ILL - Electronics
Electronic modules
NIM modules (Nuclear Instrumentation Module) were used for rate meters, and as a powered chassis for detector amplifiers etc.
CAMAC (Nuclear Instrumentation Module) was used extensively as the primary interface to computers.
The CAMAC at the ILL was quite unique. Elsewhere it was used only as a computer interface; exceptionally LEDs signalled a limited functionality. At the ILL this was developed into a mixture of control and display. Many units had switches and buttons, only some of which could be overruled by computer control. This allowed some off-line control to be performed, performing start/count/stop operations. In addition the scalers had either an integrated display, or a generic display which operated by stealing from the CAMAC bus cycles to interrogate units during latent periods. Preset timers used thumbwheels to set values (two digits, KLx10M) - these BCD scalers could hence not be reloaded precisely to continue interrupted sequences). A number of modules had reset buttons. There were no serial numbers on the modules, which often were permanently wired in different configurations rather than having internal jumpers. There was little written documentation, apart from the original circuit diagrams.
The French engineers sent out designs for external constructors. Klesse had additional technicians to build boards in-house. Lefevre used Schlumberger individual crate controllers initially to couple CAMAC to computers, typically using a single crate. Klesse used GE controllers and a CAMAC branch system derived from big installations to cope with multiple crates. In 1981, IN6 was the first instrument to employ a DEC CAMAC crate controller, with its asociated DMA block transfer device to link it with UNIBUS systems. This later became the standard for most instruments with DEC computers.
Axis motors and control
Motors were typically step motors or DC servo motors for the axes with large inertia. In 1971, when Dureau made the choice of step motors the available electronics lacked efficient acceleration and braking ramps and this impaired the efficiency of these motors. Later better electronics were developed and large step motors were employed even for major components e.g. IN8.
A generic controller for motors had been developed to use the Precilec encoder and reader. This used a coarse and fine coupled coils to measure their relative positions using high frequency AC. The input was switched using a combination of a rotary selector and keys; only one shaft could be read at a time. As a consequence position setting was performed sequentially. This too was reflected in the initial mechanical designs using rubber wheels pressed onto the floor by air pressure, and this too being used to brake static components.
Instrument setup at the start of each cycle was very slow using computer controlled short scans, as the computer had to service requests from multiple instruments sequentially. Many scientists were very adept at using the manual positioning keys and scaler start-stop buttons to avoid the delays in awaiting the computers control.
Some instruments (notably the British contributions built by Grubb-Parsons, D3, D15, D16) used absolute encoders which enabled simultaneous positioning. These were however more fragile than the Precilec coders which could be used even in high radiation areas notably the monochromator casemates.
Counting systems
Simple detector instruments used CAMAC scalers. The multichannel instruments sent event information to NICOLE buffer memory which was then used to update the data files on disk. For the small angle scattering instrument (SANS), D11, these data directly incremented the contents of a dual-ported memory. For the new SANS instrument, D17, a new CAMAC based memory system was created, which was then further developed for the NICOLE replacement (see The "Deuxième souffle").
VME electronics
By the late 1980s there were increasing problems in maintaining CAMAC modules (even finding replacement chips), and the distributed processing possibilities of a chassis of VME modules was attractive, offering, for example, phasing control for multi-choppers, together with control of counting. In practice the demands of such instruments were much simpler, and a master/slave organisation was finally adopted, similar to the CAMAC systems. Newer and cheaper instrument-bus systems have been developed, using smaller profile boards. Fortunately the option to develop even more expensive VXI applications (with an analogue part of a larger format board) were not adopted.
The VME boards used Motorola 68000 chipsets. These were interfaced and developed using Macintosh computers. Initially at the ILL, CERN software was used to link to the VAX instrument computers. This was later changed to OS-9 then further developed using a configuration database when the VME systems were connected via Ethernet on a dedicated physical network. Considerable use was made adding ILL designed FPLA custom daughter boards to pre-manufactured VME mother boards.
Development of arrays of 1D position sensitive detectors lead to collaborations between the detector group and the electronics to use flash ADC converters for localising positions. This started the present tendency to deliver interpolated data to the scientists. The actual derivation of the conversion to real space is not included in the current data.