4

I am reading over the history of the Nova, which claims that a key moment in its creation was a 4-bit ALU which was shown to them by their Fairchild sales rep. I assume this refers to the 9340? But other sources claim the Nova used the 74181, which was from TI.

I believe this bit of confusion is due to the 9340 being a licensed 74181? I see mentions of the 9341 being "the equivalent" to the 74181, but I'm not sure if the last digit being 0 or 1 is meaningful.

But now I see notes that the early Novas did not use ALUs at all, but were based on simpler adders and such, like the 7483. It was only the SuperNova that used more complex designs, apparently the Signetics 8260. However, spec sheets suggest the 8260 is a 74181, so I am again confused.

Does anyone have a better source on all of this with more detail?

Maury Markowitz
  • 19,803
  • 1
  • 47
  • 138
  • 2
    As mentioned here (Footnote 1), most manufacturers used their own numbering scheme, leading to great popularity of equivalence lists. After all, these numbers are order numbers in their inventory system, not names - despite engineers using them name like. – Raffzahn Jan 03 '21 at 14:16
  • 4
    "I am reading over the history of the Nova, which claims..." - where are you reading this? – Bruce Abbott Jan 03 '21 at 20:57
  • 1
    Re, "...early Novas did not use ALUs..." I guess that depends on what "ALU" means. I am pretty certain that the term "ALU" was in wide-spread use well before any '181-style chip was invented. The ALU is that part of the CPU that performs combinitorial operations on instruction operands. – Solomon Slow Jan 04 '21 at 21:24
  • @SolomonSlow - Absolutely; and when "a CPU" is several cabinets, more people are keenly aware that "the CPU" is actually made up of several different units. In particular, "control" and "arithmetic" are very obviously different. – dave Jan 05 '21 at 15:54
  • 2
    Having said that, many British computers did not have an ALU. They had a "mill" instead. – dave Jan 05 '21 at 15:56
  • Ok the article has been updated with the information below plus a few new sources. The confusion appears to be due to the 9300 series being used but not the 9341 specifically. In fact, I'm not sure they ever used a 9341 even in later models. – Maury Markowitz Jan 06 '21 at 19:54

2 Answers2

5

The 9341, 54181 and 74181 are indeed functionally identical. This is from a contemporary Fairchild databook:

Scan from Fairchild databook (taken from https://apollo181.wixsite.com/apollo181/about)

Looking through the parts list in the Nova maintenance manual posted by another-dave, the original 1969 Nova doesn't appear to use it, though. On the other hand, a 74181 is clearly visible center-right in this photo of a slightly later Nova 1200 board:

partial enlargement of the Wikipedia photo of a Nova 1200 board

Michael Graf
  • 10,030
  • 2
  • 36
  • 54
  • The link above notes that the 181's appeared on the 1200 series. So the question then is what did the original models use? I do know that Fairchild was working on a 4-bit ALU that was NOT released, perhaps this is what inspired the design even though it was not used. – Maury Markowitz Jan 05 '21 at 14:02
  • I guess they built the functionality using less integrated ICs. As you can see from the circuit diagram in the databook, the '181 isn't that complicated. – Michael Graf Jan 05 '21 at 14:53
2

In the hope that this helps (but I'm a programmer so this is all greek to me):

Nova maintenance manual has what seems to be a complete parts list in Appendix B.

SuperNova Technical Manual has schematics and parts list near the end.

dave
  • 35,301
  • 3
  • 80
  • 160