Analog and hybrid computing
Electronic analog computers were popular in the period from 1945 to the end of the 1980s. They were used to solve differential equations and to carry out simulations of physical systems. Many simulations were of a technical nature. Important fields of application were physical, electrical, and chemical systems and processes. Especially companies from the aerospace and defence industries used electronic analog computers. But even a company such as the Dutch Gasunie purchased a (hybrid) analog computer at the end of the 1970s. With hybrid computing, the links between the analog computing elements and the input values were controlled by a digital computer, for example, a DEC PDP or NOVA Eclipse. In the end, analog computing became “extinct” and was replaced by digital computing and simulation. The reasons were the increasing costs of digital computers, the flexibility of digital computers, and the limited accuracy of analog computers.
Analog calculation can also be done with mechanical and pneumatic systems. A very simple example of a mechanical calculation aid is the slide rule. Numbers on the rule are indicated on a logarithmic basis. By adding logarithms -by sliding the slide rule- the result of a multiplication can be read. Another example is the fire-control on the cruisers HNLMS Tromp and HNLMS De Ruyter. These were partly made up of mechanical components such as gears and camoids (discs with a circumference with bends according to a certain function). Below, we will discuss electronic analog and hybrid computing only.
The basic element of analog computers is the operational amplifier, abbreviated to OpAmp. By adding input signals and feedback output signals via resistors, capacitors and coils, additions, multiplications and integrations can be realised. The OpAmps and the passive components are connected to each other by means of a so-called patchboard. When simulating or calculating a mathematical (differential) equation, the mathematical parameters are represented by an electrical voltage. The development of simulated functions was made visible on an X-Y plotter (course Y as a function of X) or on a multichannel paper recorder (more signals as a function of time were made visible in parallel).
Analog computers at the LEOK
The first analog computer at the LEOK was probably purchased in the early 1960s. It was an Applied Dynamics system with 64 operational amplifiers using electronic tubes. When the calculator was used, it produced a lot of heat. Later, still in the 60s, the system was expanded with another 32 operational amplifiers. Those, however, were already built with transistor technology.
Around 1978 the old and not very stable analog computer was replaced. The analog computer was ‘handed over’ to the Royal Military Academy (KMA) in Breda. The system was used there for some time for teaching purposes and for simulations of the Technical Department.
The new system (1978) was a hybrid computer system manufactured by Electronic Associates Incorporation (EAI), Princeton, New Jersey (US). The hybrid computer was a combination of an analog system and a (digital) PDP-11 computer. The function of the digital computer was mainly to take care of the numerical settings (i.e. the setting of potentiometers) and to quickly change them to repeat the simulations of a system with different initial states or variable parameters. Two LEOK employees followed a course at the manufacturer in New Jersey to operate the system.
During the merger of the LEOK and Physics Laboratory TNO in 1985, the hybrid computer was moved to TNO-FEL, The Hague. The hybrid system was still used to a limited extent for the torpedo project and the UDB project. Almost all simulations were, however, carried out on digital computers from that time. After a number of years (about 1992) the hybrid computer was removed.
Applications of analog computing
Within the defence domain, analog computers are often used for applications in the field of aviation, such as simulations of missile systems. Some examples of the use of the analogue computers at the LEOK were:
- simulating control systems
- simulation of missile-target interaction (also torpedo-target interaction) and studying various homing laws
- calculating the mirror effect of tracking radars in a dynamic situation
- visualising calculated antenna diagrams
- showing the working of an Electronic Warfare (EW) system.
Two large, characteristic projects were:
- the Seacat blind guidance project on the first analogue computer ;
- the project MK37 torpedo improvement started on the first analog computer and was continued on the hybrid computer.
Both projects -described below in more detail- were examples of simulations in which parts of the hardware system to be investigated were connected to the analog computer in order to be able to perform a closed-loop simulation. The dynamic interactions of an air target or missile (respectively a sailing target and a torpedo) were simulated on the analog computer which was connected to the electronic subsystem of the missile/torpedo. Precisely this possibility of linking simulation and a hardware (sub-)system made the use of analog and hybrid computers attractive.
M44 Seacat blind guidance project
In the 1960s, the Van Speyk-class frigates were built according to English design (Leander class). The six frigates were operational from 1967 to 1989, after which they were sold to Indonesia. As weapons, they had, among other weaponry, a 4.5″ cannon and two Seacat launchers, each with four missiles for air defence against enemy aircraft. The Seacat was developed by Short Brothers and Harland in Northern Ireland. Every launcher was equipped with an aiming device with fire control, the M44 fire control system manufactured by Hollandse Signaalapparaten (HSA).
In the original English version of this weapon system, the role of the operator was very complex. The operator of the aiming device could turn the aiming device with his feet and also aim a viewer with his hands at the target. He had to bring the fired Seacat into the radar bundle with an optical system and keep it aligned with the target. The installation then sent radio signals to adjust the Seacat in the direction of the target. This proved to be a very difficult task for the operators and required a lot of training. That is why the project “Seacat blind guidance” was started.
The radar of the M44 could follow an air target. LEOK electronics were developed to automatically monitor and adjust the Seacat. Target and Seacat were simulated on the analogue calculator. The developed electronics could be connected to this model and the operation could be tested in this way. The development was built into the M44 fire control by HSA and proved to be a big improvement.
MK37 torpedo improvement project
The Royal Netherlands Navy had purchased new torpedoes around 1970 for the four 3-cylinder boats and the two Zwaardvis-class submarines. In a modernisation program, the boats were equipped with new fire control equipment (type Mk17) made by HSA. The entire system with sensors, fire control and weapons (both the old British Mk 8 and the American Mk37 mod 2 torpedoes for the Royal Navy) was evaluated in 1970 in which the LEOK collected and analysed the data.
The Mk37 mod 2 was a torpedo that was propelled by an electric motor using energy from a large battery. The advantage of electrically driven torpedoes is that they are relatively quiet. The disadvantage, compared to torpedoes with a fuel engine, is that the available energy is more limited, which leads to a lower speed and shorter range. Therefore, it was decided to convert some of these torpedoes into the Mk37C by replacing the battery and electric motor with a fuel tank and a combustion engine. This resulted in a higher speed and a greater range. The disadvantage, however, was an increase in the noise generated by the engine (self-noise) and the higher speed (flow-noise). This led to more noise in the acoustic system used to detect and track the target.
The new torpedo was evaluated by the LEOK in 1975 together with the LWS-30 sonar developed by Van der Heem. This evaluation showed that the Mk37C did not function properly in all cases. Because of the larger noise, it was harder to detect the target, an enemy ship or submarine. Often the target was lost in the tracking phase.
The Mk37 was not used by the US Navy but was by the navies of Norway and Canada. To investigate the problems, the Navy could not approach the US Navy. However, a collaboration was started with Norway and Canada. To investigate the problems and to find possible solutions, a research assignment was granted to the LEOK. The project included a large number of activities, including:
- research on the electronics of the torpedo
- theoretical calculations of the possible performance of the acoustic system in an active and passive mode
- participation in seagoing trials
- analysis of shots with exercise torpedoes, among other things by reading the recorded data from a 14-channel film recorder
- installation of an amplifier and an electromagnetic tape recorder in practice torpedoes for the recording of the acoustic signals
- advise on the exercise programs together with representatives of the Royal Netherlands Navy, and consultations with the manufacturer of the torpedoes (first Northrop, later Honeywell who took over the program).
Parallel to the research at the LEOK, the Physics Laboratory TNO (PhL) advised on the flow noise and designed a new nose for the torpedo. This development resulted in a drastic reduction in self-noise.
In this project, much use was made of the analog, later hybrid, computer. On the computer, the path of the target and the torpedo was simulated in three dimensions. From the distance between target and torpedo and the direction of the torpedo in relation to the target, the strength of the target signal to be received by the torpedo was calculated including the directional information.
The torpedo was equipped with a transducer split into four quadrants. This allowed the direction of the incoming signal to be determined in the horizontal and in the vertical plane (similar to the mono-pulse system for a tracking radar). The signal receiver and the control electronics were included in the simulation, the transducer was not. The calculated four input signals were converted to signals on the operating frequency of the sonar system via a separately developed amplifier. These signals were presented to the electronics of the signal receiver of the torpedo. This receiver then produced the rudder commands to steer the torpedo. The rudder commands formed the input for the model on the analog computer. In this way, a closed-loop system was created which could be analysed.
The simulations gained much better insight into the functioning of the acoustic system and the electronics of the torpedo. It became clear why the torpedo did not function properly in some situations. Proposed improvements to the electronics could be studied in the simulations and then applied in practice.
The suggestions for improvement were taken over and applied by the manufacturer, usually after thorough discussions. This led to new versions of the torpedo, the Mk37D and later the MK37E. A short article on the modernisation of the Mk37 torpedo has appeared in the International Defense Review in 1983. That article is based on data from Honeywell and it mentions the role of the LEOK.