Kurzbeschreibung: Hybrider Präzisionsanalogrechner RA770

Complete English translation of the original German-language document (50 pages).

---

Brief Description

Hybrid Precision Analog Computer RA 770

---

Table of Contents

Section	Page
General Information	2
RA 770 Overview	5
RA 770 Profile	7
Computing Elements	8
Programming	17
Operation and Control	19
Input/Output	25
Construction and Configuration	30
Areas of Application	35
Service	39
Technical Data	41
Sales Addresses	45

---

General Information

Despite strong competition from the digital computer, the analog computer has secured a firm place in the solution of technical problems. Its use has long since ceased to be confined to the classical field of control engineering. Starting as an electronic device that only electrical engineers could understand and operate, it rapidly developed into a computing machine that is today employed in many areas of science and technology, and not least for the treatment of purely mathematical problems.

New significance was added in recent years through the possibility of coupling it with a digital computer to form a hybrid computing system, which combines the advantages of both types of computer and has opened up new areas of application.

The analog computer differs from the digital computer in its construction and mode of operation through the following essential characteristics:

• Numerical quantities are represented by electrical voltages. Data output takes place either in the form of curves displayed on oscilloscopes, XY recorders, or multi-channel recorders, or digitally via a digital voltmeter after appropriate conversion of the analog measured quantity into digital information.

• The actual computing unit of the analog computer consists of a large number of individual independent computing elements such as integrators, summers, multipliers, potentiometers, function generators, and others.

• The inputs and outputs of these computing elements are interconnected on a patch panel, which is conveniently positioned on the front panel of the analog computer and is readily accessible.

• The processing of a given program takes place in parallel with respect to time, since all computing elements involved in the solution of a problem operate simultaneously.

  This means that the computation time required is independent of the scope of the respective problem.

• By simply varying the selectable computing speed and computing time, time compression and time expansion can be achieved, which is advantageously applied in the simulation of dynamic systems.

• The analog computer possesses the already-mentioned special computing element, the “integrator,” which permits direct, continuous integration and differentiation — without recourse to numerical methods. It is this advantage over the digital computer that gives the analog computer its special position. It is preferably applicable wherever integrations and differentiations are necessary. From this, the following areas of application can be derived:

  • Ordinary linear and nonlinear differential equations and systems of such equations.
  • Partial differential equations within certain limits.
  • Simulation of dynamic systems that can be described by differential equations.

AEG-TELEFUNKEN recognized the significance of electronic analog computers at a very early stage and began its own development work in this field as early as 1955. Since then, these products have been continuously further developed and improved. This assures the user of mature, proven technology and operational reliability.

AEG-TELEFUNKEN is today in a position to offer the interested party a complete family of analog computers from its own development and manufacture. The product range encompasses, in addition to the Hybrid Precision Analog Computer RA 770 — which is presented in this publication — the desk analog computer types RA 710 and RA 742, as well as the Hybrid Computing System HRS 860.

Approximately 500 systems have already been installed domestically and abroad and have proven themselves in many fields of science and technology.

---

Hybrid Precision Analog Computer RA 770

RA 770 Overview

The Hybrid Precision Analog Computer RA 770 is a universally applicable, fully transistorized 10 V analog computer of accuracy class 0.01%, with a maximum configuration of 142 computing amplifiers, 10 comparator amplifiers, and 84 coefficient potentiometers.

Its technical and systems engineering concept is based on extensive investigations into the main emphases of analog computer applications and economical expansion possibilities.

Starting from a simple, cost-effective basic configuration, the computer capacity can be expanded beyond the maximum build-out of a single RA 770 to a total of 3 RA 770 units. In that case, for example, 426 computing amplifiers are available, of which 90 are integrators. Computing components in plug-in design and module construction, together with a basic unit pre-wired for full build-out, make this possible in a straightforward manner.

Through the use of a reference voltage of 10 volts combined with a favorable circuit arrangement, the desired high bandwidth of the computing elements is achieved. The advantage of low power consumption — and thus the elimination of any need for air conditioning — distinguishes the RA 770, as it does all AEG-TELEFUNKEN analog computers.

The modern desk-top form offers a maximum of convenience and ease of operation, with a functional arrangement and the best possible overview of all operating controls, indicators, and output devices. No part of the computer lies above eye level or beyond arm’s reach when the programmer is in the seated position.

The designation “Hybrid Precision Analog Computer” already indicates that the RA 770 can be equipped with freely programmable digital elements for the construction of control programs, which means a far-reaching automation of the program sequence.

The possibility of use in a hybrid computing system was taken into account at the time of systems design. Via the Hybrid Coupling Unit HKW 860, the RA 770 can be interconnected with the AEG-TELEFUNKEN digital computer TR 86 to form the Hybrid Computing System HRS 860.

Comprehensive and mature hybrid software is available for this system, permitting programming at a problem-oriented language level.

Detailed documentation enables easy and rapid familiarization and provides valuable guidance on the diverse areas of application.

---

[page 5: block diagram figure — “BLOCK STRUCTURE OF THE HYBRID PRECISION ANALOG COMPUTER RA 770” — showing the main subsystems: COMPUTING ELEMENTS, HYBRID COUPLING UNIT, INPUT/OUTPUT, and OPERATION AND CONTROL, with interconnecting arrows]

---

RA 770 Profile

Basic Unit / System

• Flexibly expandable from a simple analog computer to a hybrid computing system.
• 10 V technology guarantees a large bandwidth of the computing elements.
• Low power consumption of less than 1 kW eliminates the need for air conditioning and permits connection to a standard 220 V / 50 Hz supply.
• Pre-wired for full build-out, with computing components in plug-in design and module construction.
• Up to 2 slave computers controllable from the master computer, which means tripling the computing element capacity.
• Option of parallel operation of up to 3 desk analog computers RA 710, RA 741, or RA 742.
• Interchangeability of nonlinear computing elements between the RA 770 and the desk analog computers RA 710, RA 741, or RA 742.

Control and Programming

• Central and free digital control of computing sequences.
• 9 pre-wired computing and test programs selectable by pushbutton.
• All integrators individually controllable via the digital add-on unit.
• 6 digital timers, of which 4 are each settable in two decades or one timer settable in 4 decades.
• Operating mode and program selection, as well as automatic setting of servo potentiometers, by pushbutton pre-selection.
• Continuous adjustment of servo potentiometers in all operating modes, independent of the automatic setting procedure.
• Separate, exchangeable analog and digital patch panels to facilitate programming and economical computer expansion.
• Fully shielded analog patch panel, as well as patching cords and connectors, ensure good cross-talk attenuation.
• Digital add-on unit individually configurable with a large number of digital elements.

Computing Elements

• Construction of computing elements in silicon planar technology guarantees outstanding noise and drift behavior.
• Computing amplifiers with 550 kHz bandwidth, high long-term stability, and short recovery time after overdrive.
• Computing amplifiers, potentiometers, and reference voltage protected against overload by automatic current limiting. No replacement of fuses required.
• All integrators switchable as integrators/complementary integrators, stores/complementary stores, summers, and open-loop amplifiers.
• Special integrators for use as track/store units with extremely short tracking time constants and outstanding hold characteristics. Switchable as summers.
• All integrators equipped with electronic control switches (field-effect transistors) and up to 4 time constants.
• All integrator capacitors temperature-compensated. No thermostat stabilization required.
• Special noise generator with outstanding normal distribution and power spectral density, as well as high bandwidth.
• Special resolver for coordinate transformation and coordinate rotation.
• Fast comparators, optionally with electronic or mechanical comparator switches.
• Comparator switches also usable as digital-to-analog switches, directly controllable by digital elements.
• Pre-wired shift registers and counters — forward- and reverse-shifting or forward- and reverse-counting.

Input/Output

• Built-in digital voltmeter with 4-digit address display and 6-digit value display.
• Digital printer for logging the digital voltmeter display.
• Built-in dual-beam storage oscilloscope with switchable bandwidth and wide-ranging time base.
• XY recorder in a drawer for functionally correct handling.
• Multi-channel recorder connectable.
• Electronic delay-line device for extending the range of applications.
• Hybrid coupling unit for connection to the AEG-TELEFUNKEN digital computing system TR 86.

---

Computing Elements

Thanks to its universal applicability, the Hybrid Precision Analog Computer RA 770 is equipped with all linear, nonlinear, and digital computing elements as required for analog and hybrid computing.

These computing elements are characterized by high static and dynamic accuracy and permit high computing speeds. Particular attention was also paid to ease of use and reliability (for a table of the most important specifications, see the chapter “Technical Data”).

Linear Computing Elements

Computing Amplifiers

The computing amplifier is the most important building block of an analog computer. Its characteristics determine, to a large extent, the static and dynamic accuracy not only of linear computing operations such as addition and integration, but they are also essential for numerous nonlinear computing operations.

In the RA 770, computing amplifiers of various designs are employed, each specially adapted to the respective requirements. They are the product of many years of experience in the development and manufacture of high-quality DC amplifiers at AEG-TELEFUNKEN. Special procedures are used for all computing amplifiers to counter uncontrollable voltage fluctuations and interference voltages (such as noise and drift). For computing operations with high accuracy requirements, chopper stabilization is applied using electronic choppers with their own chopper voltage generation.

In addition to the chopper-stabilized amplifiers, so-called “drift-compensated” amplifiers are available, which employ a less elaborate method for reducing the zero-point error. They represent a cost-effective alternative to the chopper-stabilized amplifier with respect to technical complexity and achievable accuracy. Areas of application include simple inversions and as input or follower amplifiers for multiplications.

A frequently occurring effect during the computing run is the overdriving of computing elements due to scaling and circuit errors. The consequently necessary continuous short-circuit stability is achieved through current limiting to the so-called short-circuit current (see Technical Data). After elimination of the overdrive cause, a short recovery time of the amplifiers is of interest. Special circuit engineering measures were also applied here. The recovery time stated in the Technical Data refers to a return from 10-times overdrive to a summing point voltage U_SP of ≤ 1 mV, i.e., to the maximum component error of 0.01%.

Integrators

The most important and most characteristic computing element of the analog computer — the integrator — is equipped in the RA 770 with all the necessary comfort in control options, speed, and accuracy as demanded today of a universally applicable analog computer.

All integrators are equipped with electronic control switches (field-effect transistors) that enable a minimum computing time of 100 µs.

All integrators can also be used as pure summers through individual switching. To enable the automatic execution of iterative computations, all integrators can furthermore be switched optionally as integrators/complementary integrators or stores/complementary stores.

For each integrator, 4 time constants are available, which can be switched — both individually and centrally by pushbutton (“10x faster” key) — to the next smaller value (next higher integration factor). The special feature of this “10x faster” key is that, in addition to switching to the next higher integration factor, the time intervals for the pause, compute, and hold times are also shortened by a factor of 10, so that time compression or time expansion can be achieved in a simple manner.

When the digital add-on unit is used, additional possibilities for integrator control arise:

• Individual controllability of all integrators in up to 30 groups, independent of the central control
• Automatic time constant switching (“10x faster” switching) for two integrator groups independently of each other
• Automatic time constant switching in arbitrary steps for: a) all integrators simultaneously b) up to 6 different integrator groups independently of each other
• Start of a new computing cycle after a compute halt within 10 µs

The integrator capacitors have negligible temperature coefficients through a special temperature compensation procedure. Thermostat stabilization is therefore not necessary. Each capacitor is individually trimmable.

Store (Track/Store)

For particularly demanding track and hold characteristics, a special integrator is available that operates on the principle of 3-switch control. With this integrator, store (track/store) units can preferably be realized.

Normal integrator and summer operation, as described above, is also possible with this element.

Summers / Inverters

The summers and inverters each consist of the previously described computing amplifier together with an associated resistor network.

The quality of this network affects the static accuracy, bandwidth, phase error, and stability of the summer or inverter. Accordingly, the network resistors have an error of not more than 0.01%, and are matched within a network so that the error within a single network could be reduced to less than 0.005%.

Furthermore, phase compensation was carried out, which means an increase in stability and bandwidth.

Depending on the circuit configuration, the network resistors permit weighting of the input quantities by the factors 1, 10, and 0.1 (see also the chapter “Construction and Configuration”).

---

Nonlinear Computing Elements

Coefficient Potentiometers

The coefficient potentiometers — designed as manual or servo potentiometers — are ten-turn types. Both grounded (one end of the winding is connected firmly to computing ground) and floating potentiometers are available at the analog patch panel. To avoid damage from excessive current loading, each potentiometer is protected by a current-limiting lamp, which provides a non-destructive overload protection.

All manual coefficient potentiometers permit highly precise setting via a precision adjustment knob, which can be secured against unintentional rotation by a locking lever. A mechanical coupling of 2 potentiometers each is provided for a total of 4 potentiometers (tandem potentiometer).

Each servo coefficient potentiometer has a setting motor, a gearbox, and a slip clutch. With a view to a long service life of the motor and potentiometer, the average setting time was established at a value of 1.5 s.

Rapid repetitive or iterative operation also places high demands on the nonlinear computing elements such as multipliers and function generators with respect to bandwidth. For this reason, the aforementioned elements consist of resistor-diode networks, whereby the bandwidth is determined essentially by the computing amplifiers connected before or after them.

Since the accuracy of these networks depends strongly on the temperature coefficients of the diodes and resistors used, special temperature compensation methods are applied for all computing elements.

Multiplier

The multiplier — which permits the computing operations of multiplication and squaring (directly), as well as division and square root extraction (in conjunction with an open-loop amplifier) — operates on the two-parabola method. The approximation of the parabola is accomplished by 16 biased diode segments per parabola branch.

If only the squaring function is desired as output (Y = +X²), a squaring function unit is available that represents a simplified, cost-effective variant of the multiplier.

Fixed Function Networks

For frequently occurring functions, special fixed function networks are available. In the RA 770, sine, cosine, arc sine, and logarithm function networks can be employed. (See table of fixed function networks.)

---

Overview of Fixed Function Networks

Function / Output Quantity	Follower Amplifier	Input Range (X)	Curve Shape	Approximation with Diode Segments
Y = sin(πX/2)	—	−1 < X < +1	Sine curve	2 × 10
Y = sin(2πX)	—	−1 < X < +1	Sine curve (full period)	2 × 10
Y = cos(πX/2)	—	−1 < X < +1	Cosine curve	2 × 10
Y = cos(2πX)	—	−1 < X < +1	Cosine curve (full period)	2 × 10
Y = +arcsin(X)	—	−1 < X < +1	Arc sine curve	2 × 7
Y₁ = +½ lg(+100X₁)	—	−1 < X₁ < −0.01	Logarithm curve	5
Y₂ = −½ lg(+100X₂)	—	+0.01 < X₂ < +1	Logarithm curve	5

---

Universal Breakpoint Function

For special nonlinear circuits, the active universal breakpoint function is available with the following areas of application:

• Double limiting
• Absolute-value generation
• Dead zone
• Sloped hysteresis

Through the use of integrated amplifiers that effect an idealization of the diode forward characteristic (so-called active diodes), highly accurate setting of breakpoint voltages is possible.

For all functions, positive and negative input voltages can be regulated independently of each other per card at two built-in potentiometers (coarse and fine adjustment). External setting via servo or manual potentiometers is possible by simple switching.

Using the example of limiting, this means: maximum values in the positive and minimum values in the negative range can be preset independently of each other.

The 2 breakpoint voltages present per card can also be coupled for the limiting of a single function.

Beyond this, special nonlinear functions can be realized with the aid of open-loop amplifiers, floating potentiometers, capacitor patch plugs, diode patch plugs, and comparators, all of which are available on the RA 770.

Noise Generator

With the noise generator, the behavior of systems under random disturbances (stochastic processes) can be investigated.

A Zener diode is used as the source for the required statistically varying voltage. A built-in regulating amplifier with a downstream modulator and low-pass amplifier (Chebyshev filter, 3rd order) provides a high approximation to the normal distribution (Gaussian distribution) and a constant power spectral density (from 0 to nearly 5 kHz) with a negligible mean value of the noise signal. A further advantage of this special computing element is the high bandwidth (0 to 4.5 kHz), by virtue of which even briefly sampled instances of the noise signal may be regarded as statistically independent.

Function Generators

For the approximation of arbitrary functions, the RA 770 basically offers 2 types of function generators, which differ essentially in the number and shiftability of the breakpoints (abscissa segments).

Variable Diode Function Generator

The Variable Diode Function Generator allows the approximation of arbitrary functions in all 4 quadrants by forming a polygon with 20 linear segments. The 20 breakpoints lie at fixed intervals of 1 V within the range between +10 V and −10 V. For each breakpoint, the slope of the approximating straight line is settable via a potentiometer. To facilitate adjustment, the potentiometers are equipped with fine-drive mechanisms providing a six-fold increase in the rotation range.

Values of ±1.7 V/V and ±5 V/V are selectable for the maximum slope of the segment lines per breakpoint. This switching provides advantages because, with smaller maximum slopes — which are often sufficient for representing functions — smaller errors due to noise at the output are achievable.

Adjustable Universal Functions

The Adjustable Universal Functions allow the approximation of arbitrary functions in all 4 quadrants by forming a polygon of several diode segments with adjustment of the zero crossing, variable slope, and adjustable breakpoints.

The advantages of the Adjustable Universal Functions lie in the possibility of realizing an individual approximation of the problem-specific function using various selectable card types with 5 or 6 diode segments. Multiple cards can be coupled with one another, where normally 4 cards form one function. Through appropriate circuit design of the follower amplifier, maximum slopes between 0.3 V/V, 3 V/V, and more are selectable.

The use of these function generator cards is appropriate wherever the user wishes to implement special functions that are to remain unchanged over a longer period of time. For this reason, a larger degree of operating convenience was deliberately foregone in order to achieve a cost-effective alternative to the Variable Diode Function Generator.

Electronic Resolver

With the Electronic Resolver, the following computing operations can be performed:

• Coordinate transformation: Cartesian coordinates → Polar coordinates (X, Y → R, φ)
• Coordinate transformation: Polar coordinates → Cartesian coordinates (R₁, φ → X₁, Y₁; R₂, φ → X₂, Y₂)
• Rotation of a Cartesian coordinate system by a presettable angle
• Use as a rate resolver through additional circuit connection with RA 770 computing elements

When the resolver is not required for coordinate transformation:

• Execution of 4 multiplications with high-precision parabola multipliers including inverting amplifiers (2 × 2 dependent products)
• 6 inversions

Precise and temperature-stable diode function generators and parabola multipliers are used in the Electronic Resolver. This allows large bandwidths and an outstanding zero-point stability to be achieved. The accuracy is in the same order of magnitude as for high-precision servo resolvers.

---

[page 14: figure showing switching/connection schematic for the Comparator with inputs X, X₁, X₂ … Xₙ]

---

Comparators

The comparators of the RA 770 differ from the conventional comparator through their division into a comparator amplifier (analog-to-digital switch) and two comparator switches (digital-to-analog switches). In order to obtain optimal circuit connection options, the inputs of the comparator amplifier and the analog inputs/outputs of the comparator switch are located on the analog patch panel, while the digital outputs of the comparator amplifier and the digital control inputs of the comparator switch are located on the digital patch panel.

At the digital patch panel, it is now possible to couple arbitrary comparator amplifiers and comparator switches directly with one another, or to include them with their digital interfaces in freely programmable control circuits (see Free Computer Control), thereby establishing a connection between the analog computing elements and digital elements — a typical characteristic of a hybrid analog computer.

For systems without the digital add-on unit, two comparator switches are each permanently coupled with one comparator amplifier. For each comparator amplifier or analog-to-digital switch, an asymmetric hysteresis is settable via an associated potentiometer in order to achieve unambiguous comparator decisions.

The comparator switches or digital-to-analog switches are designed so that they are controllable optionally either from the comparator amplifier or from arbitrary digital elements of the digital add-on unit (see chapter “Operation and Control”).

As an alternative to the comparator switch in electronic design — for achieving short switching times — mechanical comparator switches (relays) are also available when ideal switch characteristics are desired.

Function Switches

Function switches serve for the manual switching of analog and digital signals, whereby the inputs and outputs of these switches are accessible at the analog and digital patch panels respectively.

Switching of analog signals

The corresponding function switches are latching changeover switches with a toggle handle and a defined center position.

Switching of digital signals

These function switches are changeover switches with spring-loaded pushbuttons, which can be locked in the depressed position.

Stepping Switch

This is a decade counter with controlled relays, which is particularly suited to the following functions:

• Automatic parameter variation: with 2 parameters: up to 10 different values per parameter; with 1 parameter: up to 20 values.
• Automatic selection of computing elements, e.g., for the purpose of printing out computing results.

The stepping switch is freely programmable, so that a wide variety of control programs for the above-mentioned functions can be assembled at the digital patch panel.

---

Digital Elements

These serve for the configuration of the RA 770 as a hybrid analog computer and permit arbitrary modifications and additions to the computing programs settable at the central control unit (see also chapter “Operation and Control”).

The range of application of the RA 770 is thereby considerably extended, since the individual programming of iterative procedures — as required, for example, in the solution of boundary-value and optimization problems — is made possible.

Via a separate digital patch panel (see also chapter “Programming”), the digital elements listed below can be interconnected to form digital control programs.

With this selection of digital elements, all necessary control circuits can be programmed.

Further special elements are obtained through combinations of the above-mentioned basic units, e.g.:

• AND or OR operations with positive outputs by simply cascading inverters
• Differentiators by interconnecting a JKT/RST flip-flop with appropriate logic gates

---

Overview of Digital Elements Usable in the RA 770

Type	Symbol	Function	Configuration
INV	—	Inversion of a logic signal (inverter)	1 input, 1 output
NAND2	—	“AND” operation (conjunction) with negated outputs, 2 inputs	2 parallel outputs
NAND4	—	“AND” operation, 4 inputs	4 parallel outputs
NOR2	—	“OR” operation (disjunction) with negated outputs, 2 inputs	2 parallel outputs
NOR4	—	“OR” operation, 4 inputs	4 parallel outputs
FF	—	Flip-flop with pre-store. Usable as JKT or RST flip-flop depending on circuit connection and programming	4 inputs (R, S, T, T̄); 4 outputs (2 × A, 2 × Ā)
MF	—	Monoflop (monostable multivibrator). Pulses with a pulse width settable from 10 µs to 1.2 s	2 inputs (S, T); 2 outputs (A, Ā); 6 control inputs and 1 potentiometer for pulse width setting
SR	—	4-bit shift register for parallel and serial input, forward-shifting	4 flip-flops
SR2	—	4-bit shift register for parallel and serial input, forward- and reverse-shifting	4 flip-flops
ZL	—	Counter, decoding 1 of 16, parallel input, forward-counting	4 flip-flops
ZL2	—	Counter, decoding 1 of 16, parallel and serial input, forward- and reverse-counting	4 flip-flops
Free Diodes	—	e.g., for extending the number of inputs for NAND2 and NAND4	—

Programming

The programming system of the RA 770 is designed for the broadest possible range of applications. The type and arrangement of the patchable inputs and outputs of the computing elements on the patch panel were chosen to ensure the greatest possible availability for the most varied problem solutions. For example, of the maximum 142 computing amplifiers, 60 are implemented as so-called open amplifiers (important for implicit computation techniques) and are converted by simple patching with short-circuit plugs into integrators, memory elements, summers, or inverters. Since the summing points of all open amplifiers and of a further 10 inverters (with fixed feedback) are also accessible at the patch panel, it is possible to increase the number of their inputs — for integrators and summers — and to extend inverters into summers by adding free resistor networks (see also the chapter “Construction and Equipment”). The most common circuits can be realized with extensive use of plugs, i.e., with only a small number of patch cords.

The analog and digital patch panels are spatially separated from each other and are interchangeable. This offers advantages in that, for example, unnecessary duplicate programming can be avoided. Frequently, identical digital programs are combined with different analog programs — e.g., when the same optimization strategy is applied to different analog programs. The reverse case also occurs: when the analog program remains unchanged but the digitally programmed optimization strategy is modified by exchanging the digital patch panel. When extending the computing system with slave computers, the advantage of separate analog and digital patch panels also has a positive effect, since control of the program sequence at the slave computer can likewise be carried out from the digital control program of the master computer.

[page 19: cover/title pages — figure only]

---

Analog Patch Panels

The interchangeable analog patch panels consist of drilled metal plates with 1,872 patch sockets, which can be easily inserted onto the socket field using a parallel guide. Programming is carried out with shielded plugs and patch cords; the patch panel itself establishes the common ground connection.

The patch panels are divided into 10 individual sub-panels, which are largely identical to one another. This represents a great simplification for the user, since essentially only the layout of one sub-panel needs to be learned. The sub-panels feature a meaningful color code for the patch sockets and contain unambiguous identification labels for the computing elements.

In addition to the terminals of the computing elements, the analog patch panel also provides various control sockets, free input networks, multiple sockets, arbitrarily usable cross-connection lines to the digital patch panel, to slave computers, or to external devices, as well as the lines and outputs for the input/output devices. Also already built in are the connections for data transfer when operating in a hybrid computing system. (For an individual listing, see the chapter “Construction and Equipment.”)

Digital Patch Panels

The interchangeable digital patch panels are made of plastic and have 1,768 patch sockets. Programming is carried out with unshielded plugs and patch cords. A likewise convenient color code and mnemonic letter identifiers enable rapid learning of the socket arrangement. In addition, various control lines for the analog computer and the input/output devices, multiple sockets, and arbitrarily usable cross-connection lines (e.g., to the analog patch panel) are available. (For an individual listing, see the chapter “Construction and Equipment.”)

[page 20: analog patch panel diagram — figure only]

[page 21 (internal): figure — Digital Patch Panel DPF770 diagram]

---

Operation and Control

Selection (Addressing)

For the most effective possible processing of problems on the analog computer, in addition to a large selection of high-quality computing elements and good patchability, a wide variety of operating and control options is necessary.

The capabilities of the Hybrid Precision Analog Computer RA 770 in this regard can be summarized as follows:

• Centralized selection of computing elements for measurement, display, and recording by output devices
• Large selection of hard-wired computation programs and test operations
• Centralized time selection and switchability of time constants for setting computation time and computation speed
• Controllability of the computation sequence
• Automatic and manual setting of servo potentiometers
• Straightforward setting of the Variable Diode Function Generators
• Computer control through freely programmable digital elements and control signals for arbitrary extension of the control options mentioned above
• External computer control (e.g., in hybrid computing systems)

Both the centralized selection of computing elements and the display of their output voltages are carried out digitally.

A total of 600 elements can be selected in one master computer and two slave computers, i.e., 200 per computer (potentiometers, integrators, summers, inverters, function generators, measurement sockets, supply voltages). Bistable flip-flop circuits store the selection positions entered by pressing keys on the digital control unit or entered externally. The selected position is displayed on the address field of the digital voltmeter (DVM) and by the illuminated selection keypad.

This constitutes true feedback from the selected computing element, meaning that the associated identifier letter appears in the address field of the DVM display only when the selected computing element is actually present.

In addition to manual selection, the following selection modes are also possible:

Automatic Selection With this mode, arbitrary computing elements or groups of computing elements are selected automatically in succession. In the process, the selection clock rate is automatically switched from 0.5 Hz to 2 Hz when automatic recording by a digital printer is to take place.

External Selection This selection mode is used primarily for operating the RA 770 in a hybrid computing system (e.g., HRS860), where the selection is carried out by the digital computer.

---

Computation Programs and Test Operations

In the interest of convenient programmability, it is expedient to provide the most frequently used computation programs and test operations in hard-wired form. The RA 770 provides a total of 9 programs for this purpose, which are selectable by a simple key press on the digital control unit (see figure).

Continuous Computation All integrators compute continuously until they are placed into the “Pause” or “Hold” state by a key press or an external signal.

Computation with Hold All integrators compute for a selectable time and then remain in the “Hold” state until a special “Continue” command is given by pressing the “Continue” key.

Repetitive Computation All integrators repeatedly cycle through the sequence Pause–Compute–Hold–Pause–Compute–Hold, and so on. If additionally the “Progr.” key (= external program control) is pressed, the restricted settability of the hold time (see Time Selection) is cancelled, and arbitrary hold times can be set by correspondingly programmed digital elements at the digital extension unit.

Iterative Computation For iterative computation procedures, it is possible to split a computing circuit into two sub-circuits, which are controlled alternately by two separate timer systems.

• Automatic Iterative Computation Two separate groups of integrators automatically and alternately cycle through normal and complementary cycles in the sequence:

  Pause 1 – Compute 1 – Hold 1 / Pause 2 – Compute 2 – Hold 2

• Iterative Computation with Manual Start This computation mode corresponds to “Automatic Iterative Computation.” However, after the completion of a sub-cycle Pause–Compute–Hold, the computer remains in the “Hold” state and does not proceed to the next sub-cycle until a “Continue” command is given.

Single Computation This program is started after actuating the “Progr.” and “External” keys on the digital control unit, with the aid of the digital extension unit. The cycle Pause–Compute–Hold is executed once. This is followed by a pause phase, which is terminated by the next start command. This program sequence is particularly suitable for photographic recording.

Null Balancing of Computing Amplifiers The precise zero-point setting of the computing amplifiers substantially affects computation accuracy. It is therefore necessary to perform a null balance at regular intervals. By actuating the “Null” key, the balance can be performed to a minimum of residual voltage at the amplifier output without any additional patching effort at the patch panel. The balance potentiometers are arranged centrally and are easily accessible.

Static Program Checking This special operating mode allows the checking of the computing circuit patched on the analog patch panel — among other things, for correct setting of the coefficient potentiometers.

Dynamic Checking The “Dynamic Checking” operating mode serves to verify the functional readiness of the integrators and provides information about the deviation of integration capacitors from their nominal values.

---

Time Selection

The execution of repetitive and especially iterative computations at high computation speed and accuracy imposes special requirements on the control of integrators. Therefore, all times required for the various operating modes are determined by digital counters. The time base is a central quartz oscillator at a frequency of 100 kHz, from which the basic clock rates for the two mutually independent timer systems — each with 3 timers for controlling the normal and complementary sub-cycles — are derived by frequency dividers. In each sub-cycle, the following times are settable:

Pause, Computation, and Hold Times (for all computation programs) — 2-decade setting

Operating Modes: Pause and Compute	Time Ranges	Resolution
	100 µs – 10 ms	100 µs
	1 ms – 100 ms	1 ms
	10 ms – 1 s	10 ms
	100 ms – 10 s	100 ms
	1 s – 100 s	1 s

Operating Mode: Hold	Times
	100 µs
	1 ms
	10 ms
	100 ms
	1 s

Pause, Computation, and Hold Times (for computation programs “Repetitive Computation” and “Computation with Hold”) — 4-decade setting

In special application cases, the previously described setting of computation time to 2 decimal places is insufficient. There is therefore the option of using a 4-digit setting of computation time for the programs “Repetitive Computation” and “Computation with Hold”:

Operating Mode: Compute	Time Ranges	Resolution
	100 µs – 1 s	100 µs
	1 ms – 10 s	1 ms
	10 ms – 100 s	10 ms

Operating Modes: Pause and Hold	Times
	1 s
	1 s
	1 s

---

Control of the Computation Sequence

The automatic execution of the selectable hard-wired computation programs can be influenced at any time by selecting the following operating modes:

Pause All previously started computation sequences that are currently in the “Pause,” “Compute,” or “Hold” states are terminated and replaced by the “Pause” operating mode. In the “Pause” operating mode, integrators and memory elements adopt the initial conditions programmed at the analog patch panel.

Compute The previously set computation program is started.

Hold All computation sequences that are currently in the “Compute” state are interrupted and transition to “Hold.” The integrators and memory elements retain the output voltage present at that moment.

Continue A computation sequence that was automatically directed into the “Hold” state (see section “Computation Programs and Test Operations”) is restarted.

Overload Hold Upon occurrence of an amplifier overload, the computation sequence is directed into the “Hold” state, regardless of the selected computation program. This overload signal is also available at the digital extension unit and can be used for triggering digital control circuits.

Overload Signal Latching Every overload of a computing amplifier is in principle latched and indicated centrally by a red lamp (for individual indication of the affected element, see the chapter “Input/Output”). By pressing the lamp — which is also implemented as a pushbutton — the overload indication can be cleared, provided the overload has been corrected. Latching of the overload signal is important in the case of brief overloads not immediately recognizable by the programmer, which may nonetheless affect the solution accuracy of a computation.

---

Setting the Servo Potentiometers

The operating functions required for this are designed so that a simple, clear, and fast setting of the computing coefficients is possible. The operator performs the following operations in sequence at the central control panel:

• Selecting the “Potentiometer Setting” operating mode by pressing the “Pot.” key on the digital control unit.
• Selecting the relevant servo potentiometer.
• Entering the coefficient value to be set on the 4-digit entry keypad (see figure). The keys latch in and illuminate simultaneously. This means the setpoint value specified to four decimal places remains visible throughout the entire setting procedure and can — if necessary — be changed digit by digit.

[page 23: figure — Entry Keypad (key panel) for servo potentiometers and function generators]

• Starting the automatic setting procedure by pressing the “POT” key on the keypad. The address and the set value are displayed on the digital voltmeter.

If a setting procedure could not be completed successfully within a maximum of 3.5 s due to a fault or operator error, a red warning lamp illuminates above the “POT” key. By pressing the warning lamp — which is also implemented as a pushbutton — the setting command is cancelled and a new setting procedure can be started.

Independently of the automatic setting procedure, each selected servo potentiometer can be adjusted continuously with a joystick at a selectable speed.

Since this can also occur during the “Pause,” “Compute,” and “Hold” operating modes, the operator has the option of performing parameter variations during a running program not only at manual potentiometers, but also at servo potentiometers.

A further convenience for the programmer is the ability to set servo potentiometers directly according to the output voltage of computing elements.

By taking over the output quantities of arbitrary computing elements into memory elements and subsequently entering them as setpoint values, an automatic conversion of computation results into coefficient settings at potentiometers is possible.

---

Setting the Variable Diode Function Generators

For setting the Variable Diode Function Generators, the same entry keypad as for the servo potentiometers is conveniently used. In this case it serves to select the 20 breakpoints of a function. The output of the function generator to be set is simultaneously connected to the digital voltmeter, so that function values can be read off with great accuracy.

For displaying the function on the built-in oscilloscope or for output on the XY plotter, a special time-deflection generator (ramp) — patchable at the analog patch panel — is available as the time base.

---

Free Computer Control

With the digital elements listed in the chapter “Computing Elements,” as well as the clocks, timers, and control signals of the digital control unit that are brought out to the digital patch panel for free programming, arbitrary modifications and extensions of the control options described previously can be implemented.

Some key applications:

• Optional control of the entire computation program sequence either autonomously or jointly with the centralized program control through the digital control unit
• Individual control of the operating modes and time constants of individual integrators and arbitrary groups of integrators
• Control of the comparator switches (DA switches) by freely available binary signals, independently of the comparator amplifier (AD switch)
• Automatic parameter variation and selection of computing elements
• Extension of the automatic control options of input and output devices
• Transport of control data when operating in a hybrid computing system (e.g., HRS860)

The following technical prerequisites are fulfilled at the digital extension unit (see chapter “Construction and Equipment”):

• Freely programmable, synchronized digital elements, which provide the capability of running arbitrarily many control programs in parallel with one another
• Central working clock for the synchronous sequential logic, selectable via a rotary switch
• Additionally, the following fixed clock rates freely accessible: 10 µs, 100 µs, 1 ms, 10 ms, 100 ms, 0.5 s, 1 s, 2 s
• Special manual clock for convenient step-by-step testing of control programs
• All 6 timers and central control lines for integrator control freely accessible
• Central reset key for initializing all flip-flops, shift registers, and counters; can also be combined with the “Pause” key on the digital control unit

---

Input/Output

Lamp Indicators

A comfortable input/output arrangement makes an important contribution to a good human/machine interface.

In the broadest sense, this encompasses all facilities that:

• Signal the operating states of the analog computer
• Provide information at any time about the current state of the problem computation
• Optically display and document computation results
• Accept and output data as external devices for internal or external further processing

The Hybrid Precision Analog Computer RA 770 fulfills all of these requirements through the facilities described below.

The overload display is implemented such that overloads of computing amplifiers are not only indicated visually but can also intervene directly in the computation sequence (see the chapter “Operation and Control”). To enable easy identification of the overloaded computing amplifier, its address is displayed in a central indicator lamp panel. For the comparators, a status indicator is provided in the central lamp panel (for the master computer) and in the digital extension unit (for up to 2 slave computers). The respective lamp illuminates when binary “1” appears at the output of the comparator amplifier (AD switch). Also indicated in the central lamp panel is the failure of the reference voltage and of all necessary supply voltages. For the digital elements, the switching state is signaled by indicator lamps mounted directly on the rear of the cards. This provides the programmer with convenient testing of control programs.

Digital Voltmeter

The digital voltmeter (DVM) is an integrated component of the RA 770 and serves for the digital value display of all selectable computing elements together with their address.

As can be seen from the figure, the display shows:

Address: 4 digits

The meaning of each digit:

• 1st digit: Identifier letter of the selected computing element (true feedback)
• 2nd digit:
  • 0 = computing element in the master computer
  • 1 = computing element in the 1st slave computer
  • 2 = computing element in the 2nd slave computer
• 3rd digit: 0…9 — address of the individual sub-panel of the analog patch panel in which the computing element is located
• 4th digit: 0…9 — address of the computing element within the individual sub-panel of the analog patch panel

Value: 6 digits including sign

The measured value display ranges from ±1 mV to ±12 V; the measured quantity is displayed in decimal form as a multiple of a machine unit (10 V). The measurement sequence of the DVM can occur in two modes:

• Periodic: measurement is performed automatically — even during the “Compute” operating mode — at a constant rate and displayed at a speed adapted to visual persistence (e.g., with programs “Continuous Computation,” “Computation with Hold,” and “Static Checking”)
• Aperiodic: at the end of each computation phase or setting procedure, a measurement command is generated automatically or via a programmable command

Each aperiodic measurement command can simultaneously trigger a print command to a digital printer that is also integrated into the system; this printer outputs the complete information displayed on the DVM on a paper strip in one line.

Reaching or exceeding the machine unit (value display > 1.0000) is specially highlighted by printing in red.

Digital Printer

[page 25: figure showing DVM display format — figure/diagram with text label annotations: Address, Sign, Measured value; example display “P0 12 +0.2345”]

Dual-Beam Storage Oscilloscope

The oscilloscope is an indispensable output device for displaying and evaluating computation results, particularly when they are present as time-domain traces.

Therefore, a dual-beam storage oscilloscope was incorporated as a fixed component in the system concept of the RA 770. It is positioned centrally, clearly visible to the programmer in the seated position.

The most important features that make this oscilloscope particularly suitable as an output device for the RA 770 are:

• 2 channels, each with 17 sensitivity ranges
• Time deflection from the computer (external) or internally selectable in 24 ranges
• Bandwidth switchable from 500 kHz to 50 kHz
• Persistence time adjustable between 0.2 s and 60 s
• Storage of a written signal between 15 s and > 8 hours
• Brightness modulation with a deflection range of ±10 V
• Beam on/off control according to computer operating modes
• Usable as a service oscilloscope via Y-amplifiers with differential inputs and high gain, as well as an internal horizontal deflection amplifier
• 13 cm flat-face CRT with measuring graticule

[page 27: figure only — oscilloscope panel photograph]

---

Two-Coordinate Plotter (XY Plotter)

XY plotters belong to the standard input/output devices of an analog computer and allow the recording of computation results that are present as time-domain traces and can be observed on the dual-beam storage oscilloscope.

For convenient functional handling, the XY plotter is housed in a drawer of the RA 770. It enables recordings in DIN A3 format with high plotting speed and a clean, practical writing system using disposable cartridges.

Five sensitivity ranges each for the X and Y deflection allow the representation of all problem solutions at selectable drawing scales.

Connection of multi-channel recorders (e.g., an 8-channel recorder) is also provided for.

---

Electronic Delay-Time Unit

This device serves as an external device for extending the range of applications of the RA 770.

The flexibility of this device — which operates as a sampling system with digital storage of measured values — is so great that only the main applications can be cited here:

• Delay element for delaying the input signal before re-output by a presettable time
• Circulating memory for periodic output of arbitrary time-domain traces
• Controlled intermediate memory for iterative computation and for clocked recording — adapted to the recording device — of rapidly progressing one-time processes
• Analog shift register for the simplified simulation of a sampled-data controller

---

Hybrid Coupling Unit

To extend the application spectrum of the RA 770, a coupling to the AEG-TELEFUNKEN digital computing system TR86 is provided. Through such a coupling to the Hybrid Computing System HRS860, the inherent advantages of both computer types can be combined.

The Hybrid Coupling Unit has the task of performing interface adaptation and the exchange of computation data and control information between the two computers.

Construction and Component Layout

The following characteristic properties define the structural concept of the RA 770:

• The base unit is pre-wired for full expansion from the outset. This allows the computer’s complement of elements to be expanded in economically small increments.
• All computing elements are plug-in or drawer-type modules for straightforward maintenance. The user can perform component changes at the installation without assistance.
• Flexible population with computing elements for optimum adaptation to the problems at hand.
• Desk-form construction with functionally appropriate arrangement of computing and operating elements.
• The selection, control, and address system is designed for expansion by two satellite computers.
• Parallel connection of up to 3 desktop analog computers of type RA 710, RA 741, or RA 742, and interchangeability of nonlinear computing elements between the RA 770 and those desktop analog computer types — allowing optimum utilization of available computing resources.
• Straightforward expansion to the hybrid computing system HRS 860 by coupling with the AEG-TELEFUNKEN in-house digital computing system TR 86 (see also the brief description of the HRS 860).

The following illustration gives an overview of the arrangement of the principal structural units accessible from the front of the RA 770. On the rear of the computer, directly behind the analog programming panel, a magazine block accommodates the Integrators/Stores/Summers and the Summers/Inverters. The circuitry and control elements for these computing elements are housed on so-called panel cards, where one panel card is designed for 2 Integrators/Stores/Summers or 2 Summers/Inverters respectively. Additional magazine slots are provided here for:

• Servo potentiometers
• Comparators
• Free inverting amplifiers
• Further nonlinear computing elements

---

Arrangement of the Principal Structural Units of the RA 770

[page 30: figure only — labeled diagram of the RA 770 front-panel layout showing:]

• Storage oscilloscope
• Magazine slots for nonlinear computing elements
• Analog programming panel
• 8 function switches
• 16 manual potentiometers
• Digital add-on unit with digital elements and digital programming panel
• 2 drawers for accommodating nonlinear computing elements
• Drawer for accommodating the XY recorder or insert for nonlinear computing elements
• Base cabinet / drawer
• Desk surface
• Setting keyboard for servo potentiometers and function generators
• 5 inserts for nonlinear computing elements

---

RA 770 Maximum Configuration Overview

The following overview shows the maximum population of an RA 770.

In the fully expanded configuration, an auxiliary cabinet of type RA 776 accommodates additional inserts for computing elements, all of whose inputs and outputs are accessible at the analog programming panel of the RA 770. This overview is limited to a certain standard configuration, as it is also broadly reflected in the assignment and labeling of the corresponding jacks on the analog programming panel. The quantities listed below can be varied relative to one another (details on request).

Analog Computing Elements

Element	Quantity
Computing amplifiers, total	142
— of which Integrators/Stores/Summers	30 ¹⁾
— of which Summers/Inverters	30 ²⁾
— of which Inverters	18 ³⁾
— of which Inverters associated with nonlinear computing elements (44 free available)	64 ⁴⁾
Manual coefficient potentiometers	16
Servo coefficient potentiometers	68
Parabolic multiplier networks	18
Quadratic functions: sin ½πx	18
sin nπx	36
cos ½πx	36 — alternatively or mixed
cos nπx	36
2 arcsin x	36
Settable universal functions	24
Universal break-point functions (limiting, absolute value, dead zone, sloped hysteresis)	8 — alternatively or proportionally mixed
½ lg 100X	16
Noise generators	8
Variable diode function generators (20 diode segments)	2
Electronic resolvers (or 8 parabolic multipliers)	30 ²⁾ ⁵⁾
Free input networks	10
Switches: Comparator amplifiers (A/D switches)	20
Comparator switches (D/A switches)	18 ⁶⁾
Function switches	2
Step switches (10 steps)	—

Digital Elements

Maximum 24 plug-in units of the following elements, freely combinable.

Element	Elements per plug-in unit
Inverters	16
NAND (2 inputs)	8
NAND (4 inputs)	4
NOR (2 inputs)	8
NOR (4 inputs)	4
JKT/RST flip-flops	4
Monoflops	3
Shift register, 4 bit, forward-shifting	1
Shift register, 4 bit, forward- and backward-shifting	1
Counter (1 of 16), forward-counting	1
Counter (1 of 16), forward- and backward-counting	1
16 free diodes	1

Interconnection Lines

Type	Quantity
Analog cross-connection lines	40
Digital cross-connection lines	40
Inputs for A/D converter	16
Multiplying inputs of D/A converter	16
Outputs of D/A converter	16
Interrogation lines (of which 15 freely programmable)	24
Control lines, freely programmable	36
Interrupt lines (of which 10 freely programmable)	16

Input/Output

Device	Quantity
Digital voltmeter	1
Digital printer	1
Two-beam storage oscilloscope	1
Two-coordinate recorder (connection option for multi-channel recorder)	2
Delay-line unit	as desired

By parallel connection of 2 additional RA 770 units, the element complement listed above is tripled.

---

Notes on the RA 770 Maximum Configuration Overview

1) 30 open amplifiers which can be configured as integrators, complementary integrators, stores (track/store), complementary stores, or summers.

2) This quantity is made up as follows:

• 8 open amplifiers with input network 1, 1, 1, 10, 10, S (S = summing junction) configurable as summers.
• 2 open amplifiers with input network 1, 1, 1, S configurable as summers.
• 16 open amplifiers with input network 10, 10, S as inverters, expandable to summers by means of 16 free input networks 1, 1, 10, 10, S.
• 4 open amplifiers with input network 10, 10, S as inverters, expandable to summers by means of 4 free input networks 1, 10, 10, 10, S.

3) Of these, 10 inverters have the summing junction brought out and can be expanded to summers with free input networks.

4) This quantity is made up as follows:

• 24 inverters as input and output amplifiers for nonlinear networks, all freely available.
• 16 inverters as input and output amplifiers for variable diode function generators, of which 8 are freely available.
• 24 inverters as input and output amplifiers for 2 electronic resolvers, of which 12 are freely available.

5) Of these: 8 free input networks 10, 10, 10, S and 2 free input networks 1, 1, 1, S included in the base unit.

6) Of these: 8 accessible at the analog programming panel and 10 accessible at the digital programming panel.

---

Fields of Application

Mathematics, Physics

The hybrid precision analog computer RA 770 is outstandingly well suited for use in science and engineering. The following examples provide a broad overview of the wide-ranging application possibilities of the RA 770:

• Ordinary differential equations: linear, nonlinear, with constant and variable coefficients; systems of ordinary differential equations (coupled ODEs)
• Partial differential equations insofar as these can be reduced to ordinary differential equations
• Linear algebraic systems of equations of moderate size
• Variational problems: eigenvalue, boundary-value, and optimization problems (parameter optimization)
• Simulation of dynamic processes

Mechanical Engineering

• Simulation of machine tools
• Simulation of vibration systems
• Simulation of the performance capabilities of motor vehicles and rail vehicles
• Replication of crank drives
• Simulation of the diesel injection process
• Production quality control — runout deviation of motor vehicle wheels
• Real-time simulation of electronic braking controllers for motor vehicles

Electrical Engineering

• Determination of impact stresses in electrical machines
• Calculation of electron trajectories in electron microscopes
• Treatment of transformer problems
• Treatment of electrical and mechanical networks
• Calculation of synchronous and asynchronous machines
• Correlation direction-finding of multiple transmitters
• Calculation of optimal filters

Control Engineering

• Simulation and measurement of transfer functions
• Replication of closed-loop control elements
• Measurement of frequency response
• Dimensioning of controllers
• Optimization of control loops
• Identification of process parameters

Chemistry and Process Engineering

• Reaction-kinetic problems
• Simulation of reactors
• Calculation of mass-transfer processes in chemical engineering
• Process simulation
• Optimization of hydrogenation processes

[page 36: figure only]

Aerospace and Biology

• Calculation of optimal space flight trajectories
• Representation of airfoils and their streamlines
• Control maneuvers of television and broadcast satellites
• Investigation of pharmacokinetic problems
• Metabolic models (e.g., iron metabolism)
• Heart and circulatory models
• Analysis of galactose turnover in liver patients and healthy subjects

Measurement Data Processing

• Measurement of characteristic quantities of stochastic processes
• Detection of characteristic points of a signal, e.g., maxima, minima, zero crossings, inflection points, etc.
• Filtering of analog signals
• Frequency analysis of analog signals
• Calculation of mean value and variance
• Replication of sampled-data systems
• Construction of synchronous converters and coordinate transformers
• Graphic evaluation of measurements from strain-gauge rosettes
• Representation of static and dynamic stress states
• Perspective representation of computation results
• Automatic processing of gas chromatograms

New application areas are opened up for the RA 770 through coupling with the AEG-TELEFUNKEN digital computing system TR 86 to form the hybrid computing system HRS 860.

The combination of the advantages of both types of computer substantially increases the capability of the system and creates a cost-effective alternative to purely digital large-scale computing installations.

The HRS 860 has proved its value to date, for example in:

• Automatic optimization procedures and simulation of complete systems in control engineering
• Processing of iterative procedures for the solution of partial differential equations
• Processing and analysis of biosignals in medicine
• Simulation of beam-guidance systems in high-energy physics
• Education at technical colleges and universities
• Development and testing of new computing techniques and programming systems

---

Service

The activities of the manufacturer in supporting the customer with all questions concerning the procurement and maintenance of a computing installation deserve to be regarded as equally important as the product itself.

The prerequisites for this are present in outstanding measure within AEG-TELEFUNKEN.

Development, manufacturing, and central sales operations are united at the company’s headquarters in Konstanz, and by virtue of the close communication this affords, they guarantee technically mature products and effective customer consultation and support.

Technical Service

The Technical Service provides advice on questions of site planning and installation of computing installations for the customer’s computing center.

The well-developed network of branch offices (addresses on the last page) guarantees customer-oriented support. All branch offices have well-trained staff both for customer consultation and sales, and for maintenance and service of delivered installations.

Training, Documentation

Training of customer personnel in the programming and operation of the computing systems is likewise of great importance. To this end, AEG-TELEFUNKEN maintains a central training institute in Konstanz with a wide range of courses.

Should the customer decide to service the computer under its own management, AEG-TELEFUNKEN offers the possibility of preparing the customer’s own personnel for their tasks at the central training institute in Konstanz.

Mature documentation and a spare-parts stock assembled specifically for this purpose provide valuable support.

Analog and Hybrid Computing Center

Further support is offered by the Analog and Hybrid Computing Center in Konstanz. Specialists there assist customers in their problem work in order to make optimum use of their computer.

It is also of inestimable value for the user, for example, to be able to discuss technical details of the computing installation directly with the development engineers. Sales staff will be pleased to arrange such a contact.

---

Technical Data

This chapter contains the most important characteristic data of the RA 770. The data are excerpted from the chapter “Technical Data” of the Technical Manual for the RA 770, which contains important boundary conditions and more detailed information. The specifications given are based on measurements using the standard measurement methods of AEG-TELEFUNKEN (in part according to SCI). Unless otherwise stated, data were measured at the computer.

---

1. System

Principal Dimensions

Parameter	Value
Height	1348 mm
Width	1948 mm
Depth (with desk surface)	942 mm
Weight, fully populated	approx. 530 kg

Power Supply

Parameter	Value
Mains connection voltage	220 V ± 10%
Frequency	47–63 Hz
Power consumption, total	approx. 21 kW

Data of the Inverter with 20 kΩ Feedback:

Parameter	Value
Bandwidth (−3 dB), 20 V peak-to-peak	500 kHz
Bandwidth (−3 dB), 1 V peak-to-peak	770 kHz
Phase error at 100 Hz	0.008°
Phase error at 1 kHz	0.08°
TIDE at 100 Hz	1.4 mV
TIDE at 1 kHz	14 mV

Computing Components

Parameter	Value
Input/feedback resistors	20 kΩ / 200 kΩ
Initial condition error (within one network)	5 kΩ / < 0.005%
Computing capacitors	5, 1, 0.5, 0.1, 0.05 µF
Computing capacitor error	< 0.01%, 0.2%, 0.45%, 1.0%, 1.0%

Reference Voltages

Parameter	Value
Voltage	±10 V
Current capacity	±270 mA
Balance error	±0.002%

Operation and Control

Parameter	Value
Fixed programs	9
Selectable elements	3 × 200
Timer base frequency	100 kHz (crystal)
Timer range	100 µs – 100 s

---

2. Analog Computing Elements

Computing Amplifiers (Chopper-Stabilized)

Parameter	Value
Stabilization	electronic chopper
Output voltage	< 12.5 V
Output current	225 mA
Short-circuit current to ground	< 50 mA
Short-circuit current to reference	< 55 mA
DC voltage gain	approx. 3 × 10⁸
Long-term drift (referred to summing junction)	≤ 5 µV/24 h
Temperature drift (typical value)	0.6 µV/°C
0-dB frequency	< 3 MHz*
Recovery time to U_out ≤ 1 mV after 10× overdrive	3 ms

Integrators/Stores

Parameter	Value
Time constants	k = 1, 10, 100, 1000 s⁻¹
Drift in “Hold” (k = 1 s⁻¹)	10 µV/s
Drift in “Operate” (k = 1 s⁻¹)	40 µV/s
Reset time (+10 V to 0 V, ±0.1%) after switching “Operate”→“Initial condition”	0.18 ms
Switching time	2 µs

Integrators/Stores (Track/Store)

Parameter	Value
Number of electronic control switches	3
Time constants	k = 1, 10, 100, 1000 s⁻¹
Drift in “Store”	20 µV/s
Static error of output voltage in “Track”	< 0.01%
Settling time of output voltage from 0 to 10 V ±0.05%	100 µs

Summers/Inverters (Chopper-Stabilized)

Parameter	Value
Bandwidth (−3 dB) of 20 kΩ/20 kΩ inverter, 1 V peak-to-peak	550 kHz
Bandwidth (−3 dB), 20 V peak-to-peak	450 kHz
Phase error at 100 Hz	0.01°
Phase error at 1 kHz	0.09°
TIDE at 100 Hz	2 mV
TIDE at 1 kHz	18 mV
Noise at summing junction (full bandwidth)	< 50 µV rms

Computing Amplifiers (Drift-Compensated)

Parameter	Value
Stabilization	differential bipolar transistor
Output voltage	±10 V
Output current	±20 mA
Short-circuit current to ground	80 mA
Short-circuit current to reference	130 mA
DC voltage gain	> 10⁵
Long-term drift	300 µV/24 h
Temperature drift	25 µV/°C
0-dB frequency	3.5 MHz
Recovery time to ≤ 2 × 10⁻⁴	0.5 ms

Manual Coefficient Potentiometers

Parameter	Value
Resistance	5 kΩ
Resolution	1/0.02%
Setting error	0.01%

Servo Coefficient Potentiometers

Parameter	Value
Resistance	5 kΩ
Resolution	< 0.0125%
Setting error (typical)	0.015% FS
Setting time	0.1 s* – 0.3 s/V*

*) Values measured outside the computer

---

Parabolic Multiplier Network

Parameter	Value
Static error for X = E, 0 < Y < E	0.015% FS
Dynamic product error	0.02% FS
Bandwidth (−3 dB)	140 kHz
Phase error at 100 Hz	< 0.06°
Temperature drift	0.5 mV/°C

sin ½πx, sin nπx, cos ½πx, cos nπx

Parameter	Value
Static error for −0.5E < X < +0.5E — sin ½πx / cos ½πx	< 0.1% FS
Static error — sin nπx / cos nπx	< 0.3% FS
Bandwidth (−3 dB)	140 kHz
Phase error at 100 Hz	< 0.06°

2 arcsin X

Parameter	Value
Static error	< 0.1% FS
Bandwidth (−3 dB)	140 kHz

½ lg 100X

Parameter	Value
Static error	< 0.5% FS
Bandwidth (−3 dB)	140 kHz

Universal Break-Point Function

Parameter	Value
Adjustable limiting range	−10 V … +10 V
Static limiting error	0.1 mV/V
Zero offset with temperature change	< 0.05 mV/°C

Noise Generator

Parameter	Value
Power density constancy up to 4.5 kHz	±1%
Deviation from normal distribution	3%
Total temperature drift (10°C–40°C)	±25 mV

Variable Diode Function Generator

Parameter	Value
Number of diode segments	20
Switchable maximum slopes	±1.7 V/V; ±5.0 V/V
Setting error	< 0.05% FS
Temperature drift (per SCI)	< 1 mV/°C

Settable Universal Functions

Parameter	Value
Number of diode segments	5 or 6 per card
Switchable maximum slopes	±0.3 V/V; ±3.0 V/V and greater

Electronic Resolver

Polar coordinates → Cartesian coordinates:

Parameter	Value
Total static error	< 0.2% FS
Bandwidth, φ = const., R variable	> 180 kHz
Noise (per SCI)	< …

Cartesian coordinates → Polar coordinates:

Parameter	Value
Static error, magnitude	0.2% FS
Static error, angle	< 0.03% FS
Bandwidth, φ = const., R variable	> 140 kHz
Noise (per SCI)	< 0.01% FS

Comparator Amplifier

Parameter	Value
Input sensitivity, adjustable from	1 mV … 25 mV
Hysteresis, adjustable from	1 mV … 50 mV
Switching time	5 µs

Comparator Switch

Parameter	Value
Switching time, electronic	< 2 µs
Switching time, mechanical	600 µs
Maximum interrupting time, electronic	100 µs
Max. switching current, electronic	50 mA
Max. switching current, mechanical	200 mA

---

3. Digital Add-On Unit

Potentials:

Signal	Level
Binary “0”	0 V, tolerance ±1 V
Binary “0”, external input	0 V, tolerance ±1 V
Binary “1”	+8 V … +12 V
Clock pulse	8 … 10 V (1→0 transition, 50 ns/V)
Operating frequency	50 kHz typical

---

4. Input/Output

Digital Voltmeter

Parameter	Value
Display range	±15 V
Error (up to 10.5 V)	0.01% ±0.01%
Input resistance	200 kΩ
Conversion time	< 10 ms

Digital Printer

Parameter	Value
Print digits per line	10
Print speed	3 lines/s
Coding	BCD

Two-Beam Storage Oscilloscope

Parameter	Value
Screen area	8 × 10 divisions
Storage persistence, adjustable from	0.2 … 60 s
Storage time, adjustable from	15 s … 8 hours
Sensitivity	0.1 mV … 20 V/div. in 17 ranges
Time deflection (internal)	0.1 µs … 5 s/div. in 24 ranges
Bandwidth, switchable	500 kHz / 50 kHz

XY Recorder

Parameter	Value
Format	DIN A3
Writing speed	max. 75 cm/s
X, Y measuring ranges	5 each (0.5 mV/cm … 5 V/cm with attenuator)
Input resistance	1.27 MΩ

Electronic Delay-Line Unit

Parameter	Value
Storage capacity	max. 1000 words × 10 bits
Delay time, adjustable — a) internal clock	1 ms … 1000 s
b) external clock	0 … ∞

Hybrid Coupling Unit

Parameter	Value
A/D channels	16 or 32
D/A channels	max. 30
Resolution ADC, DAC	14 bits incl. sign
Conversion time ADC	5 µs
Conversion time DAC	5 µs typical (max. 10 µs)
Conversion accuracy ADC, DAC	±0.01% ±½ LSB
Cycle-time counter with register	7 bits (programmable)