Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!ames!ucsd!sdcsvax!ucsdhub!hp-sdd!hplabs!sdcrdcf!trwrb!trwind.ind.TRW.COM!robert
From: robert@trwind.ind.TRW.COM (Robert W. Snyder)
Newsgroups: comp.sys.amiga
Subject: Problems with Hard disk and 4.3 motherboard?
Message-ID: <301@trwind.ind.TRW.COM>
Date: 23 Feb 88 02:18:08 GMT
Reply-To: robert@trwind.UUCP (Robert W. Snyder)
Organization: TRW Information Networks Div
Lines: 70


I am posting this for a friend any replies can be sent to me and I
will forward.  Robert
-----------------------------------------------------------------------
I have just purchased an Amiga 2000, Amiga 2090 controller board,
Segate ST251-1 hard disk and a Micron 2Meg board.  I installed 
everything and stuff seemed to work OK, then...  My system began
to guru for no apparent reason.  Suspiciously, it always gurued
during hard disk access.  Even when I was only copying files from
place to place.

My mother board is rev 4.3.
The driver software is dated end of Nov.  Works with overscan pics.
The disk was installed as dh0: (no partitions)

I came up with a simple test scenario:
	Run 4 normal cli's and in each execute a batch script 
	with lots of 'dir opt a' on dh0:.

After some experimentation I found out the following:
	(with micron board installed):
	(1)	Boot workbench, fastmemfirst, binddrivers.
		The test will guru the machine (typically in < 5 minutes).

	(2)	Boot workbench, binddrivers.
		The test will run a long time (> 1/2 hour) without problems.
	
	(with an Amiga 2meg board):
		The test will run a long time (> 1/2 hour) without problems
		in either scenario 1 or 2.

	(without an expansion mem board):
		The test will run a long time (> 1/2 hour) without problems
		in either scenario 1 or 2.

So I figured OK, the micron board might have problems with the 2090
board DMAing into its memory (binddrivers after 'fastmemfirst' causes
the driver buffers to be allocated in the expansion mem right?).

So I put the Micron board in a friends 2000 (mother board rev 4.2) 
his 2090 controller (older) and his ST251 drive (all the same 
software), then tried scenario 1.  Guess what?  It worked fine.  

I couldn't keep trying scenarios on my friends machine so I went 
home and tried some longer ones on my machine.  This was done
without the expansion memory!  Doing the test for long time
periods caused the system to guru also.  It gurued after about
45min to 2 hours after starting the test. (I ran the test 3 times)

I have contacted Micron, they had heard of some similar problems
with the rev 4.3 motherboards and their expansion memory, but have
not been able to test it since they cannot seem to get 
an Amiga 2 Meg board and an Amiga 2090 controller.

I believe that Micron is trying to get the problem fixed.
But what concerns me more is that the system Guru's *without
the expansion memory*. (BTW it passes the Commodore system diagnostics)

Has anybody heard of this problem before?
Does anybody have tests which check DMA to/from the 2090 board?
Any ideas?

Jeff
(Hmmm...  Maybe I shouldn't have upgraded from my reliable 1000)
--------------------------------------------------------------------------

-- 
Robert Snyder       Disclaimer  --  nobody claims dis, but me
TRW Information Networks Division 23800 Hawthorne Blvd, Torrance CA 90505
USENET: {ucbvax,ihnp4,sdcrdcf}!trwrb!trwind!robert
ARPA: robert@trwind.TRW.COM