From wikipedia: "From 1956 with a retail price of $47,000—equivalent to about $408,000 today... a desk computer ... approximately 740 pounds (340 kg) and was mounted on sturdy casters" and "The single address instruction set had only 16 commands. Not only was the main memory on magnetic drum, but so were the CPU registers, timing information and the master bit clock, each on a dedicated track. The number of vacuum tubes were kept to a minimum by using solid-state diode logic, a bit-serial architecture and multiple usage of each of the 15 flip-flops." - https://en.wikipedia.org/wiki/LGP-30

(Note that this is the very machine behind The Story of Mel - http://www.cs.utah.edu/~elb/folklore/mel.html and http://en.wikipedia.org/wiki/The_Story_of_Mel)

Anyone up to tackling this with a CPLD and a serial EEPROM?? It can run ALGOL!

A scan of the programming manual of the Librascope General Purpose 30:
http://ed-thelen.org/comp-hist/lgp-30-man.html
"For fifty seconds the STANDBY light is on to indicate that tubes are at half filament power. For fifty more seconds the STANDBY TO OPERATE light is on indicating that the tubes are at full filament power and that the drum motor is energized. When the operate light comes on, the D.C. voltages are applied and the computer is ready for use."
Via https://plus.google.com/u/0/+AlexandreK ... iaqYfVzCMQ



(20Kbit EEPROM in a 3-pin package: http://www.digikey.com/product-detail/e ... ND/1768308)

Great find. Thanks for posting about it.

I do have a soft spot for serial architectures; they represent some fine engineering. Similar to the MiniCPU-S project I started some time ago, but haven't returned to finish.

May have to consider the challenge you posed. Will do some more reading on the subject and consider an approach.

Looks like Oscar Vermeulen (of PiDP-8[*] fame) is now some way to making an LGP-30 replica - perhaps this too will be sold as a kit.
http://obsolescenceguaranteed.blogspot. ... gp-30.html

Indeed, Oscar mentions Christopher Bachman from randomvariations.com is working on a CPLD implementation of an LGP-30!

[*] (Over the weekend we soldered up our jointly-owned PiDP-8 and got some life out of it, although it needs another visit to debug.)

Fascinating... thank you, Ed.

Wasn't the Kenbak-1 a "bit-serial architecture", too?

Anyone have any links to bit-serial explanations/examples that a non-EE like myself might be able to grasp and understand?

TIA... Cheerful regards, Mike

Looks like you're right about the Kenbak-1. Many early machines were bit-serial as it uses much less logic - a one-bit ALU operates on any word length one bit at a time. The words pass through a bit at a time. A major cycle involves many clock ticks: one for each bit in a word and then usually one or two more for housekeeping.
See http://www.kenbakkit.com/
and see this talk transcript and the links within: http://www.computerhistory.org/atchm/in ... nkenbaker/

(Indeed there was a cost-reduced PDP-8 model which was bit serial - it was somewhat cheaper and a lot slower, and not very successful)

Not sure what to point you at for learning - possibly the materials in that Kenbak- kit link will help?