able IC op amps.
REV. C
–9–
AD [ ANALOG DEVICES ]
