Date of Production: 1991

CPU Type: VAX of type "SOC" (System-on-Chip), KA660, 35ns cycle time

Random access memory: 40MB

Max. ranom access memory: 64MB

Internal IO Bandwidth: Q-bus: 3,125 MB/s, DSSI: 5MB/s

Chassis: BA430, containing 12 Q-bus slots and 4 drive shelves, one power supply

Supported operating systems: OpenVMS, NetBSD



Story:

The VAX 4000-200 (DEC-internal codename "Spitfire") was the last VAX system completely built around the DEC's popular bus called "Q-bus". The processor board named KA660 was the fasted one for Q-bus based systems and was heavily used for upgrades of older microVAXen and VAXservers (Models 3300/3400/3800/3900 and MicroVAX-II/III/III+).

The so-called VAX SOC chip is a "System-on-Chip" version of the CVAX-processor which was, in addition to above named field upgrades, used in the VAXstation 4000 VLC, the MicroVAX 3100 Models M30 and M40 and the VAX/VAXserver 4000-200. The CPU uses 6kB of L2-cache, runs with a cycle time of 35ns and also includes the floating point unit (CFPA) of the CVAX-processor. It was produced using DEC's "CMOS-3" process (1 micron, three metal layers).

The CPU-board is equipped with the "Single Host Adapter to Interface DSSI" (SHAC) chip and a "Second Generation Ethernet Controller" (SGEC) to decouple DSSI and the network traffic from the Q-bus and to make full usage of the DSSI-performance in terms of bandwidth which is higher than the one offered by the Q-bus.

This particular system was my very first VAX-computer!. It came to me via ebay in 2003, not knowing anything about its history before being sold on ebay. It probably was used in a networked or DSSI-clustered environement, as it neither contained DSSI harddisks nor an internal tape drive. The drive shelves were convered with filler panels. Besides 8 MB of memory, one single additional KFQSA-card was mounted, which might have been used for DSSI-Clustering or external DSSI pheripherals.

The BA400-series chassis is in my point of view very well designed. The plastics enclosure can be removed, so that the chassis can be placed in a 19" cabinet. Just removing the plastics enclosure turns it into a rack-mountable VAX :) In contrast to the VAX 4000-300 using a BA440 backplane and chassis, the 4000-200 uses the BA430 backplane and chassis. They differ amongst others in terms of CPU/memory board location and connector types. The Compared to the BA430 backplane, the BA440 backplane uses dedicated CPU and memory boards slots to get higher performance compared to the qbus used in BA430.
The BA430-backblane used for the VAX 4000-200, which is a 4-layer backplane (two signal layers, one power layer and one ground layer), offers the possibility to connect a SCSI controller (qbus-based or DSSI-to-SCSI bridge) to a SCSI-device installed in the right-most device shelf of the chassis via an external SCSI-cable attached to a dedicated backplane connection. This offers the possibility to install an RZ-type SCSI disk drive or a TZ-type SCSI tape drive into the VAX rather than using a DSSI-device or a TK50/TK70 tape drive.

This VAX was my opening gate to the VAX platform, learning alot about it and getting into touch with its native operating system OpenVMS and - as an alternative - the UNIX BSD-derative called NetBSD.

Configuration:



Using an Emulex UC08 SCSI controller on the Q-bus, a Fujitsu 5,25" SCSI harddisk and a Cipher 995S reel tape drive (aka TSZ07, DEC labeled) became accessible. The Fujitsu disk became the NetBSD system disk. In parallel, OpenVMS was installed on a RA92 disk using a KDA50 SDI Q-bus controller. I used this configuration until 2005 and really had fun with that VAX!
Between 2005 and 2010, I moved several times during my studies and did not have time to play with it.
In 2012, I digged it out again in order to use it primilarly as a test system for peripherals such as harddisks and external tape-drives.

After powering it up again after several years of hibernation, it turned out the Fujitsu harddisk was bad. Moreover, while working with it regularly again, I could observe who the CPU-board slowly died. In the beginning it did not always pass the power-up tests and with time, it became worse and worse so that it never passed the tests in the end. Thankfully, a kind computer fellow offered to me a working board if I would give him my bad CPU-board. Since then, my VAX works happily again!

Since 2012, openVMS 7.3 is installed on a 2GB large DSSI-disk (RF73) and a KLESI-controller (DSSI) is installed, which offers possibility to make use of a TU-81plus reel tape drive. 

More Pictures of the VAX 4000-200: