Since August, 2019, when I purchased a Computer Automation Alpha LSI-2 minicomputer on eBay.com, I have been scouring the internet for information, hardware, and software for the almost forgotten Computer Automation “Naked Mini” Alpha 16 and Alpha/LSI-2 minicomputers from the first computer revolution in the mid-1970’s.
Restoring a 44 year old minicomputer is not for the fainthearted. First, you need to realize that my particular choice for restoration is not the easiest target. Computer Automation has been out of business since 1992. Even though Alpha LSI mini’s were manufactured for nearly 10 years, and tens of thousands were sold, they were mostly used as “components” in other products, and were not well known as stand-alone minicomputers.
It all started with Beginnings…
18 Comments
I really appreciate your endeavor here and envy the exposure you had to this mini back in its day. I’m just learning about the LSI-2, having acquired one that was shoe-horned into a Princeton Gamma Tech PGT-1000 termial. The PGT-1000 console was sold as part of a scanning electron microscope system. It was apparently used to display spectral analysis data generated from within the SEM.
My power supply (71-10044-00) is only partially functional and I am currently working on it to restore the -12VDC power rail. In such a state, the system will not attempt a boot due to some pretty tight power fail circuitry. Information, parts, software, and knowledgeable users, such as yourself, are indeed hard to come by.
This PGT-1000 was also fitted with two 8″ floppy drives, complete with two half card driver boards plugged into the LSI’s Maxi-Bus, and two 8″ floppies that I have not been able to read yet. A keyboard and CRT monitor are also interfaced to this LSI-2 via some custom boards, apparently produced by or at the request of Princeton Gamma Tech.
As an enthusiast of this type of computing, I look forward to hearing more about your project and software development. Please feel free to contact me so we can exchange information, photos, and or findings.
Great work!
Hi Andrew,
Sorry for the delay in responding – I have been very busy with my day job and haven’t had time to keep the blog updated.
I just posted an update so I hope that gives more information.
David
By the way – here is the bitsavers.org archive of LSI 2 schematics including the power supply!
http://bitsavers.trailing-edge.com/pdf/computerAutomation/schematic/00-53502.pdf
David
This was the first computer I ever programmed. It was bought to my high school in 1976 just when I starter my last year. It was hugely expensive and very modern!
Hi,
I am glad to read that someone else appreciates the qualities of the Alpha LSI-2, too. In the early 1980s, at the Institute of System Theory of the Berlin Technical University, we were using three of them for measurement tasks. For what was then considered to be number crunching, we had a DEC PDP-11/44. When comparing the Alphas with some of the DEC peripherals, the Alphas pioneered some features the PDPs were still lacking. For instance, we had 4 core memory cards, 16 kBytes each, in each chassis. All you had to do was to put them into contiguous slots, no DIP switches had to be adjusted. Try that on a PDP-11…
Still, there were some hardware problems with the Alphas:
– In the analog power supplies the total current of about 20 A was rushing through one not-so-wide P. C. B. track (next to the regulation transistor) which consequently had a tendency to burn away.
– There was an optional 16-bit I/O card that had, by design, a hardware bug in its interrupt logic. Whenever an interrupt request was signaled by some other card (or the clock…), it produced a glitch on its own interrupt request line. Sometimes the bus arbitration logic detected that glitch and asked for an interrupt vector which it never received. As a result, occasionally an interrupt was vectored to address 0 which is the Power Fail interrupt… Fortunately, this bug could be fixed. There were enough spare gates on that board.
– The floppy disk controller, which actually was far more advanced than that of the DEC PDP series, signaled the “I/O done” interrupt a bit too early. In the device driver we had to introduce a small fixed waiting time to overcome that.
We were running Alpha DOS on these machines. We also had the source code of DOS which we modified slightly. I still have the code of the modules I had modified (as a souvenir). We were programming in Fortran and I made the file I/O runtime module more usable. Now it was possible to open and close files for a channel number many times – without running out of memory space for runtime overhead data.
All the best for your project! Hope to hear from you.
Best regards,
Bernhard
Hey Bernhard,
Thanks for your detailed comment and kind words. I like that you can remember all of the quirks of the Alpha LSI machines so clearly!
I have a 16 bit IO board and also 8″ floppy disk controllers. I hope to get both working, so any details on your fixes from the past would be appreciated. I have schematics for both boards.
I would also like to get DOS working – if you have any software you can share either as files or as a scan it would be appreciated. There is a lot of binary code in disk images on bitsavers.org, but very little documentation.
Please check out the update I just posted.
David
Hello David,
I took a close look at the schematics of the 16-bit I/O interface as it is available at bitsavers.org.
It looks different than the one we had then. Obviously there were different versions of 16-bit I/O interfaces. The ones we had sent some MaxiBus input signal (perhaps PRIN) through some gates for internal purposes and then tried to AND out the original signal to obtain IUR. Since every gate has a delay, this was a silly design mistake. You can find out if your cards are OK by hooking up a logic analyzer to IUR-. If glitches appear in the rhythm of your system clock as long as the cards are plugged in then they have this problem.
What I wrote about the floppy disk interface related to those cards with order numbers that ended with …696. Whenever an I/O operation was completed, we had to wait for a short time until the next operation could be initiated. This had to be done by software in the floppy disk driver and in OS:ILD. If we did not, floppy disk operations were not reliable.
I have the source code of DOS (with my changes) and some utilities but no means to assemble and link them.
I wrote a document describing my changes to DOS (in 1984) which I will translate to English in the near future.
Best regards, Bernhard
Hello David,
at last my translations of the documents (in German) are finished.
If you gave me an email address that permits me to append files I could send to you what I have.
Best regards,
Bernhard
Hello Bernhard,
I find your comments here quite interesting, especially the floppy drive timing quirk as I have one of the 696 cards. I’m working on getting a naked mini LSI/2 up and running with an Alpha 16 front panel I slapped onto it. I’m hoping you might be willing to share your disk operating system code and modification document with me? Like David I am very interesting in anything relating to these systems.
Thank You,
Andrew Hendershot
Hello Andrew,
I am terribly sorry that I overlooked your e-mail for such a long time.
If you give me a mail address that permits me to append my source files I will send to you what I have.
Best regards,
Bernhard Auer
Dear Sir,
Unfortunately I cannot help you much at all.
But I can tell you that these computers were used a lot on Phillips Petroleum Company Norway (PPCON) installations in The North Sea, in the 70s and 80’s. Us PPCON technicians were trained in details on hardware and software troubleshooting, replacing any discrete components and doing software patches. On the computer boards itself and the large number of process I/O boards.
They were supplied by Daniel Industries in Houston (now part of Emerson) for their fiscal metering systems for oil and gas.
PPCO/PPCON is now named ConocoPhillips. May be you can find knowledgeable enthusiasts from Daniel and/or PPCO.
I wish you good luck and would be happy so see how you are doing.
Best regards,
Stig Arvid Knutsson
Kuala Lumpur
Hi,
How is your restoration progressing?
I have a very similar project with a CA Alpha.
/Mike..
Hi Mike –
It is moving forward much faster lately – sorry for the long delay in posting these updates.
David
Hi,
I was a post grad at Chelsea Collage electronics department, University of London from 1972-75.
I had an Alpha 16 to myself and programmed in assembly language.
I was working on speech analysis and synthesis.
Amongst other things I wrote a fast Fourier transform routine.
I had access to the two extremes of computer power, the Alpha at one end and the university CDC6600/7600 at the other.
I will watch your project with interest and would like to help in any way I can.
Regards Neil
Hi
I worked at Computer Automation from 1980 to ~1992 on LSI-2 & LSI-4 based systems. Initially involved with building the systems in the Iriish factory then customer support
Its great to see the systems being revived!
Best of luck with the work and if I can help or remember anything I’d like to help
Hi,
Interesting, reviving an LSI 2 computer!! I am recently retired from Michelin after 40 years, my first job in Michelin was at their technical training center and included teaching board level repair on the LSI-2 and LSI-4 computers. Later I worked as an electrical engineer writing assembly language for the LSI-2s that were used in a large rubber processing facility. Funny after all these decades I still remember the instruction sets!
Hi,
Nice to find one of these beauties! I was a language student from 1977 on and got interested in phonetics. I started the studies, and we had one of these.in the phonetics laboratory. It was controlled by a Teletype 33 with an integrated punched tape reader/writer. Somebody from another university had written the machine code that made the Alpha tick. I knew absolutely nothing about computers at the time. After a couple of years of studies, I started my work as an assistant at the department. One of the first tasks was to find a new punched tape reader for the Alpha. However, that was not a simple task in 1981 because the local sellers had not sold them for years.
Well, we ended up purchasing an 8-bit, Z80 based microcomputer with floppy disks to replace the tape reader. Nobody could, however, make the microcomputer and the Alpha communicate. Fortunately, the microcomputer turned out to be able to replace the LS-II and do much more. Moreover, the decibel level of the lab dropped a lot after we turned off the Teletype and the Alpha. The new PC also replaced our Z80-MCB industrial computer that was programmed to control our Swedish-made speech synthesizer. So, within a year, I had made a revolution in the laboratory, making most of the old technology redundant.
The side effect was that I learned programming enough to drop my language career and work with computers and programming for the rest of my working life. Alpha LS-II was a turning point for me, even though I only used it for less a year or two.
Markku –
Thanks for your comment. I will be posting more updates on the Alpha/LSI-2 this summer as I have a lot of new progress with the systems. I will also be revamping the blog to better organize the posts, and I hope to start uploading videos of the project to a YouTube channel. I have more Z80 system progress as well. Thanks for your support.