As will be evident, the evolution through the different designs was triggered partly by changing technology, partly by external influences, but also by the political and financial developments within the Benelux Cross Antenna Project itself.
Initially the team lead by Christiansen, Högbom and Seeger worked on a large telescope for a wavelength of approximately 75 cm. Once there was an organisational frame-work and some money could be spent on staff and on consulting engineers, a first (Christiansen and Högbom 1961) design was discussed in the BCAP council. The design was an advanced version of Mills' cross antenna (Mills et al. 1958).
The choice of wavelength was relatively arbitrary, probably based on the successful observations of others at 75 cm. Since the aim was to obtain an instantaneous beam of 1 minute of arc or better, the overall size of the cross was planned at 5 by 5 km. As an illustration of its size, Figure 1 shows the cross antenna projected on the map of the city of Leiden. This was one of the designs presented in December 1961 at the OECD meeting of experts in the field, mentioned above.
The elements constituting the East-West and the North-South arms were cylindrical paraboloids, 30 metres in size, the surface of which was made of stretched steel wires. The paraboloids were to be tilted along East-West axes only. Figure 2 shows a cross-section in two extreme operational positions. The instrument would, at any one time, only be capable of observing inside a narrow strip along the meridian, but the entire Northern sky would be accessible.
The conceptual mechanical lesign had been drafted in consultation with Hooghoudt, who had also been involved in the construction of the Dwingeloo telescope. In one of his notes Hooghoudt mentions 25 million Dutch guilders (NLG) as a first estimate of the cost of this Benelux cross for 75-cm wavelength.
In the course of 1961 an alternative design, operating at 21-cm wavelength, was drafted and a first version of it was presented at the OECD meeting in Paris. It would consist of a large number (100 or more) of parabolic dishes. One of the reasons for considering this was the disappointingly high cost of the first design, once it had been properly calculated by a firm of consulting engineers. In a comparison of the 21-cm and the 75-cm cross antennas Högbom (1961) evaluates the technical and some of the scientific merits of the two systems.
Four attractive features are mentioned as advantages of the 21-cm array over the 75-cm one:
- The guaranteed international protection of a wide band around 1420 MHz (21 cm) against the uncertain future of the 408 MHz band in this respect.
Spectral versatility of the 21-cm array in the sense that changing to another wavelength could well be considered, while it would be nearly impossible for the 75-cm design; thus it would be possible to get information about the spectra of radio sources.
- The 21-cm design would allow measurement of polarisation while the cylindrical 75-cm design would be limited to one polarisation only.
- At 21 cm there would be the possibility of observing the Hydrogen line radiation.
One of the disadvantages mentioned is the longer integration time required at 21 cm because most radio sources are considerably fainter at that wavelength than at longer ones. However, unlike the first design the second one would allow tracking sources for a considerable time so that this effect would be more than compensated, albeit at the expense of telescope time. A final decision between the 75-cm and 21-cm options was deferred by the BCAP council until a more detailed version of the 21-cm design would be available.
Thus, the second design of a cross antenna was worked out in more detail and a more mature version was described by Christiansen, Erickson and Högbom in 1963. An artist's impression of this design is presented in Figure 3. For the shorter wavelength the cylindrical paraboloids with stretched wires as surface were no longer appropriate. So the elements chosen were parabolic dishes of 30 m diameter with a wire-mesh surface. A larger dish of 70 m diameter was foreseen in order to provide the short interferometer spacings not available in the array. The parabolas would be mounted equatorially and be steerable to all Northern declinations. Their range in hour-angle would be limited to 4 hrs.
Since this cross antenna was designed for a shorter wavelength, its overall size could be smaller (1.5 km) than the 75-cm one, while its resolution was slightly better. On the other hand, there were going to be one-hundred of these parabolic elements, requiring at least 2500 interferometric combinations to be made.
In addition to the 30-metre dishes, one large telescope of 70 metre diameter was foreseen to provide the short spacings which could not be realised between the other elements. The entire array was to consist of two parts: an inner array of 750 m dimension, where the telescopes were packed together and an outer part, of 1.5 km with a much sparser population of telescopes. The resolution of the entire array was to be better than one arcminute, that of the inner array three arcminutes at 21 cm. A schematic diagram of this design is given in figure 4.
It was planned to complete the inner (3 arcminute) array first and to experiment and observe with it for a while before the outer telescopes were included. In the early nineteen-sixties the complexity of the electronics and of the electronic data processing of this radio telescope were seen as an enormous, perhaps even a mindboggling, challenge.
Although the final choice between the two designs of the Benelux Cross- antenna had been deferred, all work of staff and consultants during 1962 was concentrated on the 21-cm cross antenna. In September the BCAP council formally confirmed the choice on the basis of a draft proposal to the Belgian and Dutch governments.
In addition to scientific and technical papers this proposal contained a cost-estimate for the cross of 30 million guilders, and a planning which foresaw completion of the 3-arcminute cross by the end of 1965 and of the complete telescope in 1967.
The proposal also contained a report on possible locations for the Benelux Cross Antenna, all near the border between the Netherlands and Belgium. The preference for the border region was partly political, but it had practical elements as well: border regions tend to have less industrial activity and fewer roads than the more central areas of countries, making it easier to obtain natural protection against radio interference.
Some of the locations had also been candidates for the 25-metre telescope which was ultimately built in Dwingeloo in the North-East of the Netherlands. In October 1962 the proposal for the Benelux Cross Antenna was sent to the government representatives to the BCAP, who submitted it to the two Governments in February 1963.
While the staff of the BCAP worked on designing feeds and electronic components for the new cross antenna, some of them (Högbom, in particular) started thinking about modifications which would simplify the electronics considerably. Thoughts about the possibility of a one-dimensional East-West array of telescopes were triggered by Ryle's (1962) success with earth-rotation aperture synthesis and his well advanced plans for the Cambridge One-Mile telescope. See Brouw, this volume, for a description of the principle of aperture synthesis.
Clearly, it was politically unwise to switch to a completely new design of the Benelux Telescope at the time at which a detailed design had just been submitted to the Governments. For this reason, Jan Oort, as president of the BCAP council, carefully introduced these new ideas to the Council as `A possible re-arrangement of telescopes, which would lead to achieving three-arcminute resolution considerably faster, ... and giving the possibility of ultimately achieving an even higher resolution than would be possible with the current design'.
The `re-arranged' array of telescopes proposed by Högbom early in 1963 consisted of a 3100 metres long East-West array of 28 paraboloidal telescopes, which would yield a resolution of 17 arcseconds using the technique of earth-rotation aperture synthesis. When complemented with another array of 6 telescopes 1600 metres further East or West, a resolution of 10 arcseconds could even be achieved.
The proposed arrangement is depicted in figure 5. Further improvement of resolution could be achieved rather simply by adding a few more telescopes at a later date. The smaller number of telescopes (28 or 34 compared to 100 in the 1- arcminute cross) would decrease the number of correlations between telescopes drastically. This would open possibilities for a multi-channel receiver system for observing the 21-cm hydrogen line in the future. Also the smaller number of telescopes would make using this array at other wavelengths much more convenient.
The individual telescopes had to be steerable over 12 hours of hour-angle (as opposed to 4 hrs in the 100-dish cross-version). This would make their construction more expensive, but the smaller number of dishes required would probably compensate for that. Disadvantages of the new versus the `official' design were: lower speed (most observations would take 12 hours or more), and poorer sky-coverage (resolution below a declination of 15 would be bad in the North-South direction). Also, it would be more difficult to observe short-term variability of sources, but, at the time, this does not seem to have been an important issue. At least it hardly gets mentioned in the memoranda and documents.
It was clear that the building budget of this more modest array would certainly not exceed that of the cross antenna proposed to the governments. Hooghoudt produced details of the design of the paraboloids and a cost-estimate of 30 million guilders later in the year 1963.
For Jan Oort, an agonizing period of waiting for a reply started with the submission of the plans for the BCAP to the Dutch and Belgian governments. It did not take the Dutch government very long to give its approval, but in Belgium the situation turned out to be more complicated, on the government-level as well as among the Belgian astronomical community.
Another international astronomical issue was at stake: the European Southern Observatory (ESO). It appeared that simultaneous participation in ESO and in the BCAP would not really be possible in Belgium. As, on the other hand, the Dutch astronomers, in particular Oort and Blaauw who were both involved in the establishment of ESO, believed that Belgian participation in ESO should definitely not be imperiled, a compromise solution was sought and found: Belgium would not invest in the construction of the new radio telescope, but it would continue to contribute its share of the preparation cost of the BCAP and later of the operational cost.
This compromise was discussed between the prime-ministers of the two countries in February 1964. It was confirmed officially three or four months later. Fortunately, the decision of the Dutch government to contribute half of the cost (as quoted in the proposal) of the Cross Antenna Project, was not adversely influenced by the Belgian withdrawal of its share of the construction. The money earmarked for building the BCAP was still available. Of course, this was only half of what was originally budgeted, since the other half would have been contributed by Belgium.
Nevertheless things became very much simpler politically all of a sudden:
- There was money in the bank to build a large radio telescope, although the design would have to be adjusted to a smaller budget.
International political complications were of little influence on the progress of preparations.
- There was no longer any reason to locate the new telescope near the border between Belgium and the Netherlands, the realisation of which had turned out to be a tall order.
Instead, one could try to find a site not too far from the Dwingeloo telescope and its electronic laboratories.
At first two possible sites within a very short distance from the Dwingeloo telescope were considered. In discussions with the Dutch National Forestry commission, the owner of the area around the telescope, it became clear that a second, much larger, installation near Dwingeloo would be difficult to realise because a significant enlargement of the car-free area around the observatories would interfere too much with the touristy nature of the region.
However, the Forestry commission was very helpful in suggesting another site near the village of Westerbork, 25 km removed from Dwingeloo. The presence of what had been the `concentration camp Westerbork' during the second World War, in the late sixties still in use as housing for immigrants from the South Molukkan islands in Indonesia, accounted for the fact that tourism had not been promoted in this area.
Plans to house the remaining inhabitants of the camp in nearby villages and towns were in an advanced state, and the local authorities were pleased to receive some additional pressure from the planned construction of the telescope to vacate the camp completely. Within a few months, in the fall of 1964, permission was obtained from all authorities concerned to build the new radio telescope.
The Synthesis Radio Telescope, at that time still a Belgian/Dutch venture, received its full name `Westerbork Synthesis Radio Telescope' after the final choice of location and at the explicit request of the Council of the village of Westerbork. Ever since World War II the name of the village had been identified by the general public with the camp in which many thousands of Dutch Jews and Gypsies had been kept prisoner before their transport to concentration camps in Germany and Poland.
Understandably, the village was very eager to see its name connected with a new prestigious scientific facility. As we saw above, preferences tended to change in favour of a linear East-West array of dishes as the year 1963 progressed. Once it became clear, early in 1964, that the new telescope would have to be built for a smaller amount of money, it was evident that an adaptation of the design presented to the two governments (the 100-dish cross antenna) would not be possible without seriously compromising the scientific goals of the telescope.
The 28-dish East-West array, that had been considered in 1963, was too expensive as well but this design could be adapted to make it affordable. So design work started on a more modest version of the line-array proposed by Högbom. Since the total budget would be close to half of what was foreseen for the Benelux Cross, some sacrifices had to be made, obviously. With respect to the array of 3 km length which would fully synthesise a field in 12 hours time (the third design, discussed above), some of the resolving power was given up by shortening the array to 1.6 km and the observing speed was reduced: a full synthesis of a crowded field would require several 12-hour observations rather than just one.
The design which was put out to tender, consisted of ten fixed dishes of 25 metre diameter and one similar dish movable along a rail-track 300 metres in length. A plan to build and assemble the telescopes as a series was part of the call for bids. The lowest bidder, the ship-building company Wilton-Feyenoord, was prepared to build the array of telescopes for a price considerably lower than was budgeted.
The board of the Netherlands Foundation for Radio Astronomy (NFRA) did not lose time to decide to modify the contract in such a way that a twelfth telescope, a movable one, was included for the price foreseen in the budget. By this action the speed of the instrument was increased by a factor two. The considerably smaller number of telescopes compared to the original Benelux Cross, and the relatively small number of 20 independent interferometers - the result of the equidistant spacing of the elements of the fixed array - had a number of advantages.
The possibility to add a spectral-line mode at a later date has already been mentioned, but general simplicity and stability were of paramount importance, especially in the early days. Right from the start the array was designed with full polarisation capabilities. One of the reasons to have the dishes equatorially mounted was to facilitate polarisation observations: the sky would not rotate with respect to the telescopes. In the sixties and early seventies this was still an important consideration in view of the limited speed and capabilities of electronic computers.
As mentioned before, constructing the telescope and its electronics had become a purely Dutch affair early in 1964. The telescope was still planned to be operated as a joint Belgian-Dutch venture, however. This meant that ultimately an organisation would have to be founded similar to, but not identical with the Netherlands Foundation for Radio Astronomy (NFRA).
Part of the staff working on the new telescope was paid out of BCAP funds and was working in Leiden, another part was NFRA staff, working in Dwingeloo. Since the primary objective was to get the telescope built, nobody worried too much about the international complications in the operational phase. However, during the years 1964 through 1966 it became more and more evident that the semi-international arrangement foreseen would ultimately not be viable.
Finally, in June 1967, exchanges of notes between the Belgian embassy in The Hague and the Netherlands Foreign Ministry (July 1966 - June 1967) led to the conclusion that Belgium considered the cooperation in this project terminated.