Google-apper
Hovedmeny

Post a Comment On: Ken Shirriff's blog

"Inside the vintage 74181 ALU chip: how it works and why it's so strange"

10 Comments -

1 – 10 of 10
Anonymous Anonymous said...

I seem to remember some similar stuff in the high loop of the IFR service monitor, the 1200, 500 and 1500 had the same one I think. This was a unit that generated RF in coarse steps between 2 and 3 GHz, iirc, and was implemented as two boards in a module packed with ECL ICs. I'd never seen ECL before and if i have since don't remember it.

That would have been a box you would have loved, the IFR service monitor.

March 19, 2017 at 9:45 PM

Anonymous Len said...

It looks like this ALU (or related) was used in at least one arcade machine in 1980 - as part of a 12-bit processor - by a company named Cinematronics:

https://ia800503.us.archive.org/3/items/ArcadeGameManualRipoff/ripoff.pdf

March 20, 2017 at 9:56 AM

Blogger Unknown said...

There is another explanation of the '181 here:
http://web.eecs.umich.edu/~jhayes/iscas.restore/74181.html

March 20, 2017 at 4:52 PM

Anonymous Anonymous said...

Why do s0 and s1 seem backwards? They are in the standard order they should be, counting up in binary.

Daniel

March 20, 2017 at 6:13 PM

Blogger Unknown said...

For your discussion of carry lookahead, I'm assuming C0 = Cin. Is that right?

March 27, 2017 at 8:12 AM

Blogger Joe Zbiciak said...

A later TI part, the TMS320C80 DSP, implemented a 3-operand boolean unit with selectable carry chain. You could provide an arbitrary 3-operand truth table (8 bits), along with optionally selecting in the carry chain. This allowed for even more interesting hybrid operations such as (A + B) AND C.

Granted, the C80 was not a standalone ALU, and is quite a bit more obscure I imagine. But, it's the first thing I thought of when you started listing some of the curious functions the 74181 offers.

May 30, 2017 at 11:35 AM

Blogger Unknown said...

Hey Ken!
Thanks for the great write-up! I've spent some time duplicating the block diagram with individual logic gates and have built up a couple of prototypes! let me know if you'd like for me to send you one!

https://hackaday.io/project/25596-mega-one-8-one

thanks
Dave

June 22, 2017 at 5:21 PM

Blogger Tony Cox said...

Comparing to the '181 datasheet (Nat Semi in my case) the logic diagram has the P and G outputs swapped compared to your lovely interactive graphic version. Which one is correct?

November 22, 2017 at 1:34 PM

Blogger Paul said...

Hi Ken,

Great blog..brings me back quite a few years.

I can state with authority that the Prime (spelled Pr1me) computers that were TTL all used the 181 (what else would you do?)

That would be the P400, P750, P850 primarily.

The P9950 (which I worked on) was the first machine built from (10K) ECL and used the 10181. Fairly soon the LSI level bumped up and I never used them again.

-paul

August 15, 2018 at 1:18 PM

Blogger Amaury said...

hi Ken,

this chip, with the 182, should be done in that google thing, probably in the fastest TTL-compatible process that can be manufactured (act ?)

https://fossi-foundation.org/2020/06/30/skywater-pdk

August 17, 2020 at 1:31 PM

You can use some HTML tags, such as <b>, <i>, <a>

Comment moderation has been enabled. All comments must be approved by the blog author.

You will be asked to sign in after submitting your comment.
Please prove you're not a robot