Nielsen3

Nielsen3 / Gray Case

This is an interesting Kenbak-1. Or at least it seems that way as it's been studied more than some others. It seems like sort of a "Frankenstein" pieced together from discordant parts. The internals show it to be a very early "beta-prototype" Kenbak-1, but at some point it was put in an unpainted gray bottom case. And an unique "CTI Serial Number" sticker was put on the back, which has an incorrect model number. No other examples of a CTI Serial Number sticker exists.

Only two of these "Rev A" earlier printed circuit boards have been discovered, and neither were ever sold by Blankenbaker to a customer.

This computer was apparently included in the big shipment of computers and parts that was shipped to CTI when Kenbak Corporation closed in 1973. Robert Nielsen purchased it at auction when CTI went bankrupt, and used it in his electronics trade school, until he sold to Herbert Eisengruber for his Nova Scotia Computer Museum.

Front/Top

3608 x 1501

Back/Top

4032 x 3024

Front Top

4032 x 3024

Gray Bottom Case

3018 x 1779

Bare Top Case

4032 x 3024

Bare Bottom Case

4032 x 3024

Component Side of PC Board

4032 x 3024

Bottom side of PC Board

4032 x 3024

Backlit bottom of PC Board.

4032 x 3024

Closeup Matrix Photo of Top of Printed Circuit Board

This will only probably be visible on a wide screen monitor, not a phone screen. Not sure why left-most photos are yellow.

Top/Front Running

Many remaining original Kenbak-1's are not running. I think only 4 are still running.

Closeups of interesting or unique parts.

Blowup of Clock Circuit

3024 x 4032

The two dark brown mylar capacitors on the left (next to the orange electrolytic capacitor sticking off the board) set the clock rate in the multivibrator circuit. This computer has a clock speed near 1.6 MHz, probably due to the choice of capacitors. The first couple computers Blankenbaker made (prototype, and the two Rev-A computers) used different capacitor combinations, but the "Rev B" production computers all used small orange ceramic disc capacitors, producing a standard 1 MHz clock rate.

The Mylar Clock Capacitors

3024 x 4032

This may be why this particular Kenbak-1 has a faster than usual clock rate. These brown Mylar capacitors are different. There was one in the first prototype, two in this computer, and all other capacitors in clock circuits were ceramic disc capacitors. Production computers used 0.0015 microfarad capacitors. The code "471" would signify 0.00047 microfarads, making the clock faster. The code "103" would signify 0.01 microfarads which would make the clock slower, so that can't be the value code. The resistors and transistors are the same, so breadboarding the components to check clock may be confirmation.

Shift Registers

4032 x 3024

Power Supply

4032 x 3024

Power Supply

4032 x 3024

Power Supply without Capacitor

4032 x 3024

The white rectangles were high power current shunt resistors to measure current.

Checking large Capacitor

4032 x 3024

Measuring the 50 year old capacitor

Power Supply Transistors

4032 x 3024

Multiple washers used as standoff between heat sink and pc board

3024 x 4032

The shift register memory chips

IC55 leg tie

1769 x 887

Several IC legs were pulled out of the socket, and tied to the leg next to it. I suspect this was an input which was not connected, but needed to be connected to an output. Early TTL IC's would oscillate or draw heavy power if an input was left disconnected.

IC64 leg tie

1370 x 775

IC123 leg tie

990 x 603

Some Curious Details about this Unusual computer.

The Gray Case Bottom?

The bottom half of the case is gray, not the usual blue. I asked John Blankenbaker in 2013 if he remembered a gray bottom case, and he didn't recall it, but said "I don't remember about the gray bottom on the case. That was the natural color of the case as purchased on the shelf from Bud. The first two computers that I shipped to customers were hand crafted. I modified the case myself. Later, the units shipped by Bud were painted with the necessary holes in them." The first owner, Robert Nielsen, says he was told they "ran out of blue cases."

CTI Serial Number 001?

While this is a Kenbak-1 labeled machine, while owned by CTI someone put on a strange CTI serial number sticker saying "Serial No. 001". No other machines, Kenbak-1 or CTI branded, had a CTI serial number sticker. Maybe CTI was just experimenting with how to re-label the Kenbak-1 to re-brand it for mass production. This serial number sticker also calls it "Model 5005" which should say "Model 5050." Maybe this sticker was just an experiment by the CTI team trying to decide how to rebrand the computers to their own product, but then decided against using these stickers.

Inside the Case, things get More Interesting:

Looking inside, it was noticed that the printed circuit board is marked "Kenbak 10000 Rev A." This is different. No other Revision A board had ever been noted before this. The prototype computer was printed with "Rev " without any letter after it, and all the other production machines said "Rev B." Most people thought Blankenbaker went directly from his Prototype to the production machines, and even called the prototype a "Revision A" board, even though it didn't include any letter. But now it's apparent that Blankebaker actually created a middle version, sort of a "Beta-prototype" after the first prototype was made, and before the production machines (Revision B) were made.

Since this computer was noted to have a "Rev A" printed circuit board, it was discovered that the computer in the Computer History Museum in Mountain View California, Serial 216, also has a "Rev A" circuit board. These are the only two computers with this "Rev A" board found, and interestingly, neither one of them was ever sold by John Blankenbaker. The Serial #216 was kept byJohn Blankenbaker in his attic until 1986, when he needed to send a working computer to Boston's The Computer Museum, and this Nielsen3 was never sold, it was shipped off with all the parts and partially completed computers to CTI, when Kenbak Corporation folded.

"Rev " prototype

"Rev A" beta-prototypes?

"Rev B" production computers

Can we Verify when these "Revision A" boards were Made?

Luckily, John Blankenbaker didn't buy all of his integrated circuits in one large purchase. He made a purchase before making the prototype, then purchased some more before the Revision A computers, then purchased more before the Production Revision B computers. Looking at date codes, we can see evidence of that. Through high resolution photos of several computers we can read the date codes on TTL integrated circuits. This is detailed in the Comparison of Original Kenbak-1 Computers page, but here's a bit of that information:

Prototype computer: Latest Date Code 7103 (mid-January, 1971)
Nielsen3 Revision A: Latest Date Code 7129 (mid-July, 1971)
Serial #216 Revision A: Latest Date Code 7129 (mid-July, 1971)
Serial #183 Revision B: Latest Date Code 7143 (late-September, 1971)
Nielsen1 and 7 Revision B: Latest Date Code 7337 (mid-September, 1973, but some may be 1974 dated. assembled after Kenbak folded)

This is good evidence the two "Revision A" machines were after the first prototype, but before the production machines.

John Blankenbaker probably never wanted to sell these "imperfect" "Revision A" machines. He only sold the perfect "Revision B" boards as he was a bit of a perfectionist. An early development revision of a product, after the first prototype, but not ready for production, is often called a "Beta Prototype" and that's a reasonable name for the "Revision A" computers.

Time to Ask John Blankenbaker about this "Rev A" PC Board:

The "Rev A" circuit board was noted around 2013, but the mystery didn't really start to "ache" until 2022, when I finally decided it's time to ask John. It's been a long time, 51 years. Would he remember how many different "tries" it took to get the PC board right? And how many of these "Rev A" beta-prototype circuit boards did he make? Did he order just two of the second-try circuit boards, or more? Unfortunately, he couldn't recall these details. Too much time had passed. These details, indeed, are apparently lost to history.

So how did this early "Beta Prototype" get the Gray Case?

It is easy to switch cases on a Kenbak-1. Just a few screws, and the PC board and front panel lifts out. The power supply can easily be swapped out - it only requires a few wires to be desoldered. But once the PC board is soldered to the front panel, it's very hard to remove. Removing it would require extensive desoldering and resoldering of many wires in a delicate small area. And removing the front panel shouldn't ever be necessary, since any malfunction of front panel switches or lamps would be easy to replace just that part. So was this "marriage" between the gray case and the circuitry made by Blankenbaker from the start, or later. A hint lies in some "blue-paint" on the front panels.

Blue Paint suggests Case was Changed

Here is a close-up of the back side of the front panel. There is clearly blue paint on the front panel, right where it would touch a bottom case. It's unknown if this is because it was attached to a freshly painted blue-case (John painted the first few cases blue himself) or more likely because glue from the switches ran down the panel, sticking the painted panel to the bottom case. Regardless, it's a safe bet that this computer once had a fully blue case. Perhaps they needed all the blue cases for the important "Production Computers" to sell.

One Last Interesting Finding:

In examination of the Nielsen3 computer, one other interesting thing was noted. There is a curious cut of a printed circuit tracing, which was then reconnected by soldering in a jumper. This is something any electronic designer will recognize as an attempt at debugging a non-working circuit, where the connections are in doubt, but the location of the error is not known. This is not something done when a previously working circuit stops working, due to a component burning out or a connection going bad. It's only done the first time, when trying to figure out why a new printed circuit board is not working. This suggests John Blankenbaker was debugging the printed circuit design, trying to fix something, when he wasn't quite sure where the problem is.

In this photo, underneath the horizontal white wire, is a printed circuit trace, which is cut in two places. The cuts are highlighted with the red arrows. White scratches are visible on the PC board substrate.

But after cutting this, the maker realized that connection was not the problem, so a short piece of wire was soldered to the two sides, fixing the signal line to how it previously connected.

This is not a procedure which would be done in just fixing a circuit which previously worked, and stopped working. This is a procedure that is done after a new printed circuit board is made but doesn't work. And once one printed circuit board is made working, the correcting changes can be made directly to the following boards, without having to make a change, and undoing that change.

So to repeat, if more than one "Revision A" board was made, it seems probably that this board was the first one John Blankenbaker got working, as it shows signs of back and forth debugging changes. Once this circuit board was working, later boards would be fixed with simple cut traces in some spots, and jumpers in other spots, but would not have a trace which is both cut, and then jumped back to it's original state.

But this can't be known with certainty until the other known "Revision A" computer (currently at the Computer History Museum) has the bottom of the PC board examined. High resolution photos of the component side of the board exists, but no photos show the bottom of the circuit board.

Google Sites

Report abuse