Call for final Elf99 design input

From: Ethan Dicks <erd_at_infinet.com>
Date: Thu Dec 3 13:59:21 1998

>
> The key to the elf design was minimalism and low cost.

True.
 
> Substituting a latch and LEDs for a binary display instead of the til311
> is cheaper and more buildable.

True, but the TIL's are so cool.
 
> using modern cmos and rams help.

At least in terms of not having to dig up 1822/2101's, yes.
 
> The basic elf was far to minimal and frequently expanded to get desired
> functions. The basic design did not easily permit that as it didn't
> decode memory or IO addresses.

Also true. One of it's big limitations. At the time, however, to
add even a 4-bit patch and more RAM would have been a major expense,
for the RAM, that is. The latch would have been under $1.

 
> The RCA VIP or the 18S020 Evaluation board allowed far more flexibility
> with relatively little more logic. Thei cost was they had some rom.

I have a picture of that board in my RCA circular. I lusted for one
when all I had was an Elf.
 
> This is a reproducable design. UT4 fits in 512byts of a 2716, the rest of
> the ram can be done with byte wide parts. The 1852s can be kept and the
> rest were common 40xx series. With 4k of ram and UT4 (or similar)
> programs like PILOT, TinyBasic, or some of the other neat software with a
> terminal.

What was the memory map?
 
> A much simplifed machine using 1802, 2 1852, 2 4028, 2 4042, 1 62256,
> 1 2716 and 1 6116, some glue logic for reset, run, runp and the same
> serial scheme as 18S020 would give 32k of contigious ram, 2k-32byts in the
> >8000h area and the remaining 1.5k in the 2716 could hold any number of
> things along with the .5k ut4 monitor. This would be a very useful
> system that could accomodate expansion for IO (more ram??????).

That sounds very much like what I'm designing here. Not identical,
but similar. I would need to know lots more before I could even attempt
to make it compatible.

-ethan
Received on Thu Dec 03 1998 - 13:59:21 GMT