Contents
How it all began Buying an antique computer Box arrived First inspection Disassembly and repair Lights On! How old? PCB Inventory Next activities Right or wrong? Related Links Contact
My Rolm 1666B mil-spec Computer
military designation: AN/UYK-64(V)
Contents How it all began Buying an antique computer Box arrived First inspection Disassembly and repair Lights On! How old? PCB Inventory Next activities Right or wrong? Related Links Contact |
|
...sometime in the late 70's, early 80's. I was a young computer engineer working in a small company doing fascinating programming jobs. There were no PCs and no Microsoft Windows at that time (what a relief...). The equipment being used at that time were either mainframes (IBM 360s and similar heavy iron) or 'Minicomputers'. Don't take 'Mini' in the sense of today's PDAs or alike. The typical minis were Digital Equipment's PDP-11 and Data General's Nova/Eclipse. A well-equipped Nova 3 could easily fill two or three cabinats and weigh 100 kilograms or more. At the beginning, software was punched on card-stacks or papertape.
Diskettes were 8 inches and
stored a whoppy 160 kilobyts. The larger harddisk-drives were about the
size of laundry machines with capacities of around 10 megabytes.
This was the time I fell in love with the Nova's elegant and very clever instruction set. Four general-purpose 16-bit registers, a genial indirect adressing mode-bit and an adress-space of 32K (77777 octal) words. The instruction set was so intuitive that it easily could be memorized (in Octal-code). Should there be no terminal around, instructions could be entered via the 16 data and address toggel-swiches on the Nova's frontpanel. Everything you need for degugging was a Start/Stop, Display/Deposit and Single-Step switch. The military/ruggedized variant of the Nova was the Rolm 1602. An almost indestructible aluminum piece of engineering artwork. The most advanced Rolm was the 1666B, an equivalent to Data General's line of 'Eclipse' superminicomputers. At that time I wouldn't imagine that I ever would own such a treasure (which, besides, went over the counter for $218,000 per piece at that time).
...when you grow older more and more often you get such sentimental moments... Recently I was surfing the web for reminiscences to the good 'ol computer times. I sympathized to buy some antique computer equipment (people buy Louis-Quatorze chairs for a lot of money and nobody will smile about that...). My favourites were either some original Altair/IMSAI or a real DG Nova. However both of these are very rare and people who own an IMSAI dont'want to give it away anymore. So one Sunday night in my basement hobby-room I discovered this ebay entry:
THIS was the one I was waiting for! I had to
have that one. Unfortuntely the guy (yes he's name is 'Guy') who sold the
Rolm lived halfway 'round the globe (Sydney, Australia). Surprisingly the
airfreight rate was reasonable. I did know the forwarder (BAX Global, both
thumbs up) from a previous shipping and so I put an initial bid on the Rolm.
Because noone else increased the bit, I won. Money was transferred via PayPal
(which took almost a week until it arrived at Guy's account. Meanwhile Guy
screwed and glued a plywood box together. The transport was uneventful.
Even my concerns about customs was unfounded.
Box arrived, Power and Documentation needed!
A week later I picked up the 60-Kg box at Frankfurt cargo center. What I still needed now was a 240-to-115V transformer. I ordered a 1000W one, factory-new on ebay for just Euro 60. Now there was one last hurdle to take. I got a couple of cables with the Rolm but no power cable. The question now was how the power is wired to the 7-pin socket (J1). The owner of PWA Electronic was very helpful and mailed the pinouts immediately. After some Google searching I found www.tpub.com This people sell a lot of outdated military literature. And voila- I found the complete maintenance and repair handbooks. For a few $ I odered the CD which arrived a few days later.
This
was my second Christmas and of course I had to have a look inside the black
box ;-) However I did not follow my temtation to just hook up the power
cord. From collectors of antique radios I learned that the electrolytic
capacitors will leak and dry out after so many years. If such an electronic
circuit is powered up then it will smoke up immediately or even explode.
So, one of my first looks went to the power supply. By loosening eight self-retaining
screws the power supply can be taken out in a few seconds
Rear (*1) side of power supply
(and Rolm) with blower and diffusor to both sides of the box
Sidenote: *1: actually by manufacturer/military
terminology this is the front side. But this end really looks more like
a vacuum-cleaner than a computer. So I will use front and rear in a reverse
sense.
.
Not
a single wire has to be removed if the power supply is pulled out. The contact
fingers of the mainframe easily slide into the sockets of the power module.
Lots of amps are flowing thru this massive bus bar (lower front). Regard
the two guide pins located right and left above the socket. Proper alignment
always is assured even if the power supply is inserted in a hurry (because
of friendly fire maybe ? ;-))
This capacitor (220uF, 50V) was the only component
which was obviously defective. When I removed the transparent coating a
smell like old fish escaped. I exchanged it by a new, smaller one of same
capacity, voltage and temperature range (mil-spec parts are good for up
to 105 °C).
Frontside with all sockets (*1) If all plugs are installed, it is not easy
to remove them again, because they are so close together that one hardly
can turn the bayonet shells with fingers. For the purpose of installation
and removel we used to use special pliars at that time. The 'S' in 1666B-S
indicates this is a later semiconductor memory model (the earlier models
were of core memory).
*1: the sockets alone cost more than a current
personal computer
Inside
of the box. All boards are installed, power supply (right) is removed. All
boards are multilayer (yes, multilayer was a known but expensive technology
in 1984) and covered by an aluminum plate. The chips are thermically bonded
to this aluminum plate which distributes the heat to both sides. Using Allen-screw-operated
wedges, the board-sandwiches are tightly clamped to the grooves on the sidewalls.
This way the waste heat is transferred throught the sidewalls and removed
by the right and left heat exchangers which are fed by the blower on the
rear end. This construction assures a hermetically sealed confinement. No
dust or humidity can enter the internals of the computer. There even was
a special tool for board removal.
Backplane
with wiring. Note the many wires on the forward boards. I assume this could
be the connections to several multiple serial I/O interface bords. Probably
a computer used for data communiction. I have to investivate deeper in this.
Changing all the remaining capacitors just
for safety reasons is not feasible. So I slipped the power supply in and...
time has come!
Big surprise - All lights are on, fan is running
(my hands shaking for excitement). Reset and BITE do respond when depressed.
So... is this machine really computing now? - I cannot tell at the moment,
but it looks not bad.
No, not me, I mean the Rolm...
This
is Rolm is of serial number 618. The PCBs of the power supply carry 1981
as date of manufacture. The warranty seals on the case shows '25th Nov 1986'.
So it's 20 years since the case was open the last time. The front side of
the Rolm carries an 'ETM' (operation time logger). That one shows 690 hours
of operation. This confirms with the overall good condition. The mechanical
components show only very little wear (and no bullet holes at all).
This is the inventory of printed circuit boards of my Rolm. The card cage has a total of 22 slots. Between slot 11 and 12 there is an solid aluminum divider separating the card cage into two departments. All slots except slots 13 to 15 and 19 to 22 are occupied.
| Module name, Description, Options | Model # | Serial # | QC Seal Date | |
| Priority/Load Module A | 3566 | 11090 | 18.Dec. 85 | |
| Adaptor, Fixed Disc | 4055 | 384 | 17.Dec. 85 | |
| Floppy Disc Interface | 3382 | 953 | 2.Juni 85 | |
Async Line Multiplexer Options 8(4), 9(75,300,1200,2400), 90(x4) |
S617 | 193 | 18. Dec. 85 | |
Async Line Multiplexer Options 8(3), 9(75,300,1200,2400), 90(x4) |
S617 | 219 | 18. Dec. 85 | |
Async Line Multiplexer Options 8(2), 9(75,300,1200,2400), 90(x4) |
S617 | 138 | 5. Sept. 85 | |
Async Line Multiplexer Options 8(1), 9(75,300,1200,2400), 90(x4) |
S617 | 201 | 18. Dec. 85 | |
Async Line Multiplexer Options 9(75,300,1200,2400), 90(x4) |
S617 | 229 | 18. Dec. 85 | |
Processor Card Set, CPU PFP |
5711 | 807 | 18. Dec. 85 | |
Processor Card Set, CPU ROM |
5711 | 585 | 18. Dec. 85 | |
Processor Card Set, CPU Data, Options 45 (2400) |
5711 | 809 | 18. Dec. 85 | |
Processor Card Set CPU MAP |
5711 | 919 | 18. Dec. 85 | |
| unoccupied | ||||
| unoccupied | ||||
| unoccupied | ||||
| ERCC Controller | 1754 | 507 | 18. Dec. 85 | |
| High Density Memory Control | 1753A | 370 | 12. Dec. 85 | |
| Semiconductor Memory Module 512K (*1) | 2032 | 287 | 15. Aug. 85 | |
| unoccupied | ||||
| unoccupied | ||||
| unoccupied | ||||
| unoccupied |
Annotations
*1: of course kiloBITS, makes a whoppy 65 kilobyte of main memory
Next activities will be to find out if the processor is really working. Because this type ofcomputers don't have an external operating panel with lights and toggle-switches, the only way to get access to the CPU is via the system console (virtual control panel). A PC running HyperTerminal will make a good candidate for a system console. The Boot ROM of the Rolm also contains a kind of interpreter which allows to modify memory locations and accumulator contents. It could even be possible that there are hardware tests implemented - but this at a later date.
(1) Real computers don't need a Reset/Reboot-button (only Windows PC's do)
Right?
Wrong! the Rolm got Three(!) of them. A 'Power-On Reset', a 'Run Reset' and a 'BITE/Boot' button.
(2) Real computers don't crash (only Windows PC's do)
Right?
Wrong! Proof: aim with a calibre .50 BMP weapon on the Rolm and you can watch the crash ;-))
Sidenote: destruction of army electronics material to prevent enemy use shall be accomplished in accordance with TM 750-244-2.
http://www.designdata.com/rolmdg_support.html
this company still sevices Data General and Rolm computers
http://www.wildharecomputers.com
this company offers emulation software to run DG and Rolm native programs on modern equipment, even personal computers
http://www.pwa-electronic.de/deutsch.htm
besides other business activities also this German company services Rolm equipment. The owner of this company was very supportive when I asked for a pin-layout of the Rolm's power
connector. Thanks for this
If the kind reader of this lines is also interested in historic Rolm and DG computers, please don't hesitate to contact me at:
ga at loeff dot de
Or if you even could tell me where I could get (preferrably as an inexpensive surplus item) this 7-pin power plug, then you'd save my day ;-)
This is a treaded socket (threads on the outside).
I could decipher 'SAE 18-9' - maybe the thread designation. The outside
diameter is about 28.5mm. The upper five pins are about 1.2mm in diameter.
The two larger pins below are about 2.4mm in diameter