Piet Hooijmans 1918-2006

One generation earlier in electronics

Although we might think that technology developments go fast these days, with the emergence over the last 30 years of personal computers, mobile phones and the internet, we easily forget that developments were at least as spectacular one generation before. During his life my father Petrus Johannes (Piet) Hooijmans witnessed the emergence of radio broadcasting, and then became a telegraph operator in the intercontinental radio telegraphy. The post-war years of expansion finally gave him the opportunity to become an engineer, working for twenty years on a number of state-of-the-art electronic subjects, often with/under some high-profile experts. To name a few:

From 1947 to 49 he worked for Dr Bert Dammers, then group leader in the Philips Radiobuizenlab, on his way to become one of the leading Philips electronic engineers.
With Dammers he worked on the brand new, unique enneode EQ40, and was allowed to use it in the very first 567-line Philips experimental TV receiver he built himself.
At the PTT Central Lab he worked under van Duuren, inventer of the famous Telex-over-Radio (TOR) high speed radio telegraphy system, and later the first director of the Dr Neher Lab in Leidschendam.
Piet designed an experimental X-band 9GHz telephone radio link.
He then switched to the newest domain in electronics, digital logic, working under Willem van der Poel, the inverter of the PTERA and very successful ZEBRA computers. Piet designed digital equipment connected to the ZEBRA, designing discrete logic cards that were used for many years to come.
He then moved to other very new digital domains, like optical reading of machine printed text.
And he was the first to implement a parity checker for decimal giro account numbers, obtaining a patent in the process.
He closed his technical career in yet another new domain, electronic telephone exchanges, where he was omne of the interface experts between Ericsson and the PTT.

Although none of the subjects Piet worked on is unique or breaking boundaries, they were all on the edge of new technology and applications, and it is remarkable to see how every few years he became involved in yet another new sub-domain of electronics. All in all a career he could be proud of. I’ll therefore use his career as a trigger to describe some remarkable, interesting, and often forgotten engineering developments between roughly 1945 and 1970.

Telegraph Operator

Piet was born November 7, 1918 (“four days before the end of World War 1 and on the first anniversary of the Russian Revolution” as he used to say) in Den Haag, in a modest family. Although he was a bright student there was no money to do more than the MULO basic secondary education, and at the age of 15 he had to leave school, right in the middle of the thirties economic crisis. After some simple administrative and badly paid occupations, in 1936 he finally applied for the job of radio telegraph operator (“telegrafist”) with the Dutch PTT. Out of more than 1500 applications he and some 25 others were selected, who went through an intensive Morse-code training and were then stationed in the Amsterdam main telegraph office.

Around this time there were two very different forms of telegraph communication: the wired “Rijkstelegraaf” (the oldest form) that was mainly used for homeland distribution, and the radio telegraph, known as “Hollandradio PTT”. Piet was active on the latter long distance international radio telegraph connections, such as New York, Tokyo, Moscow, Buenos Aires and of course the Netherlands East Indies (Bandoeng) and West Indies (Suriname, Curacao). The transmitters were located in Kootwijk in central Netherlands, housing amongst others a 400kW long wave transmitter that was used for the first long distance radio link with the Dutch Netherlands East Indies. Later more medium wave transmitters at lower transmit power were added. The receivers were in Noordwijk (NORA) near the coast. Making a connection was still done by hand, but once this was established the actual message transmission was done through punched-tape fed machines, mostly the Siemens & Halske, that operated at much higher speeds than a human operator could achieve. On the receiving side the operators would need to listen out on the transmitting station until they heard the Amsterdam call sign, and then switch on the receive equipment. Piet did this for 10 years, and Morse code thus became his second “language” that he knew by heart and would never forget any more. Later we, his children, were always deeply impressed when he could on the spot translate any sentence in the Morse code which he would “sing” in the classical tuut-tuut-die-tuut as one hears it from a receiver.

During the war international traffic was substantially reduced, and essentially limited to hearing out of allied communications. This was, however, of little value since all of it was encrypted and/or purposely deceptive. There was also a lot of sabotaging by not writing down what was picked up from the ether. After “Dolle Dinsdag” (“Crazy Tuesday” the day in September 1944 when there was the rumour of Allied landings in the northern part of the Netherlands, and all German occupying forces fled temporarily) the public services collapsed, and Piet spent the rest of the war under cover with his family in Den Haag in order to avoid deportation and forced labour in Germany. After the Liberation in May 1945 he went back to report for service in Amsterdam, but nothing worked, all equipment was either stolen or demolished. Instead he was sent by truck and motor to Eindhoven, which was already liberated since September 1944 after Operation Market Garden. Here the PTT, with the help of the US Army, had installed a temporary radio station for its telegraph service. Piet took up his radio telegraph work again. The service was located in an empty school building near the Zeelsterweg, leading from the Philips complex at Strijp to the then military airport of Eindhoven. The transmitter was in Acht, north of Eindhoven. Connections were mainly with New York, Paramaribo in Surinam, and Curacao in the Dutch Antilles. He worked there till August 1945, when he went back to Amsterdam where services had been re-opened, with new transmitters at Kootwijk and receivers at NORA.

Picture of the Amsterdam main telegraph office on the Spuistraat around 1929, right behind the Royal Palace on the Dam square. This neo-classical building was built in 1898 and housed both the Amsterdam main Post Office and telegraph service. The Telegraph department occupied the 3rd floor, while on the attic the batteries were located. The antenna on the roof was for long wavelength reception of signals transmitted from the station at Kootwijk.

The beautiful telegraph transmitter station at Kootwijk.

The radio telegraphy receiving station NORA at Noordwijk.

The Siemens & Halske Morse-code machine transmitter (left). After the connection was set up by hand using the Morse-code key on the lower right, the machine transmitted the taped message at a 5-6 times higher speed than a human telegraph operator could. On the right the equivalent Siemens & Halske receiver, which recorded the received Morse-coded message on tape.

The international radio telegraphy desk on the third floor of the Amsterdam Telegraph Office just after the war. This was where Piet was working. On the right a telegraph operator reading the received Morse-coded message from paper strip and translating it in regular text on a typewriter.

These 4 pictures from Telegraafkantoor Amsterdam 1898-1990, PTT Telecom, 1990.

Left, picture of Piet Hooijmans in 1943, in the midst of his radio telegraph period. On the right a picture of Piet (on the right in white shirt) in the temporary PTT radio telegraph station in Zeelst near Eindhoven, summer 1945.

Picture of the final recommended circuit diagram of a low noise, high output power microphone amplifier for consumer applications (Fig 5. from the report).

Through his persistent endeavours, Ronald Dekker was able to trace the descendants of the former manager of the Radiobuizen Laboratorium. From them he obtained this picture of “An FM radio”, which is almost certainly the one described.

Philips Eindhoven

Due to the low level of activities during the German occupation and to keep them busy, the PTT allowed its employees to follow technical trainings. Piet therefore started courses in Telegraphy and Telephony Technology, since courses in radio technology were prohibited by the occupying forces. After the Liberation, however, he picked this up again and in 1946 he obtained his diploma from the Nederlandsch Radio Genootschap. It was at this time that he applied for a job as electronics engineer with Philips in Eindhoven, the PTT having refused him an internal transfer.
After some difficulties to physically get to Eindhoven for an interview (many train bridges were still not repaired and on top of that the winter of 1946-47 was very harsh, with even an Elfstedentocht, (the eleven cities ice skating tour in Friesland, the 9th one in history, on February 8) he was accepted and started May 1st, 1947 in the Radiobuizenlaboratorium I (Radio Tubes Laboratory) in the “Witte Dame” (White Lady, so called because it is an impressive white building) on the Emmasingel in Eindhoven. He worked in the team of Mr Dammers on consumer applications, mainly radio and television. Right at this time the EQ40 nine electrode tube was developed and becoming available experimentally, and several applications of it were developed. Similarly Frequency Modulation (FM) was being developed for public radio, requiring commercial FM radio receivers. Piet worked, with and under Dammers, on both.
Two formal reports with Piet Hooijmans as co-author have survived:

Report 1: O-37S-VIII
Laboratory: Radiobuizen Lab I
Subject: Balansversterker met ECC40 als phase omkeerbuis; 2 x EL41 in balans als eindtrap. (Balanced amplifier with ECC40 as phase inversion tube; 2 x EL41 as balanced power stage)
Authors: Dr Dammers, P.D. v.d. Knaap, P.J.Hooymans
Date: August 20, 1947

In the introduction the authors give an interesting indication of the reasons for their investigation, illustrating the good link that Dammers had with the product development organizations in the Philips factories:
When building the aforementioned amplifiers we have put ourselves in the position that these are targeted as examples for set makers, knowing that they use next to the Philips radio tubes also components readily available on the market. There are some issues where there is an essential difference between what is normal for set makers and what we call a typical Philips design. This difference is reflected on several topics.

The report then analyses different circuit topologies in relation to output power and noise sensitivity, where the requirements on the expensive power supply concept are the important criterion. Losses in ripple suppression versus current requirements of the transformer are important considerations. All circuits are built around the new Rimlock series of Philips tubes: the ECC40 dual triode, EL41 audio power amplifier and EZ40 double power supply rectifier. In the final version an EF40 pre-amplifier tube is added.

After a list of seven considerations related to this theme, the report is rather straightforward, with several measurement curves (looking at the hand writing, all measured by Piet Hooijmans), four alternative circuit diagrams of the basic design (1a to 1d) and a final circuit including the microphone pre-amplifier (Fig 5). Settings are given for all tubes.

Working on the EQ40

Next Piet worked on FM radio reception, where the details can be found in the following report:

Report:                  Q-46S-II
Laboratory:          Radiobuizen Lab I
Subject:                Beschrijving van een A.M./F.M. ontvanger
                                      (Description of an AM/FM receiver)
Authors:                B.G. Dammers, P.J. Hooymans
Date:                      March 5, 1949

Although Frequency Modulation (FM) radio transmission was only in development and not yet operational (FM transmissions in The Netherlands started in 1954 but in Germany already in 1950), the struggle was to find economical solutions for combined AM and FM reception. Philips had a reference design for FM-only receivers, based on the new EQ4o enneode valve, which was specifically designed for this role. The existing long, medium and short wave radio transmissions were all using Amplitude Modulation (AM), which is fundamentally different from FM. So combining them in one radio without adding additional (expensive) valves was a challenge.

The investigations of Piet and his boss Bert Dammers lead to two interesting conclusions. The first is that, based on their measurements of the FM receiver, they propose a modification of the EQ40 design: a variable winding speed of the 3rd and 5th control grids. Since this was well before the official launch of the tube, and their statements about the necessity are very outspoken, I assume that this modification has indeed been applied in the final EQ40 valve design!

Secondly, they propose a "re-definition" of the valves in the radio set, depending upon the reception standard. E.g. the pentode part of an ECH42 that acts as mixer for AM, but as IF amplifier for FM. And the EQ40 that acts as FM detector and AM amplifier. Although they haven't solved all issue by the time the report was issued in March 1949, this trend was continued and over time we see this concept being implemented in Philips radios. For details see my story on the EQ40/EQ80.

I don't know if Piet has ever realised that his boss Dammers would grow to become one of the star engineers of the Philips Electron Tube division, highly respected by engineers and management throughout the electronic divisions on TV, Audio and later also the professional divisions. Dammers became the head of the Centrale Applicatielab Bouwstenen (CAB, Central application lab components) until his early death in 1969. For more stories on the CAB and Dammers see here, here, here, and here.

The same tube in two different housings: left the EQ40 in Rimlock pinning and the glass reference pin on the left side, right the Noval EQ80. The EQ40 enneode (nine electrodes) was the pinnacle of developments of the pentode, invented by Philips. In fact it were two serially stacked pentodes in one tube, with a screen in between. By applying two phase shifted signals to the two sections the anode current was proportional to the phase difference. The tube was thus marketed as the "phi-detector" and ideally suited for FM or PM detection.

The story of the EQ40

Because of the occupation the 50 year Philips jubilee in 1943 could not be celebrated, and in 1948 the 55 year anniversary of the company was thus used. All Eindhoven personnel defiled through the city, here we (probably) see people of the Radiobuizen Lab. Piet Hooijmans should then have been amongst them.

Picture of the total IF and LF sections including the EQ40 as FM detector for the combined FM/AM receiver.

These are the conclusions on these topics in the report:
These measurements were based on tubes with so-called constant winding speed (“spoed” in Dutch) of the control grids g3 and g5. Therefore the tubes have been changed based on variable winding speed control grids. This gives a lower slope of the characteristic, with a less annoying influence on the AM detection. This is explained in the release report of the EQ40 by Radio Lab 1.

The experiences gained with this FM/AM receiver with an FM band and an AM band, have demonstrated that the use of one single mixer tube for FM and AM leads to switching difficulties (complex switches, mutual influences on bands). The Americans have solved this in general by using two separate RF channels. We have now proposed a solution which offers, while keeping the number of tubes used unchanged, a much better repartitioning of the functions such that the FM/AM reception performance can be improved. Experiments in this direction will continue.

Piet's home-built TV receiver

Next to these official occupations in the Radio Lab, Piet Hooijmans was also one of the employees allowed to participate into something quite revolutionary at the time: constructing a private television receiver! For all the details about this scheme and the television see my page on the development of television. By early 1949 Piet had his TV, including the EQ40 FM detector, working. Since March 18, 1948 Philips had a commercial TV broadcasting license, and a transmitter that covered an area of 40 km around Eindhoven. Transmissions were very sporadic, officially three times per week for 1.5 hour, but Piet only remembered the Saturday afternoon. During one of the first weekends he had his TV working at home he had invited two of his sisters to Eindhoven, which was at the time still quite a journey. Because of this special occasion, two young ladies visiting from Den Haag, Piet’s landlord specifically ordered his wife to make tea for breakfast, instead of the traditional coffee. Together they watched the program in his room on the Woenselse Markt. Piet was allowed to take his TV with him when he left Philips, and it seems he upgraded it to include a tuner function. I still have this historical piece of equipment in my possession!

The home-built television receiver as built by Piet Hooijmans in 1948-49.

More on the development of television by philips 1946-1956 here

PTT DNL Radiolab

In September 1949 he left Philips. The main reason was not the job, in fact he enjoyed himself quite a bit in the Radiobuizen Lab, but after having lived for all his life in Den Haag and Amsterdam he could not get used to the still very small and rural (read "boring") Eindhoven. He applied again for a position as electronic engineer at the PTT Central Laboratory in Den Haag, and this time he was accepted. Initially this was on the Kortenaerkade, opposite the Royal Stables of the Noordeinde Palace, but in 1955 the lab had grown so big that it needed a new building, which was constructed in Leidschendam, just north of Den Haag. From then on it was called Dr. Neher Laboratorium (DNL) and Piet would stay there the rest of his working life. He moved to Voorburg and, after he married Josephine Tromp in 1957, this is where their three children were born. In 1965 the family moved to Leidschendam where they would live the next 40 years.

A maquette of the Dr. Neher Laboratorium (DNL) in Leidschendam as it was constructed. The large tower was for radio transmission experiments, the two wings east and west housed the technical departments. The workshops and administrative offices were in the centre.

One of the two main wings of the DNL, where the engineers had their working space. Offices were on the ground and first floor.

Initially Piet worked under Dr Ir H. van Duuren, the inventor of Telex over Radio (TOR), a very successful method for encoding binary telex traffic over fading radio channels that greatly enhanced reliability of long distance radio telegraphy links. After the Second World War this system was rapidly introduced on many connections, and one of the reasons why there was quickly much less work for Piet and other PTT radio telegraph operators. In the group of van Duuren (the Radio Laboratorium) Piet worked first on highly accurate frequency measurement devices, needed for the steadily growing number of communication radio links. A description of the device is given in publications in the magazine of the Nederlands Radiogenootschap July 1951, Proceedings of the IRE, July 1952, and in August in the French publication L'Onde Electrique. In both it is mentioned that Piet Hooijmans was responsible for the design of the receiver part. In this team he also found back his old friend John Coster, who had been his radio telegraphy colleague in Amsterdam for 10 years, and had made the same moves to become an electronic engineer.

In the meantime Piet had made the next step by obtaining his HBS-B diploma in 1953, after two years of evening study together with his friend and colleague John Coster. This in turn allowed them to be selected for the internal PTT management courses at Voorlinde, the PTT management training centre in Wassenaar. [Voorlinde is today a nice museum of modern art]. After this they steadily grew through the ranks. Piet was also allowed to follow for two years courses in electronics at the Technische Hogeschool Delft (now Technical University Delft), but he never finished it (probably because the children, including myself, were born).
In 1955, at the opening of the DNL in Leidschendam, van Duuren, Piet's boss, became the new director of the lab, and his role of head of the Radio Laboratorium was taken over by Jonkheer van der Wijck. After his return from Eindhoven Piet had been living with his parents again, but in 1957 he married Josephine Tromp, a secretary at Philips Telecommunicatie Industrie (PTI) at Hilversum. They moved into a rented floor in Voorburg, around the corner of the DNL, but in 1958 obtained an appartment in the same village.

Piet Hooijmans (left) and his friend and colleague John Coster studying intensely while at Voorlinde, 1955.

Towards the end of the decade Piet seems to have been more involved in external activities, outside the DNL. To start, he spent some time at the radio transmitter station at Kootwijk, together with colleague R. Visser, where they investigated the non-linear performance of a high power transmitter module. This was triggered by the emergence of band I TV broadcast, 90-110MHz FM radio broadcast and mobile car radio (mobilofoon), which all suffered from harmonics of radio-telegraphy transmissions in the 3-30MHz band. The RF strip they measured used the newest Philips QQE 06/40 dual power pentode and indeed showed considerable (4th order) harmonic distortion when g2 voltage was set too low.

Circuit diagram of the QQE 06/40 RF amplifier of Kootwijk Radio that was analysed by Piet Hooijmans.

October 1957 Piet, together with his colleague van der Scheer, went to Philips Telecommunication Industry (PTI) in Huizen, for the release measurements on the VHF SFR 329/03 transmit/receiver and the associated STR 113/21 carrier radio-telephony system. The system, operating with two transmitters at 158,3 and 172MHz was found to be according the specifications, and the report, written by Piet, recommended PTT management to approve its acceptance.

Ku-band radio link for telephony, 1958

For the next decade Piet stayed active in radio transmission, culminating in the design and development, together with his colleague Marius van Beveren, of a 3cm wavelength (9.4GHz) radio transmission link for 240 analogue telephony channels. An internal PTT report from December 1958 gives a full description of it. These systems were designed for the backbone radio link between Goes (in Zeeland) and Roosendaal (Noord Brabant). During the big flooding of Zeeland and parts of Holland in 1953 (de Watersnoodramp) many telephone cables and switching offices had failed, and it was decided that a wireless backup system was needed. Goes, in Zeeland, was the central hub and the link Goes-Roosendaal was the bridge to the main land. This was the system designed by Piet and his colleagues. It was a classical heterodyne receive and direct frequency modulation transmit chain, with Klystrons for transmitter and local oscillator signal generation. The entire IF and baseband electronics, as well as all power supply and control loops were designed with tubes, most of them from the professional series of Philips like the E88CC, E180F, E80L and EAA91. Only a handful of OA85 semiconductor diodes were used plus 1N23C detector diodes in the waveguides to extract the 70 MHz IF.

Piet Hooijmans in 1956.

Block diagram of the 3cm X-band radiocommunication link. G1 is the Klystron main local oscillator. [Onderzoek en ontwikkeling bij KPN]

The 3cm wavelength, 9.4GHz microwave radio link for telephony backbone transmission, designed by Piet Hooijmans in 1957.

But here the story does not end, because this system would have a quite interesting life! In the book “Honderd jaar telefoon 1881-1981” that was published by the Koninklijke PTT at the 100 year jubilee of the telephone service in the Netherlands, Piet himself tells about the further history as part of the chapter on R&D that he had authored. It is explained that the system was installed in both the Goes and Roosendaal television and communication towers and, being the first in the Netherlands, used extensively for propagation experiments by, amongst others, the young ir Leo Krul. The system was never used for operational telephony connections, though, for that it was still too revolutionary in the conservative telephony community of the PTT. However, it would have a second life at the Technical University of Delft, when Leo Krul moved there to become professor of the Microwave Department, where he took with him the system that was used for many more years. Twenty-five years later the then older prof Krul would be my master thesis professor at the TU Delft when I graduated in 1984 with a thesis on radar calibration. As a last footnote it can be mentioned that Krul, van Beveren and three other colleagues jointly received the Dutch Veder Price in 1961 for “their joint development work on the experimental 3cm radio telephone communication system on the Goes-Roosendaal link”, whereas Piet Hooijmans, despite being the co-author of the system report, is not mentioned. Whatever the reason, apparently it did not bother him too much, since I have never heard him complain about it. In retrospect this was the culmination of his activities in tube-based radio receiver design, and Piet decided it was the right moment for a next step.

Enter digital technology, 1960

After this Piet made an interesting move within the DNL, possibly triggered by the lack of recognition in the Radio Laboratorium. Around 1959 or 1960 he moved into a brand new area of electronics: digital (transistor-based) design. Since 1953 the Dr Neher Laboratory was actively developing the brand new machines called computers. These were designed by the brilliant engineer van der Poel. After the initial PTERA (PTT Elektronisch Reken Apparaat, or PTT Electronic Computation Device) which had quite some reliability problems, he designed the ZEBRA (Zeer Eenvoudige Binaire Reken Apparaat, or Very Simple Binary Computation Device). This was a very successful machine, and eventually series produced by the UK company Standard Telephone and Cables (STC or STANTEC). Some 60 were sold throughout the world to Research labs and Universities. The PTERA and first version of the ZEBRA were still entirely made with tubes, but the latter gradually migrated towards transistors.

Piet’s entry into this new digital world happened when he moved from the Radio Lab to the Mathematische Afdeling (MA, Mathematical Department) under van der Poel, who had in 1956 succeeded Kosten as head of the department. The MA had two roles within the DNL: operating the ZEBRA for advanced simulation and computation processes, and secondly to develop new technologies and equipment based on advanced mathematical methods. It seems that Piet was the principal electronic engineer in the MA to design digital equipment for these roles. In the first role the MA struggled with the limited storage capacity of the ZEBRA, so when an Ampex magnetic tape recorder came available Piet designed an interface between the ZEBRA and this tape recorder, with a magnetic ring core memory as buffer. It shows the state-of-the-art digital electronic design during the early 1960s. All digital circuits (AND and OR-gates, flip flops, shift registers etcetera.) were designed from discrete components: transistors used were Philips ASY27 and OC76, the RCA 2N586 plus a few Intermetal OC470, one of the first silicon junction NPN transistors. These were made into elementary modules based on printed circuit board (PCB) cards, of which tens were used.

The PTERA computer, with in the front its inventor Willem van der Poel and behind him the first head of the Mathematische Afdeling Leen Kosten. One can see the hundreds of tubes sticking out of the panels.

The ZEBRA computer as installed in the Dr Neher Laboratorium in Leidschendam.

Front view of the ZEBRA-Ampex recorder interface as developed by Piet Hooijmans. [PTT DNL Report Nr. 68 MA, 1964]

Bottom view of the opened ZEBRA-Ampex recorder interface. Left the recorder, probably still with a few valves. In the centre two blocks containing the digital circuit boards developed by Piet. One unit with three blocks of some 25 PCBs each was able to swing out, as shown, for ease-of-service. [PTT DNL Report Nr. 68 MA, 1964]

Machine reading of stylised numbers, 1963

In parallel, because most reports are issued within one year, Piet worked on three other topics, all related to data processing by the rapidly growing PTT organisation. The first project was an experimental set-up for fast reading of machine writing. This was a technology much pursued at this time, both in the US and Europe, driven by the need to handle the rapidly increasing amount of administrative data generated by the emerging computers and digital systems. For the Dutch PTT this was especially the reading of the punched cards that were used by the Postcheque en Girodienst (PCGD) and Rijks Postspaarbank (RPS), that were both under or linked to the PTT because they used the post office services. (After a number of re-namings and mergers, both ended up in today's ING Bank). For this the Mathematische Afdeling built a test set-up, to evaluate the efficiency to detect printed text.

The definition drawings of the stylised DNL-B machine font, used for the automatic reading machine. Note Piet's signature on the document. [PTT-DNL Report Nr. 187 MA]

The number font used for the tests was the DNL-B, defined earlier by van Duuren. However, based on the experiments it was step-wise optimised during the project, the template on the left was the final result after 7 iterations between July 1961 and June 1963.
For the experiments a string of numbers on a paper strip was mounted on a rotating drum, the reading spot illuminated by two high intensity lamps. Two font sizes were used, one with 0,1" spacing and 3,4mm high numbers, one with 4mm spacing and 5,3mm high numbers. A 10x magnifying lens fed the reflected light to an optical detector formed from a 3 x 15 array of photo diodes. The effective reading speed for the small and large fonts was 1600 and 1000 numbers/sec, respectively.

As for the ZEBRA-Ampex interface, Piet designed the digital control system to translate the detector matrix output into a stream of numbers equivalent to those on the paper strip. The same approach to the circuit design was used, based on standard-sized PCBs with discrete digital circuits. Based on the stylised DNL-B font the translation from detected light to numbers was fairly straightforward, based on AND combinations of the vertical line elements defined in the font drawing. Two functions extended the total processing: first the ability to support 33% tolerance in the vertical position of the numbers, and secondly a dual-number detector, to guarantee that one one number was detected at a time.
It seems the experiments were deemed successful, since as a next step the character reading function was integrated into a much bigger system.

The experimental contravertolker set-up, which contained no less than 13 bays of electronic boards designed by Piet Hooijmans. [PTT-DNL Report Nr. 197 MA, July 1964]

The contravertolker was a considerably bigger piece of equipment than the test set-up made previously, and the reading of the machine writing was only a quarter of the functionality. This is illustrated in the diagrams below. The first steps were still the same: a punched card was fed into the reading section, which decoded the 20/27-number typed code. This was stored into a memory, after which the punching cycle started. For this the card was step-wise (a punched card contained 51 columns) drawn under the puncher, driven by the emitter pulses, which also steered the reading from the memory of the associated number. The complexity was that the 27 original positions were not one-to-one identical to the punch card columns, so this required a decoder. During punching the card was also read back (8 columns later), and the result compared to the result from the memory.

Conceptual drawing (top) of the contravertolker and the block diagram (below). Note that the flow in the top drawing is right-to-left, and in the bottom drawing the inverse. The green blocks are from the experimental set-up built earlier by Piet. [ibid]

Reading (upper right) and punching set-ups (below) of the contravertolker prototype. [ibid]

In September 1963 Piet celebrated his 25 years with the PTT since 1936 minus the 2,5 years with Philips. Although I was only 4,5 years old I still remember, because we were picked up from home by a large taxi; we didn't have a car yet at that time. Piet and Josephine were welcomed in the large central entrance of the Dr Neher Lab by his boss Willem van der Poel, and there was a festive ceremony with speeches and presents. Most of the time I spent on my knees looking out of the window at the car park full of cars, which I found extremely fascinating.
In the meantime Piet had climbed the internal ladder of the PTT from a simple telegraph operator to a respected engineer at the PTT DNL laboratory, benefiting from the growth opportunities in the rapidly expanding post-war economy and the rapid developments in electronics.

The experimental test system for reading stylised numbers, with omn the right the rotating drum and in the black cone the optics and detector matrix. [PTT-DNL Report Nr. 187 MA, October 1963]

This was the Contravertolker (Return translator). A vertolker (translator) was a device to read punched cards back to normal text. The contravertolker was intended to translate typed machine text into punched card holes. This was a function intended for the post offices, where clients of the PCGD and RPS would offer their - still unpunched - transfer cards. The PTT/PCGD/RPS had recently introduced a registration machine, which typed the coded instruction along the top of the card. For a postal cheque of the PCGD these were 20 characters (type (1), post office nr (4) account nr (8) and amount in Hfl (7)), for an RPS savings deposit or withdrawal 27 (type 1, post office nr (4), savings account nr (8), amount (7) and updated savings (7)). Once these cards were received by PCGD or RPS central offices, the contravertolker would read the typed code (using the same DNL-B stylised writing as used in the previous experimental set-up) and translate that into instructions for a card punching machine. As a means of verification, the punched code was read back and compared to the stored code that was read from the card. In case they were not identical an alarm was given.

Examples of PCGD cheques (top) and two RPS transfer cards with their respective machine-written text along the top of the cards, 20 numbers for the PCGD, 27 for the RPS. [PTT-DNL Report Nr. 197 MA]

As always, all electronics were made using the proven PCB cards with discrete circuitry that Piet had used in all earlier systems. The result was a large cabinet full of electronics, see the picture, with 13 19" drawers of maximum 26 PCBs each. From top to bottom:
A         read amplifiers
B/C     number recognition
D         control
E-J      memory
K         punching magnet activation; final comparison check
L          emitter inverter stages
M        power supply
The number recognition and storage took roughly 0,25s, but the punching process was much slower, and the total cycle time 3 seconds. It is unknown to me whether the machine was actually produced and used in the PCGD/RPS centres. On the one hand it seems that a function like this was required as long as the punched card was the basis of the entire operation. At the same time at least the prototype was bulky (the entire 19" cabinet) but especially slow due to the punching machine. With a 3s cycle time the capacity was 1200 cards/hour, and many machines would be needed. I therefore suspect that, if the machine was ever used in practice, it must have been after an engineering cycle to down-size the electronics (ICs were emerging!) and possibly a faster punching tool.

September 1963, the 25-year PTT-jubilee of Piet Hooijmans. Here Piet and his wife Josephine are received by Mr Willem van der Poel, Piet's boss and the head of the Mathematische Afdeling.

De Nummeronderzoeker, the first parity checker, 1964

The last technology Piet worked on while in the Mathematische Afdeling was another new development: automatic error detection in decimal numbers, or, what we call today, parity checking. This was a typical project for the MA, because it was a technology that required a solid theoretical basis before any implementation. The parity system Piet describes in his report adds one parity number to a 7-number giro account number, such that the sum of the eight numbers is equal to zero modulo 10. Interestingly, and probably even at that time unknown to any reader, Piet used his own giro number (he was a very early and loyal PCGD customer) 660852 as example. (Because the sum of the numbers is 27 it requires a 3 to make it 30 or 0 module-10: 6608523. With the method applied the system should be able to detect

one wrong number
the interchanging of two numbers
wrongly interpreted combinations like thirteen-thirty

This combination of requirements could only be achieved with an odd numbering system, and thus not using the standard decimal system. A first step was thus that the decimal numbers were converted to the so-called biquinary system: any decimal number A was translated to a 2-digit number b-a, with a from 0 to 5, and b=0 for A between 1 and 5 and b=1 for A from 6 to 0. (Note that 0 is coded as decimal 10). The parity check now became a modulo-2 check on b and a modulo-5 check on a. The equipment Piet developed was the verifiier of the parity, not the generator of the parity number. The equipment was intended to be used e.g. PCGD or RPS offices for fast verification of giro account numbers including their parity number that were used by clients. Since the basic giro account number used 7 numbers, the giro number with parity was 8 numbers.

Example timing diagrams for the account number 06739167, with the last 7 the parity number. Numbers are received as trains of pulses (with zero giving 10 pulses as shown). Trigger T indicates the number rank being received, Q changes state if an even number is higher than 5, while the counter direction trigger is low when T and Q have the same sign. Below that the state of the 5-state bi-directional ring counter. [PTT DNL Report Nr. 194 MA]

Not to complicate matters too much, I'll summarize the operation using the example given in Piet's report. After a detailed theoretical analysis, and based on the biquinary approach, the verifier (nummeronderzoeker in Dutch) used to up-down counters: one for the binary digit b (this was counter Q) and one for the quinary last digit a (counter C). This is illustrated in the diagram for the example of account number 0673916, which becomes 06739167 with the added parity number. The principle of the parity checker is that at the end of a processing cycle both parity counters Q and C are equal to zero as shown in the example. In that case a green light would turn on.

The theoretical part of the parity checker was much more complex than the practical implementation. The practical logical relations between the state counters T (odd/even), D (direction up/down) and counter C were derived using Karnaugh-diagrams, which could be simplified through the use of don't cares. The resulting circuit was made from 16 of the usual standard circuit blocks and connected to an Anker registration machine as shown on the pictures.
The clock frequency of the system was 10kHz, which meant that one digit was analysed in 1ms and the total cycle took 8ms, which was fast enough to consider it a real-time response.

The Nummeronderzoeker or parity checker built by Piet Hooijmans, using the standard digital circuit cards. [PTT-DNL Report Nr. 194 MA]

The parity checker seems to have been the highlight of Piet's technical activities. March 1964 a patent application was filed on his name Inrichting voor het controleren van een getal van decimale cijfers", which took nine years to be granted in June 1973. The patent was also filed in Germany. The fact that the patent and the DNL report were both only authored by Piet, suggests he did the development on his own. Especially this Nummeronderzoeker was from a theoretical perspective quite a complex story, with a solid number-theoretical basis.

The Dr Neher Lab in Leidschendam in the early 1960s.

The parity checker system in the box left next to a standard Anker registration machine used in postal offices. [PTT-DNL Report Nr. 194 MA]

Piet's Dutch patent for the parity checker concept.

Digital telephone exchange AKE13, 1966

The Ericsson AKE13 telephone switch system Piet helped to introduce within the Dutch PTT. [Picture Ericsson].

By now it was 1965, and Piet made a third move within the Dr Neher Lab. After the Radio Laboratorium and the Mathematische Afdeling, he now moved to the Schakeltechnisch Lab (SL, Switching Laboratory). Whether this was related to the fact that his boss van der Poel was moving steadily towards the Technische Hogeschool Delft, to become a full time professor there in 1967, I don't know. In any case, the Switching Laboratory was the place where equipment for the telephone system was developed, specified , or analysed. Also in this domain a revolution was taking place, the introduction of digitized telephone exchanges. It is highly likely that Piet, given his now 5 year experience with building digital equipment (which was definitely not a standard technology then as it is today) was asked to transfer to the SL to support this introduction. The system involved was the first electronic telephone exchange, the Swedish Ericsson AKE13. This was a new generation of switching systems, based on digitally programmed multi-processor cores. The Dutch PTT was the first customer of this new switch family, ordering it in 1964 for its Rotterdam regional exchange (districtscentrale).

As it turned out, the introduction of these so-called half-electronic switches was a far from obvious affair, neither for Ericsson nor for the PTT. Apparently the PTT wanted to make a big step in new technology, possibly also to support the rapidly growing harbour of Rotterdam, but in 1964 they ordered the AKE13 "off the drawing board", since Ericsson was not able to show any prototypes or demos. In fact, that company had its hands full with the development of the AKE12, which was a smaller subscriber switch with a single processor. The AKE13, ordered for Rotterdam, was a regional exchange with up to 24.000 in- and outgoing lines, requiring multiple processors to handle the traffic load. Although the AKE12 and 13 were both still in early development, Ericsson nevertheless gave a course on its new electronic switching principles from July 1965 to May 1966 , with Piet Hooijmans participating. In November 1966 Piet had compiled a summary report on the new concept, although in the introduction he observed that much data was still missing, for example on the co-operation between the processors. There were also no programs available yet.
In fact, things were not going that smooth. In 1966 Ericsson installed a first AKE12 switch in Tumba, south of Stockholm for first trials, which turned out to be disastrous. The amount of software errors had been completely under-estimated, but there were no rigorous software tests available then. So the only solution was to operate the switch "life" and accept the drawbacks of bad performance, in order to find as many errors as possible. Ericsson admits that the whole AKE13 roll-out became a very expensive exercise.
It is therefore no surprise that the AKE13 development, with not just one but four processor cores, became an even more difficult process. Although the PTT had targeted/requested it to be installed in the Rotterdam DistrictsCentrale DC II in the Waalhaven by 1967, the project slipped and slipped. It is not clear how much time and effort this involved from the DNL and Piet, but I can imagine they had regular progress meetings with Ericsson, who would explain the reasons for further delays. In the meantime, the internal technology used rapidly became outdated itself, still 100% based on the same technology Piet had been using in the DNL for the last 8 years: discrete logic on large PCB cards. As far as known at least the first AKE13 did not use any of the modern ICs that were rapidly, in large volumes and low prices coming to the market.

Top level picture of the so-called half-electronic AKE switches. The telephone signalling and traffic layer is at the top, which was still based on an electro-mechanical relay-based switching fabric. At the bottom the processor unit, 1 for the AKE12, four for the AKE13. In between the transfer unit, interfacing to the signalling and relay activation electronics. [Electronica en communicatie, 1968]

The complexity of the multi-processor AKE13 is illustrated below, with left (theoretically) m processors, and to the right the complex switching fabric to assign processor capacity on a dynamic basis and connect them to Data Storage resources at the bottom. [ibid]

The AKE13 finally operational in the PTT DC II in Rotterdam-Waalhaven, 1971. [Ericsson]

The PTT Districtscentrale II in Rotterdam-Waalhaven

After much delay the AKE13 was finally installed and declared operational in 1971, seven years after being ordered. By then the frustrations about the delays were swept aside because the PTT could still claim this was the first electronic multi-processor telephone exchange declared operational in the world. Again Piet had been involved in a new, exciting and innovative technology development. it would be his last, though, since around 1968 he was asked to come to the general management of the Dr Neher Laboratorium. He therefore did not witness the completion of this project, and I doubt he was invited to the official opening.

Head of Services,1968-1980

In 1968 he moved into general management, becoming the head of all services (“Algemene Zaken”) within the Dr Neher Laboratory under his former boss van Duuren, who was already some time the head of the DNL. AZ effectively had all non-technical services within the DNL: the personnel department, publicity, the still sizeable work shops, the library, etcetera. Although he was no longer involved in technical work, all technical reports of the lab had to be co-released by him, and he carefully screened them all for consistency and correct use of the Dutch language. Piet had a - to modern standards - huge office in the management floor above the main entrance of the lab. To come there one had to walk through a dark corridor with concrete pillars and wood on the floor. As a kid I would go to meet him at the end of the day, being told by the guardian at the gate that I was allowed to go up to his office. I would find him there sitting at his desk as on the picture, and I was allowed to read through the large pile of technical reports on his other table. After that we would make a walk through the lab, visiting some of his friends and colleagues, showing me equipment in the labs. No wonder I choose to study electronics too! He would play this role until his retirement in 1980, after a career of 42 years with the PTT and nine years before it would become KPN. The Neher-laboratorium where he worked 25 years was closed by KPN in 2001, to the regret of Piet and many others, and today houses apartments.

Piet in his new role as Hoofd Algemene Zaken of the DNL, 1968.

Josephine and Piet Hooijmans entering the main hall of the DNL at Piet's 40-year jubilee with the PTT, September 1978.

The main entrance and management building of the Dr Neher Laboratorium. Management, including Piet Hooijmans, had its office on the first floor with the brick wall. Above it is the canteen. Piet's office was the second window on the opposite side, not visible on the picture.

One of the last contributions to the PTT history by Piet was the chapter Onderzoek (Research) that he wrote for the thick book, issued at the 100-year jubilee of public telephony in the Netherlands. It appeared the year after his retirement.

Piet Hooijmans receives the Royal Order of Oranje-Nassau from de mayor of Leidschendam. Josephine is proud. 19 December 1996.

Life after the PTT

He then decided to start into yet another new direction, and picked up Law studies at the University of Leiden, as a regular student amidst the other 1000 young students that start each year. His thesis, under the supervision of prof. Dr. Veldkamp, the former Minister of Social Affairs, is on the “Simplification of the Organization of the Social Security” in 1985. (Apparently no-one has followed his recommendations, because since then the Dutch social security system has not improved in terms of transparency and simplicity.) After his graduation he used his newly acquired knowledge to give for almost ten years free legal advice in Leidschendam, for which he eventually received a Royal Knighthood in 1996.

My father, Piet Hooijmans, died peacefully on June 9, 2006 at the age of 87.

Pieter Hooijmans, 2016-2022

References

To my regret I started diving into the technical history of my father after he had passed away. Of course we had talked about some elements of his work, and when I was young I regularly visited him at the end of his day in the DNL, but we never discussed technical things in the detail I would have liked, in hindsight. Fortunately, however, Piet was a very structured man, who kept everything he deemed interesting. And so I have a pile of reports he had written, which are the basis of the story told above. Similarly I have a pile of books and leaflets related to the PTT and Philips. All this forms the basis of this story, supplemented by my own memories and data found on the internet.
Reports

Philips Radiobuizen Lab I Report O-37S-VIII, Balansversterker met ECC40 als phase omkeerbuis; 2 x EL41 in balans als eindtrap. (Balanced amplifier with ECC40 as phase inversion tube; 2 x EL41 as balanced power stage), Dr Dammers, P.D. v.d. Knaap, P.J.Hooymans, August 20, 1947
Philips Radiobuizen Lab I Report Q-46S-II, Beschrijving van een A.M./F.M. ontvanger (Description of an AM/FM receiver), B.G. Dammers, P.J. Hooymans, March 5, 1949.
Staatsbedrijf der Posterijen, Telegrafie en Telefonie, Dr Neher Laboratorium, Praktijkverslag, De harmonischenproduktie van een zendversterker, (Harmincs generation of a transmitter amplifier), R.J. Visser, P.J. Hooijmans, 1957.
Staatsbedrijf der Posterijen, Telegrafie en Telefonie, Dr Neher Laboratorium, Keuringsverslag, Keuring van VHF zend/ontvanginstallatie SFR 329/03 en draagggolftelefonieinstallaties STR 113/21 van N.V. Philips Telecommunicatie Industrie te Huizen op 10, 11 en 14 oktober 1957, P.J. Hooijmans, November 1957.
Staatsbedrijf der Posterijen, Telegrafie en Telefonie, Dr Neher Laboratorium, Beschrijving Nr. 44 RL, Straalverbindingsinstallatie voor de transmissie van 240 telefoonkanalen, werkend op een golflengte van 3,2 cm (Radio relay system for the transmission of 240 telephne channels operating at a wavelength of 3,2cm), M. van Beveren, P.J. Hooijmans, December 1958.
Two books, one containing all technical drawings.
Staatsbedrijf der Posterijen, Telegrafie en Telefonie, Dr Neher Laboratorium, Beschrijving Nr. 68 MA, De koppeling van een magnetische bandrecorder aan de Zebra (The connection of a magnetic tape recorder to the Zebra), P.J. Hooijmans, February 1964.
Two books, one containing all technical drawings.
Staatsbedrijf der Posterijen, Telegrafie en Telefonie, Dr Neher Laboratorium, Beschrijving Nr. 187 MA, Leesinrichting voor gestileerde cijfers (Reading system for stylised numbers), P.J. Hooijmans, October 1963.
Staatsbedrijf der Posterijen, Telegrafie en Telefonie, Dr Neher Laboratorium, Beschrijving Nr. 194 MA, De nummeronderzoeker. Een inrichting voor het controleren van beschermde postrekeningnummers (The number controller, device for controlling of protected postal account numbers), P.J. Hooijmans, April 1964.
Staatsbedrijf der Posterijen, Telegrafie en Telefonie, Dr Neher Laboratorium, Beschrijving Nr. 197 MA, De contravertolker, P.J. Hooijmans, July 1964.
Staatsbedrijf der Posterijen, Telegrafie en Telefonie, Dr Neher Laboratorium, Mededeling Nr. 163 SL, De elektronica in de automatische telefonie. Samenvatting van een lezing die op 20 januari 1965 door P.J. Hooijmans werd gehouden tijdens een studiebijeenkomst van hoger technisch personeel (in militaire dienst) (Electronics in digital telephony), P.J. Hooijmans, February 1965.
Staatsbedrijf der Posterijen, Telegrafie en Telefonie, Dr Neher Laboratorium, Beschrijving Nr. 93 SL, AKE, een half-elektronische telefoon-stelsel van L.M. Ericsson - Deel I: Algemen aspecten, P.J. Hooijmans, November 1966.
Staatsbedrijf der Posterijen, Telegrafie en Telefonie, Dr Neher Laboratorium, Beschrijving Nr. 93 SL, AKE een half-elektronische telefoon-stelsel van L.M. Ericsson - Deel II: AKE12, P.J. Hooijmans, September 1967.

Articles

Nauwkeurige direct aanwijzende frequentie meetinrichting van 30Hz - 30MHz, L.R.M. Vos de Wael, Tijdschrift van het Nederlands Radiogenootschap, Vol. XVI, Nr. 4, July 1951.
Direct-Reading Frequency Measuring Equipment for the Range of 30 CPS to 30 MC, L.R.M. Vos de Wael, Proceedings of the I.R.E., Volume 40, Number 7, July 1952, pp.807-813.
Dispositif permettant d'effectuer avec une grande précision des mesures de fréquence dans la gamme 30 c/s .. 30 MC/s et indiquant directement les résultats, L.R.M. Vos de Wael, L'Onde Électronique, Vol. XXXII, Nr. 305/6, pp.351-356, Aug. 1952.

Books

Telegraafkantoor Amsterdam 1898-1990, PTT Telecom, 1990.
Over en uit, Driekwart eeuw radiocommunicatie 1900-1975, B. van Hoogland, J. Tours, PTT Telecom, 1990.
D. van de Nieuwe Giessen, Onderzoek en ontwikkeling bij KPN, Een geschiedenis van de eerste honderd jaar, KPN Research, 1996.
Honderd jaar telefoon, De geschiedenis van de openbare telefonie in Nederland 1881-1981, onder redactie van J.H. Schuilenga, J.D. Tours, J.G. Visser. J. Bruggeman, Den Haag, 1981.
Chapter 7 Onderzoek written by P.J. Hooijmans.

Update history

First published January 2017
October 2022
The entire section of the period in the Dr Neher Lab has been re-written, based on my research for the page on the first digital computers. This forced me to dive into the technical reports of my father, which in turn resulted in a complete re-write with much more interesting details.