Computer history: DEC systems 1989 – 1994
DEC systems (period 1989 – 1994)
On July 13, 1989, the VAX 11/750 for scientific calculations (VC) was replaced by a VAX 6310 which was better suited to deal with the central VAX/VMS capacity for scientific calculations. and the processing of the VAXCluster work. A new type of disk was also started, the SA600 disk array with RA90 (1.2 GB) disks. The VAX 11/750 was handed over to the TNO’s Dienst Grondwaterverkenning (Groundwater research) in exchange for CDCnet equipment that was left after decommissioning their CYBER 810.
At the end of 1989, TNO-FEL’s financial department predicted that the VAX 8350 would be too small to run all planned FELBIS applications by mid-1990. At the end of June 1990, the management, therefore, decided to replace the VAX 8350 with a VAX 6410. At the same time, the VAX 6310 was upgraded to a VAX 6410 for scientific applications. A few years later, TNO’s Board of Directors decided that the whole organisation had to use the central SAP system. FELBIS, therefore, had to be transferred to the central system within a year. The IBIS project had to accomplish this transition. This project did not go without a struggle as the FELBIS administrative reports were user-friendly. Moreover, a number of the FELBIS applications such as equipment management were not supported by the SAP system. The users at FEL-TNO felt like going back years in time. As a project logo the ibis hieroglyph from pharaoh times was used; very appropriately as the ibis symbolised the Goddess of writers and calculators in old times.
Fewer operator interventions
The new Equipment management and IT services -Apparatuur, IT-Voorzieningen en Instrumenten group (AITV&I)- convinced the FEL management that the system management of the central computer systems had to be automated. Apart from the backup jukebox connected to the CDC 4380, AIT&V proposed the installation of a DEC TA90-tape unit. The TA90 could automatically exchange twelve IBM-3480 cartridge tapes (200 Mbyte a piece) using two tape cartridge stackers. A friendly contact at IBM told us that a new version of the 3480 tape cartridge unit would be equipped with a data compression option. This would allow up to 600 MB per cartridge for backup data. We increased our negotiation pressure on the DEC sales representative. In the end, TNO-FEL was the first one in The Netherlands that had a TA90E unit installed (as DEC called the original IBM unit with data compression) for the same price as the original offer by DEC for a TA90. DEC did not hide the original manufacturer name well. On the inside of the unit door, an original IBM disk with controller software could be found. Using more ‘inside’ information, a new type of combined disk/tape controller, the HSC5X-DA became part of the well-negotiated contract as well.
Our first experiences with this new HSC5X-DA controller were bad. After two days, the controller virtually breathe one’s last: the disk units and the tape unit slowed down by a factor of 1,000 or even did not respond anymore. A replacement controller card worked only one day before showing the same type of problems. On the same day, the similar controller in the other VAX 6410 showed the same problem. DEC had not that many spare controllers available in their European spare parts and repair centre in Nijmegen. Moreover, the repair centre could not reproduce the problem. In the US, some other customers experienced the same kind of problem.
Finally, it was discovered that the controller card contained the complex VAX6000-CPU of which multiple CPU test pins were not connected to a ground signal level. Whenever the free-running pin picked up some weird electric signal, a CPU-interrupt occurred. As there was no interrupt software code available to handle the situation, an endless loop occurred. The problems were resolved after a couple of very troublesome weeks.
In the end, the users at TNO-FEL had access to very modern central VAX facilities.