logo

Besides the Hell company and Siemens-Halske, there have been other manufacturers of teleprinter machines that use the Hellschreiber principle. Presented below are the ones that I am aware of. Some manufacturers may have filled the void caused by interruption during WW2 of exports from Germany to countries outside the German/Axis influence sphere, and it taking 2-3 years after WW2 for war-torn German companies to resume production.



By January 2019 this page had grown to 250 photos and diagrams. It had become rather large (ca. 24 MB download size), which caused long download times for some users. I decided to group the manufacturers listed above into two separate pages of roughly equal size: one for manufacturers in the UK (five of those listed above), and one for the rest of the world. This should be almost transparent to you. Please use the (unchanged!) list of manufacturers above to get to them, and update your bookmarks - if necessary.

©2004-2020 F. Dörenberg, unless stated otherwise. All rights reserved worldwide. No part of this publication may be used without permission from the author.


Latest page update: January 2020 (added info about a fourth surviving "Telewriter").

Previous updates: October 2019 (uploaded ref. 7V about the NVA/RTF Feld-Hell, updated associated text and added fig. E14), September 2019 (added ref. 7W).


red-blue line

Other mfrs


red-blue line

GENERAL POST OFFICE (GPO)

In the British Empire, the 1904 "Wireless Telegraphy Act" granted full monopolistic control of radio waves to the General Post Office (GPO). The GPO licensed all senders and receivers (telegraph, telephone, radio, teleprinter). It also owned, operated, and maintained all transmitter installations, including the punch-tape senders for Hellschreiber and regular teletype/teleprinter broadcast service. Since 1935, the GPO had already been providing the actual Hellcasts for Reuters news agency. The London Press Service (LPS) was inaugurated in August of 1945 as one of the "gray" propaganda services that was operated by the Central Office of Information (COI), ref. 30. It was the COI (successor to the war-time Ministry of Information, MOI) who, with technical backing from the GPO, proposed to introduce a Hellschreiber service for the LPS. Until then, LPS only had Morse telegraphy broadcast service.

In 1947, the GPO issued a request for tender on behalf of the Foreign Office / Colonial Office for a new Hell-printer and a "thermionic relay" contract. Ref. 19A, 20C. The "relay" is an electronics box with vacuum tubes ("thermionic valves") for a Hellschreiber-tone-pulse detector and a printer-solenoid driver amplifier. The request included full specifications and all official design drawings (including for the printer motor, specified to the Klaxon Company Ltd). The GPO obtained a tender from three companies in December of 1947 (ref. 19A):

  • Communications Division of Marconi's Wireless Telegraph Company Ltd. in Chelmsford/Essex. Marconi had been merged into Cables and Wireless Ltd. in 1929 (shortly thereafter renamed Cable and Wireless Ltd.), and was taken over by the English Electric company in 1946. Cable and Wireless (so, probably Marconi) had performed Hellschreiber transmission/reception experiments and signal bandwidth measurements in 1938/39. Ref. 28A, 29. Marconi already negotiated Hellschreiber patent licenses in 1935 with Siemens & Halske via Telefunken Gesellschaft für drahtlose Telegraphie m.b.H. (Telefunken for short). Ref. 36. Siemens was 50% owner of Telefunken, until bought out by joint-venture partner Allgemeine Elektricitäts Gesellschaft (AEG) in 1941. Marconi and Telefunken had general "exclusive territory" agreements.
  • Post Office Factories Department. The factories generally covered the assembly and manufacture, repair and reconditioning of all Post Office equipment and machinery. Until 1941, the factories were under control of the PO Stores Department.
  • Coventry Gauge and Tool Company Ltd. The Coventry tender included a "relay" from Pye Telecommunications Ltd. in Cambridge. The latter company was created early 1944 as a subsidiary of Pye Ltd.

GPO

Both Marconi and Coventry Gauge & Tool / Pye were awarded a contract. As stated above, this was a "new" contract. So there was a preexisting contract: printers from Coventry Gauge & Tool, and "relays" from Pye (ref. 19A). Early-model GPO-procured printers were also supplied to Reuters' news agency, when the supply from Siemens-Halske in Germany was interrupted when WW2 started (ref. 12A, 23A). As printers for both 12-line and (newer) 7-line Hell-font were required, the printer helix ("print wheel", "printer wheel", "marking wheel") could easily be exchanged (ref. 12A).

In total, at least 73 printer/relay sets were ordered under the "new" contract. Of these, 50 were from Coventry Gauge & Tool / Pye, and 3+20 from Marconi (ref. 19B, 20B, 23B). The Coventry units were intended for Europe and the Middle East, the Marconi units for India and the far East (ref. 20A). In November of 1948, acceptance of the 20 Marconi sets was delayed, pending implementation of a "considerable" list of necessary modifications. Marconi agreed to recall, modify, and improve the 20 printers that were on hold at the GPO, and replace a few old model printers that were already fielded (ref. 18D).

The Marconi printer and relay were Marconi-design. The printer was compatible with the GPO-designed relay, but the Marconi relay was not compatible with the GPO-designed printer. I.e., Hell-equipment sets from Marconi and from Coventry Gauge & Tool / Pye were incompatible! The Marconi relay was actually an off-the-shelf general-purpose model for high-speed telegraphy (ref. 20C). Note that both Hell-printer suppliers were obliged to use the GPO-specified motor from Klaxon Co. Ltd. Ref. 19A, 20A.

The following diagram illustrates the relationships between the government and industry partners involved in the GPO contract for Hell-equipment:

GPO

Figure K1: Relationships between British government agencies and industry regarding Hellschreibers and Hell-casts

(sources: ref. 12A, 18A, 19A, 23C, 23D)

The table below shows the prices for the various equipment items:

GPO

Figure K2: 1948 prices of British Hell-equipment

(sources: 1 = ref. 18A, 2 = ref. 19B, 3 = ref. 19D, 4 = ref. 19E, 5 = ref. 23E)

To put these prices in perspective: in 1946, the average salary in the British "metalworking, engineering and shipbuilding" industry was £22.40 per month (US$98.56 in 1946), ref. 27. Based on general inflation data (ref. 21), this would be equivalent to ca. £806 in 2016 (ca. €930 and US$1005). A Morse operator made about £300-400 a year in 1949 (ref. 23F). So, at the time, a Hell-printer represented several months’ salary of an average worker – as did German Hell-equipment in Germany. A teletype/teleprinter set, including "adaptor" [ = 2-tone filter unit], cost £350 (ref. 26A, 27A), which is significantly more than a set of Hell equipment. For pricing data on German and Swiss Hell-equipment, see the "Hell equipment prices 1937-1952" page.

The GPO-printer is, of course, a British "clone" rendition of the original and ubiquitous German Siemens-Hell Hellschreiber printers for news casts. The design and specification most likely came from the GPO's Research Station (part of the Engineering Dept. of the GPO's Radio Branch) at Dollis Hill in northwest London (near Willesden). Ref. 31. Dollis Hill is also where the GPO tested the Siemens-Hell system (both 12-line and 7-line) in 1934 (ref. 12A) and where the famous "Colossus" code-breaking computers of Bletchley Park were built (1943/44).

According to a 1947 manual of the GPO (ref. 12B), three Hell-printer versions were developed, and produced in significant numbers (i.e., not prototypes):

  • No. 1 Mark 1 (serial nr. 7-47).
  • No. 1 Mark 2 (serial nr. 49-73).
  • No. 1/T Mark 2 (serial nr. 74 and above).
  • Note: GPO engineers evaluated a Marconi printer (pre-dating the “new” model) with serial nr. 322 (ref. 20A); the highest serial nr. that I have documented is 229.

The referenced manual states that printers with serial numbers 1 to 6 and 48 "are of German origin, or are obsolete models". Apparently, there either was no model "1/T Mark 1", or few or none were produced...

The "T" indicates that printer components were treated or selected for operation in the hot-and-humid Tropics. A fair amount of motor insulation failures were observed in tropical climates (ref. 23A), but the motor design was not under control of the printer manufacturers. Note that the "relays" also needed to be fully tropicalized. E.g., by using capacitors with higher voltage rating, and sealing of transformers to avoid water absorption (ref. 19C). GPO engineering concluded that the Coventry printer was tropicalized "as much as possible and to a slightly greater extent than the examined [ = pre-new] Marconi model" (ref. 20A).

The following photos are of the label on an "early model" GPO-printer (ref. 12A, 12D). There is no model designator (e.g., "No 1") marked on the label: it reads "HELL PRINTER G.P.O. SERIAL .."

GPO

Figure K3: Label on an early model GPO Hell-printer - only "Hell printer" and "G.P.O. Serial" indicated

(source: Fig. 3 in ref. 12A, Fig. 12.38 & 12.39 in ref. 12D)

Serial nr. 31 has model "No 1/T" marked on its label, though based on the serial nr., it should be a model "No 1 Mark 1". Possibly, it is a regular model no. 1 that was modified by Marconi to be suitable for operation in tropical climates (ref. 18D). Serial nr. 192 and 229 are model "1-T", not "1/T", but those designators may have been interchangeable.

GPO

Figure K4: Equipment labels of three GPO Hell-printers


All GPO/Foreign Office combined, there were three versions of the Marconi printer (ref. 18D):

  • Ca. two dozen printers per the GPO design and specification.
  • Ca. 20 printers, based on that GPO design, but incorporating some of Marconi's own improvements.
  • A "new" model (ref. 23C, 24C). Announced changes (in response to Reuters’ experience and field experience at British overseas Posts, compared to the original GPO design, and corresponding GPO engineering assessment (ref. 24C):
  • Strengthened ink roller supporting arm, and this roller now has a double-race ball bearing [not considered a problem area by GPO].
  • Self-aligning bearing on the drive shaft of the tape feed [not considered a problem area by GPO].
  • Improved position-locking of the pinch-roller spring of the tape feed [not considered a problem area by GPO].
  • Motor governor resistor changed from screw-in type to clip-in type [actually a GPO suggestion, also implemented by Coventry].
  • Wiring to the motor cleaned up [also implemented by Coventry].
  • Access to motor brushes improved by moving other components [can only be marginal, as the GPO design fixes the motor location close to the front panel].
  • Finer motor speed control of the motor (1½% vs. 4% per notch) [motor manufacturer Klaxon Co. Ltd. is unwilling to make the change. Note: the motor of the German Hell-printers has a superior, continuously variable centrifugal speed regulator, i.e., no incremental notches. The GPO considered it too complex for use in the GPO printer].
  • Printer magnet assembly improved, to avoid shorts experienced with the original G.P.O. design.
  • Mounting of the top cover plate of the printer module was changed to avoid distortion.
  • Added clearance between top & front cover plate of the printer module [also implemented by Coventry].
  • “Oilite” bearing [ = porous bronze plain/journal/sleeve bearing] on the printing spindle shaft and on the driven tape feed roller, to avoid seizing [GPO has no experience with such bearings in the tropics, and they may be less accurate].
  • Detection of remote-control start/stop tone pulse made more robust to inadvertent starts & stops when receiving noisy signals [not considered very important by the GPO].
  • To provide satisfactory ventilation, louvers are incorporated in the back of the unit and air inlet holes in the bottom.
  • The holder for 4 inch diameter paper tape rolls has not been adapted to 5½ or 8 inch rolls, as this is not practical with the GPO-specified attachments points [Note: a 4 inch roll suffices for 2 hrs of continuous printing; at the time, the longest Hellcasts of the London Press Service were 1½ hrs, ref. 20A].
  • All printers are now tested for speed control, and correct printing of received Hellschreiber signals.

GPO

Figure K5: Front view of an early model GPO Hell-printer - cover of the printer module removed

(source: Fig. 12.39 in ref. 12D)

Below are photos of GPO printers with serial number 12, 31, 32, 192, and 229. I have not been able to determine whether they were manufactured by Coventry Gauge & Tool or by Marconi, though most likely by Marconi. There are no manufacturer identification markings on the outside of the units, nor on the inside of the unit that I was able investigate with the cover removed. Also, based on the common GPO design, the Coventry and Marconi printers were "practically identical" with "no serious advantages on either side" (ref. 20A).


GPO Hell-printer model 1 with serial nr. 12

The photos below are of a 1943 GPO Hell Printer:

GPO

Figure K6: Front view of G.P.O. Hell-printer model 1 with serial nr. 12

(source: ref. 40)

GPO

Figure K7: Front view of G.P.O. Hell-printer model 1 with serial nr. 12

(source: ref. 40)

GPO

Figure K8: Top view of G.P.O. Hell-printer model 1 with serial nr. 12 - cover removed

(source: ref. 40)

GPO

Figure K9: Bottom view of G.P.O. Hell-printer model 1 with serial nr. 12 - cover plate removed

(source: ref. 40)


GPO Hell-printer model 1/T with serial nr. 31

GPO

Figure K10: Front view of Hell-printer model 1/T with serial nr. 31 - cover of printing module removed

(this printer is in the collection of the Science Museum in London, photos with permission)

The printer "box" measures ca. 28.8x20.2x17 cm (WxHxD, 11.25x8x6.75 inch), excluding the protrusions (printer module, paper tape holder, tape feed mechanism, toggle switch, etc.). The holder for the roll of paper tape is copied directly from the 1934 Siemens-Hell printer model T.empf.12b. Compared to the German original, the GPO-printer has a paper tape feed mechanism with three rollers instead of two. As the German arrangement works flawlessly, the reason for adding a third roller is questionable. The printer uses standard 15 mm wide Hellschreiber paper tape. In the photo above, the retainer spring to the left of the printer solenoid (the dark orange component at the center) is missing - compare to the black & white photo above. The ink roller above the printer spindle is also missing.

GPO

Figure K11: Right-hand view of GPO Hell-printer model 1/T serial nr. 31 - cover removed

(at center: the horizontally installed motor with the slow/fast speed regulator knob)

GPO

Figure K12: Left-hand view of GPO Hell-printer model 1/T serial nr. 31 - cover removed


In the photo above, the jack marked "DC input" is connected directly to the solenoid of the printer mechanism. On the left is circuitry for remote on/off control - also copied from the 1930s/40s Siemens-Hell Presse-Hell printers models T.empf.12 and T.empf.14. The components marked "B" and "A-TH" are electromechanical relays. The "TH" stands for "thermal", to indicate that relay "A" is the thermal time-delay relay. A tone pulse of at least 0.5 sec would turn the motor on, whereas a tone of about 10 sec (8 sec in Siemens-Hell printers) would turn the motor off. Remote control operation is active when the main switch is in the "start stop" position. Note: this is only start-stop for the motor, and is not related to start-stop synchronization of sender and printer.

GPO

Figure K13: Toggle switch for selection of motor operation


GPO

Figure K14: Rear view of Marconi Hell-printer model 1/T serial nr. 31 - cover removed

(the component with four gray disks on the left is an L.T. (low tension) selenium bridge-rectifier)

GPO

Figure K15: Top view of GPO Hell-printer model 1/T serial nr. 31 - cover removed


GPO

Figure K16: Bottom view of GPO Hell-printer model 1/T serial nr. 31 - cover removed


The long green component in the photo above is the speed governor resistor, rated at 30 watts. It is screwed into a small light bulb socket. Per ref. 23A, its power rating was insufficient and it tended to overheat. It was replaced with a clip-on type resistor in Marconi's "new" printer model.

The model 1-T with serial number 192 and 229 both have a 10-start printer spindle. Adjacent helix-starts overlap by 50%. The actual spindle hub is slid onto the spindle shaft. This makes it easy to exchange spindles, for compatibility with both the 7-line and the 12-line Hell font (ref. 12A). In the German Hell-printers, the spindle hub is an integral part of the shaft, and not easily exchanged. The GPO spindle hub, excluding the spindle threads, has a diameter of 22 mm. The threads of the spindle have a height of about 1.5 mm.

GPO

Figure K17: Close-up of the printer spindle of model 1/T serial nr. 31


GPO

Figure K18: 3-prong power plug of Hell-printer model 1/T serial nr. 31


The printer's mains power plug fits into a switched power outlet at the lower right-hand corner of the front of the type W11 printer amplifier:

GPO

Figure K19: Printer detector/driver-amplifier "thermionic relay" model W11, serial number 1038

(this item is in the collection of the Science Museum in London, inventory nr. 1978-146, photo with permission)

The above "relay" measures ca. 42x26x26.5 cm (WxHxD, 16.5x10.5x10.25 inch). The inventory documentation of this unit does not indicate if it was manufactured by Marconi or Pye. There are no manufacturer identification markings on the housing. See ref. 37 for a description of types W1 through W8 (1938-1943). The latter types were designed by the Wireless Branch ("W. Branch") of the GPO, for "a variety of uses on radio circuits": keying a radio transmitter, operating electromechanical telegraph recorders (incl. Hellschreibers), as well as for "singing suppressors and voice-operated switches" on certain radio telephone circuits. It is reasonable to assume that types W9-W11 were designed by W. Branch as well. The first type (W1 ?) was developed around 1931. Type W8 comprised four tubes (valves; types V.T. 200 and V.T.196, two each), and was the first type that specifically intended for operating Hellschreiber printers. It was a table model with a DC output current of 20 or 45 mA (into a 500 ohm load), could provide both 110 Vac and 110 Vdc to the printer, and could also be used as a tone-signal amplifier. Audio input could be passed directly to the tone detector (full-wave rectifier with two solid-state diodes) or via an audio input bandpass filter with a 900 Hz center frequency and a 600 Hz bandwidth. Alternatively, the W8 could be used as a DC-pulse amplifier.

The next three photos show another GPO amplifier. The W-type is unknown. It only has a 110 Vdc output (on the front), so it does not appear to be a type W-8.


GPO

Figure K20: Front view of a GPO printer amplifier (W-type unknown)

(source: ref. 39)

GPO

Figure K21: Front view of a GPO printer amplifier (W-type unknown) - cover flipped up

(source: ref. 39; the toggle switch to the left of the empty tube socket is for selecting 20 vs. 45 mA DC output current)

GPO

Figure K22: Bottom view of a GPO printer amplifier (W-type unknown)

(source: ref. 39)


GPO Hell-printer model 1/T with serial nr. 32

Below is a photo of the inside of serial nr. 32. Note the differences compared to the top view of serial nr. 12 above.

GPO

Figure K23: Top view of GPO printer 1/T with serial nr. 32 - cover removed

(source: ref. 38)


GPO Hell-printer model 1-T with serial nr. 192

GPO

Figure K24: Front of model no. 1-T Hell-printer with serial nr. 192 (ink applicator roller missing)

(source of 1-T SNR 192 photos on this page: ©2013 Tony Radio Collection, used with permission)

Note the different shape of the paper tape roll holder, compared to the printers with serial number 31 above, and 229 shown further below.

GPO

Figure K25: The inside of the printer module of model 1-T serial nr. 192, with the electro-magnet


GPO

Figure K26: Close-up of the printer spindle of model 1-T serial nr. 192


GPO

Figure K27: Paper tape transport mechanism (left) and down-gearing for the spindle & paper transport (serial nr. 192)


GPO Hell-printer model 1-T with serial nr. 229

GPO

Figure K28: Printer model no. 1-T Hell-printer with serial nr. 229 - ca. 1944

(this printer is in the collection of the Science Museum in London, inventory nr. 1978-145, photos taken with permission)

Note the slightly different paper tape roll holder above compared to the one of serial nr. 31 and 192 further above.

GPO

Figure K29: Close-up of the paper tape feed mechanism and printer module of 1-T Hell-printer with serial nr. 229


As noted above, the printer model with serial nr. 31 had overheating problems. Its housing is fully closed. The model with serial nr. 229 has a screened ventilation hole (ca. 6 cm diameter) in the top of the housing, a group of ten holes (ca. 14 mm diameter each) in the bottom cover plate, and ventilation louvers at the back.

GPO

Figure K30: Bottom cover of model no. 1/T Hell-printer with serial nr. 31 (left) and model 1-T with serial nr. 229


In the photos above, note the different placement of the retaining screws of the bottom cover. Also, in serial nr. 229, the jack for "DC" (input to the printer magnet) on the left-hand side of the unit is placed much closer to the front panel than the "DC Input" jack of serial nr. 31.


The Claxon Company Ltd.

The motor in the GPO-printers is made by The Klaxon Co., Ltd. As the name clearly suggests, Klaxon's primary product was car horns (i.e., claxons). The company was founded in 1909. It had the telegraphic address "Klaxonet, London". In the exhibitors listing of the 1937 British Industries Fair, the following Klaxon products are listed: fractional-H.P. motors and geared motor units, generators, regulators, relays, transformers, grinding machines, industrial signals and sirens, fire and burglar alarms, staff locators, electric and hand operated horns, push buttons, electric sign flashers. A 1938 Klaxon catalog lists "air-raid warning devices for internal and external situations", gongs, sirens, whistles, and other sound- emitters". In 1961 it was listed as a manufacturer of fractional-HP motors, warning signals and windscreen wipers.

GPO

Figure K31: The motor of the GPO model 1-T serial nr. 192, including 25:1 down-gearing from 2500 rpm

(photo source: ©2013 Tony Radio Collection, used with permission)

There are several addresses associated with the company: Klaxon Co. of Birmingham, 36 Blandford Street, London, W1, and Klaxon Ltd., 201 Holland Park Avenue, London, W11. The latter may have been just an import registration address, e.g., for foreign subsidiaries. This address also appears on fractional-HP motors, geared motor, and synchronous motors. The GPO-designed motor model EK5GB1-W3 is a series-wound AC/DC type.

Even though the initial order was only for 50 motors, Klaxon had large problems delivering them. Due to non delivery of the motors (a mere £5 item), the printers from Coventry Gauge & Tool were delayed. In turn, London Press Service was forced to delay its transition to Hell-service to Europe and the Middle East by at least six month, to the end of 1947. This became a political issue (due to the missed savings for the government), and the GPO even looked into getting motors from Germany (ref. 26D, 26E)! In March of 1947, Mr. Gentry, Sales Manager of Klaxon, blamed the delays on the company being inundated with orders [for their automotive motors & horns], the order from new customer (Coventry Gauge & Tool) being placed late (in August of 1946 = 6 months ago...), Klaxon prefers to provide motors to long-time client Marconi [ = competing printer], but Klaxon is willing to support any other company to make the motors. Ref. 26A, 26B.

Further delays were incurred mid-1947, when Coventry Gauge & Tool requested Klaxon to make modifications, as the delivered prototypes did not comply with the required wiring standards (ref. 26C). Another change was made at some point: the motor speed governor had two (carbon)brushes in the Mark 1 model, but only one in the Mark 2 machines (ref. 12B).

In August of 1947, Mr. Gentry (meanwhile General Director) blames all delays on the production output being limited by "government interference and fuel supply restrictions". As to producing a mere 50 motors, he states: "Can’t do it! Just can’t do it!" (ref. 26B).

Note that Pye also had delays in deliveries of their relays/amplifiers, but those were due to labor shortage and sickness (ref. 26F).


GPO Hell Sender

The following photo shows a GPO "Hell Sender". The actual model/type is unknown. On the left hand side, there are two outputs: "7-line" and "12-line". The latter has a plug in it. These outputs correspond to the 7-line and 12-line Hellschreiber font. Also on the front are two exchangeable notched disks and associated switch contacts. The front disk appears to capture the pixel sequence of three characters, which could be a station identifier. From the available photo, it is unclear if the disk in the rear also captures several characters. Possibly one disk is for characters in 7-line format, and the other in 12-line format. This sender is for continuously sending a fixed character sequence. The unit has three toggle switches on the front, for "space mark", "Motor on", and "osc[illator]". The latter implies that the unit had a built-in tone oscillator.

GPO

Figure K32: GPO "Hell Sender" with serial number 47

(source: ref. 41)


Y-SERVICE

The German Wehrmacht started using the Hellschreiber system in 1935 (wired and wireless). The British radio interception service appears to have been oblivious to this until late 1940 or early 1941 (§1.3 and 1.7 in ref. 32, ref. 33). Even when the signal type was correctly identified, there was no suitable printing equipment available: the standard commercial Hellschreiber operated at a different speed. There were two radio-interception stations that specialized in teletype/teleprinter and Hellschreiber traffic: the Foreign Office (FO) station at Knockholt (near Sevenoaks, in the countryside, about 30 km southeast of the City of London), and the Metropolitan Police (MePo) station at Denmark Hill in Camberwell (about 5 km southeast of the City).

The British "Wireless Interception" service (WI-service, phonetically abbreviated to "Y" service) was responsible for the monitoring of enemy radio transmissions. The radio-intercept stations were known as Y-stations. The service dates back to WW1, and was run by the Royal Navy: Naval Intelligence Department I.D. 25, also known as "Room 40". In 1920, the service was transferred from the Navy to the Foreign Office (FO), and "Room 40" was renamed to Government Code and Cipher School (GCCS). In 1939, the GCCS moved to Bletchley Park (BP) in Milton Keynes/Buckinghamshire (about 80 km / 50 miles north-west of London), and was renamed to Government Code Head Quarters (GCHQ). Ref. 35.

During WW2, the Y-service covered radio-telephony, Morse telegraphy, and Non-Morse (NoMo) transmissions, whether encrypted or not. NoMo traffic included teletype/teleprinter and Hellschreiber. The Y-stations were operated by a number of government agencies (the branches of the armed forces, the Metropolitan Police, and the General Post Office) and the Marconi company. Some stations only had direction-finding (D/F) capability.

During the course of WW2, the service grew from a few Y-stations, to a global network of small and large stations. They were located in the UK, the Middle East, Far East, North Africa, mainland Europe, and offshore. Intercepted encrypted signals were either analyzed locally, or transferred (by dispatch riders on motorcycle or via teleprinter) to BP. Sometimes BP is referred to as "Station X" (i.e., station nr. 10), though that actually refer to a small Special Intelligence Service (SIS) wireless station (MI6 Section VIII) that was originally located at Barnes in west London (south of the Thames), and temporarily moved to BP.

A "satisfactory universal Hellschreiber machine" was designed and manufactured in small numbers during the spring of 1942 (ref. 12C, §2.7 in ref. 32). It is unclear by which organization. It is also unclear if there was just one such development initiative, or multiple in parallel: early 1943, the Foreign Office research department (also) decided to build a universal machine that was to cover all speed ranges (ref. 22A). The latter had a printing mechanism that was copied form the German commercial machines. The universal-DC motor with "complicated centrifugal speed governor" was replaced with a type of synchronous AC-motor that was normally used in teleprinter equipment of Creed & Company Ltd., with an adjustable speed regulator comprising a pair of friction disks.

The Metropolitan Police Y-station at Denmark Hill had equipment workshops in West Wickenham at least as early as 1939 (§2.6 in ref. 32). At some point, the FO's Knockholt station was expanded with an equipment research & development operation (laboratory and workshop). However, this may not have been until 1943 (ref. 22A). This later became the Foreign Office Research & Development Establishment (FORDE).

During 1942-44, RAF 192 Squadron (§9.3 in ref. 32) and the British/American Noise Investigation Bureau (NIB, ref. 34A) investigated a special format of Hellschreiber signals that were transmitted by the Luftwaffe's rotating radio-navigation beacons of type "Bernhard". The FO's Knockholt facilities assisted 192 Sq with subject knowledge and the required tone-filter.

No photos or detailed descriptions of these universal Hell-printers are available.

If you have any information about this "universal Hell-printer", please contact me!




CREED & COMPANY

Creed

The British Creed & Company was founded in 1912 and its name shortened to "Creed & Co." in 1916. In 1924, they entered the teleprinter market and were absorbed into the International telegraph & Telephone Corp. (ITT) in 1928. Ref. 16A.

Late 1950, Creed developed "… a simple facsimile transceiver, designed to exchange, with an identical machine over relatively short distances, brief messages written and recorded on Teledos tape" (ref. 16B). It is indeed closer to a direct-printing black & white fax system, then a Hellschreiber in the strict sense. However, it does use a spinning "stylus" (from the Latin word for "pointed implement") at both the sending and the receiving (printing) side.

The transceiver (sender/printer) was Creed Model TR.105. Its successor, model TR.105/1, basically had redesigned electronics and no manual gain control.

Creed

Figure P1: Creed Transceiver Model TR.105

(source: ref. 16B)

On the sending side, a message is hand-written on a segment of 3/4 inch wide Teledos tape. The height of the written message is 5/8 inch max. This tape has an enameled surface. Writing is done with a regular pencil, not harder than HB-grade. The pencil-written message is pulled between the spinning stylus and the (curved) platen. The tips of the 4-pointed stylus sweep across the moving tape. As the revolving stylus scans the moving tape, its electrical resistance to earth/ground varies according to whether the points are touching blank tape or message markings. The resulting pulses are transmitted to a second machine via phone lines.

Creed

Figure P2: Stylus and curved platen


Note that this method actually dates back to 1864! The patented "copying telegraph" of Bernhard Meyer had a scanner/sender that used a metal tablet on which text was written with non-conductive ink (Creed's Teledos tape and conductive pencil "ink" is actually the opposite). Scanning was done with a swinging platinum stylus. Meyer's electro-chemical printer used paper tape and a 1-turn spindle.

At the receiver, a blank Teledos tape is passed under the stylus. The incoming message pulses from the (pone)line are amplified into voltage surges at the recording stylus. These surges are high enough to burn through the lacquer of the blank tape and leave dark markings on its surface. These markings are identical with the pencil markings on the transmitter tape. Electrical contact with the stylus is maintained via a slip-ring. This is, in effect, a thermal printer that uses "dry-electrolytic action".

The Creed Teletape system appears to be a direct retake of the late 1940s Teletape system developed by Western Union (ref. 16C, 16D). Western Union's electro-sensitive tape was called Teledeltos rather than Teledos, and printing was also electro-thermal. The stylus is identical in both cases. The Western Union' stylus spun at 1800 rpm.

Creed

Figure P3: Scanning-stylus of the Western Union Teletape system

(source: ref. 16C)

Creed

Figure P4: Western Union Teletape transceiver with built-in amplifier & control unit

(source: ref. 16C)

Creed Model TR.105 has separate motors for the stylus and tape feed. At the receiver, the motors are started automatically when tone pulses are received. "Black" is represented by a keyed tone of 5000 Hz (+/- 200 Hz). Bandwidth was 2500-7500 Hz. This is too much for a standard analog public telephone network but a local network would be fine. Tape is fed at 1 inch/sec (1.5 m/min). Equivalent scanning resolution is 100 lines/inch.

Creed

Figure P5: Functional schematic of Teletape Transceiver Model TR.105

(source: appendix A of ref. 16E)

For the given tape feed speed (1 inch/sec), this implies a stylus speed of 1500 rpm. Model TR.105/1 has four vacuum tubes: CV2136, CV491, 6AU6, and CV1535. The unit operates on 110-145 or 200-225 Vac, 50 Hz (selectable in steps of 5 volt AC). Power consumption is 28 W (standby) - 90 W (sending). Life expectancy of the stylus was sufficient for one roll of Teledos tape. Ref. 16B.

Creed

Figure P6: The inside of Teletape Transceiver Model TR.105 (rear view)

(source: ref. 16B)


Creed

Figure P7: The inside of Teletape Transceiver Model TR.105 (view from right-hand side)

(source: ref. 16B)


BRITISH PORTABLE PRINTER / KEYBOARD-SENDER "TELEWRITER" YB.02251

In August of 2011, I received a message from David H. Jones in England. He recalled examining a rare start-stop Hellschreiber variant at the end of 1945, while he was working at the Post Office Research Station at Dollis Hill (north London). He was kind enough to write down his quite detailed memories (ref. 43A). The machine was portable, British-made, had a 3-row keyboard (no figure-keys), and an electrochemical Hellschreiber printer mechanism. The machine appeared to have been part of a small batch of pre-production prototypes. Many years later, in April of 2018, the plot finally thickened when I was contacted by Peter Prest. He owns a machine that appeared to match David Jones's description, though adapted to series production and with a 4-row keyboard. Subsequent research has turned up two additional machines.

TLWRT-GTL

Figure R1: Telewriter - case opened

(source original unedited photo: Science Museum in London, inventory nr. 1968-586)


TLWRT-GTL

Figure R2: Telewriter - case closed

(source original unedited photo: Science Museum in London, inventory nr. 1968-586)

The equipment label on the carrying case in the photo above shows the following information:

  • YB.02251 is the part number. "YB" refers to the section "Signal Stores, Automatic Telegraph, Line Transmission Equipment and Cryptographic Equipment" of the British Army Ordnance Store catalog. I.e., it is a military teleprinter.
  • Telewriter is the model designator.
  • G.T.L.: unknown.
  • SSerial № 28.

The machine shown above is one of four known to still exist. It is in the collection of the Science Museum in London. It was donated to the museum in 1968 by B. Robertson, ref. 43K. The second machine carries serial nr. 109, and is in the collection of the Signals Museum RAF in Henlow/UK. The third machine is mounted on a board, rather than in a numbered carrying case. It is in the private collection of Peter Prest. In his analysis (ref. 43J), Peter construes that the machine was probably conceived in 1940, based on the operating instructions dating from the end of 1945 (ref. 43B), and assuming standard product development and procurement processes. The existence of a fourth machine was brough to my attention late 2019. It has been in a private collection in the UK since 1984 and has serial number 25.

None of the three machines carry any identification of the manufacturer, unless "G.T.L." on the case label is a reference. However, research so far has not found a plausible explanation for this abbreviation.

TLWRT-GTL

Figure R3: Telewriter

(source: left - adapted from Fig.1 in ref. 43B; right - unedited original photo: ©2018 Peter Prest; used with permission)

Note the hinged clear plastic hood to the left of the paper tape spool in Fig. R1 and R3. Most likely, its purpose is to protect the fragile printer wheel and the distributor rotor. Such hoods are not uncommon for conventional teleprinters of the same era, where they may also provide some noise reduction.

The unit is actually quite small: it measures a mere 13½x6½x9½ inch (WxHxD, ≈34x16½x24 cm). The case is made of dark brown paxolin (one of the trade names for phenolic resin bonded paper laminate, like pertinax). The machine weighs 24½ lbs (≈11 kg). It is powered by 12 Vdc, provided by two 6 Vdc / 16 Ah batteries.

A regular Hellschreiber printer uses a spinning helix (spindle) to generate an inked point that continuously sweeps across the width of a paper tape. The moving tape is tapped against the inked spindle, in the rhythm of received pixel pulses. See the "Printing Hellschreiber signals" section of the "How it works" page. The Telewriter uses a different approach: it creates such a sweeping point with a printer wheel (a.k.a. "pecker wheel"). It has 5 tangential springs that are evenly spaced. At the end of each spring tab, there is a small chisel-shaped "pecker". It measures approximately 0.3 x 1 mm. See the photo below. This suggests that printed characters are about 5 mm wide (twice as wide as standard Hellschreiber). Such a printer wheel prints a single line of text, unlike standard 1930s/1940s Hellschreiber spindles (see the "Synchronization" section of the "How it works" page). This is consistent with the Telewriter using a start-stop system to synchronize sending and receiving machines, whereas standard Hellschreibers of that era use no synchronization.

The Telewriter does not use ink. Instead, it uses an electro-chemical process. A water container is inserted into the left hand side of the machine. Chemically impregnated paper tape passes over the wet wick in the top of the water container. This moistening "activates" the paper, just before it reaches the printer wheel. The tape has to be moist, so as to conduct electrical current. Electrochemical telegraphy printing on paper tape dates back to the early 1800s by Samuel Thomas von Sömmering (a remarkable anatomist, physician, anthropologist, paleontologist, and inventor) in Prussia/Germany. It was improved upon several decades later (e.g., Alexander Bain, ref. 43C, 43D). It is unknown which chemical compound was used to impregnate the paper tape of the Telewriter. Rudolf Hell's original 1929 Hellschreiber prototypes used standard yellowish potassium ferrocyanide (prussiate of potash, "gelbes Blutlaugensaltz", unlike the red ferricyanide). Often, ammonium nitrate was added as a deliquescent (to keep the paper damp). Passing current through the yellowish salt solution causes electro-oxidation that decomposes the salt solution into a compound called Prussian Blue ("preußisch Blau", "Berlin Blau"). The impregnated paper tape only turns dark blue at the electrode with the highest potential. Typically, a potential of about 1 volt suffices (which may explains the 30 ohms resistor in series with the printer wheel). The tape is bleached at the electrode with the lower/negative potential. Heating the compound, mixing it with a strong acid, and exposing it to UV light, causes toxic cyanide gas to be released. A description of how to impregnate tape and a recipe for the chemical solution can be found in ref. 43L.

Prussian Blue dye was used since the early 1700s, including for dyeing the cloth used for the uniforms of the Prussian military - hence its name. It is also gave its characteristic color to "blueprints": copies of technical drawings, based on a photochemical process involving Prussian Blue, widely used in the decades preceding the modern photocopier.

TLWRT-GTL

Figure R4: Close-up of the "pecker wheel" with its five spring-loaded tips

(source unedited original photo: ©2018 Peter Prest; used with permission)


Note that the five tipped springs leafs are evenly spaced, but only around a part of the circumference of the printer wheel, almost as if a sixth one is missing:

TLWRT-GTL

Figure R5: Principle of the Telewriter's electrochemical paper tape printer


Close-up of the Telewriter printing wheel, single character sequence in slow-motion

(©2018 P. Prest; used with permission)  --  if player controls not visible: move mouse cursor over image


The original reels of Telewriter paper tape carried the British Army Ordnance Store number YB 03858 (ref. 43K). Width of the paper tape is 11/16 inch (≈ 17.5 mm). This is slightly wider than the 15 mm standard Hellschreiber tape for printing two identical parallel lines of text, and much wider than the 9.5 mm tape used in single-line start-stop Hellschreibers that were made by Hell/Siemens-Halske. The paper tape transportation drum below the printer wheel has a diameter of 1.25 inch (≈ 3.2 cm).

The Telewriter keyboard has a 4-row 38-key QWERTY layout with the keys A-Z, 2-9, space, + - / and "." The letter "I" is also used for "1", and the letter "O" for zero. For legibility of the print-outs, only capital letters are used. The presence of the space key implies that the Telewriter is a start-stop teleprinter system: the sending and receiving machine are asynchronous. They are (momentarily) synchronized for each character, by a start pulse that is sent as part of the each character. Synchronous Hellschreiber machines do not have (nor need) a space character: they are not start-stop systems. Instead, they rely on sufficiently equal motor speeds, combined with always printing two identical parallel lines of text, one above the other. One line is always legible, even if speed differences cause the slanting of the printed text. Start-stop systems (esp. over radio, with noise, fading, and multi-path echoes) are vulnerable to false (or omitted) start-pulse detection.

TLWRT-GTL

Figure R6: Four-row keyboard layout of the Telewriter machine

(source unedited original photo: ©2018 Peter Prest; used with permission)


The letter type on the keys closely resembles that of a 1930s portable typewriter of the British Oliver Typewriter Company. The keys are metal-rimmed, which is typical of typewriters of that era (Remington, Bar-Lock, Underwood, etc.).

The Telewriter uses the following bit-map font:

TLWRT-GTL

Figure R7: The font of the Telewriter - 5x5 pixels in an 8x7 field

(source: based on by signal tracing in a Telewriter machine by Peter Prest)


TLWRT-GTL

Figure R8: Comparison of Telewriter with standard Hellschreiber fonts and speeds


As with all pre-1970s keyboard teleprinter systems (i.e., pre Hell-80), the machine has an electro-mechanical font "memory". It comprises an individually coded key bar for each key of the keyboard, a set of 30 cross-wire installed just below the key bars (see Fig. R9 and R10), and a scanning rotary switch with 56 contact studs (one for each pixel of the 7x8 font matrix, see Fig. R11).

TLWRT-GTL

Figure R9: Bottom view of the Telewriter machine

(source unedited original photo: ©2018 Peter Prest; used with permission)


TLWRT-GTL

Figure R10: Fixation and tensioners of the keyboard cross-wire ends (same on the right hand side of the keyboard)

(source unedited original photos: ©2018 Peter Prest; used with permission)


TLWRT-GTL

Figure R11: The rotary distributor of the Telewriter machine

(source unedited original photo: ©2018 Peter Prest; used with permission)


The "scanning switch" distributor, with it turned brass studs embedded in black thermoplastic, may have been made by Painton & Co. Ltd. of Kingsthorpe/Northampton (est. 1935). Ref. 43A.


TLWRT-GTL

Figure R12: Painton & Co. Ltd. "Winkler" switch wafer and 1952 advertising


The 7x8 font matrix shown in Fig. R7 above, shows that first two columns are identical for all characters. These columns contain the start pulse of 8 pixel durations. This leaves a 7x6 sub matrix for printed character pixels. The last column is always blank, and provides the required spacing between printed characters. Of the remaining 7x5 sub matrix, the top pixels (row 7) are also always blank. The reason for this is unclear, as a start-stop Hell-printer does not need this: it only prints a single line of text, not two identical lines that require spacing between lines. In the end, the font memory only accommodates a 6x5=30 pixel sub matrix. This is why there are 30 cross-wires installed below the key bars of the keyboard. Note, however, that the five pixels of the bottom row are not used in the actual font of Fig. R7. Most likely, these pixels (and associated cross-wires) were reserved for making the font more legible at some point in the future - as was done in the "real" Hell fonts.

TLWRT-GTL

Figure R13: Going from full character encoding to "keyboard selector + rotary distributor" implementation



Each of the 30 cross-wires is connected to the corresponding contact stud of the rotary distributor:

TLWRT-GTL

Figure R14: Connections between keyboard cross-wires and rotary distributor

(distributor wiper shown at the rest position = stud 4; note: on the machine, stud 9 is at the top)


The contact studs that correspond to the start pulse pixels are hard-wired to the electrical "common" of the local key board and printer mechanism, and to the "L1" terminal of the 2-wire line that is connected to a remote Telewriter machine.

Each keybar has key-specific protrusions:

TLWRT-GTL

Figure R15: Keybar for the letter "E"


When a key is engaged by pressing it, the protrusions make contact with (only) the cross-wires that correspond to the pixels of that particular character. As in Creed teleprinters (in particular Creed model 3X), the keybar is perforated to save weight.

When the machine is turned on, the motor is running continuously. As described in detail further below, when a key is pressed, the wiper of the rotary distributor makes exactly one revolution at a speed that is equivalent to 160 rpm = 160 characters per minute = 160/60=2.67 characters/sec. Each time the wiper passes a stud contact that is connected to "common" via the cross-wires and the selected key bar, a DC voltage pulse causes a pixel dot to be printed by the local printer and simultaneously by the connected remote printer. When the distributor wiper transitions from one stud to the next, the contact is "make before break" (i.e., both studs are briefly connected). While printing a character, the paper tape smoothly advances over the width of a character and then stops, like the rotary distributor.


TLWRT-GTL

Figure R16: 12 Vdc motor of the Telewriter machine in situ - with 90° gearing (left) and centrifugal governor (right)

(source unedited original photo: ©2018 Peter Prest; used with permission)


The motor has a centrifugal governor that is mounted on one of the two motor shafts. As visible in the photo above, the black disk of the governor has two slip rings on the shaft side. Two spring-loaded carbon brushes ride on these rings. Inside the governor, the rings are connected to the two contacts of the centrifugal switch. The governor has a diameter of 5 cm (2 inch).

TLWRT-GTL

Figure R17: The centrifugal governor with its cover - for comparison, the governor of a 1930s Creed 7B teleprinter

(source unedited original photo: ©2018 Peter Prest; used with permission; source photo Creed governor: ref. 43E)

The manufacturer of the motor is unknown. Another such motor, known to be of WW2 vintage, has Air Ministry markings (embossed "crown + AM"), as well as "Type CM3 12 V" marked on one end bell of the housing.

TLWRT-GTL

Figure R18: The 12 Vdc shunt field motor of the Telewriter machine - with the 90° gear box attached

(source unedited original photo: ©2018 Peter Prest; used with permission)


TLWRT-GTL

Figure R19: Top view of the drive mechanism

(source unedited original photo: ©2018 Peter Prest; used with permission)


TLWRT-GTL

Figure R20: Drawing of the top view of the drive mechanism

(note: lever L2 does not interact with the drive mechanism; possibly it was used to hold coaxial shaft E during assembly of the machine)


Most of the driven shafts are supported by "Oilite" type bearing bushings (UK: bushes). The bearing sleeves are made of sintered porous bronze and are impregnated with oil, which makes them self-lubricating.


TLWRT-GTL

Figure R21: Drawing of the front of the mechanical drive unit with printer and distributor



TLWRT-GTL

Figure R22: Interaction of keyboard bar 2 and slip clutch A

(photo and original drawing: ©2018 Peter Prest, used with permission)

The fixture F1 (green), the Bar 2 Follower (orange), and springs Sp2 and Sp3 are also visible in Figure R23 below, through the bottom cut-out in the rear panel.


TLWRT-GTL

Figure R23: Rear view of the mechanical drive unit

(source original unedited photo: ©2018 Peter Prest, used with permission)

Peter Prest analyzed the motion sequence of the machine with the help of a low-speed motor drive that replaced the original DC motor. Let's go through the mechanical drive sequence for local operation only. Refer to Figures R19 - R23 above, and the schematics in Figures R24 - R26 below:

  • The machine is in the powered down state (Master Switch "off"):
  • A 12 volt /16 Ah battery is connected to the + and - terminals.
  • Terminals R1 and R2 are not bridged (they are only strapped for line operation, with one or more remote machines).
  • The machine is turned on with the Master Switch, which is a 4PDT switch (four simultaneous toggle switches, MS1-MS4):
  • The contacts of MS1-MS4 toggle (see circuit diagram in Fig. R25 below, which shows switch positions for Master "off").
  • The motor starts and runs continuously (switch MS4 is closed).
  • The motor speed is regulated by the centrifugal governor switch S3 (see Fig. R17), which is placed in series with the DC motor's armature. The motor's field winding is energized permanently (as opposed to permanently energizing the armature and using the field winding for speed regulation).
  • Via a small 90° gear box (see Fig. R16, R18), followed by skew gears A and B, shaft A (located in Compartment 1) is now driven continuously. Ref. Figure R20.
  • At the far left end of shaft A (i.e., in Compartment 2), gear E permanently drives gear F. Hence, shaft D is also driven continuously.
  • Note that the coaxial hollow shaft C (colored red in Fig. R20) does not rotate yet! It is prevented from turning with shaft A, as slip clutch A is held by Catch A. Hence, shaft B (of the rotary distributor) does not rotate either.
  • Likewise, coaxial shaft E (colored green in Fig. R20) is at rest. It is prevented from turning with shaft D, as slip clutch B is held by Catch B. Hence, shafts F ( = printer wheel) and G ( = paper tape transport) do not rotate either.
  • Switch MS3 is closed, so the printer system is enabled.
  • The drive system remains in this idling state until one of the keyboard keys is pressed:
  • The hinged keyboard Bar 1 is installed across the front of the keyboard mechanism (see the left hand image of Fig. R10), and passes underneath the SPACE key. If a key is pressed, the entire Bar 1 pivots, and its lower edge moves towards the front of the machine, away from the keyboard.
  • The movement of Bar 1 pulls keyboard Bar 2 (the orange item in Fig. R22) also towards the front of the machine ( = to the left in Fig. R22). This causes the Bar 2 Follower to pull on the lower end of the Catch A Arm (blue in Fig. R22), such that Catch A releases Slip Clutch A (red in Fig. R22). As Bar 2 moves forward, the bend in the upper branch of Fixture 1 forces the Bar 2 Follower forward and downward. Spring Sp2 ensures that this follower follows the curved contour of Fixture 1.
  • Shortly after Slip Clutch A is released and now rotates with Shaft A, the Cam B follower of Lever L1 (yellow in Fig. R10) drops off the vertical edge of Cam B, as the lever is pulled by Spring Sp1. The lower end of this lever moves towards the front of the machine until it is stopped by Peg 1 that is mounted on Bar 2. The force of the spring is such that Bar 2 is kept in its forward position, and thereby Bar 1 in its downward pivoted position. Therefore, the pressed key remains pressed for the duration of the motion sequence, while preventing all other keys from being pressed. The forward motion of Bar 2 causes its follower to slide past the bottom end of the Catch A Arm. The latter then pops back against the adjustable Catch A End-Stop of Fixture 1, as it is pulled by Spring Sp3. This re-arms Catch A, for disengaging Slip Clutch A from Shaft A again at the end of the drive sequence.
  • The keybar (see Fig. R15) of the pressed key descends onto the 30 keyboard cross-wires (see Fig. R9, R10, R14). Each keybar has key-specific protrusions that touch a subset of these wires, and connect them electrically to keyboard "common". This wire subset corresponds to the pixels that make up the character that is selected with the pressed key (Fig. R7, R14). Each of the 30 cross-wires is connected to one of the 56 studs of the distributor (though wires nr. 26-30 are held in reserve and are not used). Via the keybar protrusions, only the key-specific pixels are enabled.
  • As soon as, and as long as, Slip Clutch A is released, the coaxial shaft C rotates with the continuously turning shaft A. Via skew gears D and C, shaft B ( = rotary distributor) now also turns. The distributor wiper (which was at rest at distributor stud nr. 4) rotates clockwise (looking at the front of the machine), and in sequence, makes contact with each of the 56 pixel studs.
  • The rear part of Shaft B carries Cam A. When idling, the follower of the cam-driven contact S1 rests in the notch of Cam A and S1 is connected to terminal R1 (see Fig. R24-R26). As shaft B starts to turn, S1 leaves the notch, is disconnected from R1, and makes contact with the 500 Ω resistor. At this point in the drive sequence, the distributor wiper passes "start pulse" studs nr. 5-12. These studs are hard-wired to keyboard "common". Hence, the electro-magnet is energised via switch MS3 and cam contact S2. This causes the spring-loaded Catch B to release Slip Clutch B
  • Soon after the motion sequence starts, the cam briefly actuates cam contact S1 (see Fig. R20, R25, R26). This briefly connects one side of the two series-connected solenoids of the electro-magnet to +12 volt: via the 500 ohm resistor, cam contact S1, switch MS3, and cam contact S2.This connects the lower side of the solenoids to the negative terminal of the 12 volt battery. The resulting energization pulse causes the spring-loaded armature of the electro-magnet (see Fig. R23) to pull Catch B, which releases Slip Clutch B. In turn, this enables coaxial shaft E (green in Fig. R20) to rotate with the continuously turning shaft D. As shaft E rotates, Cam C changes the state of cam-driven contact S2. This causes the electro-magnet to be de-energised. I.e., the electro.magnet is only energised by a short pulse. Catch B is re-armed as soon as the solenoid energization pulse subsides. However, this catch cannot disengage Slip Clutch B until the end of the drive sequence.
  • The two solenoids of the electro-magnet are marked "500" in the original circuit diagrams (Fig. R24, R25). In the machine of Peter Prest, these components are actually marked "580 Ω" and "ST.39362 A". The buzzer relay is a British Post Office type, marked "4671 ACF W43.1", and its metal cover is embossed "AN".
  • Coaxial shaft E drives shaft F ( = printer wheel) via gears G and H. It also drives shaft G ( = paper tape transport) via worm gear I and worm wheel J (see Fig. R20, R21). During the drive sequence, the printer wheel makes one revolution, whereas the paper tape transportation drum is rotated over the width of a printed character.
  • During the remainder of the drive sequence, the wiper contact of the rotary distributor touches contact studs nr. 14-56 followed by nr. 1-3. Based on the pressed key, specific contact studs are connected to keyboard "common", via the protrusions of the associated keybar. Each time the distributor wiper passes over such a "pixel" contact, the scanning "pecker" of the printer wheel creates a blue pixel on the paper tape.
  • Near the end of the drive sequence, the Cam B follower of Lever L1 is lifted again by Cam B (see Fig. R22). This causes the bottom of the L1 lever arm to move towards the back of the machine. This allows the keybar spring to push Bar 2 also backward, and release the pressed key by letting Bar 1 pivot back to its default position. At the same time, the rounded edge of the Bar 2 Follower rides over the bottom of the Catch A Arm (momentarily pushing the back end of Bar 2 downward), before reaching its initial position again.
  • Catch A was already re-armed, shortly after starting the drive sequence. It will disengage Slip Clutch A from Shaft A when the Catch grabs the notch of that Slip Clutch again, at the end of the latter’s revolution. Likewise, Catch B disengages Slip Clutch B from Shaft D.
  • Both of the cam-driven switch contacts (S1 and S2) are also back in their starting state.
  • This terminates the drive sequence, and the machine idles until the next key is pressed.

The video/audio clip below goes throuhg the entire character set:

Telewriter character set - rotary distributor pulses, converted to tone pulses

(audio recording: ©2018 Peter Prest; used with permission)  ---  if player controls are not visible: move mouse cursor over image


The sequence of the recorded 38 characters follows the Telewriter keyboard rows from left to right: Q through P, A through /, + through •, 2 through 9. Note that the Telewriter uses bursts DC pulses, not tone pulses. For illustration purposes, the DC pulse sequences were converted to tone pulses with a tone frequency of 1500 Hz.

Below are three video clips that show the motion of the distributor wiper, the printing wheel and paper transport drum below it, upon pressing one of the keyboard keys. To be able to show the entire drive sequence in slow motion, the original DC motor was replaced with a (whining) stepper motor.

Front of the Telewriter - covers removed, slow-motion single character sequence

(©2018 Peter Prest; used with permission)  ---  if player controls are not visible: move mouse cursor over image


Close-up of the Telewriter distributor - single slow-motion sequence

(©2018 P. Prest)  ---  if player controls are not visible: move mouse cursor over image


Angle view of the Telewriter - covers removed, slow-motion single character sequence

(©2018 Peter Prest; used with permission)  ---  if player controls are not visible: move mouse cursor over image


Top view of the Telewriter drive system - covers removed, slow-motion single character sequence

(©2018 Peter Prest; used with permission)  ---  if player controls are not visible: move mouse cursor over image



When two Telewriter machines are connected, the following overall system circuit diagram applies:

TLWRT-GTL

Figure R24: Telewriter circuit diagram - two machines interconnected

(note: switch contacts MS1-MS3 shown for Master Switch MS in "ON" position; MS4 is in the motor circuit - not shown)

The above diagram is based on the official complete circuit diagrams shown in Fig. R25 and R26 below. To make it easier to follow and understand the power and signal flows, the motor circuitry is not shown, and the layout has been rearranged. The diagram shows two Telewriter machines in the "line operation" configuration: in both machines, the terminals R1 and R2 have been strapped, and the two machines are interconnected via their L1 and L2/E terminals.

Let's assume that one machine is powered up and idling, whereas the second machine is powered off. Each machine has a "call buzzer". In both machines, the buzzer is normally "off". In the idling machine, this is because switch contact MS2 is open (Master Switch is ON), as is relay contact a1. The associated buzzer relay cannot be energized, because switch contact MSN1 is open (Master Switch is ON). In the second machine, switch contact MS2 is closed (Master Switch is OFF), and relay contact a1 is open. Each time a key is pressed on the idling machine, it will go through a drive sequence as detailed above. Each time its rotary distributor makes a revolution, the start-pulse contact studs and the pixel contact studs of the pressed character (none for SPACE) cause the minus terminal of the 12 volt battery to be connected to the L1 terminal of the operating machine. Via the line wire, this terminal is connected to the L1 terminal of the "sleeping" machine, and the lower side of its buzzer relay. The plus 12 volt terminal of the operating machine is connected to upper side of the buzzer relay in this "sleeping" machine (via the L2/E line, the R1-R2 strap, and the closed MS1 contact). So, each pulse of the rotary distributor in the operating machine energizes the buzzer relay in the sleeping machine and briefly activates the buzzer. The buzzer pulses will prompt the operator of the "sleeping" machine to turn that machine on, which disables the buzzer (MS2 is open). At that point, the system configuration is basically that of two idling machines. Standard operating procedure after turning the machine on, is to advance the moistened tape from the water container wick to the printer wheel, by pressing any key a sufficient number of times (ref. 43B). At the same time, this would signal "ready to receive message" to the operator of the calling machine, as we will see next.

When a key is pressed on either of the two idling machines, that sending machine will go through a complete drive sequence:

  • The start pulse in the sending machine will cause the solenoids of the local electro-magnet to be briefly energized.
  • The solenoids in the receiving machine are energized simultaneously. They are effectively connected in parallel with those in the sending machine: via the L1 and L2/E line connections, the R1-R2 strap, the (still closed) cam-driven S1 contact, switch contact MS3, and the cam-driven S2 contact. The energization pulse causes Catch B to release Slip Clutch B. The printer drive mechanism in this machine (Slip Clutch B, all downstream gears and shafts, the printer wheel and the paper tape transport) will go through the standard drive sequence.
  • In the receiving machine, no key is pressed. Hence, its Slip Clutch A is not released and the rotary distributor remains at rest! Only the printer mechanism is activated!
  • The motor in the receiving machine is powered by the local 12 volt battery, but the current for its electro-chemical printer comes from the battery of the sending machine, via its distributor.
  • Upon completion of the drive sequence, both machines are in the idling state again.

TLWRT-GTL

Figure R25: Telewriter circuit diagram - aluminum placard on the machine

(source unedited original photo: ©2018 Peter Prest; used with permission)


The circuit diagram shown in the 1945 instruction manual is slightly different (and improved):

TLWRT-GTL

Figure R26: Telewriter circuit diagram - from the Work Instructions (manual)

(source: Fig. 5 in ref. 43B)

In Fig. R26, compared to Fig. R25, the "pecker wheel" is called "printing wheel", the 56 studs of the rotary distributor have been renumbered, all switch and relay contacts are labeled ("MS" refers to the 4DPDT On/Off Master Switch, whereas S1 and S2 are cam-driven), correct symbology is used for those contacts, the buzzer is connected between the battery and the motor's RF choke inductors, and switch MS1 and buzzer relay contact A1 are swapped (Figure R26 reflects the actual wiring).

The 30 ohm resistor in series with the printer wheel provides current limiting, in case the printer mechanism is inadvertently operated without paper tape while the machine is configured for line operation. I.e., R1 and R2 are strapped, there is no line resistance, and only the 250 mA fuse and the series resistor provide protection.

To the right of the printer/distributor drive unit, there is an interconnection panel with the main switch:

TLWRT-GTL

Figure R27: Panel with main switch, fuse, and terminals for phone line & batteries

(source unedited original photo: ©2018 Peter Prest; used with permission)


The panel has two phone line terminals that are labeled "L1" and "L2 or E". Communication between machines was via a 2-wire phone line, or a single-wire phone line with return via ground/earth ("E"). This could be a point-to-point or an "omnibus" circuit (a.k.a. "party line", where the line is shared by multiple parallel terminals). As signaling was done with DC pulses, the phone lines could not include transformers. Up to four additional 1.5 volt cells (of an 8-cell battery) could be connected to the machine, to boost the line voltage so as to obtain sufficient signal current in the remote receiving machine(s) when operating with a line resistance above 1000 ohms (ref. 43B). When no additional cells are used, the terminals marked "EXTRA CELLS" are bridged. A 250 mA fuse is in series with the additional cells. The fuse holder is made by Belling & Lee and has a "10H/9613" military stores part number. The terminals marked "R1" and "R2" are to be bridged for line operation. See the schematics (Figs. R25 & R26 above) and the operating manual (ref. 43B). Note that the "+" of the 12 volt battery is connected to L2/"earth", not to L1!


TLWRT-GTL

Figure R28: Top of the "incoming call" buzzer, made by Sun Electrical Co. Ltd., and 1918 SunCo advertising


The water container measures 3.5x2.75x1.09 inch (WxDxH, ≈ 8.9x7x2.8 cm). It is made of six separate pieces of yellowish celluloid-like translucent plastic. The material is about 0.085 inch thick (≈ 2.2 mm). The top and bottom pieces are L-shaped, see photos below. The side is a single strip that was softened and formed into the outline of an "L", and is sandwiched between the top and bottom pieces. The ends of the strip overlap by about half an inch (≈ 1+ cm), and are glued together. There are three strengthening pieces installed on the outside of the container: a round piece underneath the circular base of the wick holder, a rectangular one underneath the spring-loaded container retaining clip, and a small round piece for the cap chain retainer. The first piece and metal base are fixed from below with four brass screws that are cut off and filed flat on the outside. The retaining clip is fixed with two brass screws and nuts. The screw cap and wick holder are made of nickel plated brass. The flat wick is about 0.7 inch wide (≈ 18 mm). The wick holder base holds a round plate with two projections that prevent the plate from turning. The plate holds a retaining clip with ends that are folded upward. The wick is passed upward on one side of the clip, across it, and down again on the opposite side of the clip, back into the container. The wick is long enough such that both ends lie on the bottom of the container. The plate can be pulled out, to refill the container with water, or to replace (or adjust) the wick. The wick is actually a long rectangular strip of coarse cotton-like cloth, folded over lengthwise, with the open edges sewn to form a long, closed flat envelope. The envelope is filled with a long strip of material, possibly felt.

TLWRT-GTL

Figure R29: Top and bottom of the water container with screw cap - note the wick holder and the coarse cloth wick

(source unedited original photo: ©2018 Peter Prest; used with permission)


TLWRT-GTL

Figure R30: The water container, with the wick holder cap screwed on

(source unedited original photo: ©2018 Peter Prest; used with permission)

In Figure R21 above, note that the just-moistened paper has to travel a certain distance to the printer wheel. Probably the width of a dozen or so printed characters. This is why the Working Instructions (ref. R43B, §6) state that before actually using the machine, the operator must "depress any key several times until the damp part of the paper reaches the peckers."


TLWRT-GTL
TLWRT-GTL

Figure R31: Setting up instructions - placards on the machine's paper stand (see Figure R3)

(source unedited original photo: ©2018 Peter Prest; used with permission)



BRITISH "DAILY EXPRESS" PRINTER

During World War 2, the British "Daily Express" newspaper used a British-made Hellschreiber Printing Unit to monitor the Hellcasts from the German news agency DNB. Ref. 42. The manufacturer of the machine and the associated amplifier is unknown.

If you have any information about this printer, please contact me!

Daily Express

Figure S1: The Daily Express "Hellschreiber Printing Unit" and associated amplifier

(source: ref. 42)


Daily Express

Figure S2: Close-up of the printer - lid removed from ink roller and from paper tape holder

(source: ref. 42)


Daily Express

Figure S3: Close-up of the printing mechanism and tape

(source: ref. 42)


REFERENCES


External links last checked: January 2016, unless stated otherwise.


Note 1: due to copyright reasons, this file is in a password-protected directory. Contact me if you need access for research or personal study purposes

red-blue line