24 Nov
2009
24 Nov
'09
6:04 p.m.
------------------------------
Message: 20
Date: Mon, 23 Nov 2009 22:50:31 +0000
From: Pete Turnbull <pete at dunnington.plus.com>
Subject: DEC VT100 character generator
To: ClassicCmp <cctalk at classiccmp.org>
Message-ID: <4B0B11B7.4020304 at dunnington.plus.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Several months ago, someone was looking for an image of the 23-018E2
character generator ROM for a VT100. A generous reader has given me an
image, which I've uploaded to my website at
http://www.dunnington.u-net.com/public/DECROMs/
Better late than never, I hope!
--
Pete Peter Turnbull
Network Manager
University of York
------------------------------
Pete,
That matches the rom I typed from the vt100 tech manual datasheet for
MESS. Are you sure someone dumped it from a real chip and didn't just
submit the rom I typed up (which is now floating around the 'net)? The
rom is marked as a 'bad dump' in MESS, because I was almost sure the
typed one is very slightly wrong due to the fact that the sum16 of the
one I typed does not end with "00", which seems to have been standard
DEC practice for roms at the time. Hence I was hoping someone with a
working or scrap vt1xx series board would dump the real thing.
The trouble with most people dumping the original chip seems to be that:
A. it is soldered to the vt1xx board
B. it has uninverted CE (I think...) and a few other pinout oddities
(see schematic on bitsavers)
P.S. the second optional character rom (only selectable if you have an
AVO board installed, or a VT102/VT131 which has AVO builtin), labeled
23-094e2 is also not dumped. This one, fortunately, is socketed on
systems which have it, and I believe has a normal pinout. I think it
contains european characters and formatting/word processing characters.
It might also contain some or all of the technical font used on the
later vt3xx+ systems as shown here: http://vt100.net/charsets/technical.html
P.P.S. the main cpu roms from a vt1xx with the word processing romset
installed are also not dumped. I have no idea what the numbering on
these is though. Two of the CE pins alternate in binary form for the
chips to allow them to be 'self decoding' and inserted in the four
sockets in any order! cute, but makes dumping them a bit harder.
--
Jonathan Gevaryahu
jgevaryahu(@t)hotmail(d0t)com
jzg22(@t)drexel(d0t)edu
24 Nov
24 Nov
7:16 p.m.
A character generator ROM is unlikely to have a fixed value checksum,
because there usually isn't a "spare" location in a character generator
ROM in which to stuff a contant to make the checksum come out to the
desired value.
Eric
25 Nov
25 Nov
12:17 a.m.
On 25/11/2009 02:04, Jonathan Gevaryahu wrote:
[ stuff which was below a signature line so it didn't come out ]
The ROM image came from from someone I've got data from before, who
fixes things, and says he dumped it from one he desoldered from a
working VT. I have no reason to doubt him.
The idea that DEC made a practice of arranging the checksum to come to
zero or end "00" is news to me. I'm fairly sure that's not the case.
There must be many exceptions, if only because there's often not a spare
location to store the necessary constant to adjust the checksum.
Having worked with DEC stuff when it was current, even though it was a
long time ago, I'm quite familiar with the ROMS, including in particular
Q-Bus devices and VT1xx devices, because I used to repair them, to
component level, for a living. In fact that's why I started collecting
the images.
Using different polarity chip selects on masked ROMs is common practice.
Many manufacturers did that, and it's not hard to get round. If your
programmer can't deal with it, just bend the pin carefully and use a
test lead to tie it high or low.
--
Pete Peter Turnbull
Network Manager
University of York
2:50 a.m.
On 25/11/2009 02:04, Jonathan Gevaryahu wrote:
[ stuff which was below a signature line so it didn't come out ]
The ROM image came from from someone I've got data from before, who
fixes things, and says he dumped it from one he desoldered from a
working VT. I have no reason to doubt him.
The idea that DEC made a practice of arranging the checksum to come to
zero or end "00" is news to me. I'm fairly sure that's not the case.
There must be many exceptions, if only because there's often not a spare
location to store the necessary constant to adjust the checksum.
Having worked with DEC stuff when it was current, even though it was a
long time ago, I'm quite familiar with the ROMS, including in particular
Q-Bus devices and VT1xx devices, because I used to repair them, to
component level, for a living. In fact that's why I started collecting
the images.
Using different polarity chip selects on masked ROMs is common practice.
Many manufacturers did that, and it's not hard to get round. If your
programmer can't deal with it, just bend the pin carefully and use a
test lead to tie it high or low.
--
Pete Peter Turnbull
Network Manager
University of York
Pete:
Apologies for doubting you then, I'm just surprised the rom matches.
There actually IS spare space in the character generator rom for the
vt100 since bytes xxF xx0 xx1 xx2 xx3 xx4 xx5 xx6 xx7 xx8 are used for
rows 0 thru 9 of a given character xx(yes byte xxF comes first, I don't
know why DEC did that but it says so in the vt100 tech manual), but
bytes xx9 xxA xxB xxC xxD xxE are unused and can have any data stuffed
into them.
I'll mark the rom as good in MESS; who dumped that rom by the way? I
want to give proper credit (unless they wish to remain anonymous).
--
Jonathan Gevaryahu
jgevaryahu(@t)hotmail(d0t)com
jzg22(@t)drexel(d0t)edu
5458
days inactive
5458
days old
3 comments
3 participants
participants (3)
-
eric＠brouhaha.com -
jzg22＠drexel.edu -
pete＠dunnington.plus.com