I have a memory error and the MV-processor won’t start since it needs a portion of the memory for scratchpad. If any of you reading this has the schematics for the MV/3200, MV/3500 or MV/3600 I would very much like to have a copy. I am also looking for the FE password for the on-board diagnostics.
If you have a MV/3200, MV/3500 or MV/3600 complete system or system board I am very interested in getting in touch with you. You will find contact information at the bottom of the page.
Today Rikard Sohlberg, a former colleague from Data General Sweden, visited to get a shot at running CEO for the first time in 15 years:
Today I took the commuter train from Upplands Väsby to the city of Uppsala some 48 km to the north. At the Uppsala train station I was picked up by Karl-Magnus Pettersson who has a Nova 3/12 that has been sitting in his garage since 1995. In addition to the dual cabinet with the Nova 3/12 CPU that also has an expansion chassis, was Karl-Magnus documentation of his software written in Fortran 77, 17 (yes, seventeen!) 5 Mb Disc Cartridges, some spare boards and an extender board! Great! Although there is a way of moving the CPU board to the top slot and wire wrapping a couple of wires from slot one to slot 12 to be able to troubleshoot the CPU board, having an extender board makes troubleshooting any board a lot easier. And getting an extender board is like striking gold!
After loading the system onto the trailer we drove to Upplands Väsby where the Nova 3/12 now is in my storage room.
Thank you very much Karl-Magnus.
How great is this! A Nova 3/12 with “Blinkenlights” on its way in! Awhile back, a gentleman called me and asked if I wanted his Nova 3/12? I guess you know the answer. His system was used for real-time data collection and evaluation and was purchased in 1979.
As you can see from the picture, it is a two-bay system with one 6045 Phoenix 5+5Mb disk drive (one 5 Mb fixed platter + one 5 Mb cartridge disk) and a paper tape reader in the CPU cabinet plus two Diablo 44 disk drives in the peripherals cabinet drive (one 5 Mb fixed platter + one 5 Mb cartridge disk per drive).
I have been told that the CPU has the Triple Option installed which consists of Hardware Multiply and Divide, Memory Parity Generator/Checker and Memory Management and Protection Unit (MMPU). The system comes with complete documentation, about 20 spare boards and 17 disk cartridges.
Today I received a letter from Ed McManus from Marlborough, MA with a Data General CEO keyboard template and a DG pin.
“When I retired from DG/EMC (29 years) they gave me the original Nova (S/N 1). I gave it to the Computer Museum and it is still chugging away out there.”
Thank you so much Ed! If any of you out there has a CEO manual they can part with; I’d be happy to take care of it. (You will find email addresses at the bottom of this page.)
Microsoft Office? Watch my video and weep! Data General had a fully fledged office package ten years ahead of you!
It is time to run Data General’s office automation system, CEO or Comprehensive Electronic Office for the first time in close to 30 years.
You guessed it; I don’t remember CEO´s command buttons so I had to try them all to send my message. (If any of you have the CEO keyboard template and is willing to part with it, I’d love to have it!) My first e-mail message was sent almost 30 years ago in 1994 using CEO at the DG office in Huvudsta, a suburb of Stockholm, Sweden.
Click on picture to start the video running CEO
Having the MV/3200 made it possible to see what systems were generated and available in the :SYSGEN directory.
A starter system was present named SYS.PR and starting the Technical Maintenance software on the MV/4000 allowed me to start AOS/VS using the option “Run a specified program” and selecting “:SYSGEN:SYS.PR” and hey presto, my MV/4000 started AOS/VS!
Using SED (I know, VSGEN is interactive so I can list, view, add and delete devices but SED makes it easier to get a complete overview and you can edit the .CSF file directly) I checked the configuration file (.CSF) for the default startup system; AOSVS_7.65.PR, that came with the MV/4000. It turns out this system is generated with the IAC-16 starting at line 0 which is CON0 or the operators console and this was the reason I was not able to use CLI when I first started the MV/4000. Since I don’t have the two cables to connect the Terminal Connection Box to the back of the MV/4000, I need to use the asynchronous port on the Multi-Function I/O Controller board for my CON0. Running “XEQ VSGEN/DEFAULT=MV4000DC_IAC” I get the IAC and disks preset and just needed to change the starting line of the IAC to 2 and the number of lines to “0 – 15”.
The power supply now runs fine and the MV/3200 starts every time but with a tweak; since the battery in the timekeeping circuit MK48T02B-12 is flat, every power on test stops at the character “D”. Since the MK48T02B, in addition to being an ultra low power SRAM and a real-time clock, it also is the power-fail control circuit which fails on power up. In order to power up, power must be applied, the date and time set and then leave the system on for a few minutes. Rapidly power off and then on makes the power on tests pass.
I have ordered and received a new MK48T02B-12 but I am waiting for a de-soldering gun that I have ordered to arrive before replacing the chip.
The joy of starting up the MV/3200 was short lived. After switching the system off after running for an hour, the system would not power up anymore. Complete silence when switching on the power so the obvious place to look was the power supply.
A visual inspection reveled than an electrolytic capacitor had the pressure release valve blown:
Faulty capacitor replaced:
Using a PC power supply splitter cable and a pair of automotive light bulbs I built a power supply load device. Switching supplies usually needs some kind of load in order to start oscillating. The light bulbs used are 12VDC, 10W why a load of 0,83A is imposed on the 12VDC supply and 0,35A on the 5VDC supply. The great advantage using light bulbs is that they also provide a visual indication. In my case, they would turn on and off as the power supply tried to start.
Since I don’t have the schematics (my always so very helpful friend Bruce is looking for the drawings), I had to revert to a hands-on troubleshooting and by “hands-on”, I mean literally hands-on! First I used freeze spray to cool down an area of 5x5cm until the light bulbs would go dark for a bit. I then used my fingertip and touch each transistor until I found the faulty one. When I touched the transistor, the power supply stopped oscillating due to the capacitive load my body through my finger imposed on the transistor. Using a magnifying glass to inspect the transistor lint covered in black dust or sooth was clearly visible sitting between the pins of the transistor. This combination created a path for current to flow between collector to base and emitter to base disrupting proper operation of the transistor.
If you consider doing what I just did, make sure that you only touch components that are insulated or you might end up dead! No kidding! The primary side of this switching power supply has a positive 167 volt DC side and a negative 167 volt DC side. Enough to make a person kick the bucket!
Since my initial cleaning clearly was not properly done, I hit the power supply board with the big artillery:
Hot water and dish washing detergent in the kitchen sink using a soft paintbrush to gently wipe the surface of the board. Only enough water to cover the surface of the board and making sure water is not being sucked up into the inductors or any switches. Cleaning must be done very rapidly so I only allow the board to remain in the water for a maximum of 60 seconds. Next step is a quick rinse under the hot water tap to rinse away any detergent. Third step is to immediately soak the board in isopropyl alcohol for a few seconds, pull it out and shake the board, hold it on its sides vertically to make the alcohol drip off and finally use a heat gun to dry the board. Use extreme care with the heat gun! You can easily melt the solder on the board as well as frying the components!
The final step was to use a dental brush to brush away any residue between the pins of the transistor that would not function correctly:
Since I don’t have an extender board for the MV/4000 DC I am unable to troubleshoot the faulty memory board so I bought a replacement board. I installed the board and the system now boots into AOS/VS without the memory errors. There is still an issue with AOS/VS or rather CLI32 with does not respond to commands.
The hardware is now running fine but without AOS/VS on 5.25 inch diskettes I will have to pause any work on the system until I have figured out a way to transfer AOS/VS from cartridge tape to diskettes.