15 Feb
2013
15 Feb
'13
8:48 a.m.
Hi all,
my Name is Alex, me and a friend of mine have been actively
participating in a computer collection, mostly repairing
stuff and bringing it back to life. We have numerous old apples, vaxes,
decstations, suns, next etc and some
real weird old stuff ;)
Lately we discovered a DG nova standing in a dark corner of a storeroom.
We have absolutely no peripherals, just
the bare machine. It has 2 CPU boards, a general IO card, 2 Memory
boards and some unknown third party controller.
The Machine as it sits now has 2-3 problems (at least):
1: Somehow the "desposit next" switch does not work as expected by not
incrementing the adress sometimes. Doing a deposit-examine next-deposit
works, so thats a minor problem.
2: Simple programs run fine in single step mode. Press RUN and the
machine seems to hang. You can not even STOP anymore, you have
to reset the machine. Annonying, but to test the simple stuff single
stepping also works....
3: The biggest problem is: The ACs have numerous bits stuck on 1. It is
the same bit pattern for all ACs. We can store and
deposit the non-stuck bits, but all others remain a 1. With this
behaviour its really really hard to write any meaningful asm to
test the machine ;)
Please see the attached Link for pictures of the machine and the
bit-pattern of the AC. Since the AC Deposit works as expected
and the error shows only when examining we are positive that the error
is not in the AC Registers itself but somewhere where
the switch location is actually shifted into the AC....or something like
that. We are also struggling a bit to get all the documentation
in a meaningful relation. So if you have any pointers on what to do or
check, please chime in.
If the machine gets back to life i have plans on emulating a storage
device to run one of the available os-es. Should be doable with
a microcontroller and some logic glue.....but it looks like a long way ;)
See actual pictures of the machine here
http://bigalpha.ath.cx/nova/pics/ori/index.html
where you can also see the bit Pattern of the AC.
I kinda wonder why its
1 11X 111 X11 1X1 11X
I guess this should point us nearer to the point of failure.
best regards
Alex
10:02 a.m.
When I got my Nicolet, I had a similar sounding problem.It turned out to be a missing
clock.Of course, yours may be soemthing different but you aregoing to need some schematics
and a scope to track down.( I did mine without a schematic but I was lucky ).Dwight
Date: Fri, 15 Feb 2013 14:48:07 +0100
From: RolexTM at gmx.de
To: cctalk at classiccmp.org
Subject: DG Nova repair etc.
Hi all,
my Name is Alex, me and a friend of mine have been actively
participating in a computer collection, mostly repairing
stuff and bringing it back to life. We have numerous old apples, vaxes,
decstations, suns, next etc and some
real weird old stuff ;)
Lately we discovered a DG nova standing in a dark corner of a storeroom.
We have absolutely no peripherals, just
the bare machine. It has 2 CPU boards, a general IO card, 2 Memory
boards and some unknown third party controller.
The Machine as it sits now has 2-3 problems (at least):
1: Somehow the "desposit next" switch does not work as expected by not
incrementing the adress sometimes. Doing a deposit-examine next-deposit
works, so thats a minor problem.
2: Simple programs run fine in single step mode. Press RUN and the
machine seems to hang. You can not even STOP anymore, you have
to reset the machine. Annonying, but to test the simple stuff single
stepping also works....
3: The biggest problem is: The ACs have numerous bits stuck on 1. It is
the same bit pattern for all ACs. We can store and
deposit the non-stuck bits, but all others remain a 1. With this
behaviour its really really hard to write any meaningful asm to
test the machine ;)
Please see the attached Link for pictures of the machine and the
bit-pattern of the AC. Since the AC Deposit works as expected
and the error shows only when examining we are positive that the error
is not in the AC Registers itself but somewhere where
the switch location is actually shifted into the AC....or something like
that. We are also struggling a bit to get all the documentation
in a meaningful relation. So if you have any pointers on what to do or
check, please chime in.
If the machine gets back to life i have plans on emulating a storage
device to run one of the available os-es. Should be doable with
a microcontroller and some logic glue.....but it looks like a long way ;)
See actual pictures of the machine here
http://bigalpha.ath.cx/nova/pics/ori/index.html
where you can also see the bit Pattern of the AC.
I kinda wonder why its
1 11X 111 X11 1X1 11X
I guess this should point us nearer to the point of failure.
best regards
Alex
noon
G'day Alex -
Your e.mail begs many question.
What documentation do you have?
Are you following the checkout procedures in the DG hardware manual?
What "simple programs" can be run?
What does "reset the machine" mean? Does it mean use the Reset switch,
power off the machine, or...?
The original Nova is an interesting whose big weakness was its power
supply (mainly the higher-voltage memory Vinh and Vmem supply).
Note that console accumulator functions exam and deposit may be
functionally tested without having a memory board in the computer(!).
As Al said, the Nova uses data paths of 4-bit "nibbles" for most
internal transfers - even to and from main core memory(!). Your symptom
appears a common single-bit failure mode.
I will contact you off-list with followup ...
Bruce Ray
Wild Hare Computer Systems, Inc.
bkr at WildHareComputers.com
On 2/15/2013 6:48 AM, Alex wrote:
Hi all,
my Name is Alex, me and a friend of mine have been actively
participating in a computer collection, mostly repairing
stuff and bringing it back to life. We have numerous old apples, vaxes,
decstations, suns, next etc and some
real weird old stuff ;)
Lately we discovered a DG nova standing in a dark corner of a storeroom.
We have absolutely no peripherals, just
the bare machine. It has 2 CPU boards, a general IO card, 2 Memory
boards and some unknown third party controller.
The Machine as it sits now has 2-3 problems (at least):
1: Somehow the "desposit next" switch does not work as expected by not
incrementing the adress sometimes. Doing a deposit-examine next-deposit
works, so thats a minor problem.
2: Simple programs run fine in single step mode. Press RUN and the
machine seems to hang. You can not even STOP anymore, you have
to reset the machine. Annonying, but to test the simple stuff single
stepping also works....
3: The biggest problem is: The ACs have numerous bits stuck on 1. It is
the same bit pattern for all ACs. We can store and
deposit the non-stuck bits, but all others remain a 1. With this
behaviour its really really hard to write any meaningful asm to
test the machine ;)
Please see the attached Link for pictures of the machine and the
bit-pattern of the AC. Since the AC Deposit works as expected
and the error shows only when examining we are positive that the error
is not in the AC Registers itself but somewhere where
the switch location is actually shifted into the AC....or something like
that. We are also struggling a bit to get all the documentation
in a meaningful relation. So if you have any pointers on what to do or
check, please chime in.
If the machine gets back to life i have plans on emulating a storage
device to run one of the available os-es. Should be doable with
a microcontroller and some logic glue.....but it looks like a long way ;)
See actual pictures of the machine here
http://bigalpha.ath.cx/nova/pics/ori/index.html
where you can also see the bit Pattern of the AC.
I kinda wonder why its
1 11X 111 X11 1X1 11X
I guess this should point us nearer to the point of failure.
best regards
Alex
4181
days inactive
4181
days old
3 comments
4 participants
participants (4)
-
aek＠bitsavers.org -
Bruce＠Wild-Hare.com -
dkelvey＠hotmail.com -
RolexTM＠gmx.de