I have made a lot of progress in getting my newly acquired DECstation 5000/240 going again. Before doing anything I took it apart and opened up the power supply to inspect the electrolytic capacitors. They all looked fine. I then removed everything I could from the machine, including the CPU daughter card, and powered it up. The fans turned, which was good, and I checked the ripple with my oscilloscope. So the power supply looked fine.
I put back the CPU and one memory board, leaving everything else out, including the video controller. I connected a terminal emulator to the 3rd port for the alternate terminal and powered it up. I got nothing on the terminal, and the diagnostic LEDs all lit up and stayed lit. That was not a good sign. The one good thing was that the two green LEDs on the CPU daughter card lit up, which is supposed to indicate that it has established communications (with what I am not sure).
There followed a lot of messing around, trying the different jumpers, swapping the memory board, adding the video controller, reseating the EEPROM, nothing worked. Then I tried gently rocking the CPU daughter card in its socket. That worked. Hopefully it was just not seated well, and not a sign of some bad joints.
At that point I got the following output:
KN03-AA V5.2b 3/misc/kbd ?STF (4: Ln#0 Kbd self test) 3/misc/mouse ?STF (4: Ln#1 Pntr self test) ?RTCsi/cntl >>
So, it was rightly complaining that there was no keyboard and no mouse. It was also complaining that the real-time clock was not working, this will be the battery in the Dallas chip that has expired.
Unfortunately, I couldn’t enter commands via the terminal emulator. This is usually a problem with the combination of how the PC serial port works and the various intervening connectors to get to the port on the machine. I fiddled around with connectors for a bit. While doing this the machine suddenly powered down, after being on for perhaps an hour. I only got it back after unplugging the power cord. I have no idea what caused this. It happened again later when I put the whole system back together again to take the photograph of it working.
I decided not to bother investigating the serial port for now and just went to find the proper keyboard cable and a DMAG monitor cable to connect to my sync-on-green LCD monitor. Once I could enter commands I entered the cnfg command and got output similar to the following (re-typed)
3: KN03-AA DEC V5.2b TCF0 (32MB) (enet: 08-00-2b-39-f4-b3) (SCSI = 7) 0: PMAG-JA DEC V5.3a TCF0 (CX 00 d=8,24)
So it looks like I have a 32MB module, which is great! I found that I have three 32MB modules and two 8MB modules. Unfortunately it seems you can’t mix them.
I then installed the other options. The TURBOchannel Extender did not show up in the output of the cnfg command. I am not sure if that is normal or not.
I have saved the most intriguing item for the last bit of this post. The mystery option turns out to be some kind of experimental video board. It contains an FPGA and a large C-Cube chip. When I use the cnfg command to get the summary configuration it says
1: JV01A-AA DEC x0.06 TCF0 (MultiMedia Engineering)
When I use the cnfg command to get the details about it I get the following:
1: JV01A-AA DEC x0.06 TCF0 (Audio, Decomp, Comp) The usual suspects: Ken Correll, Tim Hellman (a.k.a. ‘The Lab Boys’) Bernie Szabo, Victor Bahl (Code ‘R’ Us) Bob Ulichney (Key grip) Greg Wallace (The big cheese)
Someone on the classiccmp email list told me that it is probably a Jvideo board, a prototype built by DEC SRC for the J300 series of video and audio adapters. He pointed me to the Digital Technical Journal, Volume 7, Number 4; in fact two of the people mentioned above authored one of the papers. One of them also wrote further articles on video rendering and video conferencing in Volume 5, Number 2.
Here are some pictures of the board:
Rob's Old Computers 