The subject of this article is about a specific machine currently owned by Erik Klein. To clarify, this article is actually not very related to the Computer Museum of America, except to analyze how an electrical engineer named "Tom Crosley" was able to make the Kenbak‑1 a more usable machine in 1974, relevant to "the story of the Kenbak‑1" in general.
In the March 1974 edition of the Amateur Computer Society Newsletter, Tom Crosley wrote:
"My computer is basically a Kenbak‑1, to which I have been making additions. Right now I am working on the memory addition; then will come the CRT, second TTY, and mag tape. Even with only the original 256 words of memory, I have enjoyed writing programs for the Kenbak-l, such as a line-by-line text editor tor the TTY; games (e.g., Nim) and a Turing machine simulator. After adding the 2K of memory, allowing much larger programs, I still don't feel the slow cycle time will be a drawback; since most of my applications will be TTY-oriented, most of my programs will still be I/O-bound. I would be interested in corresponding with anyone else who has a Kenbak-l and has made or is thinking of making additions to it."
In June 1974, Tom Crosley was published again:
"To the baslc Kenbak‑1, so far have added interrupt system plus real-time clock (1 second interrupt ); am just completing a full duplex TTY controller (at first I used serial I/O for the TTY); am adding a paging register to select one of (initially) 16 128-word pages (only in effect for addresses 204-376; the lower addresses - and all registers in memory - will be available independent of the page register). Am making use of the "don't care" bits of the NOP instruction to add 15 I/O instructions which will be single word (data set up in registers)."
Tom will be using a TV Typewriter as a CRT terminal, and two mag-tape transports.
Per www.kenbak.com, Crosley bought his machine from an advertisement in Scientific American in 1971 "It was a major purchase -- $750 [...] I was only making about $11,000 a year." In communications with Crosley, Thomas Jones (of kenbak.com) was also told:
"Crosley combined his electrical and computer skills by building an ASR-33 teletype interface for his Kenbak‑1. This solved the 17 year long mystery of why Erik Klein's computer had some an extra integrated circuit soldered into the usually empty IC99 spot, and had extra wires added. Tom says he could load in a short bootstrap loader, and then could automatically load in programs via the ASR-33's keyboard or paper tape."
Crosley's surviving machine was sold on eBay in 2004https://www.kenbak.com/provenancedetailshttps://forum.vcfed.org/index.php?threads/what-to-do-with-my-kenbak-1.671/post-671. It was bought by Erik Klein, who went on to post the first detailed information about the Kenbak‑1, on his website. From this, we have several detailed photos of the computer, which I've edited to correct perspective and provide an interactive overlay of the front and back of the board below:
It is possible to cross reference these against the gerber files created by Thomas Jones.
After staring for long enough, the following can be noted:
Crosley's Kenbak is the only one documented to have its slot cover missing, it can be imagined that their teletype cabling was fed through the unused "punch card" slot instead.
In my opinion, a memory expansion circuit would not have been possible, excluding that banked memory possibly could have been done. The machine code language of the Kenbak‑1 asserts only 256 addressable bytes. If Crosley had redesigned the Kenbak‑1 into essentially a new computer (with incompatible machine code language), using his original Kenbak‑1 as a base, then his desire to expanded the memory to 2K would be more plausible, but this would have minimally involved adjusting the length of certain instructions to accept 2-byte memory locations, and also adjusting the behaviour of the program counter to match. As mentioned by Crosley in 1974, memory access would have also been slower, if each additional connected shift register in the amended design had to be cycled through in order to access a specific bit of memory.
There is no visible indication that the RAM expansion got very far (or at least no attempt seems to have been made to install such an expansion). There are many signs of rework on the board, but the 1404A shift registers appear to either have been not reworked, or reworked to the same quality as originally done by John Blankenbaker. There is some excess solder on both connected pin 6s of the 1404A chips, but that pin is labeled NC for "Not Connected" in the available data sheet of the AM1404A. If a wire did used to be connected here, it could only have been to steal ground.
In my analysis of the "Crosley" Kenbak‑1 photos posted on vintage-computer.com, I estimate that (involving the strangely-present IC99), and output achieved by . , there are some stray wires still attached in some places, but it seems an attempt was made to return the historic machine to its original condition.