Specialized Computing Device Based on Integrated Circuits
Abstract Generated abstract
This paper describes a specialized analog time-pulse computing device built on integrated circuits for calculating the reflux ratio in a chemical production control system. The design uses nonlinear function blocks for square-root extraction, a combined square-root and multiplication block, a summing operational amplifier, and a division-multiplication unit based on pulse voltage dividers operating in self-oscillating closed time modulators. The article outlines the operating principles and circuit choices for these units, including the use of integrated amplifiers, Schmitt triggers, logic elements, emitter followers, and field-effect transistor modulators to meet accuracy and temperature requirements. The completed device is reported to operate with an error not exceeding 2.5 percent over 5 to 50 degrees Celsius, for positive input signals up to 10 Hz and output voltages up to 3.5 V.
Full Text
Specialized Computing Device Based on Integrated Circuits
O. V. Botvinnikov, V. L. Kim, I. G. Smyshlyaeva, N. V. Trikhanova
(Presented by the scientific seminar of the Department of Computer Engineering)
A specialized computing device for calculating the reflux ratio in a chemical-production control system according to the formula
\[ z=\frac{\sqrt{x_1}+\sqrt{x_2}+\sqrt{x_3}+x_4\sqrt{x_5}}{x_4\sqrt{x_5}} \tag{1} \]
is an analog computing device of the time-pulse type.
Time-pulse computing devices with pulse voltage dividers possess great flexibility and provide high static accuracy. With their aid it is possible to perform various mathematical operations on periodically recurring time intervals \(Q\) and dc voltages \(U\), in the general case solving the problem of modeling a function \(z\) of many arguments \(x_j\)
\[ z = F(x_1, x_2, x_3,\ldots, x_n). \tag{2} \]
The block diagram of the device is shown in Fig. 1. The following designations are introduced in the block diagram:
1, 2, 3 — nonlinear-function blocks for reproducing the square-root extraction dependence;
4 — nonlinear-function block simultaneously reproducing the square-root extraction dependence with multiplication;
5 — operational amplifier performing the summation operation;
6 — division-multiplication block.
Fig. 1
The required accuracy of reproduction \((\delta = 2.5\%)\), over a wide temperature range \((t^\circ = 5 \div 50^\circ \mathrm{C})\), and the chosen method of construc-
...of nonlinear-function blocks and division-multiplication blocks makes it possible to use integrated microcircuits as the element base.
Nonlinear-function block for extracting the square root
For constructing nonlinear-function blocks, a circuit with cascaded connection of linear pulse dividers (1, 2, 3) was selected. The functional diagram of a device for reproducing the square-root extraction function is shown in Fig. 2. The following designations are introduced in the diagram:
Fig. 2
- TsPI is an interval-selection circuit consisting of a direct-current amplifier and a rectangular-pulse shaper with controllable duty factor.
- IDN is a linear pulse voltage divider consisting of a switch and a low-pass filter.
- \(U\) is an isolation amplifier with transmission coefficient \(K = 1\).
The closed circuit formed by IDN-1, U1, TsPI, IDN-2 is a closed time modulator. Let us consider the static equations. At the output of IDN-1 the voltage is equal to \(U_0\Theta\). At the output of IDN-2 a voltage is generated, called the processing voltage \(U_{\mathrm{proc}}\), equal to \(U_0\Theta^2\). If the difference between \(U_{x5}\) and \(U_{\mathrm{proc}}\) is denoted by \(\Delta U\) and \(\Delta U \to 0\), then \(U_{x5} = U_0\Theta^2\); hence
\[ \Theta = \sqrt{\frac{U_{x5}}{U_0}} . \tag{3} \]
With the appropriate selection of the characteristics of the component elements of the closed time modulator, stable self-oscillations arise in it. The duty factor of these oscillations \(\Theta\) is proportional to
\[ \sqrt{\frac{U_{x5}}{U_0}} ; \]
If one more linear pulse divider, IDN-3, is connected to the circuit in Fig. 2, as shown in Fig. 3, then it is possible to reproduce a dependence of the form
\[ U_z = U_{x4}\cdot \Theta = U_{x4}\sqrt{\frac{U_{x5}}{U_0}} . \tag{4} \]
It is obvious that if, instead of the voltage \(U_{x4}\), a reference voltage \(U_0\) is applied to the output of IDN-3, then the output voltage \(U_z\) will be equal to \(U_z = \sqrt{U_{x5}}\).
The square-root extraction unit has the following parameters:
Maximum input and output voltage \(U = 3.5\ \mathrm{V}\).
Maximum error at normal temperature \(\delta \le 0.3\%\).
Maximum error when the temperature changes \(\delta \le 0.4\%\)
from 10 to \(75^\circ\mathrm{C}\).
Figure 3
Multiplication-division unit
The functional diagram of the unit is shown in Fig. 4. As can be seen from the figure, the division-multiplication unit consists of nodes analogous to the nodes of the square-root extraction units and has an analogous operating principle. The circuit also operates in the self-oscillating mode.
The unit has the following parameters:
Maximum input and output voltage \(U = 3.5\ \mathrm{V}\).
Maximum multiplication error: \(\delta = 0.1\%\).
Maximum division error: \(\delta = 0.3\%\).
Temperature range \(t^\circ = 1 \div 50^\circ\mathrm{C}\).
In the division-multiplication unit and in the functional converters the following integrated circuits are used:
1TSh221G — Schmitt trigger;
1UT221V — DC amplifier;
1KTO11B — emitter follower;
1LB211V — logic inverter;
1UT401B — operational amplifier.
Figure 4
Computing amplifier
The computing amplifier performs the operation of summing four quantities. The input resistances of the computing amplifier are the resistances of the pulse-divider filters of the nonlinear-function blocks; the output resistance is the resistance of the pulse-divider filter.
of the division–multiplication unit. The basis of the circuit is an operational amplifier.
The operational amplifier is built according to a circuit with an MDM (modulation–demodulation) channel. The main stage selected is an integrated operational amplifier 1UT401BV (1UT402A may also be used). The amplifier input current, drift, and input-current difference are large, which makes it impossible to sum voltages from sources with an internal resistance of more than several hundred ohms.
Fig. 5
To reduce the input current and the input-current difference, an integrated emitter follower 1KT011 is connected to the input of the stage, [4], Fig. 5. In such a circuit the input-current difference is
\[ \Delta I_{\text{in}} = 10 \div 100 \text{ nA}. \]
The input current does not exceed
\[ I_{\text{in}} = 1 \div 1.5 \,\mu\text{A}. \]
The AC-signal amplifier in the M-DM channel is implemented on the integrated circuit 1US191. The modulator is built on field-effect transistors with a \(p\)-\(n\) junction of the KP103 type [5, 6].
The amplifier has the following parameters:
static gain \(K = 1000\);
input current \(I_{\text{in}} = 1 \div 1\,\mu\text{A}\);
input-referred zero offset \(\Delta U_{\text{zr}} \le 300\,\mu\text{V}\);
maximum output voltage with negative feedback applied \(U = 6\text{ V}\).
The specialized analog computer operates with an error not exceeding 2.5% over the temperature range from 5° to 50°C. The frequency range of the input signals may be from 0 to 10 Hz; the input signals are positive, and the maximum value of the output signals does not exceed 3.5 V.
REFERENCES
-
E. P. Ugryumov, V. B. Smolov. Time-Pulse Computing Devices. Leningrad, Energiya, 1968.
-
G. Korn, T. Korn. Analog and Analog-Digital Computing Machines. Moscow, Mir, 1967.
-
G. Haske, G. Korn. Computing Technology. Handbook, vol. 1. Moscow, Energiya, 1964.
-
E. A. Kleiner, G. A. Malitskii, Yu. S. Malyshev. Use of integrated operational amplifiers in units of digital instruments. Instruments and Control Systems, 1971, No. 6.
-
E. Mendzhitskii. Operational Amplifiers. Moscow, Energiya, 1967.
-
L. Sevin. Field-Effect Transistors. Soviet Radio, 1968.