#A2000 reset probs
11 messages in this thread
Hello.
I am having a problem with my A2000. It is equipped as follows:
– 2 MB of chip RAM – ICD AdSCSI 2080 controller/2 MB fast RAM/ Quantum
40 MB hard drive – running Kickstart ver. 37.175, WorkBench 37.67
The unit is "stock" otherwise.
The problem is this: often, during initialization, the screen goes through
its normal color changes–several shades of gray, then white. The power
LED blinks slowly. Then the screen blacks out, the LED blinks again, and a
"Software Failure" guru comes up. The numbers it gives are:
ERROR- 8000 0003
TASK – 00203210
I can reboot, and the screen goes through the gray shades. Then it goes
green and holds for 2-3 secs. Then it goes white and boots from the hard
drive.
I got the machine to die by using Quarterback Tools v1.5. The error it gave
then was:
ERROR- 8000 0003
TASK- 002662C0
I used a program on Compuserve called GURU2U and typed in the error number.
The program said the error meant:
Alert type – 68000 trap
General Cause – User trap.
That doesn't tell me anything. Does anyone have any suggestions about what
is happening? Could it have to do with the 2MB of chip RAM?
Regards, Phil Combs
Phil,
According to my GURU sheet, the 8000 0003 is probably an address error
CPU trap…
Mike Uman
[]xxi, Inc. >>>
running under AutoPilot
Hi!
What would (or could) cause such an address error in the OS (Kickstart)?
Defective KS ROM, perhaps? Defective 2 MB Agnus chip or memory chips?
Regards, Phil Combs
Hi Phil,
I believe you asked about a 8000 0003.xxxxx error and if it was caused
by defective KS-ROM, Agnus, or memory.
It's none of the above. I guessing that, since you haven't mentioned
any warning colors at bootup. At bootup the Amiga performs several
tests and reports if they pasted or failed with a series of dark grey
to white colored screens during bootup.
If the ROM was bad the Amiga would give you a solid RED screen on bootup,
if any of the custom chips are bad you would get a blue screen. If
the RAM chips are bad (or the Agnus is lose) you will get a green
screen.
The 80000003.xxxx error is normally due to bad programming, a program
playing with pointers, or corrupted code in a program.
The 68000 CPU can only access memory on even bountries ($34, 36, 68,
etc.) To access data in an odd location say at $35, the 68000 accesses
$34 which ALSO gives $35 (together they are 16-bits). To access $35 plus
$36, requires an access to $34 & $36. (This is 32-bits which requires
two accesses to complete). Trying to access $35 directly will guru the
*CPU* causing the 8000.0003 message.
Later CPUs (68020, 68030, 68040) could access the odd locations, so
allowing direct access without a CPU crash. Some programs written to run
on these 32-bit CPU Amigas will access odd addresses which is safe for
them, BUT will crash on a 68000 Amiga.
As for bad programs, some 68000 programs will have a hidden bug that will
cause it to access an odd address (crash), or these programs might change
pointors (accidently) for its, or other programs causing then to point
to an odd address.
There are always exceptions, but it think you have a bad program, or
a corrupted program crashing the Amiga.
Ron.
– ap'ing with AP. (Eagles may fly, but weasles aren't sucked
into jet engines.)
Hi, Ron. Actually, I _do_ sometimes get a green screen. I have noticed
that I can do a reset (CTRL-AMIGA-AMIGA) and the screen goes through
several grays, then black, then several grays, then green, then white, then
it boots.
Regards,
You say that you some do get a green-screen, BUT it continues to boot?
I don't think I've heard of that. Once it puts up the green-screen the
system should "hang".
I removed my Agnus from its socket a few times, I wasn't aware that doing
so would slowing weaken the pins on the Agnus. After awhile, if I moved
the A500, I head to open the case and push the Agnus back into a tighter
fixing position (I had it fixed).
It might be a bad RAM chip. Sometimes a single bit inside the chip can
become "froze" so that no matter what is written to that location it
gives the same value as was presence at powerup. It would be invisible
to the system until a program (or data) is run in that location.
Get one of the memory testing programs from the file area, and run it
with nothing else loaded or running. It should notice and report a
location that is froze.
Also, it could be hardware. Cards that "add themselves" use a program/
protocal called AUTOCONFIG(tm).
The Autoconfig changed for the A3000 & A4000 series to allow configuring
of boards to the higher 32-bit addresses. Some Zorro-II boards don't
know about the new 32-bit space (esp. 32-bit memory/RAM) and try to DMA
into it (they can only DMA into the lower 24-bit address space.)
Ron.
– ap'ing with AP. (Eagles may fly, but weasles aren't sucked
into jet engines.)
Hi, Ron. I did some checking and problem isolation, and this is what I
found. . .
I removed my AdSCSI 2080 card (it has 2MB of RAM) and booted from a floppy.
The problem went away. I replaced my 2.0 ROM with the 1.3, reinstalled the
card, problem went away. Reinstalled the 2.0 ROM, problem reoccurred.
Swapped the SIMMS on the AdSCSI with new ones, problem was still there.
_DISABLED_ the memory on the SCSI card, problem went away.
I talked with ICD, and they tell me they have a new 4.0 ROM/software update
coming out RSN (mine is 2.01). This is supposed to fix _many_ bugs they've
uncovered since my card was released. As it turns out, I think the memory
doesn't Autoconfig right with KS 2.04 and my card.
Sigh. Now I have to wait for several weeks until the new ROM is available,
and my problem _should_ go away then. I hope.
Thanks for your input. It helped me understand something about the
workings of my machine.
Regards, Phil Combs
Your welcome.
So, the AdSCSI 2080 card works without RAM, but needs an ROM update to
use RAM?
Ron.
– ap'ing with AP. (Eagles may fly, but weasles aren't sucked
into jet engines.)
Well, let's just say that, with the RAM on the card disabled, everything
works A-OK. With the RAM enabled, I can do a "three-finger salute" anytime
and the computer will not reboot properly. I have exchanged SIMM modules,
but there is no change.
The system works Okay, and the RAM works Okay, under 1.3.
Regards, Phil Combs
Phil,
You may want to check with ICD to see if you have the latest ROMs and
configuration software for the HD controller. If I remember correctly
there was a problem with some of their software and 2.04.
Dean
DKB Software
Thanks, Dean. I thought I might have to contact them–I suspected
something like that. Regards, Phil Combs