#A2000 Problems
6 messages in this thread
Hello!
Maybe someone here can cast some light on the problem I am encountering
with my machine.
I have recently exchanged my A500 against an A2000, which is fitted with
an ECS Agnus, GVP Impact II SCSI Controller with 2MB of Expansion RAM,
Quantum LPS 105 disk and Kickstart 2.04, Revision 37.175. The Revision of
my WB 2.04 is 37.67, the A2000 board is a B2000 Rev 4.4.
The first problem I have is that I get 80000003 and 80000003 software
failures by the truckload. Especially DPIV and the Superbase Pro form
editor give me the most trouble. The SB form editor always crashes the
system when I exit it with an 80000004 error. I never had that problem on
my A500 which was also fitted with the 2.0 (In fact I ripped the Kick ROM
out of the A500 and put it into the A2000). So I figure somewhere deep in
the system is something is wrong. Commodore Switzerland and several
dealers have been unable to trace the problem.
The second thing that has me puzzled is that when I launch SnoopDos to
check on my system, I see that the task "Workbench" does a "CD Workbench:"
every about 2 – 3 seconds. This behaviour starts as soon as I launch a
program from the workbench and then close it's drawer. I kept fiddling
with my User-Startup file and checked on all tasks launched in WB-Startup,
because I thought that this was caused by some Program I launch at system
startup. But this also happens when I boot off a clean, unmodified WB2.0
Workbench diskette. Is this behaviour normal, if not what is causing it?
I'd greatly appreciate any hints concerning the two problems.
Greetings
Max
Across the Atlantic on AutoPilot…
Max;
Last first. The "cd Workbench:" is normal.
The 80000003 is a dead end alert (unrecoverable) and specifically illegal
memory access (odd address). The 80000004 is (again dead end): illegal
instruction.
It is really a bit hard to guess at what could be wrong. Since the
programs both work on the A500 it is not likely that there actually is an
instruction that makes an odd memory access (68000). I would first
consider that the memory board is flakey.
Other things that you might want to look at:
The ROM for some 2000's must have a jumper installed across two pins (has
to do with the specific trace layout on some of the 2000 motherboards). I
don't know if the absence of that jumper would completely prevent the
machine from booting or not but I suspect that some machines might
actually try to operate (I am pretty sure that the line is a high order
address line).
You might take a look at the VTick signal and see if it is clean. In any
event there is a jumper that will let you use either the retrace pulse or
the power supply clock. I would suggest trying both (I have a 2000 here
that will not work unless the jumper is moved to the vertical retrace
position).
-bill
Hi Bill,
thanks a lot for the info. The wire across the Kick ROM has to be cut on
my board, it says so in the installation sheet that came with the Kick ROM
and the machine will not boot with the wire intact.
This jumper that you refer to reagrding the VTick signal, were is that
located on the board? (I don't have a board layout and there are jumpers
by the dozens on the board)
The memory on the GVP board test ok, so does the memory installed on the
mother board.
Thanks for your help
Max
Across the Atlantic on AutoPilot…
Max;
The Jumper is probably labeled J300 and is under the power supply chassis
(about front center of the power supply itself on the motherboard). I
believe that it is close to U300 (8520) but that is on memory which can be
none too good. Pins 1/2 jumpered is normal (TBASE from TICK) and 2/3 is
TBASE from _VSYNC). Tick comes from the power supply and provides a
better reference for the calendar clock but it is also susceptible to
failure.
-bill
Bill,
Thanks for the info again. May a defective 8520 play a part in the
problem? Because when I switch the two 8520s, the system clock stops
running, so one of those chips must have a problem.
Greetings
Max
Across the Atlantic on AutoPilot…
Yep, the 8520 (U300) generates TBASE.
-bill