Latest page update: 3 January 2021 (added Fig. 24, with text and references)

Previous updates: 12 June 2020 (added Fig. 17 and text)

©2004-2021 F. Dörenberg, unless stated otherwise. All rights reserved worldwide. No part of this publication may be used without permission from the author.

red-blue line



Fig. 1: Block diagram of the Hell Feldfernschreiber

(adapted from Appendix 3 in ref. 5C)

The electronic circuitry of the Hell-Feldfernschreiber comprises four stages, each with an RV 12 P 4000 vacuum tube ("valve"):

  • Tonsummer – tone-oscillator stage (here: 900 Hz tone)
  • Vorstufe – pre-amplifier stage (here: volume control, 900 Hz bandpass filter, and amplifier)
  • Endstufe – final stage (here: tone rectifier/detector and amplifier/driver for the electro-magnet solenoid of the printer)
  • Reglerstufe – regulator stage (here: motor speed).

These stages are shown left-to-right in the schematic below.


Fig. 2: Simplified schematic of the standard Hell Feldfernschreiber

The Feldfernschreiber is powered by 12 volt DC. The power input is passed to the 3-position main switch via a line filter (not shown in the schematic above) and a fuse ("Schmelzsicherung", S). A 5 Ω series-resistor can be switched in series with the fuse, while the main switch is in the "Bereit" (ready/standby) position. This voltage-drop resistor accommodates an early model of a Hellschreiber-specific transformer-rectifier power supply unit that had poor voltage regulation with small loads. According to the 1941 manual (ref. 5), the switch is not necessarily installed in Feld-Hells that were built later. The 12 volt is used to power the heater filament of all tubes (hence the "12" in the tube designator RV 12 P 4000), and the motor of the motor-generator ("dynamotor"). The generator provides an anode voltage of 150-180 volt DC (165 volt nominal, 25 mA) to the tubes.

There are three audio interfaces:

  • Leitung La Lb/E - phone line (bi-directional)
  • Empfänger - input from the high-impedance audio output (headset) of a radio receiver
  • Mithören - monitoring output to a headset.

All interfaces are transformer-coupled, i.e., galvanically isolated. With one exception: the audio input from a radio receiver. However, a radio with vacuum tubes has a transformer-coupled audio output. So, there is no need to have a input transformer. All transformers have a static shield that is connected to ground. Additionally, the transformers are placed in a shielding box that is also connected to chassis ground. Shielded wiring is used where it is prone to picking up noise or hum.


The 900 Hz tone is generated with a simple LC-oscillator ("Tonsummer"). The "L" part of the "LC" circuit is provided by the inductance of the oscillator's output coupling transformer T1 ( = = "Summerübertrager"). The capacitance "C" is placed in parallel with the "L" (parallel resonance).


Fig. 3: Schematic of the 900 Hz tone oscillator + on/off keying

In fact, with its center-tapped coil, it is a standard Hartley sine-wave oscillator (invented in 1915 by Ralph Hartley while at Western Electric):


Fig. 4: Standard Hartley-oscillator configurations

(the Feld-Hell uses configuration a)

When a key is selected via the keyboard, the continuous 900 Hz tone is keyed on/off by the associated track of the character drum ("Geberwalze"). See the "How it works" page. The tone can also be keyed with the "Morse" telegraphy key of the keyboard. The resulting tone pulses are output via transformer T2 ( = = "Leitungsübertrager") to the telephone line interface and to the tone-pulse detector.

The choice of the 900 Hz tone frequency is a compromise. Conventional "Plain Old Telephone Service" (POTS) land lines support have a bandwidth that is compatible with "voiceband": the typical 300-3400 Hz frequency range of the human voice. So, 300 Hz is the lower cut-off frequency of telecom circuits in most countries. Also, phase distortion at low frequencies depends on the number of line-repeaters, hence this distortion can be large on long circuits. Cross-talk becomes an issue at frequencies 1000-1200 Hz, where telephone receivers tend to be most sensitive. Lastly, 900 Hz is not unpleasant to the ear. Ref. Harris et al. (ref. 10).


Transformer T2 basically couples the keyed tone from the 900 Hz tone-oscillator in the forward path to the tone-filter, the pre-amplifier, tone-pulse detector and the final amplifier. However, this transformer has a a third winding. This is the bidirectional interface to a phone line. The audio input from a radio receiver is simply connected across the secondary side of the transformer. As explained above, there is no need to have an isolation transformer at this input.

The signal across the volume control potentiometer (W24 in Fig. 5) is the combination of input signals from three sources:

  • the 900 Hz oscillator
  • a radio receiver
  • a phone-line

Fig. 5: Schematic of the volume control and 900 Hz band-pass filter

Note that this implies that text that is transmitted by a Feld-Hellschreiber, is not only output to the phone line or a radio transmitter: it is also printed locally! This creates a record of the transmitted text. The same mechanism also means that full-duplex communication is possible. Obviously, simultaneously sent and received signals are printed on top of each other. This was actually used during the standard speed adjustment procedure upon establishing a communications link (ref. 1).

The phone line interface ("Leitung" connector La - Lb/E), has an impedance of 800 ohm. This may at first appear to be incompatible with what many people consider to be the (fixed) impedance of standard POTS telephone land-lines: 600 ohm. However, keep in mind that phone lines only have 600 ohm impedance at one single frequency: about 1300-1400 Hz. For the 900 Hz Hellschreiber tone, the impedance of a standard phone line is actually 800 ohms! At the high end of the phone-voice bandwidth (3400 Hz) the line impedance drops to about 350 ohm. See ref. 2. The quality and impedance of the connected phone line is not at all critical, as long as the end-to-end attenuation of the 900 Hz tone is less than 46 dB (5.3 neper). The "Leitung" port is bi-directional: keyed tones are output to the phone line, and tones from an other Hellschreiber are received from this phone line. The output level of the keyed 900 Hz tone is at least 2.2 V at the La - Lb/E port (ref. 3). The phone-line transformer is dimensioned such that the line-current does not force it into saturation (note: a standard phone lines carry 48 - 60 volt DC).

Most Feldfernschreibers are equipped with a 12-pin round connector on the front panel. If the corresponding plug is inserted, the keying contacts of the character drum are diverted to this connector, and the tone-oscillator is no longer connected to transformer T2. In this configuration, the transmitted text is not printed locally.

The output of the potentiometer output is through an LC "Tonsieb". I.e., a bandpass filter. It has a center frequency of 900 Hz. As for the tone oscillator, the "L" part of the "LC" circuit is formed by a transformer. In this case, the T3 ( = "" = "Zwischenübertrager") coupling to the pre-amplifier. The filter is turned off by manually switching-in a "spoiler" resistor between the audio gain potentiometer and the filter.

Per ref. 4A, this filter has a bandwidth of 150 Hz. Per the 1941 manual (ref. 5A-5D), the bandwidth is 100 Hz. Note that at that time, filter bandwidth was defined as 0.5 neper (4.34 dB) down from the filter's pass-band, instead of the 3 dB that is used in modern days (ref. 6). I measured a 3 dB bandwidth of 130 Hz for my first Feld-Hell machine.


Fig. 6: The bandpass curve of my first Feld-Hell machine (after re-tuning the center-frequency of the filter)

The output of the tone-filter is amplified by a straight-forward 1-stage tube amplifier:


Fig. 7: Simplified schematic of the pre-amplifier

The pre-amp has an impressive gain. This is to be expected, as the RV 12 P 4000 vacuum tube has a μ of 4000. At the headset-output port ("Mithören" - "to listen in" or "monitor"), the 900 Hz tone begins to clip at 15-20 Vp (36 Vpp !). A signal of 55 mV at the line-input produces at least 1.8 volt at this output (ref. 3).


The output of the pre-amplifier is coupled to the tone-detector via transformer T4:


Fig. 8: The detector and printer-solenoid driver of the Feld-Hell

The detector is a full-wave rectifier. The diode rectifier comprises two "Kupferoxydul-Gleichrichter": cuprous-oxide-on-copper diodes, see the "Components and Construction" page. These first-generation solid-state diodes have an excellent low forward-voltage (0.2 volt, similar to germanium diodes); germanium and silicon diodes did not yet exist in those days. However, they also have a low reverse-voltage of 6 volt. The Feld-Hell uses Siemens-Halske Sirutor 5b diodes, in which 5 such diodes are put in series. The rectifier is followed by a simple RC-filter to reduce the ripple of the rectified signal.

The rectifier-plus-filter converts the tone-pulses (locally generated or received via radio or phone-line) into DC-pulses. They are passed to the control-grid of the "normally off" final amplifier tube. The grid current passes through an RC-network (Rg, Rv, C in the diagram below). This network provides a simple automatic gain control ("Schwundregelung", "selbsttätige Pegelregelung"; ref. 7), with a dynamic range of 5 neper (≈ 43 dB). The voltage drop across Rg provides a current-limiting negative grid bias. The time-constant of the parallel circuit of Rg and C is consistent with the duration of several Hell-pixels. As a result, only the highest momentary signal amplitude opens the grid of the tube. Signals with a smaller amplitude (e.g., atmospheric interference, echo, reverberation) are blocked, and short fading is compensated. The filter Rv + C suppresses pulse type signals that have a duration that is shorter than that of Hell tone-pulses. Ref. 4A (section 7, p. 7), ref. 4B (section IIIa, p. 16), ref. 6, ref. 4C (pp. 22-23).


Fig. 9: Automatic Gain Control via variable grid bias

(source: Fig. 7 in ref. 52)

The final amplifier tube is the solenoid driver. The solenoid is connected between the 165 volt DC high-voltage rail, and the anode of the tube. When a tone is received, the tube is keyed "on" by the detector. This causes anode current to flow, which energizes the solenoid. In turn, this causes the armature of the printer solenoid to push the paper tape up against the turning spindle of the printer mechanism. Note that the solenoid is not pulled down to ground level (zero volt).


Fig. 10: Anode circuitry of the tube that drives the Feld-Hell's printer solenoid

The printer solenoid comprises 15000 windings of 0.06 mm enameled copper wire ("lackisolierter Kupferdraht", CuL) and has a nominal DC-resistance of 4090 ohm. According to Rudolf Hell (section 6 (p. 6) in ref. 4A), a dissipation of at least 0.4 watt is required to reliably actuate the printer hammer. This implies 10 mA at 4 kOhm. Other literature (e.g., p. 328 in ref. 56, 1939) states that the solenoid measures 4200 ohm and requires 10 mA at 165 V. This appears to exceed the 3 mA nominal anode current rating of the RV12P4000 tube (and 1.5 W max anode dissipation). Note that the typical/intended application of RV-tubes is an amplifier ("Verstärker"). So this 3 mA current value is for the normal tube operating point - as an amplifier. The RV12P4000 has typical pentode characteristic curves. See the blue lines in Fig. 10 below. The maximum anode current is actually a little more than 10 mA!


Fig. 11: Characteristic anode voltage & current curves of the "RV12P4000" tube

The anode dissipation has to remain below 1.5 watt. The red curve in Fig. 11 shows that, at the maximum current level of ≈10 mA, the anode voltage must remain below +150 volt. The anode voltage in the Feld-Hell machine is provided by a motor-generator ("dynamotor"). It generates +165 volt DC nominal, with an acceptable operating range of +150 to +185 volt. Yes, this is above the stated +150 volt limit. But it is not a problem: the printer-solenoid is placed between the anode and the generator voltage. See Fig. 9. One side of the solenoid is always connected to the generator voltage, the other side can be pulled down to a low voltage via the tube.

When energized, the voltage drop across that solenoid is 10 mA x 4 kΩ = 40 volt. Hence, for the generator voltage range of +150 to +185 volt, the anode voltage of this driver-tube is actually only +110 to +145 volt (+125 volt nominal, see the orange dot in Fig. 10). Perfect! Turning the printer-solenoid current on/off is a switching function, not amplification. So the solenoid-driver tube does not have to be operated at the normal amplifier operating-point (the magenta dot in Fig. 11). Note that the standard printer-solenoid current of Hell-printers of the era was not 10 but 20 mA. So, a special solenoid had to be developed for the Feld-Hell printer.

Fig. 11 shows that the tube can be operated with 10 mA at the anode dissipation limit, if Vg = Vcontrol-grid-to-cathode = 0. The "fixed grid-bias" block in Fig. 8 is connected between +12 volt and ground. When the capacitor is charged, zero input signal results in a Vg that is 12 volt below the cathode voltage. Hence, the tube is "off". When the output voltage of the detector's pre-amp reaches +12 volt, the tube is completely "open", and up to 10 mA can flow through the printer solenoid.


The motor of the Feld-Hell machine is a shunt wound DC motor. That is, the field windings are connected in parallel with the armature windings:

Berhard station

Fig. 12: Basic types of Direct Current (DC) motors - classified by placement of the field winding

The governor field compounds the field of the armature windings. When the current through the shunt (governor) field windings is increased, then the associated field flux is also increased, and the motor's armature (rotor) must turn slower in order to produce the same amount of back-EMF. Conversely, decreasing the shunt field current causes the motor speed to increase. This property is used to regulate the motor rpm. That is: to keep the motor speed constant. This ensures that received text is printed on a straight line. Note that this does not ensure that the printed text line is horizontal! For the printed text lines to be horizontal, the local motor must not only be constant, but also be controlled to the correct value ( = the same speed as the motor of the opposite Feld-Hell station). Note that text that is transmitted by the Feld-Hell machine is also printed locally, and is always automatically straight and horizontal: the same motor drives the character drum and the printer spindle.

A separate vacuum tube tightly regulates the speed of the +12 volt DC motor. The anode current of the tube passes through the motor's governor-field windings. One side of these windings is permanently connected to the +165 volt of the generator. The tube, and hence the current, is switched on/off by a centrifugal switch arrangement that sits on top of the the motor shaft. It comprises a spring-loaded governor weight, and a switch-contact.


Fig. 13: Centrifugal speed regulator and associated control circuitry

As the motor speed increases, the centrifugal force moves the governor weight away from the motor shaft. Via a lever, this causes the distance between the switch contacts (C1 in Fig. 14 below) to decrease. The contact closes when the speed exceeds the nominal speed setting (manually adjustable). This causes the tube's grid voltage to be pulled up to +12 volt. Now the anode current flows, the governor field flux increases, and the motor slows down rapidly, until the switch contact opens up again. At that time, the anode current is shut off, the field flux drops, and the motor speeds up again. This cycle is repeated continuously. Such switching between two limit values is referred to as "bang-bang control" in control engineering parlance. The regulator maintains the rpm to within 0.5 % of the rpm set-point (ref. 3, 8). An LC-filter is installed across the primary switch contacts. This fully suppresses EMI from LF to VHF frequencies. There is no arcing across the contacts. Likewise, the motor's brushes are filtered with capacitors.


Fig. 14: "Bang-bang control" motor-generator rpm regulation

By design, the unregulated rpm (at least 9000 rpm!) of this motor is much higher than the required nominal rpm (3600 rpm). "Unregulated" means that the shunt winding is not energized. An electro-mechanical protection is built in, to avoid reaching this destructive rpm in case of failure of the regulator tube. A second contact of the centrifugal switch (C2 in Fig. 14 above) closes at a speed somewhat higher than the normal speed. When this safety contact closes, the regulator tube is completely bypassed, and the motor's governor field is tied directly to +12 volt. This limits the speed to about 4500 rpm ( = tape speed about 60 cm/min instead of 47 cm/min). Unlike the primary contact, the overspeed contact has no circuitry for suppression of arcing and EMI .

Why not simply connect the nominal speed contact directly to the "cold" side of the governor windings? That is, hook it up just like the overspeed contact. This would save an expensive tube and associated parts. Indeed! But this would cause constant arcing of the switch contacts (as with the overspeed contact), and the contacts would wear out quickly. The unreliability would have been unacceptable. By using the tube, there is never more than 12 volt across the contacts and no inductive load. The overspeed switch has 165 - 12 = 153 volts across is, and the inductive load of the governor windings.

Rudolf Hell's patent of this electrical centrifugal governor ("Elektrischer Fliehkraftregler") in 1949 (German patent nr. 803577).



Fig. 15: My schematic of the Hell Feldfernschreiber (drawing nr. 24a-32 (a1/a2))

(a very high-resolution version of this diagram is here - 7.3 MB)

As can be confirmed by tracing the circuit diagram above, when the Hellschreiber's main switch is in the "Bereit" (ready/standby) position, the red signal lamp extinguishes when one or both of the motor’s carbon brushes is not in contact with the commutator, or if there is an open connection in that circuit.


Fig. 16: Internal wiring diagram of the Hell Feldfernschreiber's Printer/Keyboard/Character-drum Unit

Pins 3 and 4 of the 12-pin round connector on the front of the amplifier box are used to key a CW transmitter (see the "Interfaces to a transmitter/receiver" page). It is easy to verify correct connection: insert the 12-pin plug into the round connector, connect an ohm-meter across pins 3 and 4 of the plug, and push down the Morse-key of the keyboard ( = the key with the green dot). The ohm-meter should show close to zero ohms.

When the 12-pin plug is inserted into the connector, the thick center pin of the plug opens a switch contact that is part of the connector. When closed, this switch contact bridges pin 2 and 3. However, I have come across a Feld-Hell amplifier box (from 1941) that is different: the round connector does not have the integrated switch contact that is actuated by the center pin of the plug. Instead, it has a separate, manual switch. The amplifier box must be opened (and the module with the vacuum tubes removed) to access this switch and toggle it. It is easy enough to identify such an amplifier box: the shaft of this switch is mounted into a through-hole the front of the amplifer box, and is visible from the outside; see the yellow circle in the left-hand image of Figure 17.


Fig. 17: Manual switch instead of the switch contact actuated by the center pin of the 12-pin plug


Fig. 18: Another, further simplified, schematic of the Hell Feldfernschreiber

(source: Appendix 4 in ref. 5C; click here for full size)


Fig. 19: One more schematic of the Hell Feldfernschreiber

(source: Fig. 16 in ref. 3; click here for full size)


The rear of the Feld-Hell's electronics box has a simple removable cover. Removing this cover exposes a two-row pertinax circuit card, with point-to-point wiring. All components on the Feld-Hell's circuit card and the vacuum tube sockets have a small round sticker on them, with a letter + number (W.. for "Widerstand" ( = resistor), C.. for "Condensator" ( = capacitor), D.. for "Drossel" ( = choke coil / inductor), GL.. for "Gleichrichter" ( = rectifier diode) etc.). This identifier corresponds to that of the same component in the schematic. Very helpful during manufacturing, troubleshooting and repair! Likewise, all the solder lugs on the pertinax circuit card and the interconnect blocks have a number printed on them or next to them (or embossed next to them). This number corresponds to that of a signal line in the schematic. Note that numbering and naming is not necessarily consistent between all available schematics, and there is variation between models.


Fig. 20: rear-view of the Amplifier & Interconnect Unit of my first Feld-Hellschreiber - cover removed

(several capactors were replaced with modern equivalents)


Fig. 21: Rear-view of the Amplifier & Interconnect Unit of my first Feld-Hellschreiber - cover removed

(the components and solder lugs are numbered per the schematic)


Fig. 22: Rear-view of the Amplifier & Interconnect Unit of my second Feld-Hellschreiber

(here, the stickers on the components only have a number, not a "letter + number")


Fig. 23: Rear-view of one more Amplifier & Interconnect Unit

The unit in Fig. 23 above was manufactured by "Radio H. Mende & Co. GmbH" in Dresden, rather than by Siemens-Halske. Capacitors C47 and C18 (top right-hand corner) appear to have been replaced with SiKaTrop (Siemens Keramik Tropenfest) capacitors. They are packaged inside a porcelain tube with soldered sealed ( = moisture-tight) metal end-caps. They were trimmed to the specified value, by removing metal deposit on the outside ("Ausserbelegung") of the tube, ref. 9. They are virtually "unkaputtbar", and remain reliable even after 80+ years.

Starting in 1941, a shortage of raw materials began to develop in Germany. This was due to the enormous demand by the military industrial complex and interrupted supply flows from abroad. This included tin. Ref. 14A, 14B, 14C. Standard solder for copper wire and component-leads was a 60/40 tin-lead alloy. Hence, such solder became exceedingly scarce and the radio-electronics industry started to look at electrical welding methods (both resistance and fusion welding, and the combination thereof) as an alternative to soldering. Depending on the materials to be jointed, the welding electrodes were made of copper, carbon, or blombit (a hard copper-silver alloy). Due to the high arc-welding temperatures, the welds have a characteristic small ball/bead of molten material, unlike a low-temperature tin solder joint where only the tin is melted. The same welding technique was also used in some Torn.E.b receivers:


Fig. 24: Component lead-wires welded to solder lugs on a Torn.E.b circuit board - beaded welds clearly visible

(source: adapted from photo by K. Tiefenbacher, used with permission)

The resulting welds (wire-to-wire, wire-to-lug) were electrically just as good as soldered joints. Of course, welded joints could not simply be de-soldered with a soldering iron. So, component replacement required the clipping of wires. Towards the end of the war, copper scarcity also led to some components having iron lead-wires instead of copper. This caused quality issued with both classic soldering and welding.

Both Siemens-Halske and AEG/Telefunken made miniature spot-welding equipment. Ref. 14D, 14E, 14F, 14G.


Fig. 25: Siemens-Halske "Schweißgriffel" welding pencil and "Schwingelektrode" vibro-electrode welder

(source: adapted from ref. 14F)


Fig. 26: Advertising for "tin-saving Siemens small welder" - transformers, pencil, vibro-electrode, and pliers

(source: 1944 Siemens-Hungary, ref. 14H)

As stated above, the pertinax circuit card has point-to-point wiring and the components are wire-ended. This is basically the second evolutionary step in component interconnect technology:

  • Wire-ended components directly wired to each other ("no board", mid-1800s through the 1930s),
  • Wire-ended components munted on generic discrete-wire boards (1930-1940s),
  • Single-layer-single-sided printed circuit boards (PCB) with through-hole component installation (since the early 1940s),
  • Two-sided-multi-layer substrates (mainstream since the mid-1980s).

These days, substrates can be flexible and even transparent. Note that single- and multi-layer printed-circuit technology actually dates back to the very early 1900s (ref. 11, 12, 13)! However, this technology did not become mainstream until commercialization of the transistor in the 1960s. Main-stream mass-produced boards typically have 4-8 layers. Low-volume or very special purpose boards can have dozens of layers, e.g., the 64-layer (about 1 inch thick!) motherboard of the 1980s version of the digital fly-by-wire flight control computer of the F-16 fighter aircraft.

PCBs are also referred to as Printed Wiring Board (PWB) or Etched Wiring Board. They are product-specific, and consist of traces of a conductive substance (typically copper foil) on an insulating ( = dielectric) substrate or "board". Boards have evolved from resin impregnated paper laminate (e.g., the ubiquitous Pertinax®) with solder lugs or pins, to glass-fiber reinforced polymer sheets and other specialty materials. As is the case with many technological inventions, the PCB cannot be attributed to a single inventor. A notable early milestone are cut or die-stamped metal foil patterns bonded to paper laminate board (flexible, with the equivalent of plated through-holes or vias). This was patented in 1903 by Albert Hanson (Berlin, UK patent). Other technologies are painting, spraying, and dusting of conductive substances, chemical deposition (electro-plating) and photolithography (etching). Associated names are Arthur Berry (UK, 1913), Max Ulrich Schoop (Switzerland, flame/thermal spray, 1918), Charles Ducas (electro-plating and the concept of multi-layer, US, 1925), and Paul Eisner (Austria, UK patents in 1941/43).


red-blue line