OP27
3. Sudden motion in the vicinity of the device can also
“feedthrough” to increase the observed noise.
2
1/2
Voltage Noise
+
(
)
2
Total Noise = Current Noise × R
+
(
)
S
4. The test time to measure 0.1 Hz to 10 Hz noise should not
exceed 10 seconds. As shown in the noise-tester frequency
response curve, the 0.1 Hz corner is defined by only one
zero. The test time of 10 seconds acts as an additional zero
to eliminate noise contributions from the frequency band
below 0.1 Hz.
2
Resistor Noise
(
)
Figure 4 shows noise versus source-resistance at 1000 Hz. The
same plot applies to wideband noise. To use this plot, multiply
the vertical scale by the square root of the bandwidth.
5. A noise-voltage-density test is recommended when measuring
noise on a large number of units. A 10 Hz noise-voltage-
density measurement will correlate well with a 0.1 Hz to 10 Hz
peak-to-peak noise reading, since both results are determined
by the white noise and the location of the 1/f corner frequency.
100
50
1
OP08/108
UNITY-GAIN BUFFER APPLICATIONS
2
OP07
10
When Rf ≤ 100 Ω and the input is driven with a fast, large signal
pulse (>1 V), the output waveform will look as shown in the
pulsed operation diagram (Figure 3).
1 R UNMATCHED
S
5
5534
e.g.R = R = 10kꢅ, R = 0
S
S1 S2
2 R MATCHED
S
During the fast feedthrough-like portion of the output, the input
protection diodes effectively short the output to the input and a
current, limited only by the output short-circuit protection, will
be drawn by the signal generator. With Rf ≥ 500 Ω, the output is
capable of handling the current requirements (IL ≤ 20 mA at 10 V);
the amplifier will stay in its active mode and a smooth transition
will occur.
e.g.R = 10kꢅ, R = R = 5kꢅ
S1 S2
S
OP27/37
R
S1
R
S2
REGISTER
NOISE ONLY
1
50
100
500
1k
5k
10k
50k
R
– SOURCE RESISTANCE –ꢅ
S
Figure 4. Noise vs. Source Resistance (Including Resistor
Noise) at 1000 Hz
When Rf > 2 kΩ, a pole will be created with Rf and the amplifier’s
input capacitance (8 pF) that creates additional phase shift and
reduces phase margin. A small capacitor (20 pF to 50 pF) in
parallel with Rf will eliminate this problem.
At RS <1 kΩ, the OP27’s low voltage noise is maintained. With
RS <1 kΩ, total noise increases, but is dominated by the resis-
tor noise rather than current or voltage noise. lt is only beyond
RS of 20 kΩ that current noise starts to dominate. The argument
can be made that current noise is not important for applica-
tions with low to moderate source resistances. The crossover
between the OP27, OP07, and OP08 noise occurs in the 15 kΩ to
40 kΩ region.
R
f
–
2.8V/ꢀs
OP27
Figure 5 shows the 0.1 Hz to 10 Hz peak-to-peak noise. Here
the picture is less favorable; resistor noise is negligible and current
noise becomes important because it is inversely proportional to
the square root of frequency. The crossover with the OP07
occurs in the 3 kΩ to 5 kΩ range depending on whether bal-
anced or unbalanced source resistors are used (at 3 kΩ the IB
and IOS error also can be three times the VOS spec.).
+
Figure 3. Pulsed Operation
COMMENTS ON NOISE
The OP27 is a very low-noise monolithic op amp. The outstanding
input voltage noise characteristics of the OP27 are achieved mainly
by operating the input stage at a high quiescent current. The input
bias and offset currents, which would normally increase, are held
to reasonable values by the input bias-current cancellation circuit.
The OP27A/E has IB and IOS of only 40 nA and 35 nA at 25°C
respectively. This is particularly important when the input has a
high source resistance. In addition, many audio amplifier design-
ers prefer to use direct coupling. The high IB, VOS, and TCVOS
of previous designs have made direct coupling difficult, if not
impossible, to use.
1k
OP08/108
500
5534
OP07
1
2
100
OP27/37
1 RS UNMATCHED
50
e.g.RS = RS1 = 10kꢅ, RS2 = 0
2 RS MATCHED
Voltage noise is inversely proportional to the square root of bias
current, but current noise is proportional to the square root of
bias current. The OP27’s noise advantage disappears when high
source-resistors are used. Figures 4, 5, and 6 compare OP27’s
observed total noise with the noise performance of other devices
in different circuit applications.
e.g.RS = 10kꢅ, RS1 = RS2 = 5kꢅ
RS1
RS2
REGISTER
NOISE ONLY
10
50
100
500
1k
5k
10k
50k
R
– SOURCE RESISTANCE –ꢅ
S
Figure 5. Peak-to-Peak Noise (0.1 Hz to 10 Hz) as Source
Resistance (Includes Resistor Noise)
–12–
REV. A
ADI [ ADI ]
