Comparison of Original Kenbak-1 Computers

Much can be learned by internal exam of the original Kenbak-1 computers. Design changes and integrated circuit date codes tell us how the design evolved, and when computers were made. The following becomes apparent:

The first prototype computer was quite different, with many circuit modifications and corrections, and IC's made early 1971.
The two Revision A computers had many fixes in the PC boards, and were made from components from mid 1971.
The Production "Revision B" computers had later IC date codes from late 1971,
Blankenbaker made some minor changes to a few resistors, such as pull-up resistors as circuit revisions evolved.
Blankenbaker also experimented with different capacitors on the clock circuit between the 3 different revisions, but standardized components for the "Rev B" production computers.
One computer with a CTI labeled front panel contained a much earlier computer with earlier date codes, and serial number 203 on the PC board. They must have taken an earlier Blankenship assembled computer, but switched out the front panel to sell as their own.
Some computers with the CTI front panel contained different black capacitors, different wires, and late 1973 parts, so were apparently assembled by CTI after they purchased production rights to the Kenbak-1 in 1973.
A proposed algorithm for determining order of assembly is as follows:

The first prototype was definitely made first, in early 1971.
The two "Rev A" (beta prototype) computers were made next, in mid to late 1971, with hints Nielsen3 could have been assembled and debugged first.
Most production "Rev B" computers had serial numbers on their PC boards and this serial number likely denotes the order they were assembled (except for serial number 216, which is actually an early "Rev A" beta prototype.) Any CTI labeled machine with an early parts and Kenbak-style stamped serial number falls into this category, as it likely was just a rebranded later by CTI (like Nielsen8.)
The CTI front panel machines with IC date codes from mid 1973 and the black electrolytic capacitors were likely assembled by CTI after Kenbak Corp. closed down.
The later CTI assembled machines are difficult to order, but at least one has the serial number "501" stamped on the circuit board, denoting a new 500-series serial number started by CTI. If others are found to have 500-series serial numbers, that may denote order or assembly.

Comparison of Printed Circuit Boards of Extant Original Kenbak-1 Computers.

Click on the arrows to stop transition of the gallery above.

Evolution of Printed Circuit Board Revisions:

The prototype had a very different board design, and even the power supply board was different. It not only required a lot of rewiring (jumpers and trace cuts), but several components (a transistor, a capacitor, and a couple diodes) had to be added above the PC board to get it operating. The prototype's power supply was also different, it outputted three voltages, +5 volts, -12 volts, and -5 volts. Later revision power supplies did not supply the -5 volts, and a 5 volt Zener diode and stepdown resistor was placed onto the logic board to generate the negative 5 volts right on the logic board.

After this first prototype was working, before moving directly to the production boards, we see that Blankenbaker made a revised PC board which was labeled "Rev A" which had most of the errors and problems corrected. It was greatly improved, but still not quite perfect - it required a couple jumpers and modifications to work. Only two of these "Rev A" computers have so far been found, and neither one of them was sold by John Blankenbaker, who probably only wanted to sell the "perfect" final "Rev B" boards. One of these "Rev A" computers was apparently shipped to CTI Educational Products in South Carolina, as part of the large hoard of old parts and partially completed or unsold computers, and the other computer was the one Blankenbaker kept in his attic until 1986, when he needed to to send a working machine to Boston's The Computer Museum for the early PC contest. So it is likely Blankenbaker didn't think the "almost perfect" "Rev A" board was good enough to sell to customers, even though it functioned well. It seems to be a "second-try" or "Beta Prototype."

Evolution of Component Styles and Dates:

While interesting changes of components are visible in different computers, the date codes on the TTL chips give objective evidence of age. The latest date code of any IC gives a earliest date the board could have been made (unless a failed IC was replaced, which happened in some of the Nielson computers while used in a technical/electronics school.) Examining these date codes, along with PC board revisions, we deduce a few facts. First, it seems that John Blankenbaker didn't just buy all the IC parts as a lot of 50 all at once. There does seem to be earlier IC's clustered in the earlier computers, and older IC's in presumably later made computers. Second it's clear that some of the CTI computers have IC's as late as August 1973, after Blankenbaker sold out to CTI. This suggests CTI purchased their own IC's, and assembled some computers themselves, after they took over the product. And while they wanted to rebrand the computers from the "Kenbak-1" name to the "CTI-5050" name, it's unlikely that they made any of their own PC boards. All the production PC boards had the "Kenbak" name right on the board, and it would have been very easy to change the "Kenbak" name on the mylar sheets. I suspect they didn't produce any additional PC boards, and just used the remaining PC boards Blankenbaker supplied to them.

The CTI machines also used different electrolytic capacitors, black ones, rather than the orange used by Blankenbaker.

Table of Differences and Hypothesized Order of Production:

Taking into account the known information about PCB revision and IC date codes, and other differences, we can piece together a suspected order of production. Unfortunately, most of the extant computers lack adequate internal photos to complete this table. It seems likely that the serial number sequence does represent order of manufacture for almost all computers, except serial number 216, because this computer was definitely one of the earliest prototypes, the PC board was not stamped with the same serial number, and it's likely John Blankenbaker just attached the last serial number label (50th in his sequence) to the case just prior to sending it to Boston's The Computer Museum in 1986.

Google Sites

Report abuse