3 May
2017
3 May
'17
4:49 p.m.
Finished building a prototype mezzanine board and testing with the IM6100 (PDP-8 on a
chip) in the 560Z board today. Full writeup:
http://www.glitchwrks.com/2017/05/03/gw-osi-ram1-universal-ram
There are a couple of tweaks required, but I should be doing a first production run of
boards before long. This board is:
* 128 KW, 12-bits
* Optional memory management/bank switching
* No hard-to-find/expensive ICs (8T26, et c.)
* All through hole
* Includes lamp register mezzanine header
* Largeish prototype area
The lamp register is on a mezzanine for a few reasons. First off, octal vs. hex grouping!
It really throws me off to try and interpret one on the other. Second, it allows remote
mounting on a ribbon cable, i.e. board can be mounted to a front panel. Third, I'll
likely design a 7-segment version.
Thanks,
Jonathan
3 May
3 May
5:08 p.m.
On 05/03/2017 04:49 PM, Systems Glitch via cctalk wrote:
Finished building a prototype mezzanine board and
testing with the
IM6100 (PDP-8 on a chip) in the 560Z board today. Full writeup:
The lamp register is on a mezzanine for a few reasons. First off,
octal vs. hex grouping! It really throws me off to try and interpret
one on the other. Second, it allows remote mounting on a ribbon
cable, i.e. board can be mounted to a front panel. Third, I'll likely
design a 7-segment version.
Is the IM6100/6120 still easy to find? I'dve thought that the supply
was exhausted by now.
Octal for me is quite familiar for 12 bits; I can't recall ever seeing
a 12-bit word that didn't use it. On the other hand, 16 bits is a
natural for hex--and only gets confusing when it's in octal.
--Chuck
5:37 p.m.
Is the IM6100/6120 still easy to find? I'dve
thought that the supply
was exhausted by now.
One of my *US* parts supplier apparently has qty 500 in stock for the IM6100s! With the
failure rate I've been seeing, figure at least 350 of them are viable, maybe more.
I've also had a bunch come out of the woodwork from other hobbyists with this
project.
Octal for me is quite familiar for 12 bits; I
can't recall ever seeing
a 12-bit word that didn't use it. On the other hand, 16 bits is a
natural for hex--and only gets confusing when it's in octal.
I want to say the only 12-bit work I've done has always been octal (PDP-8), but 8- and
16-bit has been a mix. I can switch between them, but it's kinda like using vim and
$graphical_work_editor -- you use vim on a weekend long hack session and you keep hitting
`ESC:wq` in $graphical_work_editor!
Thanks,
Jonathan
6:03 p.m.
On Wed, May 3, 2017 at 8:37 PM, Systems Glitch via cctalk
<cctalk at classiccmp.org> wrote:
Nice. They are definitely not as in-demand as the IM6120.
Is the
IM6100/6120 still easy to find? I'dve thought that the supply
was exhausted by now.
One of my *US* parts supplier apparently has qty 500 in stock for the IM6100s! With the failure rate I've been seeing, figure at
least 350 of them are viable, maybe more. I've also had a bunch come out of the
woodwork from other hobbyists with this project.
I picked up a *tube* of them about 15 years back and have never come
across a project for them (but I do have real DECmates of all the
varieties, so I need both IM6120 _and_ IM6100 chips for spares).
I have done a little with OSI machines (I've long wanted to fiddle
with a C3P because a family friend had one when I was growing up) and
have been watching the development of the your 560Z board, even if I
don't have a place to stick one at the moment.
-ethan
4 May
4 May
9:53 a.m.
On 05/03/2017 05:37 PM, Systems Glitch wrote:
I want to say the only 12-bit work I've done has
always been octal
(PDP-8), but 8- and 16-bit has been a mix. I can switch between them,
but it's kinda like using vim and $graphical_work_editor -- you use
vim on a weekend long hack session and you keep hitting `ESC:wq` in
$graphical_work_editor!
Sometimes the choice of radix makes sense. At one point in my career, I
was reading both 60 bit octal dumps and 64 bit hex dumps. In both
cases, the radix made sense. The 60 bit words had a natural bit
grouping of 3 (6-bit characters, 6 bit opcode, 3 bit register numbers)
and the 64-bit one also was suited to hex (8 bit opocdes and register
numbers as well as characters). The only gotcha was that addresses were
bit addresses, so that calculating byte addresses with a byte index
involved a 3 bit shift and word addresses involved a 6 bit shift.
Halfwords involved a 5 bit shift, which didn't work for any common radix.
--Chuck
2571
days inactive
2572
days old
4 comments
3 participants
participants (3)
-
cclist＠sydex.com -
ethan.dicks＠gmail.com -
systems.glitch＠gmail.com