Wirkungsweise des ungarischen Kommandogerätes GAMMA-JUHÁSZ

Complete English translation of the original German-language document (28 pages).

Acquired by Switzerland from 1937 onward and manufactured under license by HASLER: the Hungarian fire-control computer GAMMA-JUHASZ. On the left, the two telescopes used to track the aircraft.

Mechanical Analog Computers for Heavy Anti-Aircraft Guns

ca. 1935–1955

Operating Principles of the Hungarian Fire-Control Computer GAMMA-JUHASZ

André Masson, Langenthal

General GUISAN, Commander of the Swiss Army, takes an interest in the accuracy of the GAMMA-JUHASZ calculator. “Generals’ Shooting Course” at the Zuoz firing range, 1940 (just under four years after the first anti-aircraft training course, which was still conducted by the Artillery Division).

Archeo-Informatics, Military Gear-Driven Computers ca. 1935 to 1955: Structure and Function of the GAMMA-JUHASZ Fire-Control Computer

Heavy anti-aircraft guns were used in an attempt to engage aircraft at great altitudes. In Switzerland, up to the end of World War II, 278 heavy 7.5 cm guns were in service, with effective ranges of up to 7 km, along with 101 acoustic locators for early detection, 114 large rangefinders (3 m base) for distance measurement, and 88 GAMMA-JUHASZ fire-control computers.

(Figures from the Flieger-Flab Museum Dübendorf, touring exhibition.)

Even at high muzzle velocities, projectiles were in flight for 15, possibly 20 seconds (in extreme cases, with greatly reduced hit probability, up to 25 seconds). During that interval, slower heavy bombers might travel 2 km — the guns on the ground therefore had to aim far ahead so that a hit was possible at all. The much faster shells from the guns and the slower aircraft had to be at the same place at the same time, and at that point the projectile’s explosive charge had to detonate. Where, then, must the gun aim? When must the shell explode?

To solve this problem, ballistic computers were required — the so-called “fire-control computers” (Kommandogeräte). They were put into service in various countries years before the world’s first digital computers. Mechanical analog computers were used for this purpose — true gear-driven computers weighing approximately 400 kg. The task was not simple: predicting the intercept point requires information about the exact course, altitude, and speed of the aircraft (radar did not yet exist at the time). The various variables — time of flight, distance of the intercept point from the gun, and distance from the aircraft — are all interdependent: to find the intercept point one must know the time of flight, and the time of flight only becomes apparent once the intercept point is already known.

The Swiss Army first ordered one or two US-made SPERRY devices to gain experience. Apparently it was unsatisfied with them; other types were tested, including a Hungarian product, the GAMMA-JUHASZ. This was procured, and from 1938 onward 32 units were delivered. It was subsequently manufactured under license by HASLER and continuously developed further. An extrapolation capability for curved flight paths was added, and later the device was extended with the necessary capabilities to process radar data as well. Radar tracking allowed the aircraft to be acquired through fog, clouds, or at night.

Mathematical Foundation

There are major differences between the SPERRY and GAMMA-JUHASZ devices in the fundamental approach to how the aircraft’s path is captured and mathematically evaluated — that is, in how the prediction of the future intercept point is made.

SPERRY: The aircraft’s position is immediately captured in rectangular coordinates (east/west and north/south). The velocity is also measured in rectangular components, as is the prediction of the lead — the flight path to the intercept point with the shell. This has the advantage that

[page 2 continues:]

addition is very simple: intercept point (in both coordinates) = aircraft position plus lead. The calculated intercept point is then converted back — rather laboriously — into angular values, since the guns require two angular values: azimuth angle and elevation angle.

GAMMA-JUHASZ: Here one remains entirely with the angles already available from the measurement of the aircraft. No rectangular coordinates are needed for the position or velocity of the aircraft. The flight path is reproduced geometrically — several triangles are formed using spindles (distance adjustment) and rules, from which the desired quantities are read off:

a) From the slant range (supplied by the rangefinder) and the measured elevation angle to the aircraft (telescope on the fire-control computer), the flight altitude and the horizontal map distance to the aircraft are read off purely mechanically from a reproduced triangle.

b) The map distance drives the “distance carriage,” on which two “flight-direction arms” are mounted, each fitted with a lead spindle. From these, the position of the aircraft, the aircraft’s compass heading, and the distance from the aircraft to the intercept point are derived. Steel bands measure the distance from the intercept point to the fire-control computer. It is all a geometric representation of the flown course, projected onto the map at a scale of 1:40,000.

c) Using a further geometrically reproduced triangle and a steel band to measure the third side of the triangle, the variables are converted from the fire-control computer to the center of the four guns, which stand at some distance to the side (the so-called parallax correction).

In essence, therefore, no computation in the numerical sense takes place anywhere in the GAMMA-JUHASZ; rather, the flight path is reproduced geometrically, lengths and distances are measured off, and electric motors are run until the correct angles are set.

Once non-linear relationships arise, scale-accurate geometric representations are no longer adequate. For example: after the guns are fired, the velocity of the projectiles is rapidly reduced by air resistance. All such relationships are extensively captured numerically in the GAMMA-JUHASZ by means of shaped solids (Formkörper). These are complexly shaped metal bodies whose surface represents a numerical relationship between three variables. The body is rotated about its axis on the one hand, and translated along the axis on the other — and a stationary feeler pin touches the surface from the side and measures its distance from the axis of rotation. Such shaped solids are used extensively in the GAMMA-JUHASZ. For example, the lead distance of the aircraft during the projectile’s time of flight (a pure multiplication: velocity times time) is also executed with a shaped solid, even though direct lever mechanisms for multiplication exist.

At a single point, rectangular coordinates are used: the wind is decomposed into a component “along the shot” and “transverse to it,” leading to corrections of the already-computed values for the guns. The crosswind affects the azimuth angle; the wind along or against the direction of fire changes the elevation and the fuze setting. The corrections are again stored in shaped solids.

Functional Units in the Geometric Device GAMMA-JUHASZ

(Somewhat simplified for clarity)

Measuring the aircraft — flight simulation in the computer:

Two operators stand at the telescopes and keep the aircraft continuously in the crosshairs. One rotates the elevation (altitude angle) by means of a handwheel, the other the azimuth angle: the entire device is at all times precisely aligned along its longitudinal axis toward the aircraft and continually follows it.

Once the initial rapid slewing to the aircraft has been accomplished, during calm tracking even a single operator could adjust both angles simultaneously with two handwheels (according to the original 1938 Hungarian manual). According to the more recent Swiss Regulations of 1950, two operators are needed for the two angles. At great distances with low-flying aircraft, the two telescopes obstruct each other’s view.

From the rangefinder, information about the slant range to the aircraft arrives. In the reproduced distance triangle (vector parallelogram), the flight altitude and the map distance are thereby also known. The elevation angle set with the telescope is marked in the distance triangle. The distance carriage rolls on rails inside the device and positions itself at a location corresponding to the current map distance to the aircraft.

[page 4 continues:]

Riding on the distance carriage are the measuring roller and the flight-direction arm. The former rolls on an inner intermediate floor that is fixed to the ground and does not rotate with the entire device. The measuring roller is dragged along and remains parallel to the flight direction of the aircraft. This flight direction is transmitted by gear from the measuring roller to the flight-direction arm, which can rotate 360° as required. The point at lower left in the flight-direction arm symbol represents the aircraft; the point at upper right represents the intercept point, which is extrapolated forward in the direction of flight by the lead distance. A taut steel band measures the distance from the fire-control computer to the intercept point (not sketched here; see the original gear drawings later). Every 250 meters of flight path traversed by the aircraft, the measuring roller sends an electrical pulse to the velocity meter.

In the shaped solids, the time of flight is determined from the altitude and the distance to the intercept point; this is needed for calculation of the lead, as well as the elevation angle of the gun barrels (including the “fall” of the projectiles during their time of flight) and the required fuze setting.

In the corrections, numerous refinements are applied: longitudinal wind, crosswind, parallax corrections because the guns are not in the same location as the fire-control computer, and manual corrections in azimuth, elevation, and fuze setting made after observation of the burst. These corrections are described on pages 17–19.

How well it is possible to determine from the ground whether the shell detonated too early or too late, or whether an angle is off and which one, remains an open question for now.

For better comprehension, the diagram above is drawn in simplified form only. The GAMMA-JUHASZ also takes into account things not shown in the diagram above:

There are two flight-direction arms, because the fuze setting requires a different distance to the intercept point than the calculation of the angles: owing to the loading-time delay, the end of the time prediction already occurs some seconds before the final angle prediction (the moment of firing). See page 27.

There is not only one but two horizontal velocity meters, as well as a third for the vertical velocity (already present in the first prototype that Switzerland purchased).

A further movable carriage simulates the triangles needed for the correction due to the parallax settings — again with a steel band to measure a third triangle side. After taking up position, the distance, angle, and height difference between the fire-control computer and the guns can be entered by hand. Two additional movable carriages serve the shaped solids for determining the projectile’s time of flight and fuze-setting time.

From approximately the devices delivered in 1944 onward (designated Device 43, a HASLER in-house development), an additional extrapolation for curved flight is provided; see pages 21/22.

Numerous readouts of distances, angles, and directions help the crew to maintain an overview or to detect errors.

Practically all quantities are electrically “tracked.” In addition to the mechanical precision parts, there exists an entire world of electromechanical components: contacts, relays, micrometers, motors, brakes for the motors, limit switches and limiters, fuses, etc. See the following section on “follow-up servos,” pages 8 to 11.

Functional Diagram — Partial View of the Central Section with the Distance Carriage

The large circle is the ring gear by means of which the entire computing device is rotated about the vertical axis (handwheel 1 at upper right, diagonal). The distance carriage runs on four rollers between three long gear rollers, which transmit three variable values: at the top, the normal lead vh.t, and additionally vh.k for the loading-time delay; at the bottom, the flight angle. The lead is set on the obliquely positioned flight-direction arms with the spindles; two steel bands (guided over rollers at the left margin of the image) measure the distance from the fire-control computer Kg to the intercept point. Between the flight-direction arms is an equally sized gear for setting the angle; below it (hatched) is the measuring roller, which unrolls the aircraft’s path on a fixed plate that does not rotate with the device. The measuring roller drives motor 34 for the flight angle via micrometer 35. In this image excerpt, five different points are visible, each representing in a different context the aircraft, its position, or aspects of its motion. At the left edge of the distance carriage is the segmented pin 13, which, when the altitude-distance triangle is being set, pushes the carriage away by a distance corresponding to the horizontal distance ekm (k for map, m for measuring point). The height motor 55 (at the left margin, hm) rotates the vertical spindle 56 until the height ruler 52 no longer commands “upward” or “downward” at contacts 53, 54 — at that point the altitude is correctly set in the long spindle and can be used elsewhere for further computations. The rangefinder distance is “tracked” at the handwheel at upper left and sets the correct slant range to the aircraft at pin 13 on the diagonal distance spindle 51.

[page 7: figure only — complete mechanical diagram, drawn 12 February 1941, taken from one of several instruction and operating manuals at the Flieger-Flab Museum in Dübendorf (see picture credits)]

Electrical Follow-Up Servos — Precise Copying of Analog Values

Occasionally, as signals pass through the gear trains, a new power source must be tapped. Tracing the signal paths on the overall diagram, one counts: the signal from the azimuth-angle handwheel (telescope tracking) drives the rotation of the entire device directly, and must be routed in total through nine gear pairs and two differential gear trains. For the position angle (also telescope tracking), five gear pairs must be driven. For the flight altitude, driven by an electric motor, the count is 13 gear pairs and two differential gear trains.

In many devices of other manufacture, new power sources are typically tapped by means of manually operated “follow-up pointers”: the computer indicates the current value of a variable on a rotatable ring, and a person must manually keep a second ring adjusted so that, through all motion, marker remains on marker. The GAMMA-JUHASZ has only a single follow-up pointer — at the point where the value for the slant range to the aircraft, received from the rangefinder via a magnetic synchro display, is taken in. Three rotatable, star-shaped coils in an alternating magnetic field have exactly the same arrangement at transmitter and receiver; with this arrangement, only the signal can be forwarded — no torque for driving the subsequent gear trains can be extracted. At the follow-up pointer, a person carries out this tracking by means of a handwheel — the computer then has received the signal with the correct value, even though the direct signal flow from the rangefinder to the subsequent computing gear trains is broken at the follow-up pointer and not continuously connected.

All other quantities in the GAMMA-JUHASZ are electrically “tracked,” usually by the principle of the electrical “micrometer”: the quantity to be copied is brought in its mechanical motion to an electrical center contact, which is movably positioned between two nearby end contacts that are controlled by — and control — a motor. The position of the center contact cannot be influenced by the motor. When the center contact touches one or the other end contact, the motor changes its variable “upward” or “downward”; when no further change of direction is demanded, the motor’s output has found the correct value of the variable. The principle is explained in the three following illustrations.

Mechanical drawing, two images: Kdo. Fl.-u. Flab-Trp. — Gr. II, drawn 12 February 1941. Electrical drawing: “Model 43,” drawn 5 November 1946, K+W Thun.

If a GAMMA-JUHASZ device were ever opened and disassembled, the electrical contacts of this follow-up control would have to be examined more closely (contact gap, possibly mechanical play of the components, insofar as this can still be measured after 70 years; the motor brake, mentioned below, should also be sought and understood). Measures against unwanted oscillation of the electrical follow-up are not yet known.

Illustration below:

Three electrical follow-up servos are visible in the excerpt. The rangefinder distance sets the length of the diagonal spindle 51 (manual follow-up via the follow-up pointer). The elevation angle α is set from the telescope by hand, via a shaft coming vertically from below, one-third of the image width from the left edge. This sets the segmented, universally displaceable pin 13; the movable distance carriage has assumed the correct distance. The fine horizontal height ruler 52 indicates the flight altitude and supplies this quantity to micrometer 53. The height motor 55 at upper left (number partially cropped) adjusts the vertical spindle 56 to the correct value of the flight altitude, which is then taken up by further gear trains.

The same principle applies for the azimuth lead: on the upper, obliquely positioned flight-direction arm 60, 61, V’ denotes the lead point (intercept point if firing were to occur “now”). M’ is the measuring point (the aircraft), always referenced to the map projection. The angle Gamma is the lead in azimuth. The fine, slightly oblique ruler 47 brings Gamma to micrometer 48; the lead motor 49 serves the output shaft going downward to the right with the correct value for the azimuth lead. Motor 49 and ruler 47 are fixed and do not themselves travel with the distance carriage.

[page 8 continues:]

The measuring roller (hatched, below the carriage) brings the flight direction to micrometer 35; the flight-direction motor 34 must convey its information via the long lower gear roller to the fixed part of the device. The three dashed wires lead to a “flight-direction hand switch” (possibly: whether the aircraft is approaching from the left or from the right).

Illustration below:

Three micrometers are also visible here: the subject is the corrections arising from the parallax values — that is, the positioning of the guns G (battery center) relative to the fire-control computer Kg. The parallax carriage 91 travels in accordance with the distance ekv between the fire-control computer and the lead point V’, driven by motor 97. The horizontal distance from the guns G to the lead point V’ is greater or less than the distance Kg–V’ and is fed via a tension band to micrometer 102, 103, which controls motor 104 (distance plus correction). Its output feeds into the three shaped solids for determining elevation, time of flight, and fuze setting. Finally, the correction of the azimuth angle due to the parallax angle (Psi) is applied via micrometer 98 and motor 99 to the long gear roller, entering as a correction into the azimuth angle to be set at the guns.

Electrical Section of the Follow-Up Servos (illustration below, Model 43)

The excerpt from the electrical diagram shows the follow-up servos for five quantities; from upper left to lower right these are: parallax distance, Psi motor (traveling on the parallax carriage, with contact rails), distance correction due to parallax, elevation lead, azimuth lead. The motor drawn with two stator coils for the two directions of rotation consisted in Models 38 and 40 of two separate motors meshing with the same gear — one for “upward,” one for “downward.” From Model 43 onward, a single motor covers both directions. The excerpt corresponds to roughly one-tenth of the complete electrical overall diagram. Visible at the bottom in each case are the three connections of the electrical micrometer contacts: the center element will touch one of the contacts K1 or K2, left or right. Two of the follow-up servos have these contacts drawn to the side, because there are limit switches with electrical interruption when a numerical value (length or angle) is exceeded or falls below its limit. The current through the micrometer contacts is not the motor current itself, but merely a current through a relay coil. When the relay is activated by the attraction of a crank lever, two circuits are simultaneously switched on: one for the motor (drawn with separate stator coils for each direction of rotation), and a further coil, drawn centrally below the motor, to be interpreted as an electrical solution for a brake — the motion must be stopped instantly when no further motion is demanded. Only while the motor is actively rotating is the brake released magnetically. The capacitor for spark suppression indicates that the motor current is greater than the brake current.

The motors operate on 80 V DC (wires labeled +80 V and -80 V — this does NOT give 160 V). The lighting requires 12 V AC, generated from 110 V AC. The winter heating delivers 100 W or 190 W, plus 250 W or 475 W (possibly cumulative), selectable from 80 V DC or 110 V AC, using the civilian power grid wherever possible.

[page 10 continues:]

The numerous shaped solids have feeler pins whose position is also electrically tracked. Only in the chain of the many final corrections are the various disturbances all added directly with a mechanical lever linkage, and only the final sum is electrically tracked.

HASLER version: tracking of the feeler pins on rotary bodies, drawn 11 April 1940:

At lower left is the shaped solid, which rotates and is translated along its axis depending on the input quantities. The feeler pin is mechanically coupled to an electrical center contact, which determines the direction of the two tracking motors: upward or downward.

The two threaded spindles are connected by a gear and push the mounting carrying the outer contacts (without binding) upward or downward, depending on which motor is active. The contacts of the micrometers are made of tungsten.

There will never be rest here: one of the two motors will (almost) always be running.

Not shown is the path by which the tracked quantity leaves for further use (for example, from one of the three gears at the top).

Note also the lever ratio at the feeler pin, and the contact pressure weight.

In connection with the noise of continuously running or constantly switching motors, a quotation from http://www.bunker-kiel.com/marine-flak-brigade-i/marine-flak-abteilung-ugruko-251schönwohld/1-251-nordmark/ (there, Chapter 4):

“In the course of 1944, the Hazemeyer device was removed and taken away for an unknown reason. Instead we received a new device, the L 40 (?), which we did not find as good, because it had very loud motors, so that one could no longer hear low-flying aircraft approaching. Moreover, it had no slip rings but only a thick cable on a gallows, so that during operations we had to wind the device back after several rotations to prevent the cable from tearing.”

Photographs: Exterior and Interior Views

The following photographs are taken from fine original documentation received, according to a handwritten note, by FW Kögel from Mr. Zadory, dated Zuoz, 25 August 1938 and 18 October 1938. In later years the reproductions become noticeably poorer (possibly because the booklets for the troops needed to be produced in larger editions). The paper is slightly wavy from the gluing-in of the original photographs. No French-language text as in later editions. Probably produced in Hungary. The schematic drawings are of later origin.

[page 11: figures only]

The device is constructed in stacked layers. At upper right, the data transmitters for the guns: azimuth, elevation, and fuze setting, with coarse and fine readout of values.

[figure caption page:]

The aircraft is tracked with the two telescopes; the whole device swings laterally (hand crank below the right telescope). At very low angles the telescopes obstruct each other. The eyepiece of the telescope is always horizontal (more comfortable).

[figure caption page:]

Layer-by-layer construction of the individual computing gear trains. Information must be transmitted mechanically at each stage — assembly and disassembly of the layers is delicate. In the center, the distance carriage. HASLER version 43.

[figure caption page:]

View of the topmost layer with the many fine corrections. On the right, the three transmitters for electrically sending the output values to the guns. Servomotors = electric motors for the “electrical tracking” of variables (pp. 8–11).

[page 12: figures only]

View of the distance carriage. Four wheels at the extremities. At center top, the two motors for the flight direction: they control the two flight-direction arms, on which the lead is plotted in the correct direction. Via a tension cable (not visible), the distance from the fire-control computer to the intercept point is measured to determine the time of flight. At upper right, a gear that meshes into the long gear roller while traveling, in order to receive the lead value (drive).

[figure caption page:]

Below the distance carriage (large, bright circle at the bottom), the measuring roller projects downward to a fixed base plate. Here the path and direction of the aircraft’s course (compass heading) are plotted. Every 250 m of flight path there is an electrical pulse, from which the horizontal velocity of the aircraft is determined. The vertical velocity is computed differently: the change in altitude is determined at constant time intervals.

[figure caption page:]

The distance carriage is installed, approximately in the center. Visible are the two flight-direction arms and one of the long gear rollers for transmitting information to the traveling carriage. In the foreground, large (i.e., high resolution), the handwheel for setting the distance to the aircraft (the only follow-up pointer).

[figure caption page:]

A later attempt to make the fire-control computer (further development HASLER 43, housing has become more angular) mobile and directly connected to the rangefinder, similar to devices in other countries. A configuration unknown to the troops.

[page 13: figures only]

Velocity Measurement (Horizontal)

Two identical velocity meters are available as alternatives. The output of one feeds directly into the lead computation; the output of the other drives a pointer so that the value can be read off — a (averaged) value can then be entered by hand, and this enters the further computation. Both measuring systems determine the velocity of the aircraft from the information of the “measuring roller,” which unrolls the aircraft’s path (map projection) to scale. Every 250 m, the measuring roller closes an electrical contact, which always concludes the old velocity measurement and initiates a new one. Diameter of the measuring roller: 9.95 mm.

Velocity is measured as follows (cf. the two schematic drawings below):

There are two centrifugally governed (!) speed-stabilized electric motors: the “time motor” runs continuously; the “auxiliary motor” is switched on and off via contacts. The time motor is rigidly coupled to the “drive wheel” at the far right, which completes one revolution in 12 seconds. There is furthermore (simplified) the “measuring wheel” in the middle, which is laterally displaceable, and the “setting wheel” at the end, which passes on the completed velocity value to the lead multiplication — aircraft velocity times projectile time of flight (or merely pointer movement in Measuring System I).

The auxiliary motor, via a step disc and a control arm, dictates how the measuring wheel passes the measured value laterally further. Before each individual pulse of the measuring roller, the measuring wheel is pressed against the drive wheel and rotates linearly with time. When the measuring roller delivers the electrical contact, the measuring wheel is released from the drive wheel, arrested with respect to rotation, pressed against the setting wheel on the other side — which in turn takes over the angle from the measuring wheel — and is then also arrested. Finally, the measuring wheel is freed with respect to rotation, falls back to its zero position, and is again pressed against the drive wheel. The auxiliary motor is switched off at the cam wheel by electrical contact (Phase IV), and everything waits for the next pulse of the measuring roller. For a slow aircraft, the measuring wheel remains pressed against the drive wheel for a longer time; for a fast aircraft, only briefly. — For further explanation, see pages 17–22 of Document 15 (1941), archival reference numbers in the picture credits at the end. The words “right” and “left” are reversed there with respect to the diagram reproduced here.

Measurable velocity range for the aircraft:

Model	Min	Max
First devices, 1938	27 m/s	150 m/s (= 100–540 km/h)
Kgt. 40, 43	40 m/s	200 m/s (= 150–720 km/h)
Kgt. 50	—	300 m/s (= 1080 km/h)

Maximum possible horizontal lead:

Kgt. 38, 40: 2,050 m
Kgt. 43: 2,700 m
Kgt. 50: 4,500 m

Vertical velocity range:

Kgt. 40: +15 to −145 m/s
Kgt. 43: +20 to −200 m/s

Time interval between measuring-roller pulses, for v = 27 / 40 / 150 / 200 m/s: 9.25 / 6.25 / 1.66 / 1.25 seconds.

In Document 15 (1941) p. 17, the revolution time of the step discs is given as 1⅔ seconds, likewise in the left-hand diagram (1941); in the photograph further below as 1.5 seconds (1938). Strange — should the disc have become slower in later versions? These times determine the duration of the velocity measurement — not the measurable velocity range.

Two diagrams of the horizontal velocity meter: on the left, the electrical section, with the measuring roller at lower right, which delivers an electrical pulse every 250 m of flight path. Above it, the spiral “cam disc” (on the left in four positions), which in the right-hand image (mechanical section) can be seen in the center at the top, attached to step disc II. In Phases I through III, the auxiliary motor and step discs rotate.

The measured value is pushed stepwise to the left from the active velocity meter to the output, where the two lead distances (aircraft velocity times time of flight, with and without loading-delay time) are determined by means of shaped solids. The lead distances are used in the flight-direction arms to measure off the distance to the intercept point. The vertical velocity is determined elsewhere.

[page 14: figures only]

Photograph of the Velocity Meter in Its Technical Form

On the left, with a horizontal axis, the time motor. On the right, with a vertical axis, the auxiliary motor. Note the term “brake magnet” at the top — the hypothesis from the electrical diagram was correct: the motors must be stopped immediately when no further motion is demanded. At lower left, the two pointers (red and black), which display the output of the two velocity-measuring systems in a window on the housing. A relay for the auxiliary motor is identified; attached to the auxiliary motor (right) is a further one. With the control lever at the far right, a selection can be made: automatic, manual (with running manual entry of velocity), or fixed velocity value. Behind the cam discs are the displaceable gears that pass on the measured value — a frozen angle of rotation. The control arm shifts the gears depending on the position of the cam discs (also called “step discs” in the diagrams).

Image above: In the right half, the velocity measurement (two motors clearly visible), whose result is converted in the two multipliers (shaped solids, displaceable along the gear rollers) in the center of the image into the lead distance.

[page 15 continues:]

Why Were Two Identical Velocity Meters (with Different Outputs) Provided for the Device?

First find:

From personal notes, “Page 7,” 1938, Document 18, 19, or 20: the vertical velocity measurement should be deactivated if possible, in order to protect the expensive and sensitive device. Possibly the aircraft velocity is initially determined automatically, and once the readout is constant, it can or should be switched to “constant velocity” or “manual operation” (an increase or decrease in velocity would be immediately visible at the pointer). The automatic velocity meter (horizontal) is thereby switched off and protected.

Second find:

Found in KÖGEL, Ref. 1, p. 66: “If the automatically determined values for horizontal and/or vertical target velocity fluctuated, average values could be set manually.” Cf. also the hypothesis immediately below.

Third find:

From Device 40 onward, only one horizontal velocity meter is available. And it operates with only one motor — the separate drive of the step disc with the entire sequence control has been eliminated (not understood).

The following points regarding velocity measurement remain to be clarified or verified:

How the setting wheel at the output takes over the new value from the measuring wheel (upward or downward), without itself first going to “zero” and thereby causing large fluctuations in the subsequent computation, was not understood.

The electrical circuit for controlling the magnetic brake on the auxiliary motor was not understood.

The always-mentioned “multipliers” fundamentally perform a division, since a high velocity corresponds to a small angle of rotation of the measuring wheel. Rather than (in effect) multiplying by time, a division by the angle of rotation is performed. In addition, the shaped solids must add in a fixed time of approximately 1.67 seconds: while the auxiliary motor is active and the step disc is rotating, measuring time — pressing time against the drive wheel — is missing.

The velocity measurement becomes unclear and uncertain for fast aircraft: if the aircraft requires approximately 1.67 seconds for the 250 m of the measuring roller (at v = 150 m/s = 540 km/h), almost no measuring time remains, because the revolution of the step disc already consumes that much time! As long as the measuring wheel is not in contact with the drive wheel during its revolution, no angle is captured either. Even at slightly slower aircraft speeds, the errors in the velocity measurement must become very large — possibly to the point of being unusable — because the beginning and end of the measurement duration coincide with the run-up and braking of the auxiliary motor, giving rise to uncertainties at the moment of contact and release.

Hypothesis: Is that the reason why a second velocity meter for manual operation or manual averaging was needed — because automatic measurement of fast aircraft was simply no longer possible?

The scale of the velocities displayed on the measuring device appears impractical, since all relevant velocities are compressed together at the top of the scale (the pointers are not in a meaningful position).

The greatest accuracy of the velocity display/setting lies at 25–60 m/s; it becomes considerably tighter at 60–100 m/s; above that, almost nothing is visible and only imprecisely readable or settable — the pointer lies slightly off-scale. Converted precisely: 90–210 km/h; approximately: 210–360 km/h; above that, only very imprecisely or by estimation.

A list of the cruising speeds of aircraft of the period had to be abandoned, because everything is far too uncertain: maximum speeds at high altitude are often cited, the actual speed during operations may be entirely different, etc. Heavy bombers achieved 300 to 350 km/h in steady flight; fighter aircraft significantly more even without a dive (Thunderbolt 500, 560 km/h). Ju-52 maximum 265 or 290 km/h.

In the horizontal velocity measurement, the time required for 250 m of flight path is determined: the distance is fixed, the time is measured.

In the vertical velocity measurement, the altitude difference over which the aircraft climbs or descends within two seconds is determined (not contained in the diagram above; see the overall system diagram): the time is fixed, the distance is measured.

[page 16: figures only]

Readout of the velocity meter, scale in m/s. The relevant values are all crowded together on the right. A black pointer (System I, manual) and a red pointer (System II, automatic) are visible — when the device is switched on, they will first jump to the correct position. 1938 version.

[figure caption page:]

External connection of the GAMMA-JUHASZ fire-control computer to the outside world: power connection from the generator with DC and AC voltage, output of all measured values to the distribution box for the four guns, fire bell, telephone, earth. View from below, next to the tripod column (the rangefinder has its own connector).

Wind, Weather, and Further Corrections

(For an illustration, see page 12, lower right, and also page 25.)

The device computes on a geometric basis the three quantities that are to be transmitted to the guns — and extensively corrects them beforehand, before the values actually go to the “transmitters.” On the one hand, direct, free-hand manual interventions are possible — for example, after observing the burst clouds in the vicinity of the aircraft under fire. On the other hand, atmospheric influences such as wind direction and wind speed can also be entered numerically.

How precisely the incorrectly positioned burst clouds can actually be attributed to individual variables is difficult to estimate: if the cloud is behind the aircraft and too high — is the position correct but the fuze time was too long, or are the angles wrong, or should the aircraft’s velocity be increased in the computation? Possibly the rangefinder operator sees the situation more clearly — but he is occupied, as he must continuously set the correct distance to the aircraft.

Most corrections are made with the aid of three-dimensional shaped solids. With these, even complicated non-linear relationships can be taken into account. At the end, the corrections are captured by electrical tracking and introduced with motor power into rotating shafts, which by means of differentials change the previously computed quantities in the correct sense and by the correct amount. A differential is functionally similar to the differential gear in an automobile, but may also be manufactured more conveniently in a slightly different form. In the GAMMA-JUHASZ, they probably have the same geometry as in an automotive drivetrain.

In the following description, a symbol ** is placed at each correction input that involves a direct manual input, which is not processed further numerically within the device but is simply added to the final value at the very end (the scale at the input knob is, however, already marked numerically). In the excerpt, four motors are visible whose functions are labeled. For three of these motors — for the correction of time, azimuth angle, and elevation angle — the associated electrical micrometer for controlling the motor direction is arranged in immediate proximity.

[page 17 continues:]

Function of the manual inputs at the far right, from top to bottom:

Wind direction — Wind speed — Correction of fuze time** — Correction of azimuth angle** — Correction of muzzle velocity — Correction of “air weight” (pressure and temperature) — Correction of elevation angle**

Function of labeled shafts A through F:

A: Projectile time of flight to the intercept point (including lead), uncorrected. Controls the rotation of seven shaped solids.
B: Sum of all corrections to the projectile time of flight, electrically tracked.
C: Flight altitude (aircraft altitude plus altitude lead plus parallax altitude), controls the translation of seven shaped solids.
D: Correction of the azimuth angle due to wind (component transverse to the shot).
E: Azimuth angle fully corrected, used for correct decomposition of the wind into transverse and longitudinal components.
F: Azimuth angle of aircraft plus azimuth lead; at “F” still without parallax correction, without wind correction. 5 and 85 are differentials for mixing in further quantities.

The output (feeler pin) from each group of three channels at the top is added in a lever linkage and supplies the correction of the projectile time of flight (fuze-setting time).

The output (feeler pin) from each group of three channels at the bottom is added in a lever linkage and supplies the correction of the barrel elevation at the guns. The azimuth angle is corrected only by the crosswind (plus possibly a manual intervention).

The always-present second feeler bodies with a horizontal axis (also called “multiplication bodies”) may have something to do with the differential weighting of the three added [quantities — text ends at page 18].

corrections (conjecture) — and they allow the partial corrections to be made dependent on one additional variable each. In this way the longitudinal wind (parallel to the direction of fire) enters the computation only at this stage, whereas previously the influence of flight altitude and flight time had been combined together. Taken as a whole, the two cam bodies represent a continuous four-dimensional table ground into metal — in a sense, a pre-war spreadsheet. This was 11 years before ETH Zurich rented a half-finished ZUSE machine and completed it themselves (1949) in order to gain experience with a program-controlled machine “as the first university on the Continent.” The Z4 was in operation from 1950 to 1955 (“Bund,” 3 July 2013, p. 29). It operated with relays and had a mechanical “memory” consisting of displaceable sheet-metal cams, hand-sawn, with a capacity of 64 numbers.

The earliest traces of functional or commercially sold GAMMA-JUHÁSZ devices appear on the internet from 1934 onward.

Wind decomposition into rectangular components: The first differential following the input wheel for wind direction is connected as a difference: wind direction relative to the barrel direction. The second differential in the wind-speed path is related to the rotation of the inner ring, which causes the oblique spindle to set the total wind speed: if only the wind direction is adjusted, the inner and outer rings must rotate together — this leaves the wind speed at the oblique spindle unchanged.

Further Developments — Models 38, 40, 43, 43/50 R (year of development)

Model 40 differs from Model 38 only very slightly. In Model 40 the mechanical sight, or “collimator,” was eliminated entirely (cf. title page, above the telescope on the right, and p. 12, Fig. 6). Internally, the second velocity meter (horizontal) was eliminated; certain minor changes (socket or switch for the fire bell) are visible. The housing remains the same, i.e., round. HASLER emphasizes that the electrical section was built to be significantly less failure-prone.

Model 43 receives improved handwheels for easier input of azimuth and elevation angle (from the telescope) and of slant range (from the telemetry unit), and for the first time incorporates curve extrapolation. The housing becomes angular and the telescopes are newly mounted on top (to allow the device to be dug in).

For the handwheels, new electric motors produce variable speeds so that manual input is relieved of load: when the input quantities increase or decrease regularly, the motors follow autonomously — only when the angles gradually deviate are manual interventions still necessary. The motion of the handwheel and the electric motor are summed in a differential gear.

For Model 43/50 R, HASLER developed an intermediate level for the reception of radar data. In this arrangement, personnel had to continuously re-enter the radar values by hand using new follow-pointers on the device — almost an anachronism when one considers the extensive electrical follow-up controls that the 1938 device already possessed. Thanks to the coupling of the computer with the radar device, the old, large acoustic detection instruments (ELASKOP) and the searchlights could be taken out of service. Image of the 43/50 R: see page 27.

Selected Dates and Quotations on the Coupling of Radar with the Computer, from the book Fliegerabwehr by Hermann Schild, pp. 46–50

Development after the war progressed only slowly:

1945 First radar training for instructional personnel. English Radar Mark IV (predecessor of the later-deployed Mark VII).
1950 Trials begin with Radar Mark VII. Conversion of a prototype of the HASLER computer. Target tracking is possible with Mark VII, but not target acquisition. A target-designation radar (ZZR) is required for that.
1951 Broad-based radar trials.
1952 Radar trials are continued.
1953 24 of the total of 83 command computers are progressively converted to 43/50 R (R for Radar).
1955 and 1956 Efforts toward radarization of the heavy anti-aircraft artillery are “continued on a back burner.”
1957 The few Radar Mark VII sets are allocated to the troops.
1958–60 Further Mark VII units are procured, 12 in total. All searchlight companies are re-equipped with the ZZR TPS 1-E.
1961–63 Procurement of the 35 mm twin-barrel cannon OERLIKON with a new, electrical analogue computer and its own radar (Fledermaus, Superfledermaus). Only the fire-control device SKYGUARD (1975) operates with a digital computer.
1964–67 Retraining of heavy anti-aircraft units to the 35 mm anti-aircraft gun and Bloodhound (missiles).

[page 20]

New Handwheel Input from Model 43 Onward

Mechanical diagram of the new handwheels with motor assistance: Azimuth (top), elevation/range (middle), bank angle (bottom). By shifting gears, the input to the governor gear can be interrupted — the motor itself is also switched off electrically. The original, direct manual mode is retained. Azimuth and bank angle each have a “transmitter,” cut off at the top: the information about which aircraft in a formation is being engaged is relayed from the command device to the telemetry unit so that its range can be measured precisely. The “elevation/range” handwheel is complicated because different quantities (slant range or altitude) are accepted from the telemetry unit. (Was this a marketing feature or a technical necessity??)

Electrical diagram of the new handwheels: Range/elevation (top), bank angle (middle), azimuth (bottom). The purpose of the locking magnets is not yet clear. The wires L3 and L5 to the locking-magnet coils at the bank-angle station are labeled elsewhere as “bank-angle zeroing unit.” Why the bank-angle motor has a magnetic brake whereas the azimuth motor does not, and why the locking magnets are connected differently, remains unclear.

For all three handwheels, a manual intervention simultaneously adjusts the variable itself (directly from the handwheel to the output) and changes its rate of change, i.e., the motor speed. This creates a somewhat opaque situation — are there not flight paths conceivable for which the manual correction of the telescope angle and the change in tracking speed would need to have opposite signs? Is a fixed-gear coupling of the angular correction to the motor speed sensible and always permissible?

One can easily reason incorrectly: what is corrected by hand is the deviation of the azimuth or elevation angle from the motorized, uniform increase of the angle. The handwheel is adjusted when the curvature of the angle curve as a function of time demands it. The motor delivers the angle correctly as long as the first derivative is constant; the handwheel allows the second derivative to be taken into account. Whenever the aircraft in the telescope lags behind (runs ahead of) the motor movement, a larger (smaller) angle and simultaneously a higher (lower) speed are required. So it all works out — everything is correctly and sensibly designed. How the system settles on the initial slew toward

[page 21]

the aircraft, i.e., with large and rapid angular changes, is still unknown. Excessively high motor speeds would arise here. Possibly the motor drive must be switched off entirely at the start until the aircraft has been acquired (?).

The strange elements “between springs” (e.g., part No. 145, just below the governor gear at the bank-angle station) could be interpreted as follows: a guard to prevent the motor drive from adjusting its own speed if the operator has taken his hands entirely off the handwheel; it may be a minimum resistance so that differential gear “138” does not set the handwheel in motion. However, for the azimuth a large force must always be applied to rotate the 420 kg device (Command Device 43) — the artificial resistance would therefore need to be considerable (???).

Curved Flight

With Model 43, a new extrapolation for curved flight is introduced for the first time. The specialist from the German Army Weapons Office, Alfred Kuhlenkamp, writes in his book Flak-Kommandogeräte (1943), p. 16, that most command computers — especially foreign ones — handle only horizontal straight-and-level flight. A few, including the German Command Device 36, also solve regular straight climbing or descending flight (as does GAMMA-JUHÁSZ). No concretely solved curved flight is mentioned — though he passes over the new German Command Device 40 entirely in silence. (According to K. D. Gattnar, Jenaer Jahrbuch 2008 on the history of technology and industry, no curved-flight capability is evident for Command Device 40 either. The English VICKERS device (somewhat older) also shows no curved flight capability — possibly HASLER genuinely entered uncharted territory. Until proven otherwise: HASLER may have been the first and only manufacturer, up to the end of the war, to equip a fire-control device with curved-flight capability!)

The measuring roller plots the flight path (from azimuth and range) and aligns itself parallel to the actual direction of flight. By means of electrical follow-up (not drawn), the flight-direction motor always follows this value. The aircraft’s course angle is measured from north, whereas the flight angle in the command device is measured from the rotating line of sight to the aircraft. The difference between the azimuth angle and the flight angle (differential D1) yields the course direction of the aircraft from north. In the curve indicator this value is converted every two seconds into a rate of course change, which is then immediately re-entered manually at the follow-pointer. Multiplied by the projectile flight time, this gives the expected deviation of the flight path up to the point of impact (the angle between the tangent to the flight path and the chord to the point of impact).

This angle is added to the flight-direction angle in differential D2, and this corrected value is passed to the flight-direction arm, where the lead distances and angles are formed.

The multiplication is performed by a cam body: translation according to the projectile flight time, rotation according to the rate of deviation. The output of the multiplier provides the additional angle added to the flight angle as a result of curved flight.

Using handwheel H3, counter Z2 is set to zero during the readiness check (without an aircraft).

Whether a differential adds or subtracts depends on the direction of rotation of the gears.

The curve extrapolation is built into the gear diagram; diagram of Command Device 43: see below.

[page 21 — figure description text]

Upper-left corner, horizontal: topmost — azimuth for the flight-direction disc; second from top — azimuth at a different gear ratio, for orienting the entire device. Lower-left corner, horizontal, bottom — flight altitude. Second from bottom — slant range to the aircraft. Third from bottom — projectile flight time. At right, at the visible flight arm: 68/183/184 — setting the corrected flight angle; the angle differs from what the measuring roller indicates. 42/188/60/61 — setting the lead. The two wire cables produce the range to the point of impact, which is converted in large cam bodies (not in the picture) into flight times and contributes to calculating the elevation of the guns.

Mobile Computing

The computer (Command Devices 40/43: 325 kg / 424 kg) is manhandled onto the equipment trailer by means of carrying poles (8 persons). A second trailer is the “power unit,” containing a 2.7 hp petrol motor, a DC generator, an AC converter, cable reels, various items of equipment, and a battery set of 84 cells, 130 V, 30 Ah. A fully charged battery is sufficient for approximately 3 hours of operation (without additional heating).

Where possible, the computer’s heating should be run from the civilian power grid. If electrical heating options are insufficient in winter, two “catalytic stoves” are available for warm-air heating. The temperature inside the device should not fall below +5°C in winter or exceed +40°C in summer. Electrical heaters are available in ratings from 100 W to 665 W depending on the power source. If the battery must also supply the lowest heating level during computer operation, the operating time is reduced to approximately 2 hours. Cable lengths between the computer and the power unit (the noise of the petrol motor must not excessively interfere with the sound of aircraft): 50 m to 150 m. Operation: Eight men are required to move into and out of position. During firing operations, five men work directly at the computer, plus one device chief, plus the fire-control party, plus the telemetry team (plus reliefs…).

Illustration above: The computer on the equipment trailer. Illustration below: The power unit.

Above: Equipment trailer, weight in travelling order: 1,195 kg. Model 43 shown here. At front: tripod 80 kg. On top of the computer lie the rain and cold-weather protective covers. On the side: the four carrying poles for 8 men. Below: Power unit, weight in travelling order: 1,530 kg. Spare wheel on the roof. Electrical panel and battery set.

Appendix 1: Miscellaneous Historical Notes and Individual Findings

The counterpart to the command computers in the anti-aircraft batteries were the mechanical bomb computers in aircraft, which had to calculate the optimum bomb-release point based on the aircraft’s own position, speed, and flight altitude. These computers had to be constructed as lightly as possible (unlike the ground-based anti-aircraft equipment).

Finland possessed, before and during the Second World War, a great miscellany of different weapons for air defence. As many as 13 different types of heavy anti-aircraft guns were apparently in use, in calibres 75 mm and 76.2 mm. As a result, Finland used command computers from nine different manufacturers. Praise goes to VICKERS GB (“best computer in the winter war”) and LAMDA 1940 Germany (“best computer in WW II, actually a full fire control device”). The GAMMA-JUHÁSZ device (1936) was also used: “Licence-made in Finland by Strömberg, three sub-models: I, II and improved III.” The Italian Gala-Borletti device was described as completely useless: “worthless.” In Sweden, GAMMA-JUHÁSZ was produced under licence by “Arenco” in 1942.

Finnish General NENONEN was a mathematician, rooted in the artillery, who had studied at the Military Academy in St. Petersburg. He successfully introduced simplifications in artillery practice that allowed faster target changes. “The trajectory calculation formulas he developed are still in use today by all modern artillery.” In addition, regarding air defence, he apparently pursued an idea for how fire could be opened (or perhaps had to be opened) entirely without using command computers, which were always expensive and far too scarce. This was the so-called “3T method” — unfortunately it has not yet been possible to learn more about it. The method reportedly required large amounts of personnel and did not function properly. What was mathematician NENONEN’s idea? That would be interesting to know!

The Firm Gamma-Juhász

Zoltán and István Juhász were the two leading figures of the firm Gamma-Juhász, sons of a Hungarian member of parliament. István was the technical director; Zoltán was the commercial director of the enterprise. István Juhász is the same person as Stephen Shepherd (identical life dates 1894–1981, same photograph). “Juhász” means something like shepherd or sheep-herder in Hungarian. The firm was found on the internet on one occasion under the name “Gamma Shepherd.”

Spring 1945: (Zoltán had fled to the West with his family, and eventually ended up in Colombia, South America, where, harking back to an early 19th-century ancestor, they changed their name to Andujar.) See http://magyarnews.org/news.php?viewStory=995

Also found there: The end of World War II changed all that. They were still able to stop the Germans from taking away the Swiss, English and German precision machinery of the Gamma Works, but later on, the Russians took every serviceable instrument, machine and semi-finished product, as well as most of the raw materials. The firm had to lay off its thousands of workers. ((It is likewise said of Zeiss/Jena that the Russians cleared out absolutely everything!))

In Switzerland, the device had long been in licence production by HASLER at the end of the war. Whether Hungarian original components were still required for this is unknown.

István Juhász was apparently the director from 1921 to 1945. More than a thousand units of the command device are said to have been built, used in: (Switzerland was forgotten in this list… as was Sweden) China, the Netherlands, Norway, Finland, Austria, Italy, Iran, Argentina, Poland, and the Soviet Union.

Effectiveness Against Targets

Results for heavy anti-aircraft artillery were very modest. This was known in Germany, Vienna, etc.; the reputation of the Flak among the civilian population was poor. People asked themselves why — despite the impressive Flak towers and all the “fireworks” — so few aircraft were actually being shot down.

[page 24]

From http://www.mil-mod.de/html/fla-kanonen.html:

At the start of the war, German Flak artillery was, according to the prevailing expert opinion, the strongest, most modern, and most effective in the world. This, however, changed dramatically in the course of the war. Whereas at the start of the war “only” approximately 4,200 rounds of heavy Flak ammunition were needed to shoot down a medium bomber, by early 1944 between 15,000 and 17,000 rounds were needed to achieve the same result. The causes, beyond the poorer training of the personnel manning the guns and fire-control computers (Flak helpers, Reich Labour Service personnel, “Hiwis,” war-wounded, employees of the facilities being defended, the elderly, etc.), included the continuing technical development of Allied aircraft (altitude and speed) as well as the growing experience of Allied aircrews.

Peter Ortmanns, Flakhelfer, p. 33 (address see further below, at packbierpeter…):

In the period from 5 March to 29 June 1943, 18,506 hostile incursions were recorded. The aircraft shot down in the same period number 872, corresponding to 4.7%. In a speech on 26 June 1942 to the British House of Commons, Churchill announced that of an attacking force of 726 aircraft over the city of Essen, 35 had not returned, corresponding to 4.9%. On the basis of a comparison of various sources of information it can be stated with certainty that losses were well below 5%. According to investigations by the Luftwaffe General Staff — Quartermaster General — 35,322,260 shells of all calibres were fired for 8,706 aircraft shot down. That corresponds to 4,057 shells per aircraft shot down. Of course, the many damaged aircraft or those that crashed on landing are not included in these figures.

In Switzerland, the hit probability of the heavy anti-aircraft artillery was assessed considerably more optimistically:

Colonel Alfred Büchi (ASMZ, 1934) reports that 60 rounds are needed to inflict a destructive hit on an aircraft: http://retro.seals.ch/cntmng?pid=asm-003:1934:80=100::1051 (there p. 364). This is, however, a projection; the German figures, by contrast, are based on an evaluation of actual results.

In Germany, the internal workings of command computers were so secret that the troops had to withdraw when the armorers worked on them. http://www.packbierpeter.de/joomla/images/pdf/ortmanns2.pdf (Peter Ortmanns, p. 43)

On 13 July 1943 two British “Lancasters” crashed — one at Bouveret, one near Sion. They belonged to a formation that was supposed to fly from England west of Mont Blanc over Annecy to Turin — however, due to severe thunderstorms over France, more than 100 aircraft strayed into the Valais. They were engaged by Swiss anti-aircraft fire at the Col du Marchairuz and hit several times (in the middle of the night!). The two crashes are counted as aircraft shot down by the Swiss air defence.

The crews of both bombers are buried in the cemetery at St. Martin, Vevey. Anyone visiting their graves finds an entire military cemetery with 136 fallen British soldiers from across the British Empire.

Press report on the 70th anniversary of the two Lancaster crashes in the Valais: http://www.lenouvelliste.ch/fr/en-continu/pourquoi-deux-bombardiers-seperdirent-en-valais-498-1201352

At night, aircraft were searched for with large searchlights (in Switzerland these were last operated around Lausanne for the EXPO 64 exhibition). Acoustic detection devices were used in attempts to locate approaching aircraft in time. Particularly fine photographs of acoustic detection installations can be found at: http://www.alternatehistory.com/discussion/showthread.php?t=185434&page=23 (scroll well down; there are also images of early radar installations). A large acoustic detector ELASCOPE is displayed in the Aviation and Anti-Aircraft Museum in Dübendorf.

Appendix 2: Notes on the Command Device GAMMA-JUHÁSZ

>>> Caution with dates: The “Command Device 40” was apparently developed in 1940 — some time still elapsed before delivery. In February 1941 the diagram was still being drawn, printed, and published with the double velocity-meter, i.e., from the original Model 38. There are drawings from October 1943 that relate to Model 40, not Model 43. There was probably a distinct Type 43/50, not identical with the 43/50 R (R for Radar).

>>> Lubricating the device: Use only bone oil! p. 12, Doc. 20 (1038) (personal photo 276). Lubrication is to be renewed on average every 600 operating hours. “The ballistic cam bodies are to be kept with only a breath of grease (almost completely dry).” The reason for the winter electrical heating of the device lies in the consistency of the lubricants.

>>> Transmission of measured values to the guns is accomplished with three star-configured coils in a magnetic alternating-current field. At the transmitter (sender), the coils are set to the correct angle; at the receiver, they follow synchronously and automatically. However, only the information is transmitted; no mechanical force for further direct processing can be extracted from it. The operating personnel re-enter the information manually into the system using a “follow-pointer.” All three measured values from the command device to the guns are transmitted or displayed in both a coarse scale and a fine scale (fine scale at the gun: 100 angular units per full revolution). From the telemetry unit to the command device the process is the same, but without separate coarse and fine scales.

>>> Multiple fine corrections of the calculated values are made by means of carefully shaped “cam bodies” (Formkörper). Should a different type of ammunition, a different calibre, a different gun, etc. ever be used, the cam bodies must be replaced. The “range practice ammunition” in Switzerland had a muzzle velocity of 550 m/s; the wartime ammunition had 805 m/s. This is said to produce a comparable lead with the much slower towed-sleeve target as with wartime firing. For this purpose the large ballistic cam bodies had to be exchanged (National Library, G 3188/69, p. 13, 1940). In Doc. 15 (1941), p. 30, it is noted that the cam bodies are calculated for a muzzle velocity of “805 m/sec. (or 550 m/sec.).” Whether the daily-corrections apparatus (in whole or in part) also had to be exchanged remains unclear — that would be the next assembly group, which may have been manufactured in duplicate: (original Gamma stamp visible in the photo).

[page 26]

>>> How were the cam bodies manufactured? How can so many points of a function of three variables be computed? The first electronic analogue computers were developed in connection with rocket guidance — in Germany by Helmut Hölzer (V2, Peenemünde), in Russia for the same purpose by Sergei A. Lebedev. They were not yet available for manufacturing the cam bodies. Desktop calculators for manual computation naturally existed. Where, for instance, the deceleration of a projectile by air resistance cannot be expressed in a closed formula, manual computation becomes unpleasant. Trigonometric functions can be handled with printed tables. At ZEISS/Jena there existed a variety of devices such as “cam-body measuring instrument,” “cam-body fine-copying bench,” and “copy-milling machine for spatial curves” Mod. 31 … Mod. 40 (cf. Jenaer Jahrbuch zur Technik- und Industriegeschichte, 2008, vol. 11, pp. 56–58). In the Museum of Communication, Bern, approximately 140 large sheets with countless numerical milling specifications from HASLER are stored.

>>> Electrical matters: The generator produces (from the petrol motor) 80 V / 12 A (DC) in compound operation for the command device, but 120 V to 150 V in shunt operation for charging the battery. This means that operation of the command device and battery charging do not occur simultaneously?! Battery: 84 nickel-iron cells with 30 Ah at a discharge rate of 6 A. Charge as soon as voltage falls below 84 V. Charge at 6 A until voltage has risen from 126 V to 153 V. Warnings that only KOH (potassium hydroxide solution) and under no circumstances H₂SO₄ (sulphuric acid) may be added to these cells.

>>> Follow-up motors: In Models 38 and 40, most follow-up mechanisms consisted of two motors, one for upward movement, the other for downward movement. Reason: the motors required less current than bidirectional motors — but at the cost of more complexity in the gearing. The electrical diagram for Model 40 (below) shows that the single-direction motors have an electromagnetic “Gamma brake,” whereas the bidirectional motors have a “Hasler brake.” The differently drawn motors in Model 43 (p. 10) show that later on, bidirectional motors with two stator coils were used. Clearly visible below is the electrical limiting/cut-off of motor XIV by the movement of the parallax carriage. The electrical micrometer is in both cases located to the lower right of the motors. Model 40 was built under licence by Hasler. Title of the large electrical diagram: “Principle diagram for modified GAMMA command devices, 19.7.1941.” Motor XI still has a centrifugal governor!

>>> A. Kuhlenkamp of the German Army Weapons Office describes the GAMMA-JUHÁSZ device only very briefly and summarily in his book Flak-Kommandogeräte (1943), based on captured devices from Norway. Two passages from it: “The range or altitude is set by verbal announcement.” (?? The Swiss Model 38 already has a follow-pointer. Similarly in Kögel, p. 20/21, before 1942.) (The device) “showed a considerable sensitivity to improper operation.” It is not stated what this refers to — this would presumably apply equally to every other make.

>>> After switching off the main electrical switch, “no further adjustments may be made” (Doc. 7, p. 71). The reason is the many electrical “follow-up mechanisms” with finely adjusted micrometer contacts, which cannot operate without current. It is very conceivable that the electrical contacts may have suffered damage over the many decades. In a museum setting in particular, it is likely that attempts are made here and there to turn the handwheels.

[page 27]

>>> There are two ballistic cam bodies (B.K.) for calculating a time duration, even though the aircraft flies for the same length of time from the gun-firing moment (flight time) as the projectile until detonation (fuze time). At time A the cartridge is pulled from the fuze-setting machine; three to four seconds later (time B) the gun fires. At time C the projectile explodes.

B.K. “Flight time”: Computes “at all times,” i.e., continuously, the time the aircraft will fly until it reaches the point of impact. This yields the lead distance; from this the range from the gun to the point of impact is measured, which is fed back into B.K. “flight time” to determine the time (circular logic). This computation is taken into account up to time B.
B.K. “Fuze time”: Computes “at all times” the range to the point of impact and the projectile flight time for an increased lead distance, so that the aircraft can continue flying for the previously agreed time difference B − A beyond this point. This computation is taken into account only up to time A, after which the fuze can no longer be influenced. The angles important for the gun are still updated continuously up to time B.

>>> Model 43/50 appears to have been a distinct type, not the same as 43/50 R (R for Radar). The device connected to radar received a middle intermediate level with new follow-pointers for azimuth and bank angle, clearly visible in the image below the telescopes. When actually connected to radar, telescope tracking is no longer required. On the opposite side of the device (operator absent!!) there are follow-pointers for “radar range” and, alongside them, for “telemetry range.” There were only 12 radar sets (Wikipedia). Model 43/50 R was introduced into service with the troops from 1957.

Radar

Highly interesting study on the history of the first radar trials, with many technical details, often from the Swiss perspective, 53 pages: http://www.cdvandt.org/Jucker-early-warning.pdf. If the address is no longer functional, search for: “Vorgeschichte und erste Generation Frühwarn-Radar bis ca. 1960.” Author: Hans Jucker, Schwerzenbach. Aircraft that made forced landings were of special interest precisely for the examination of their new radar equipment: cf. pp. 6–9. Further studies on the history of technology, especially radio technology, but also other topics: http://www.cdvandt.org/handbooks.htm. Also by Hans Jucker, with very fine technical details: http://www.cdvandt.org/Lichtenstein%20radars.pdf. Adventurous and strictly secret material on the first airborne radar devices (the same case of a forced landing in Dübendorf as in Hans Jucker above, but written in a more popular style): https://www.woz.ch/0803/ein-deal-mit-deutschland

[page 28]

Photo Credits

Where several images appear on a page, they are listed here in row order.

All historical images were taken from the existing documentation on the GAMMA-JUHÁSZ command device, published in various years for the use of the troops or the equipment mechanics, archived in the Aviation and Anti-Aircraft Museum in Dübendorf (the document number corresponds to the EXCEL list of 29 July 2014).

Page	Personal photo no.	Document no.	Archive registration (Museum) Old / New	Publication year
1 6 7 9	193 / 190 / 193 / 190	15	045925 / 12271	1941
10 11 12	174	14	045922 / 12268	1947
13	285	17	045927 / 12273	1941
14	281	20	045930 / 12276	1938
15	283	20	045930 / 12276	1938
17	172	7	045821 / 12166	1950
18	220 / 257	19 / 18	045929 / 12275, 045928 / 12274	1938
20	302 / 269	15 / 18	045925 / 12271, 045928 / 12274	1941 / 1938
21	171	6	039479 / 5133	(1950?)
22	287 / 301 / 268	17 / 15 / 18	045927 / 12273, 045925 / 12271, 045928 / 12274	1941 / 1938
25	218	19	045929 / 12275	1938
26	161	—	Museum Flieger und Flab, Dübendorf, permanent display	—
27	79	—	Museum Flieger und Flab, Dübendorf, permanent display	—
—	392 / 368 / 174	7 / 11 / 14	045821/12166, 045919/12265, 045922/12268	1950 / 1947
—	221	22	046567 / 12929	1944
—	368	11	045919 / 12265	1947
—	5 images scan	7	045821 / 12166	1950
—	264	18	045928 / 12274	1938
—	401	—	Nationalbibliothek, G 3188/70	1940
—	—	—	“Hasler Werke — Schrittmacher in innovativer Technik.” Haslerstiftung, undated, p. 55. Reproduced with permission.	—

Literature

Waffen und Geräte der Schweizerischen Fliegerabwehr. Steckbriefe und Kurzorientierung. Adj Uof Alfred Kögel 1913–2004. Published by Oerlikon Contraves AG, Zurich/Switzerland. Published Autumn 2006.

Facsimile passages from Kögel’s personal sketchbook: thematically broad — weapons, equipment, and all manner of geometrical situations from the daily life of anti-aircraft batteries. Training, technical data.

In Doc. 11 (1947, booklet 1, 12265), a large gear diagram of Command Device 43 is found at the very back. It contains handwritten notes, almost certainly by Alfred Kögel (senior), that go beyond personal annotations. It appears as though Kögel had also involved himself in constructive details, judging, for example, certain features to be superfluous and unnecessary. Handwriting sample, at the time Fw Kögel: Doc. 20 (1938), title page.

Acknowledgements

For assistance, interest, discussions, expert tips, introduction to further contacts, provision of documents, etc., the following are warmly thanked: Bernd Ulmann (analogmuseum.org), Daniela Zetti (History of Technology, ETH Zurich), Anja Thiele (Deutsches Museum, Computer Science section), Ulrich Wegmann (anti-aircraft veteran), Peter Blumer (Contraves), Erich Greger (History of Technology, Jena), Klaus-Dieter Gattnar (Zeiss, Jena), Harald Grahe (Association of German Engineers, publisher). The initial entry into the subject was made possible by Alfred Kuhlenkamp, formerly of the Army Weapons Office (book: Flak-Kommandogeräte, 1943, VDI).

Special thanks go to Beatrice Heuberger, Elisabeth Bengzon, and Beat Benz of the Aviation and Anti-Aircraft Museum in Dübendorf, who retrieved documents from the museum archive and granted access to the surviving exhibits.

The researcher: André Masson, CH-4900 Langenthal

Winter 2014/15