MicroVAX 2000

This is the highly successful DEC MicroVAX 2000. Introduced in February 1987 it immediately became a huge success. The same machine also sold as the VAXStation 2000. It is just a change of a jumper inside.

The system board contains most of the functions of the computer with exception for the network interface which reside on a separate board, although the thin ethernet connector sits on the main board. It is based on the same CPU and FPU chip set as used in the MicroVAX II. The CPU is called DC333 and the FPU is DC337. This paper by Bob Supnik give some interesting information about the development of the chipset. The design specification, micro code source and the micro code documentation are also available for study.

The MicroVAX 2000 mother board includes the CPU chip, DC333, the FPU chip DC337, 2 Mbyte of DRAM memory, 256 kbyte of EPROM for boot and diagnostics. There are also a four serial port, a disc controller for floppy disk and ST506 type hard disk, SCSI controller as a tape drive interface. The main clock generator is 40 MHz which is divided down to 5 MHz which feeds the CPU and FPU chips. As the same board is used in the VAXStation 2000 the board include a 128 Kbyte frame buffer to provide a 1024 x 864 one bit per pixel display. The jumper to select between MicroVAX 2000 and VAXStation 2000 is located in the upper right. The technical manual give a lot of insights of how the system operates.

Our unit has been sitting on the shelf for many years and unfortunately during this time the battery supplying backup power to the Time-Of-Year clock had started leaking. The leaking battery fluids had also affected the main circuit board. We started a process to clean of the board from battery fluid by washing in gently with diluted acetic acid to neutralize the battery fluid, then washing carefully in water and let it dry. The the chips in the affected area (upper right corner) were removed to make sure that it was clean and no battery fluid residues were present underneath the chips. After another cleaning the chips were soldered back in place (except for the 74LS240 and 74LS244 which were replaced with new chips)

Then the power supply were checked. All dust and debris were removed. The MicroVAX 2000 system unit include a dummy load board which is used in place of a hard disk. This board were connected and power applied. But there wee just a small tick on the fan. No power output. So we started checking the schematics, and the the actual circuit. The cold side switching chip looked fine and produced a nice waveform. Applying low voltage on the hot side also indicated a perfectly working power supply. So what could be wrong? Why didn't the supply start as it should? The schematics revealed, after a deeper study, that the +5V and +12 sense signals to the PWM chip were summed together to create the error signal for the regulator. So if the there is more load on the +12V than the +5V the 5V might exceed the crowbar voltage and thus force the PSU to shut down. The dummy load board had six jumpers the enabled the +5V dummy load as well and then the PSU worked perfectly.

With a working PSU (with ripple within the limits) the next step was to start the main board on the bench. A special maintenance cable was built. Two female DB9 connectors were used.

Printer port	Computer
2	2
3	3
7	5
8 connected to 9

The baud rate was set to 9600 bps.

After a while there were some output on the terminal and the machine executed the startup diagnostics.

There are some soft errors with the NVRAM and Time-Of-Year clock but this is caused by a missing battery. But the system was not completely dead. The battery liquid hadn't destroyed it and the repair hadn't broken it either. A six layer board is quite hard to desolder components from without damaging the very thin conductors on the PCB. But since this are were related to the serial port and the FIFO for the serial port, we can conclude that the repair went well!

When assembling the machine completely and attaching the network interface board as well it turned out that something was wrong since the NI consistently gave this diagnostic message when powering up the machine:

?? NI 0011.700E V1.1

Helpful people on the net with access to "MICROVAX MAINTENANCE GUIDE EK-VSTAB-MG" helped me to understand the error code: It simply meant "Rx Error". This happened even though a proper BNC T with two terminators was on the thin ethernet output of the machine. Now I found a transciever to try at the AUI connection but there it wasn't even power to the transceiver. It turned out that the cable inside the machine had a broken wire. Fixing that made the transceiver work and it was possible to connect it to a network and have Wireshark sniff the the wire. At regular times it was sending DEC proprietary MOP boot messages.

But why didn't the thin ethernet work? Further checking the cable and then the pulse transformer showed that the latter has been damaged by the battery fluids, but replacing this didn't help either. I had to also replace the NS8392 transceiver chip to make thin ethernet work as it should.

Next step is to try to get a hard disk for this machine so that we can run VMS on it. But since working MFM disks are a scarce resource we have decided to go for an emulator.