Get documents about "Sensitivity of Operational Amplifiers Output Circuits to HF
Document Sample
Sensitivity of Operational Amplifiers Output Circuits to HF Disturbance
Hallon J., Bittera M., Kováč K., Szolik I.
Dept. of Measurement, Faculty of Elec. Eng. and Infor. Technology, Ilkovičova 3,
812 19 Bratislava, tel.,fax: +421 2 6542 9600, e-mail: kovac@kmer.elf.stuba.sk
Abstract: This article deals with hf disturbance influencing the operational amplifiers
(opamps) output circuits used in measuring systems functional blocks. This disturbance
coupled to output cabling of functional blocks, which spreads along the cabling, can cause
incorrect function of operational amplifiers and consequently also of the whole measuring
system, but other electronic systems as well.
1. Introduction
With rising number of electronic systems and their concentration in the space an environment
with non-negligible level of wideband high frequency disturbance is rising up. Single devices
and their cable interfaces are in continued contact with this environment. This situation is
especially problematic within large systems from the point of electromagnetic compatibility
(EMC) view. In these systems the dimension of cable system distribution are much larger
than minimum wave length of electromagnetic disturbance. In these systems there are many
external measuring functional blocks (input sensors electronics etc.) connected to basic
system with long cabling. Immunity to hf disturbance of such measuring block may be critical
for correct operation of whole system [3].
2. Problem description
Let us consider typical measuring chain. The measured quantity sensor is situated at
technological process place. Sensor output is amplified and matched by means of simple
amplifier stage using opamp. The transmission path, the cable with appropriate number of
conductors, follows and then measuring board input with voltage follower. The cable is
situated in the diverse electromagnetic field created by electrical devices or their cabling. It
can be placed near the broadcast transmitters or mobile communication tools. Despite the fact
that undesired disturbance transfer coupling has high-pass character and noise signal
transmission can be considerably damped in consequence of design of the whole technical
equipment, significant induced disturbing voltage can be detected on the output terminals of
the transmission path at both its ends. It means that this disturbance will occur on input
terminals of measuring system but also on supplying as well as output wires of input module.
As the influence of disturbance on input and supply terminals is well known, designers
usually use special measures to suppress it. Ones suppose mistakenly, that because of low
impedance of opamp output circuit, it is immune against the hf continues disturbance. In
following text we will deal in details with susceptibility of opamp output to this disturbance.
There are three possible ways, along which the disturbing signal fed to amplifier output
is transferred:
• it goes directly into the output of opamp and from there into other parts of opamp through
its internal circuits,
• through feedback circuits part of signal gets to inverting input of opamp, that is to input
differential stage,
91
• part of signal (specially hf) couples through cross talks between pins of integrated circuit
package and between tracks on printed circuit board to other pins of opamp..
Due to number of non-linearities in semiconductor structures and presence of capacitance
(internal compensating capacitor as well as parasitic capacitance against reference potential)
demodulating effect rises up, which creates additional DC offset at the output of opamp. This
additional offset can not be compensated by external negative feedback circuit because of the
nature of its origin.
The authors of research publications usually deal with mathematical description of input
differential stage [5] or with creation of models suitable for computer simulators (the most
often SPICE simulator) for chosen type of opamp [4] to model hf disturbance susceptibility
mechanism. Generally they work with disturbance coming to the opamp inputs. Description
of behaviour of opamp disturbed at output terminal should be much more complicated and
hardly generalizable for different type groups of opamps. So our aim is to qualitatively
describe the behaviour of wide type range of opamps under the influence of this disturbance.
3. Experimental part
Measurements were performed on the module, which schema is in fig.1. Opamp was
connected as non-inverting amplifier. Its gain was adjusted by changing the ratio of feedback
resistors values. To filter the transfer of hf disturbance through feedback circuit small
capacitor (1 nF) was placed between the non-inverting input of opamp and ground. Disturbing
signal generated by signal
generator was led through 6 dB
attenuator to output circuit of
measured opamp. For better
impedance matching low value
resistors (100 Ω) were added.
This limited the amplitude swing
of measured opamps to 2 V.
Signal from amplifier output was
filtered by RC low pass filter, on
which output the DC offset was
measured by digital storage
oscilloscope.
Frequency of disturbing signal
was swept within the range of 0.1
– 100 MHz. We made also some
informative measurements with
fast opamps up to 500 MHz.
Fig. 1 Schematic diagram of measuring configuration.
After the measurement the
measured data were transferred from storage oscilloscope to PC computer and then processed.
Measurements of frequency dependence of DC offset were made within several
configurations of supplying voltage, gain, feedback filtering and resistors values. Single
supply opamps were measured by ±5 V supply voltage for better symmetry.
For illustration frequency dependence of output offset of OP177 is shown in fig. 2. It
was measured by ±15 V supply voltage, gain of 10 and without filtering capacitor in feedback
circuit. An area A is marked in the figure. It is the area bounded by the curve and frequency
axis, so it depends on both the susceptibility bandwidth and values of offset voltage. It is used
92
as integral quantitative
parameter representing opamp
sensitivity to hf disturbance [1].
The second quantitative
parameter is maximum offset
voltage.
Our aim was to cover the
wide range of frequently used
opamps including CMOS,
bipolar and unipolar technology
by our measurements.
Measured opamps with the
Fig. 2 Frequency dependence of output offset of OP177 results of our measurements are
opamp by gain of 10. listed in table 1. There slew rate
(SR) and idle status
consumption (Icc) are typical
catalogue data and may slightly differ for circuit from various manufacturers. Umax is
maximal measured offset voltage and Area is the diagram area already explained. All shown
results were obtained by measurements within frequency range up to 100 MHz.
Programmable CMOS
opamps were measured in all Table 1: Maximal mesured offset voltage and curve area
three typical configurations. SR Icc |Umax
Type Technology Area
TS271 and TLC271 circuits µ
[V/µs] [mA] [mV]
have similar marking but TLC271L CMOS 0.03 0.018 120 32.0
distinct internal structure and TS271L CMOS 0.04 0.01 94 33.1
LM741 bipolar 0.25 1.7 108 20.5
therefore they may behave
OP07 bipolar 0.3 2.5 44 5.83
differently. So we present them
OP177 bipolar 0.3 1.6 34 44.7
both. TL061 J-FET input 3.5 0.2 530 121
TLC271H CMOS 3.6 1 92 28.3
LTC1050 chopper stab 4 1 81 7.27
4. Resulting remarks: TS271H CMOS 4.5 0.8 10 1.42
LF355 J-FET input 6 2 6 0.96
1. Measured offset LT1128 bipolar 11 7.6 29 1.63
characteristics were LF356 J-FET input 12 5 2 0.35
practically identical for all TL081 J-FET input 13 1.4 26 1.00
opamps of the same type. LF411 J-FET input 13 1.8 4 0.22
For instance we measured NE5534 bipolar 13 4 2 0.13
741 circuits of several
manufacturers. Similar results were published in [1].
2. The offset voltage is caused both by direct influencing of output circuit and by entering
into opamp input via feedback circuits. At high frequencies influence of cross talks
between separate pins of IC package may rise up. It is clearly visible in fig. 3 that
resulting curve is summation of two causes: direct disturbance influence on the output and
indirect influence via feedback circuit.
3. The offset voltage value is non-linearly dependent on amplitude of hf disturbance. The
dependence is steeper than quadratic one.
4. The polarity of offset voltage may change with frequency. It may mean, that there are
more than one demodulating element participating in this process. This implication is
supported by two curves shown in fig. 3, where offset is negative without feedback
93
filtering, when the hf
disturbance may go into
input of opamp. But it is
positive for connection with
filtering capacitor placed at
the inverting input, when the
hf disturbance can not enter
the opamp along input pin.
5. The offset voltage lowers
down with rising supply
voltage. This fact may be
explained by higher currents
flowing within internal
connection and through Fig. 3 Output offset voltage of LM741 with and without
internal nonlinearities of filtering of feedback signal.
opamp structure.
6. Opamps with higher power consumption as well as opamps with higher slew rate are less
sensitive to hf disturbance.
Our resulting remarks we created according to results of measurements made on opamps
listed in table 1. Although our list is limited we may consider them as generally valid.
Conclusions
Presented results show that opamps are sensitive also to the disturbance fed at the output
terminal. It means that the circuits consisting of opamps may be influenced by disturbance
coming through output cabling. This fact highlights the need of filtering measures against the
penetration of disturbing signals not only at input and supplying terminals but also at output
circuits. This is the only way to prevent the risk of heavy distortion of measured signals and
then following incorrect functioning of whole measuring systems.
Acknowledgement
This work was supported by VEGA Grant 1/0160/03 and science–technical project
AV/804/2002.
References:
[1] Speciale N., Leone A., Masetti G.: A classification scheme for EMI failure in opamp
circuits. Proceedings of Int. Symposium on EMC, Rome 1998, pp. 225 – 230.
[2] Masetti G., Setti G., Speciale N.: On the Key Role of Parasitic Capacitances in the
Determination of the Susceptibility to EMI of Integrated Operational Amplifiers.
Proceeding of Zurich Int. Symposium on EMC, 1999, pp. 625 - 630.
[3] Fiori F., Crovetti P.S., Pozzolo V.: Prediction of RF Interference in Operational
Amplifiers by a New Analytical Model. Proceedings of IEEE EMC Conference, 2001,
Montreal, D4-P1-07.
[4] Whyman N.L., Dawson J.F.: Modelling RF interference effects in Integrated Circuits.
Proceedings of IEEE EMC Conference, 2001, Montreal, D4-P2-05
[5] Hallon, J., Kováč, K., Smieško, V.: Sensitivity of Single Supply BJT Opamp to Pulse
Disturbance and Its Modelling. In: 12th International Scientific Conference
"Radioelectronics 2002" : Bratislava, Slovak Republic, 14.-16.5.2002. - pp. 390-393.
94
