ILL - The "deuxième souffle"
Replacement of the Nicole system (SANS, TOF)
The replacement of the Nicole system by individual PDP11/34 computers included new CAMAC counting electronics first introduced for D17; an incrementation unit took in data and updated memory modules using a separate front data bus cable. Data were read out using an autonomous unit which placed data on the CAMAC bus for DMA access to the PDP11 memory with block transfers. The CAMAC configurations exceeded a single crate, and CAMAC Branch configurations were used. The memory system (Klesse, Axmann ILL76KL336, Klesse, Mounier ILL77MU18T) had three write cycles and one read-out, though at high data rates the latter was replaced by an additional increment-write. 16 bit registers were used to control input and output ranges. For IN6 this was extended to 18bits using four word alignment to cope with the larger detector array. The memory units included first 4k, then 32k, then 128k 16 bit words which could be used concatenated as 32 bit storage. A microprocessor display controller with a plasma display screen offered visualisation of the data being acquired (Dimper). With Tektronix displays on the PDP11s more sophisticated display programs were developed, common to RT11 and RSX11 (Knowles, Nelson). IN6 was the first instrument to use the DEC CAMAC and DMA controllers which connected it to the UNIBUS. For later multi-crate systems each crate had a separate controller.
The NICOLE instruments ran the RSX11/M operating system with DECnet version 1, which allowed them to transfer data to a PDP11/55 “concentrator” which had considerable intermediate storage using DMC11 interfaces and pairs of co-axial cables. One of the functions of this machine was to write the data daily to tape for transfer to the central PDP10. This also involved modifying the data to 32-bit storage if overflows had occurred; on the instrument data these events were stored after the end of the detector array data.
Replacement of the Carine system (diffraction, three-axis)
The Diffraction Group instruments shared basic counting procedures amongst instruments; the programming was mostly performed by the instrument programmers. The full replacement of CARINE was a more piecemeal activity started later, though PDP11/40 computers were already operational on D8 and D9.
Although the Solar 16/40 computer had a better performance, the software and disk layouts were rudimentary and inflexible. By 1980 the DEC PDP11/23 had been produced, a single board computer, with commensurately lower price, using a multiplexed Q-bus rather than the DEC-UNIBUS. This could be used in conjunction with the Schlumberger CAMAC crate controller. Simple instruments like D1A used the LSI-11 together with a 2.5MB RK05 cartridge/2.5MB fixed disk (same spindle) running RT11. The RK05 cartridge was a common feature until replaced by the RL02, 10MB cartridge (pre-formatted). Instruments had an unlimited autonomy, and the set of data cartridges could be re-cycled. Data were initially copied to tape on the PDP11/55 concentrator for transfer to the central computer.
The typical instrument configuration included:
PDP11/34, 256kB memory, 4 DL11 serial i/o connections 2.5MB RK05 cartridge/2.5 MB fixed disk, DEC-CA11F Camac crate controller (+CA11FP DEC-Camac DMA controller for multi-detector memory readout) or PDP11/23, 256kB memory DQ11 4 line serial i/o and Quniverter bus adapter for Camac
To help program the Solar 16/40 computers, which had been allocated to D1B and the three-axis instruments Doris Nagel from Karlsruhe was seconded as an additional programmer. She worked with the Falaise microprocessor serial-line motor and counting units to adapt the CARINE control programs.