[GreenKeys] 8-Level TD?

Fri Nov 22 14:14:42 EST 2013

Javier,

Wow, it sounds like you and I are at a very similar stage of design with
our machines. I’ll shoot you an email off line with my arrangement. I
hadn’t honestly considered a 6-bit databus, but that’s probably because I
didn’t seriously consider having any text-based I/O early on so it wasn’t
that much of a concern. At this stage I am committed to an 8-bit bus,
however, since it lends itself very well to my orthogonal instruction set
which was designed to do as much with as little as possible. Plus I have a
nice 8-level bi-directional paper tape reader, and an ASR-33 to serve as
secondary input and punch. 8-bits also lets me specify source and
destination registers in a single byte, which is important with such a
sloooow machine. And I can theoretically carry on an ASCII conversation
with other peripherals.

I’ve also got a KSR 35 (which needs a clean-up). Perfect if I want to set
up a timesharing system with it, haha. Now to lick this serial business.
-Cory

On Fri, Nov 22, 2013 at 1:12 PM, Javier Albinarrate
<javier at albinarrate.com>wrote:

>  I should have reread the whole thing hahah was thinking faster than
> reading :)
>
> You DO HAVE an ASR33....  I misread...
>
> So you're fine with the 8 bits architecture then (although you'll need a
> lot more relays hahaha)... that's not my case :)
>
> Regards!
> Javier
>
>
>
> On 22/11/2013 15:53, Javier Albinarrate wrote:
>
> Hi Cory!
>
> Well I already spammed your mailbox with all the pesky details of my
> project :)
> In any case, for the records I suggest the following approach instead.
>
> Have lemons? make lemonade
> Forget 8 bits, who needs them? Use 6 bits with 5 bits I/O.
>
> 1- 6 bits architecture: 6 bits data 12 bits address
> 2- 6 bits instruction set with the following characteristics:
>     - You need an extremely well thought instruction set, as baudot is not
> very binary friendly
>     - Avoid special characters, like LTRS & FIGS
>     - Most common actions in the LTRs range (single word because you use
> no shift)
>     - Jumps and not so often instructions in the FIGs range (2 words)
>     - Optional extended set (using just 1 instruction in the FIGS range,
> extends another 8 instructions with an extra decoder)
>     - You need to think in detail each instruction to make it fit in the
> available ranges, and you need to remember that you can use any bit order
> for decoding, it is arbitrary (you need to tame baudot)
> 3- The brain
>     - A resettable sequencer of 10 steps
>     - A couple of muxes and demuxes, for instructions and addresses
>     - A matrix of sockets, where you plug diodes 1N4007. Voilá... you have
> microcoding. That is the core of the design, because it makes everything
> EASY and FLEXIBLE.
> 4- IO:
>     - Dulpex current loop for IO to the TTY in 5 bits. (otherwise with
> everything in a single loop it is a mess to control)
>     - 6 bits buffer out and another 6 bits buffer in using relays.
>     - You need to keep the SHIFT status and eventually insert / remove the
> LTRS and FIGS. This is just a little logic block.
>     - Option A for sequencing: fast reed relays. Option B: Dekatron with
> vacuum tubes (lots of triodes there, probably cheap russian submini triodes
> would do it).
> 5- RAM:
>     - Cheap option? Solid state
>     - Bulky option?  Latched relays (I am lucky, I have 1k of them)
>     - Weird option? Neons (many cheap russian IN2 and quite a few tiny
> russian thyratrons)
>     - Magnetic option? Core. You can get loads of ferrite cores for no
> money, but it will require a lot of solid state to drive and address.
>
> Regards!
>
> Javier
> LU8AJA
>
>
>
>
> On 22/11/2013 15:15, Cory Heisterkamp wrote:
>
>  Hey Guys,
>
>
>
> Looking for some brainstorming here. For the last couple years I’ve been
> planning on building a computer entirely out of relays, with minimal
> support from vacuum tubes, Dekatrons, etc. But no transistors or IC’s
> allowed! I have most of the design worked out and plan on using my ASR33
> for the terminal. Life would be so much simpler if the communication were
> parallel, but that feels like a bit of a cheat (I see why Flexowriters were
> so popular back in the day for computer work). I have a Model 28
> Receiving-Selector which was Teletype’s version of a serial-to-parallel
> converter. Did they ever make an 8-level version?
>
>
>
> My idea is this… Basically, my output buffer holds a byte at a time in 8
> relay latches. When ready to send my data out to the world, the holding
> magnet of the distributor could release, the distributor would effectively
> “sample” the state of each of the 8 latches, and as it returned to the home
> position, trips the reset line on the buffer to clear it and it’s ready for
> the next byte to be loaded.
>
>
>
> The trick is finding an 8-level piece of equipment to start from. Any
> thoughts?  -Cory
>
>
>
>
> ______________________________________________________________
> GreenKeys mailing list
> Home: http://mailman.qth.net/mailman/listinfo/greenkeys
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:GreenKeys at mailman.qth.net <GreenKeys at mailman.qth.net>
>
> 2002-to-present greenkeys archive: http://mailman.qth.net/pipermail/greenkeys/
> 1998-to-2001 greenkeys archive: http://mailman.qth.net/archive/greenkeys/greenkeys.html
> Randy Guttery's 2001-to-2009 GreenKeys Search Tool: http://comcents.com/tty/greenkeyssearch.html
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>
>
>
>
> ______________________________________________________________
> GreenKeys mailing list
> Home: http://mailman.qth.net/mailman/listinfo/greenkeys
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:GreenKeys at mailman.qth.net
>
> 2002-to-present greenkeys archive:
> http://mailman.qth.net/pipermail/greenkeys/
> 1998-to-2001 greenkeys archive:
> http://mailman.qth.net/archive/greenkeys/greenkeys.html
> Randy Guttery's 2001-to-2009 GreenKeys Search Tool:
> http://comcents.com/tty/greenkeyssearch.html
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.qth.net/pipermail/greenkeys/attachments/20131122/e6ae9b1b/attachment.html>