Breaking News
Blog

Creating an operating system (OS) in assembly language

Blog
2/7/2016 11:00 AM EST

 57 comments   post a comment
NO RATINGS
View Comments: Newest First | Oldest First | Threaded View
Page 1 / 6   >   >>
Max The Magnificent
User Rank
Author
Re: Vacuum tube NAND
Max The Magnificent   4/18/2016 1:12:26 PM
NO RATINGS
@Ian: ...One day he thought all his Christmases had come at once - he found 44 gallon drums full of 12AT7s, as used in digital circuits - but they all turned out to be dead.

I like a story with a happy ending ... it's a shame this wasn't one LOL

traneus
User Rank
Author
Re: Vacuum tube NAND
traneus   4/18/2016 11:56:10 AM
NO RATINGS
@Max The Magnificent: "Ha! You fell into my cunning trap -- when I come to build my HRRG (Heath Robinson / Rube Gioldberg) Mixed Technology computer, you will be in charge of designing the Vacuum Tube cabinet!"

Only if the vacuum tubes are functional, not merely decorative. Maybe some 6K7GT's as touch sensors that can't be zapped by static electricity: The control grid comes out on the top cap. I need to explore how well this would work.

PS I enjoyed seeing you at ESC Boston, and I enjoyed the meetups. Thanks to all for the work involved in putting on ESC. I won't be coming to ESC Minneapolis or ESC Silicon Valley, as one ESC per year keeps the experience fresh for me. I do have some ideas for a paper for ESC Boston next year; I have to flesh them out to see how well they work.

Ian Johns
User Rank
Author
Re: Vacuum tube NAND
Ian Johns   4/18/2016 10:44:13 AM
NO RATINGS
Here is an example of how to do logic with valves:

http://translate.google.com/translate?hl=en&sl=de&u=http://www.jogis-roehrenbude.de/Leserbriefe/Bruegmann-Digital-Roehren-Clock/Digital-Roehrenuhr.htm&sa=X&oi=translate&resnum=1&ct=result&prev=/search%3Fq%3Dfriedhelm%2Bbruegmann%26hl%3Den%26lr%3D

You could also look up "Pulse, Digital and Switching Waveforms" by Millman & Taub.

This reminds me of a story concerning a friend who grew up near the Weapons Research Establishment (WRE) in South Australia. One day he thought all his Christmases had come at once - he found 44 gallon drums full of 12AT7s, as used in digital circuits - but they all turned out to be dead.

Max The Magnificent
User Rank
Author
Re: Vacuum tube NAND
Max The Magnificent   4/18/2016 10:07:56 AM
NO RATINGS
@Traneus Rex: A NAND gate can be built from one 6AS6 dual-control pentode...

Ha! You fell into my cunning trap -- when I come to build my HRRG (Heath Robinson / Rube Gioldberg) Mixed Technology computer, you will be in charge of designing the Vacuum Tube cabinet!

 

PS It was great seeing you at ESC Boston -- have you submitted a paper for ESC Minneapolis yet?

traneus
User Rank
Author
Vacuum tube NAND
traneus   4/17/2016 8:17:38 PM
NO RATINGS
Clive"Max"Maxfield asked: "Now, here's one for you -- how many vacuum tubes would it take to create each NAND?"

A NAND gate can be built from one 6AS6 dual-control pentode. Unlike normal low-level amplifier pentodes such as the 6AK5, the third grid (suppressor grid) in the 6AS6 has as high a transconductance as the first grid (control grid) has. When both grids are at cathode potential, there is plate current. When either grid is sufficiently negative, there is no plate current.

A pentagrid converter, such as a  6BE6, can be used instead of the 6AS6.

A NOR gate can be built from one dual triode such as a 12AU7. Parallel the cathodes and the plates. Whein either grid is at cathode potential, there is plate current. When both grids are sufficiently negative, there is no plate current.

In all these cases, you need negative-shifting level shifters between tubes. Usually these use neon lamps and/or resistive dividers and a supply negative with respect to the tube cathodes.

DMcCunney
User Rank
Author
Re: EE Times answer
DMcCunney   4/17/2016 5:07:36 PM
NO RATINGS
"the first adopters outside of Thompson's developer group were the folks working on patent applications,"

I believe that word-processing application was the excuse the Unix fellows used to get the suits to fund the project. (It made sense, if you look at the difficulty of preparing the documents on typewriters.)

I don't believe Thompson and colleagues needed to get the suits to fund the initial development - they were creating an environment intended to be better for their work as software developers, which was likely in their group's charter.  But it certainly helped spread Unix through other parts of AT&T.

Of course, the "word processing" application back then would have been a mix of ed for editing and nroff for text formatting when printed.  Unix didn't get vi until UCB picked up early Unix source and forked it to create the BSD variant, and things from BSD like vi and the C shell found their way into AT&T's Unix.

It's hard to understand now how something like ed and nroff was a step up, unless you're an old timer who once did stuff on typewriters.  (And I am, and did way back when before I dove into the computer pool.  I also actually like vi, which gets me funny looks.)

>Dennis

perl_geek
User Rank
Author
Re: EE Times answer
perl_geek   4/17/2016 4:45:02 PM
NO RATINGS
"the first adopters outside of Thompson's developer group were the folks working on patent applications,"

I believe that word-processing application was the excuse the Unix fellows used to get the suits to fund the project. (It made sense, if you look at the difficulty of preparing the documents on typewriters.)

DMcCunney
User Rank
Author
Re: EE Times answer
DMcCunney   4/17/2016 4:04:40 PM
NO RATINGS
@perl_geek: Unix was originally written in Assembler, but rewritten in C so that it could be ported. (I think the transition was from a PDP-7 to an Interdata.)

IIRC, the original machine on which Unix was developed was a PDP-6.  There was one essentially unused at Bell Labs that Ken Thompson and friends could get unlimited time on.  Unix was written in Macro-11, DEC's version of Assembler, through version 6.

At the same time Thompson et al were writing Unix, Dennis Ritchie was creating C, which was designed to be a high level language efficient enough that you didn't have to write in Assembler to do things like OSes.  It was also designed to be portable and relatively easy to bring up on a different architecture.  When it was sufficiently mature, Unix was rewritten in it.  I believe about 80% of Unix back then was in C, but low level stuff that had to talk to the hardware was still in Assembler.

Yes, I believe the first port was to Interdata, thpough I recall Honeywell also being in that mix.

Unix spread rapidly within AT&T, and I recall hearing that the first adopters outside of Thompson's developer group were the folks working on patent applications, because Unix provided a better environment for what they did.

The received religion at the time was that OSes could only be written in Assembler. Unix blew that idea away.

Well, C did, at any rate.


>Dennis

sw guy
User Rank
Author
Re: In praise of assembly language
sw guy   3/31/2016 7:20:04 AM
NO RATINGS
Even when CPU has instructions to operate in memory, you may have same surprise as soon as your system is multi-core.

Max The Magnificent
User Rank
Author
Re: There is an interesting book on this subject
Max The Magnificent   2/26/2016 11:51:37 AM
NO RATINGS
@Elizabeth: Actually, it only takes five NANDs...

Curse you Red Baron!  I should have spotted that -- when I was posting this comment, at the back of my mind I was thinking "you should look for logic reductions" -- but I was in a rush and I just thought about both parts seperatly -- if only I'd giving it a tad more thought I wouldn't have thsi sheepish look on my face now LOL

Page 1 / 6   >   >>
Most Recent Comments

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed