Restoring a vintage 1975 Minicomputer

November 8, 2019

Since August, 2019, when I purchased a Computer Automation Alpha LSI-2 minicomputer on, 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…


IMSAI w/CPA Front Panel vs LSI/2

June 25, 2021

Why IMSAI Here?

I know some folks are wondering why I am discussing a pair of IMSAI computers on a blog primarily about restoring Computer Automation LSI/2 16 bit mini-computers…

Well, here is my logic – both types of systems have “programmer’s” front panels that show what is happening inside the machine, and allow you to stop, start, debug, single step, and input data. Also, the IMSAI was my second computer, in my house, (that I actually owned, instead of leased ). The LSI/2 was the first computer I had possession of – even though it was leased from Computer Automation.

The second reason is that the LSI/2 and IMSAI S100 computers are actually quite similar – except for early S100 systems having an 8 bit data bus. Therefore, easing back into assembly programming is actually easier on the IMSAI 8080 and Z80 systems because there are a large number of more advanced programming tools – assemblers, debuggers, emulators, etc. that can be run on the hardware, or on a more modern PC system.

Plus, both types of systems were primarily programmed in assembly language. I know, a lot of S100 systems with CP/M ran MBASIC-80 or CBASIC, but for systems programming, it was done in assembly – look at WordStar, Turbo Pascal, MBASIC-80, Modem7, Kermit, Xmodem, and other utilities. All were written in 8080 assembly.

Another bonus is that the 8080 and Z80 systems have tons more documentation readily available on the internet.

Finally, the 8080 and Z80 systems have the capability to run an OS – CP/M – on floppy disks or even flash or SD cards. This means that I can relearn my skills with less friction and pain, then apply them to the LSI/2 which has more a primitive work environment, at least until I can get some floppies working and disks rebuilt.

So am I backing out of the Computer Automation LSI/2 restoration? NO! In fact, since I am moving soon, I have decided I need a new place with enough space to set up all of my vintage systems and be able work on them in parallel – including the THREE LSI/2 systems I own.

IMSAI Restore continued…

So how are the IMSAI’s doing? Well, after replacing the Z80 in my Fulcrum I8080 with the Jade Z80 CPU, I was able to see some functionality from the CPU. However, depositing data and viewing memory still didn’t work correctly. However, now I could read the boot EPROM in the Jade board correctly.

Next step was to see if it was the CPU, the memory, or the CPA front panel causing issues.

I built a replica IMSAI CPU-A board with an 8080-A CPU as the simplest CPU card that should absolutely work with the front panel. It mostly ran, but I still could not write to memory. So I decided to re-chip the front panel.

By the way, it is a Fulcrum replica version of the IMSAI CPA front panel PCB in the Fulcrum branded IMSAI. This version was re-laid out (with tape) from negatives of a bare original CPA, with all of the IMSAI fixes applied. I should know – I did the tape up myself at Fulcrum in 1981. I was hoping that either corrosion or bad flip-flops or one shots was the cause of the write problem. But no luck. Time to break out the scope.

I also got some advice from the retro forums about some S100 lines that might be grounded on the Jade CPU or CompuPro memory boards that could prevent memory writes. At this point I am fairly close to being able to run code on the system. But, I had to stop work because of the pending move to a new home. More coming soon.

IMSAI Updates

June 23, 2021

IMSAI Systems Restoration

As I mentioned in my previous post, I recently acquired two IMSAI 8080 computers – one original and one Fulcrum i-8080 vintage 2nd generation. My goal was to get them running and then have CP/M running on both of them. The original IMSAI came with Micropolis 5.25 inch floppy disk drives and controller and may boot when I get it working. The Fulcrum I-8080 came with a Jade Z80 S100 cpu and Morrow Disk Jockey 2D/B floppy controller, and a Morrow Thinker Toys Wunderbus motherboard. It also had a Fulcrum VIO-X video card.

I started working on the Fulcrum I-8080. The first issue was there was no memory in the system. Dusting of my ancient S100 memories, I decided to try to find static RAM boards for the system on eBay. There is more S100 memory than any other board type on eBay, but a lot of it is dynamic or off brand stuff. I found some CompuPro RAM 16 which I thought was a good deal – 32KB boards using 6116 static memory. I ordered a couple. Then as I was doing more research I discovered that many CompuPro RAM boards were difficult to use in older S100 systems like the IMSAI. They were compliant with the IEEE S100 standard while older system boards may be “pre-S100 standard” IMSAI or Altair bus interfaces with some bus lines used differently. The IMSAI front panel could also be an issue with some memory. The Vintage/Retro Computer Forums had notes about how to modify the RAM 16 to work with IMSAI style boards but it is a pain. So I found some CompuPro RAM 20 (RAM XX version) boards that have more options exposed and are easier to configure. I also got a replica Jade Bus Probe board and an S100 extender card to help with debugging.

Now I had the boards for a running system (if they worked!!).  I tested the Fulcrum power supply and the voltages looked good. In the S100 system the linear unregulated power supply provides high current DC voltages to each board and the board is responsible for regulating the voltages to +/- 5 VDC and +/- 12 VDC as needed. This uses a lot of power and generates some heat. Vintage S100 RAM boards are especially power hungry. The RAM 20 has FIVE 7805 regulators on the board.

I installed the Jade Z80 CPU, the Bus Probe,  and a memory card. When I applied power, the bus probe and front panel lit up. However the front panel switches did not have the expected results. The computer did not enter run regardless of switch inputs. I used an extender card and my Sigilent SDS1104X-E 4 channel oscilloscope. The M1 signal was not pulsing. The CPU clock oscillator was running. Reset was working. I decided to replace the Z80 CPU first. Voila! The CPU runs now. There are still other issues, but this is a good start.

One of the blue panel switches  (C&K 7113) paddle is missing, but I was able to find a couple of new matching replacements in Australia. These switches and paddles have not been manufactured for many years and are one of the hardest parts for restorers to find.  Fulcrum could not get the darker blue switch paddles that IMSAI used (they were custom made for IMSAI) and Fulcrum’s blue paddles were lighter as you can see in the photos in my original post. Unfortunately, these new paddles are lighter blue than Fulcrum’s blue paddles and do not match the other switches. It works but I will have to keep looking.

IMSAI 8080 Computers

April 9, 2021

As if I don’t have enough projects to work on, I recently acquired two IMSAI 8080 vintage computers. The IMSAI 8080 was the “second” widely sold 8-bit personal microcomputer using the Intel 8080 microprocessor and was released in late 1975 after the MITS Altair 8800 . It was later available with the more powerful Intel 8085 and used the Altair 8800 backplane/bus design which later became the S-100 standard. The IMSAI products were  made in San Leandro, CA (Silicon Valley) by IMS Associates Inc. (IMSAI). It is estimated that IMSAI sold between 15,000 and 17,000 computers and kits.

The 8080 was sold mostly in kit form. There eventually was a full line of IMSAI personal and “business” computers. However, as more companies entered the S-100 market selling products and kits for less than IMSAI, the market became too competitive. IMSAI was unable to grow the business market quickly, leaving only the slick home/gaming computers like Apple, Atari and Commodore in the running against the S-100 hobby computers. IMS Associates Inc. went bankrupt in October 1979 and the assets were sold at auction.

I owned two IMSAI computers in the 1980’s, both of which I built myself. An electronics surplus store in San Jose, CA purchased some of the assets of IMSAI at the auction and sold parts and complete IMSAI kits for several years. They also manufactured new products, including new S-100 boards and replica IMSAI 8080 systems using some of the original designs and parts under the name Fulcrum Computer Products.

I worked at that electronics store in San Jose for a year or so during the time when they were selling IMSAI parts and systems and the Fulcrum products. Part of the work included updating original IMSAI PCB designs (hand taping from photo negatives) for the front panel and other boards to include fixes and updates and development of some of the Fulcrum products — including documenting their VIO-X intelligent video terminal S-100 board and their 8″ CP/M 2.2 distribution.

Although there were a number of microcomputer “floppy disk operating systems” developed in the early days of personal computers, by far the most successful was Digital Research’s CP/M. It was first used on the IMSAI, and later was available on virtually every 8-bit computer made in the 1970’s and early 1980’s – even the Commodore 128 and Apple II (with Z80 board). While it was primitive and had limited storage compared to today’s disk systems – it was far better and faster than paper tape or audio cartridge tape solutions. It allowed these computers to be used for more demanding applications like word processing, data bases, and programming.

Later, I used these computers running CP/M 2.2, DB2, Turbo Pascal and WordStar for several years for programming and word processing. I actually wrote and published three computer books and hundreds of market research and magazine articles with these tools.

So back to the present… I acquired two IMSAI 8080 systems and one was original IMSAI, while the other was a Fulcrum “updated replica” system. The original IMSAI 8080 looks to be in good shape and has two Micropolis 5.25″ floppy drives and controller for CP/M 2.2.

The Fulcrum I-8080 has no memory or drives, but includes a Jade Z80 CPU, a Fulcrum VIO-X and Morrow Designs Disk Jockey 2D/B floppy disk controller. I may also pick up the Fulcrum dual 8″ drive enclosure for Shugart 801 drives from the same owner. I have decided to use 8″ floppies and maybe add an IDE Flash drive upgrade later. I found a pair of Shugart 801 drives in a rack mount panel. I also found some Compupro RAM20 32k static ram for this system since it does not have any memory.

Unfortunately, one of the powder blue panel paddle switches (C&K 7113) is missing, but I was able to find a couple of new matching replacements in Australia. These switches and paddles have not been manufactured for many years and are one of the hardest parts for restorers to find. However, I have just found a similar switch from Mouser made by NKK.

For some reason I have always been fascinated with vintage computers that feature front panels with status lights and input switches. I think it may be because all of the computers I used in the early part of my career (1960s & 1970’s) were either IBM mainframes or minicomputers with these features. When microcomputers first became available, they used the same lights and switches front panel I/O approach due to the high cost of other options.

After doing some research to refresh my memory about how to get S-100 computers up and running with CP/M 2.2, I found lots of code from the RLEE archive for the Morrow Disk Jockey 2D/B floppy controller, but no prebuilt 8″ floppy disk images to use for CP/M. I decided to add a CompuPro Disk 1 board to use to bootstrap the DJ card and build a new image for CP/M. I also was able to find some 8″ IMD images for the Disk 1 system, and even a set of  CompuPro original system disks. So that is my current plan to get the Fulcrum I-8080 running with CP/M again. I even have an Adaptech  Aha-1522 ISA controller  which can be used with Dave Dunfield’s IMD utility to write 8″ disks from an older IBM PC. I am looking forward to a  nostalgic and fun weekend of  building a CP/M image.

4/20/2021 – UPDATE: I have found two Canadian collectors online who have worked with the Disk Jockey 2D/B controller and CP/M 2.2. One has built a bootable disk from scratch (in 2012) for his IMSAI and featured it on his YouTube channel. It was quite a project. (IMSAI 8080 8″ floppy disk drive project in preparing for CP/M PART 1)

The other collector was posting on a vintage computer newsgroup and mentioned he had a DJ 2D/B. I asked him if he could share a boot disk image and he was nice enough to offer to send me an actual disk.

Spring Update 2021

April 9, 2021

I have survived the COVID-19 Pandemic so far, now a full year in self isolation (only leaving my home for food shopping and Dr. appointments). Luckily I have a nice view of the Pacific Ocean about 1/4 mile away, since I live in West Los Angeles. I was lucky to get both my COVID-19 vaccinations (Moderna) in February and March 2021 and had minimal side effects (headache and fatigue). I have also been very fortunate that I was able to keep working through the whole lockdown, and my employer has been able to maintain its profitability during this time.

While I have continued to add to my Computer Automation LSI-2/20 systems with some more acquisitions from eBay, I have not had as much free time as you might expect because my day job has been quite busy and has required a lot of off hours work to keep up with the business needs. Success has its costs as well as benefits.

I plan an update in the next few weeks with some major progress on the system functionality. I have also started a few enhancement projects including a remote hex key pad to augment the front panel input keys, similar to the programmers panel on the later LSI-4 systems. I have also been working on software support systems like a cross assembler and file uploader.

So I apologize for not having more  updates on the restoration process, but hopefully there will be some more good content coming soon.



Christmas update December 2020

December 24, 2020

Well, time flies when the world is a crazy place!! Between “stay at home/work at home” orders and the unrest caused by politics and economics this year, time has flashed by like DVD on fast forward.

My day job has been very busy and has not left me much time for working on my passion project/hobby of restoring the Computer Automation LSI/2 minicomputers I have acquired.

The Good, The Sad, and the Ugly

The good is that:

  1. I found a new-in-box portable Remex RAR-5200 paper tape reader punch (serial and parallel)
  2. I have found a collector in Europe that has documentation for several of the cards below
  3. I have found some of the very scarce I/O boards including:
    • Utility I/O (parallel printer and high speed paper tape interface)
    • dual TTY module
    • 16 bit I/O
    • 64 bit TTL input card
    • transport interface – for 9 track IBM compatible tape drives like Kennedy 9000 and Pertec (both of which I used in my 1973 project) (no docs)
      1. requires a tape formatter
      2. requires a 9 track drive (and a strong back – very heavy)

“Regrets, I’ve had a few…”

Naked Mini LSI-4/90

The sad news is that I didn’t buy the full rack Naked Mini LSI-4/90 System that was offered on for $1800.00 plus $850.00 shipping.

Although I could have purchased it, I was concerned that I have never found ANY hardware documentation on the LSI/4 systems or the hard disk storage they used. I love the way it looks, but I would have probably had to scrap all of the peripherals and rack cabinet due to lack of space where I live.

ASR 33 Teletype

I also did not buy the used ASR 33 Teletype that was for sale in a lot locally earlier this year. It was actually a TWX machine plus two Model 32 Telex machines. Again, I was concerned about space, but I should have done the deal – it was only $300 for everything. Just a used ASR 33 sold on eBay a month later for over $1000.

1920’s Amateur Radio Station

HAM it up!

Growing up, one of my favorite pastimes was listening to short wave radio. In those days (the 1950’s and 60’s), there were many old multi-band “console” radios from 1940’s and 50’s available for nothing. Most people were watching the TV, not listening to radio in the evenings. I listened to foreign stations, Morse code, “satellites” like Sputnik and even distant “clear channel” broadcast stations at night – some were hundreds of miles away.

When I was about 14 living in the Seattle, Washington area, I met a neighbor who was an amateur “HAM” radio operator. He was kind enough to explain how short wave radio worked to a curious and persistent kid. I tried for several years to learn Morse code so I could get my license, but I was not able to. I had a CB license, and later got an FCC Radio Telephone Third Class License with Broadcast endorsement so I could operate the 1000 watt FM radio station transmitter (KSPC-FM) at my college, but I never got my amateur (HAM) license.

A few years ago in Dec 2006,  the FCC removed the Morse code requirement for Amateur Radio. In September,  2020, I passed my Technician Class Amateur license. In October,  2020, I passed my General Class Amateur license. and in November,  2020, I passed my Extra Class Amateur license.

I now have an Amateur Radio License after 55 years! I am now W6NF (Whiskey Six November Foxtrot) and have the highest class Amateur license currently available.

ICOM IC-7300 Digital HF Amateur Radio Transceiver

Finally, my Christmas present arrived from the North Pole – a new ICOM 7300 SDR radio to use to explore my new ham privileges. Don’t let the small size fool you, this radio is the most full featured entry level rig ever offered and it is a totally digital radio. It does SSB, CW, DATA, and even the old school AM modes.

Once I was able to get an antenna up at my apartment, I was amazed at the contacts I was able to make using the new digital data modes like FT8. In one evening I had over 20 contacts from all over the world including USA, Canada, Hawaii, Alaska, Mexico, Chile, Argentina, Greece, Russia, China, Japan, and New Zealand. My signal was even received over 10,000 km away in Antarctica!


Naked Mini LSI-4 and More

October 4, 2020

October 2020 has just begun and several of my vintage mini-computer dreams have come true!

Alpha 16 Front Panel

First, I just found and bought a real Computer Automation Alpha 16 front panel from 1971 online. Believe it or not, the Alpha 16 minicomputer was the first computer I ever “owned”, or even had in my possession. In about 1972, I was working as a telecommunications engineer for Arcata Communications (one of the first private business telephone “interconnect” companies) located in the SF Bay area. I was asked to design a phone call billing system for use with a private phone system, to log calls for hundreds of phones in a business.

I proposed a solution using the Alpha 16 computer as a controller, but the company decided they didn’t want to develop and build hardware. So they released the idea to me. I partnered with a college buddy and we “leased” an Alpha 16 from Computer Automation. I developed the product idea to prototype stage in my two bedroom apartment, learning assembly programming in the process. I used a Teletype ASR33 as the console and paper tape I/O for storage.

Months later, the company approached me and offered to buy the design, which I gladly agreed to. I worked on the product for several years and it was deployed to a number of large customers. We actually used the later release Alpha LSI-2 for the production machines.

Naked Mini LSI-4/90 System

Although I never used a Naked Mini LSI-4 computer, I always wanted one. The design was so cool and very unique, with the hex LED displays and the keypads for entry. The LSI-4 was also hardware compatible with the LSI-2 but had a much upgraded architecture. It supported 64K words of 16 bit memory instead of 32K, it had multiple registers, stack operations, and more instructions (of course). There was a 9 slot chassis version with separate power supply. Plus it was faster.

I recently found an LSI-4/90, the top end model. It was used by Delco Electronics in a software development role for designing military aircraft firmware. It is on my wish list for Christmas.

High Speed Paper Tape Reader Punch

Another item on my bucket list for the Computer Automation LSI-2 restoration project has been a paper tape I/O device. I had an opportunity to buy a lot of used Teletype ASR-32 (Telex)  and ASR-33 (TWX) units but I was too slow/cautious to make a quick offer and I missed out on the deal.

Remex RAR-5200

Recently, I found a New-Old-Stock (NOS) (in the box) Excello-Remex RAR-5200 portable reader punch. It runs at 200 cps (reader) has both a serial and parallel data interface. It should be easy to interface to the LSI-2 serial port. It even has a full tech manual.

I at the same time, I was able to find some NOS mylar 1″ punch tape stock and some standard oiled paper tape as well. I also found a REMEX RRS9000BA1 “Director” paper tape reader for a great price.

I hope to use one or both of these devices to punch and load the standard paper tape software tools library binary images I have from the Internet Archive.

Sadly, I have been unable to find a Computer Automation parallel utility I/O board which was the standard interface for printers and paper tape readers, or a Computer Automation parallel distributed “Pico Processor” I/O module either. [FLASH UPDATE: Check out the December 2020 Christmas post – I found a parallel Utility card! ]

Computer Automation, Inc.

July 5, 2020

Computer Automation, Inc. was a computer manufacturing company originally based in Newport Beach, California that built minicomputers from the 1967 to 1992. The company moved to Irvine, CA in 1973. It was founded by David H. Methvin, who conceived and designed the original PDC-808 minicomputer. This 8-bit computer was sold as a Programmed Digital Controller, because the purchasing departments of some companies made it difficult to buy a “computer.”

Computer Automation PDC-808 (Courtesy Computer History Museum)

Alpha 16 Minicomputer

In 1969, Computer Automation developed the Alpha 8 and Alpha 16 line of 8 and 16 bit minicomputers. The Alpha 16 was very successful, but was a fully discrete logic (DTL/RTL/TTL) minicomputer. The processor, ALU, and registers took up three full 15″ x 17″ circuit boards (with up to 168 ICs per board) and used a lot of power.

The Alpha 16 was possibly the first minicomputer specifically designed for OEMs (Original Equipment Manufacturers) to be used inside another product.

The “Naked Mini” – The computer that’s a component.

Computer Automation created a “naked” minicomputer costing less than $1500 (in quantity) designed for use in embedded systems. The Alpha 16 became the first of the “Naked Mini” products,  a line of cheap, easy-to-use, reliable minicomputers “designed as a dependable component ready to bury into your system.” Engineers eager to “computerize” their products loved it. Their advertising tag line was “The computer that’s a component.

Alpha/LSI Minicomputer

Integrated circuit technology was developing very fast with higher levels of integration on chip. To reduce costs and remain competitive, Computer Automation soon began developing the Alpha/LSI product line.

The Alpha/LSI series of machines used increasingly more MSI and LSI components to reduce size, power, and heat, and increase the speed of the systems. The Alpha/LSI-2 was the first high production member of the family, and was manufactured from 1973 to 1985. It also had a fully micro coded processor on ROMs which enabled a large and powerful instruction set. There were several models released that were successively more powerful.

Programming a 1975 Minicomputer

July 2, 2020

In 1975, the good part about having a computer was that it could do almost anything you can imagine. The bad part is that you had to tell it exactly what to do, and how to do it.

In the early 1970’s. there were no iPhones, no video games, no personal computers, and no microprocessors. In fact, relatively few people had ever seen a computer – let alone used one. If you weren’t in a large business, the government, the military, or a large university it was unlikely that you would have access to or even ever seen a computer. There were mainframe computers (IBM, Sperry, Xerox) and minicomputers (Digital Equipment, Hewlett Packard,  Data General, etc).

In the beginning, in order to get a “programmable” computer to do something, a programmer had to write down a sequence of numeric instruction codes (called “machine language”) that the computer understood and then load them into the computer manually through a control panel or an input device (like a Teletype, paper tape reader, or punched card reader). These programs could be hundreds or thousands of instructions long. Every minute step of a program had to be coded in numbers – often in represented in octal (base 7), hexadecimal (base 16), or binary (base 2) and then loaded before it could be “run”.

That process was so time consuming and error prone that computer engineers soon thought – why not create a simpler code for programming that was easier for programmers to remember and recognize, and then have the computer translate it into the numeric codes it needed? This was called an assembler, that performed several translation steps and converted “symbolic code” into “machine code”. There were still the same number of instructions – one for each machine operation, but it was much easier (relatively speaking) for programmers to use these “mnemonic or symbolic programming language” instructions like LDA SLOC for “Load A Register from location SLOC” instead of  “B2E8 010F”, the hexadecimal code for the same LDA instruction.


0029 0800       ADDREL ENT
002A 1350              LLA 1      SAVE THE SIGN OF A
002B 3201 002D         JOR $+2    AND STICK IT ON BIT 0
002C 0150              IAR
002E 0048              TAX        WHICH IS SAVED IN XREG
002F 1350              LLA 1      MAKE IT COMPATIBLE WITH THE OTHER NO.


Above – a snipet of symbolic assembly code (center columns) and its translation to machine code (2nd & 3rd columns on left)  from the LAMBDA program loader for the Alpha LSI-2 

The mainframe manufacturers had developed “high level languages” – starting with IBM’s development of FORTRAN – to simplify the programming process ,  and allow the same code to be used on different models and even brands of computers. However, most minicomputers did not have the same resources as mainframes. They were limited in memory, processing power, and external storage devices. As a result, very few minicomputers were programmed in high level languages until several years later when costs decreased. In addition, many minicomputers were used for interactive and “real time” applications like communications controllers or process control systems, test system, or scientific measurement systems and high level languages did not have the performance needed for these applications.  That meant most minicomputer programmers used symbolic assembly  language.

In later posts we will discuss how to actually program the Computer Automation Alpha LSI-2 minicomputer.  However, for now I want to show the newest treasure I have found for the project. It is a good copy of Walter J. Weller’s classic textbook – “Assembly Level Programming for Small Computers” – published in 1975.

Walter Weller’s classic textbook

This book was a popular textbook on assembly language programming – and – it was based on the Computer Automation Alpha 16 and LSI-2 minicomputers. All of the assembly code is written for Alpha computers, and the system architecture discussed is primarily that of the Computer Automation computers. I actually used this book to learn to program my first Alpha minicomputer in the mid-1970’s.

Restoration Update – 6-2020

June 22, 2020

It has been several months since I posted on the blog about the progress of my project. I had written a series of posts and information pages, but the site crashed and I lost all that material. It was very disappointing, and I got distracted looking for a better blogging template or platform to use.

I finally found a WordPress template I like – “Hemlock” and I am now blogging again.

I will attempt to recreate the lost pages in this and the next posts, so this post may be quite long.

August 2019 – The purchase

I found an Alpha LSI 2 system for sale at a surplus company in Toronto, Canada on I was able to negotiate a good price and I purchased it. This is the unit featured on the home page and in the chassis photos in the inventory section of this blog.

The chassis was made in the 40th week of 1985, so this is a later system produced when the product was mature.

When it arrived I found it had 4 half option cards, a front panel and a power supply. No CPU, CPU options, or memory cards. Also, because the packing was not great, the weight of the power supply unit had bent the card cage chassis in several places during shipping and the cards were out of their tracks and wedged in the slots.

Cards outside of card cage slots/rails and wedged in position

The damage was not as serious as it seemed. I was able to use a pry bar and large screwdriver to pry the cards out of the card cage without further damaging them, and with hammer and wood block was able to straighten the chassis back to normal.

There were two 53988 32k x 16 bit semiconductor memory RAM half-cards and two custom half-cards with just a small number of TTL 7400-series integrated circuits and two connectors, probably for some type of digital I/O.

Chassis bent in shipping

Finding parts – CPU and option cards

I began to search on for more boards and to my surprise a number of Computer Automation boards began to appear for sale on ebay. Over the years I had seen very few items but recently a large number of components have been on offer.

By careful shopping, I was able to find LSI-2/20 CPU cards, CPU micro code ROM daughter board and CPU/TTY option cards, CORE and RAM memory cards, and a few I/O cards. Of course, I had no idea if any of these cards still worked or would function together.

I also purchased some new card edge connectors for use with the option cards and the I/O boards. It is actually hard to find some solder tab card edge connectors these days – many sizes are no longer stocked, and I had to search several online electronics surplus stores for some of these connectors.

February 2020

The first steps after getting some cards for the machine and straightening the chassis was to test the power supply. Mini-computers from the 1970’s have massive linear power supplies with huge transformers delivering 20-30 amps at 5 and 12 volts DC. They are very heavy – I think the power supply module weighs over 50 lbs.

25 amp Alpha LSI 2 linear DC power supply

I had removed the power supply from the chassis earlier to repair the card cage frame. I carefully inspected it for damage and for bad leaky filter capacitors. Bad capacitors are the bane of vintage computers. They can leak and destroy other parts and circuit boards, or they can fail after decades due to changes in their internal chemistry. They can fail partially, or completely open, or short – possibly destroying the power supply when connected to power.

Alpha LSI 2 linear DC power supply filter capacitors and regulators

However, everything looked great – like new. It is a paradox in vintage computers that older equipment from the 1970’s seldom has failed capacitors, while products from the 1980’s and even 1990’s can have many bad capacitors today – perhaps because the production volume was much higher in later years, the quality or manufacturing technology was cost optimized and not as good.

I built a current limiting rig to ensure that I didn’t burn everything up if there was a short circuit or a bad capacitor. It is simply a 150 watt incandescent light bulb in series with the 110 VAC input line. The bulb’s resistance would limit the current to around 1 amp. By the way – it is hard to find an incandescent light bulb these days! Finally, I found one at Home Depot – it was labeled as a “utility lamp” – it was clear and had a wire filament inside, so I could be sure it was not a CFL (Compact Fluorescent Light) or LED lamp.

I gathered my courage and turned on the power! With the deafening sound of a Boeing 747 accelerating for takeoff, the three 5 inch 110vac fans roared into life. No smoke or sparks!!! The series current limiting lamp glowed dimly indicating a normal “no-load” current flow (since the power supply was not connected to the motherboard) with no big shorts. I quickly turned the power back off with a sigh of relief!

Now I needed to test the power supply output. I had ordered some power resistors to use as a dummy load test so I wouldn’t risk damaging the circuit boards. Two 1 ohm, 1% 25 watt wire-wound resistors would give me 0.5 ohm at 50 watts in parallel. 5 VDC/0.5 ohm = 10 amps and 50 Watts load.

Measuring the voltage under load would provide a good indication of the health of the power supply and regulators. I got rock solid 5 VDC power at 10 amps load. Boy, those resistors get really hot fast when running at their max rating of 25 watts each!

A Mighty Wind

Finally, I needed to fix the fans… They were not bearable to be in the same room with. My iPad sound level meter showed over 80 dB — about the same as standing next to a motorcycle with bad or no mufflers. In contrast the normal room ambient noise was around 35 db.

I thought that probably the bearings were dry and/or bad from age. These fans had 110 CFM airflow rating – the highest available for a 5″ fan. Model ETRI 125 XR. I looked online and they were still being sold new – 40 years later – but for $125.00 each plus shipping. These three fans could cost me more than the computer. And there were no NOS (new old stock) or used ones on that matched this model.

Original ETRI fan and new Noctura PWM fan

Getting creative, I thought about modernization… where do people use high volume fans today – overclocked gaming personal computers! And they want them to be quiet. I looked on and found the perfect match – Noctua NF-F12 iPPC 3000 PWM, 4-Pin, Heavy Duty Cooling Fan with 3000RPM (120mm, Black) . Plus, this is a 12 vdc variable speed fan with PWM speed control, so I could build a simple temperature sensor and set the speed (and noise level) based on the internal temperature of the case. And they cost just $25.95 each with free Prime shipping.

I built a PWM fan speed controller with a $6.00 Arduino Nano clone micro-controller board, and now when I turn on the Alpha LSI, I can work on it without going deaf. And the computer is cool and happy.

March 2020

Since it arrived, I have done a lot of work on the Alpha LSI 2. Besides checking the power supply and replacing the fans, I have installed the LSI/20 CPU with Daughter Card and later the TTY/Autoload CPU option cards which I bought separately on

Checking memory bus waveforms with Computer Automation LSI 2

Since I had collected an assortment of CPU, option, and memory boards from my eBay searches, the next task was to see if anything actually would work in the chassis with the tested power supply and new fans. I tested various combinations of CPU board, daughter board, and memory boards to see if anything just worked. No luck! I got a few front panel lights, but the CPU did not respond to the front panel switches properly and kept jumping into RUN mode whenever I tried to load data into memory.

I took out all the cards again, and realized the card edge connectors might be dirty or not full inserted into the mother board. I used DEOXIT® GOLD to clean the card edge connectors.  It takes some force to plug in the CPU card since it has two edge connectors and is very large (19×24″). I tried re-seating the large cards in the motherboard sockets, and once I got everything properly installed I turned on the power and the front panel lit up completely.

Along with the CPU and daughter card (with the micro-code ROMS), I also installed a 16 bit x 32 K word dynamic RAM memory half card – a later generation that followed the magnetic core memories (4k, 8k, and 16k) that were used through the 1970’s until the price of semi-conductor memory decreased and became competitive with magnetic core memory. I had tried the core memories I have first, but they didn’t work when I was first testing and still trying to get something to run. I haven’t gone back to re-try them now that I have a winning combination of CPU boards that seem to function well.

As part of the start-up testing process, I used my 4 channel digital oscilloscope to analyze the LSI-2 MAXI-BUS signals to see if the CPU and memory access were working. The Computer Automation manuals available online and the physical manuals I had found on eBay were a great help. The standard documentation shows all of the bus wave forms and signals in full detail.

I had to re-learn the front panel procedures to get the manual data input process correct, but finally I was able to write data to memory and then read it back through the front panel console and leds.

Documentation and Software

I have also done a lot of research and have contacted a number of Computer Automation resources. I have located a number of critical manuals and software images through various Internet groups and archives which will allow me to make the system fully functional.

I found a bunch of original printed Computer Automation software manuals for sale on – located of all places in Montana. There was also a full printed LSI-2 Series Minicomputer Handbook with the LSI-2/40 supplementary manual. This later model used a different CPU Board that had additional LSI (large scale integration) chips allowing the CPU Daughter and option board circuits to be included on the main board.

In the search for information, I have found a lot of information on (part of the Internet Archive) . I also was directed to a defunct Computer Automation Museum website (from Sweden) that was partially saved on the Internet Archive. These sites had software (paper tape and disk images) and documents including hardware and software manuals and even schematics for several LSI-2 boards.

The one manual I really needed was one of the most common documents (besides the reference manual) provided with the Computer Automation computers – the standard LSI Software Manual which documented all of the software tools provided on paper tape with every Computer Automation LSI-2 computer. This included full documentation on the symbolic assemblers and object loaders used to create and load programs for the computer, plus the diagnostics and the linkable standard utility libraries.

Unfortunately, this manual was not available in any of the online archives I had found. Without this documentation, it would be very difficult to write and load programs into the system, even if I could get the standard software paper tape images I had found loaded into the computer.

Through suggestions from Al Kossow who runs the project and also works with the Computer History Museum in Mountain View, CA., I posted on a vintage computer mailing list ( and located a collector in Sweden who had this critical documentation on the LSI-2 software tools he had inherited from his uncle, and was kind enough to scan it for me. I was unable to find this standard LSI Software Manual anywhere in archives but he had a copy in Sweden. He also sent me links to a private CAI document archive at Stockholm University with other manuals including several later Operating Systems and info on the LSI-4 series computers.

June 2020 – Starting a Collection…

I have recently acquired two more Alpha LSI 2 Minicomputers. The first one was in Los Angeles – I found it on eBay – it was in rough condition, but with a complete set of boards. The person I purchased it from had bought it at a local estate sale.


This is a very early model of the LSI-2, probably made near the end of 1973 or early 1974, as most of the ICs on the CPU are dated in 7340-50 week range, and the Data Products core memory is date coded 7405.

I found the second LSI-2 while watching a video on YouTube made on Aug 31, 2017 by Curious Marc when he visited a computer collector in the SF bay area. Bob Rosenbloom’s vintage computer collection in Santa Cruz, CA. is one of the largest private collections in the USA. The video shows Bob’s incredible collection of hundreds of vintage computers and terminals.

The first thing I noticed in the video was Bob’s collection of 1970’s minicomputer front panels on the wall of his workshop. He had both the Alpha 16 and Alpha LSI-2 front panels there!

Next, while Marc was panning around I saw a complete Computer Automation LSI-2 system in a rack with two 8″ disk drive systems in the background. It looked like one drive was missing, but otherwise was in good shape. Unfortunately, I found out later that this system had already been sent to the Living Computer Museum in Seattle.

Bob is reducing his collection to focus on just early DEC minicomputers, and so is offering many of the rare items in his collection.

The remaining one I was able to buy from Bob  is in good condition and has just arrived. It even has a pair of 8″ CalComp floppy drives and some spare boards.