Last day running of a KA10 in 25th year of service

Histories
1

A seven minute video with lots of machine room noise and some blinky lights (“it’s all in the lights”). Also a paper tape mishap. And evidence of hexagonal grid routing of DEC backplanes. And a mouse, and a CAD program.

image
image611×424 34.9 KB

image
image735×486 63.6 KB

The CAD program shown in the middle of the video was written by Bryan Jensen

From the video description:

the last day of the KA10’s operation on 6-Feb-1994. It was in constant use for over 25 years and was the last one still running in the world. I worked on it in the 1970’s when I was a regional support engineer for Digital (the problem was with one of the MD10 memory banks to the left of the KA10). I had the honor of shutting it down for the last time. The CAD program shown in the middle of the video was written by Bryan Jensen, who makes an unexpected cameo appearance after the paper tape mishap! Bryan also maintained the KA10.
– Jim Kinsman

Also in the description, a link to Michael Thompson’s catalogue:

This machine appears there as:

KA10 Serial # 46 TOPS-10 Penn State University, University Park, PA
Purchased in 1968/1969 time frame for use with EIA-680 Analog Computer for hybrid computation pair. The facility was the Hybrid Computation Lab which was part of the Electrical Engineering Department at the Penn State University Park campus. The system was located in the EE West building. The Hybrid Computer Lab was removed from the EE Department and became a College-wide facility named the Engineering Computer Lab which supported general computer operations in the College. At one time there was the DEC-10, DEC-2020, VAX 11/780, VAX 11/785, VAX 8550, Harris H800, and two IBM 4341 systems along with multiple VS3100 Workstations and assorted MicroVAX systems. When the centralized computing facility was phased out (CEDCC formed) in the 1995/1996 timeframe the DEC-10 and all other systems were scrapped

via @larsbrinkhoff on PiDP10 mail list

1 Like
2

I’m still in awe that these point-wired backpanes could be made functional. Testing, and tracing and fixing any errors must have been “fun”.
Does anyone have an idea what an average production time for those had been?

3

For best wire-wrap, you use a robot:
Gardner Denver wire wrap machine

image
image600×601 86.6 KB

It says here

Top speeds on horizontal machines were generally around 500-600 wires per hour, while the vertical machines could reach rates as high as 1200 per hour, depending on board quality and wiring configurations.

Note that wire-wrap works by pressure welding on square posts: three turns makes twelve welds, electrically in parallel, so it’s a very reliable approach.

3 Likes
4

Since the wire-wrapped backplanes were computer automated (pdp8?), and all else was pre-made PCB’s. I am guessing few days to wire up and build the chassis and few days for basic testing
assuming all parts were in stock. Finding time in the production line might be another matter.

1 Like