federico@actisb.UUCP (Federico Heinz) (06/01/91)
Well, I guess that after over a month of flawless operation and
constant admiration for the machine, reality struck me for the first
time. Today I upgraded my '030 Cube to the '040 board, and therefore
transplanted all my SIMMs from the old to the new CPU board.
Expecting the machine to FLY, I turned it on, and it flew allright...
into the following error message:
Exception #2 (0x8) at 0x100416e
and there was no way of getting it to boot. I then removed half of the
SIMMs (the ones I bought to upgrade to 16MB), it worked fine. When I
left only the new SIMMs in, the machine crashed in almost exactly the
same way (different number after 'at'). The amazing thing about all
this is that the ROM monitor's 'm'-command correctly recognized the
SIMMs, but the machine wouldn't boot. Not only did the SIMMs pass the
power-on memory test, they work perfectly with the '030 board too (I
tried it with them as the only memory). I suspected this to be a
parity-no parity thing, but parity checking was disabled. Of course, I
had written the 2.0 boot block to the disk, but the thing wouldn't
even boot out of the optical.
The SIMMs that don't work are 1MBx9, 70 ns, and I'm attempting to use
them together wih the original 1MBx8 SIMMs that came with the Cube.
I'm puzzled. Furthermore, I'm scared I just threw the money for these
8MB out of the window. Any ideas about what may be going wrong here?
--
Federico Heinz (federico@actisb.uucp)PPSTL@livid.uib.no (06/01/91)
------------------------- Original Article -------------------------
Path: alf.uib.no!ugle.unit.no!sunic!hagbard!eru!bloom-beacon!ora!camb.com!spdcc!
From: federico@actisb.UUCP (Federico Heinz)
Newsgroups: comp.sys.next
Subject: Shouldn't 70 ns SIMMs work with '040?? (they don't for me, anyway)
Keywords: SIMM 70 ns 68040
Message-ID: <1429@actisb.UUCP>
Date: 31 May 91 21:26:11 GMT
Organization: Actis in Berlin GmbH, W. Germany
Lines: 29
Well, I guess that after over a month of flawless operation and
constant admiration for the machine, reality struck me for the first
time. Today I upgraded my '030 Cube to the '040 board, and therefore
transplanted all my SIMMs from the old to the new CPU board.
Expecting the machine to FLY, I turned it on, and it flew allright...
into the following error message:
Exception #2 (0x8) at 0x100416e
and there was no way of getting it to boot. I then removed half of the
SIMMs (the ones I bought to upgrade to 16MB), it worked fine. When I
left only the new SIMMs in, the machine crashed in almost exactly the
same way (different number after 'at'). The amazing thing about all
this is that the ROM monitor's 'm'-command correctly recognized the
SIMMs, but the machine wouldn't boot. Not only did the SIMMs pass the
power-on memory test, they work perfectly with the '030 board too (I
tried it with them as the only memory). I suspected this to be a
parity-no parity thing, but parity checking was disabled. Of course, I
had written the 2.0 boot block to the disk, but the thing wouldn't
even boot out of the optical.
The SIMMs that don't work are 1MBx9, 70 ns, and I'm attempting to use
them together wih the original 1MBx8 SIMMs that came with the Cube.
I'm puzzled. Furthermore, I'm scared I just threw the money for these
8MB out of the window. Any ideas about what may be going wrong here?
--
Federico Heinz (federico@actisb.uucp)
I have also posted regarding this, when I first received my 040 upgrade.
Using the ROM monitor, turn off memory check on power on and turn off
parity use if present. Power off on on again - it should work. I have
8MB 100ns 1x8 (NeXT orig.), 4MB 80ns 1x9, and 4MB 70ns 1x9 in my cube
and it works like a charm (for those of you following my hard disk soap
opera oddessy, The Formatter saved the disk - and the eight day final I
am in the middle of when the disk crashed...Many,manymany thanks to Ronald
Antony and Charles Purcell for their patience)
Of course, the machine operates at the speed of the _slowest_ memory
in the machine, thus I get a message "100ns" memory on boot - not 80 or 70.
Hope this helps.
Thor Legvold
University of Bergen
Norway
me@fiol.uib.no <- NeXTmail, er du snill.cnh5730@maraba.tamu.edu (Charles Herrick) (06/02/91)
In article <1429@actisb.UUCP> federico@actisb.UUCP (Federico Heinz) writes:
The SIMMs that don't work are 1MBx9, 70 ns, and I'm attempting to use
them together wih the original 1MBx8 SIMMs that came with the Cube.
I'm puzzled.
You can not mix (1X9) SIMMS with (1X8) SIMMS on a NeXT, even if
parity-checking is disabled. Replace your 70ns (1X9) with 70ns (1X8)
and try again.
Pascal, could this go into the FAQ, please?
-- Chuckplampani@dpls (Patrick Lampani) (06/02/91)
In article <CNH5730.91Jun1131807@maraba.tamu.edu> cnh5730@maraba.tamu.edu (Charles Herrick) writes: > You can not mix (1X9) SIMMS with (1X8) SIMMS on a NeXT, even if > parity-checking is disabled. Replace your 70ns (1X9) with 70ns (1X8) > and try again. > -- Chuck I'm not very competent yet with all this hardware stuff........ BUT I'm typing this message on a machine that has mixed parity and non parity chips. So at least on a NextStation ROM 2.1 v59 (and later I'm told), you can mix them, although you have to boot from the monitor after getting a SYSTEM TEST FAILED message. You would, of course, loose the parity checking. I know that on the station, when you put in BIGGER SIMMS, you have to put them in the lower numbered slots and the smaller SIMMS in slots 4-7. Heck, maybe it's worth a try to put the new parity SIMMS in the lower numbered slots, At least it's one last experiment until Monday when you can call someone. -Patrick Lampani
garnett@cs.utexas.edu (John William Garnett) (06/05/91)
In article <CNH5730.91Jun1131807@maraba.tamu.edu> cnh5730@maraba.tamu.edu (Charles Herrick) writes: >In article <1429@actisb.UUCP> federico@actisb.UUCP (Federico Heinz) writes: > The SIMMs that don't work are 1MBx9, 70 ns, and I'm attempting to use > them together wih the original 1MBx8 SIMMs that came with the Cube. > I'm puzzled. > >You can not mix (1X9) SIMMS with (1X8) SIMMS on a NeXT, even if >parity-checking is disabled. Replace your 70ns (1X9) with 70ns (1X8) >and try again. > >Pascal, could this go into the FAQ, please? > >-- Chuck Please double check that. From the FAQ: ]I2. What type of memory may be installed in a NeXT? ] ] and the hard drive in simultaneously. Note that parity (9-bit) ] RAMS can be used in both the 030 NeXT and the 040 NeXT. Parity ] and non-parity SIMMS can be mixed in both the 030 and the 040 ] (however, the boot on 040 boxes will generate a warning). Only ] the 040 NeXT can use the parity memory to detect parity errors ] (parity and non-parity may not be mixed if you wish to have ] parity errors detected). 2.2 (v63) ROM lets you disable parity ] checking (NVRAM preference). 2.1 (v59) does not. -- John Garnett University of Texas at Austin garnett@cs.utexas.edu Department of Computer Science Austin, Texas
tom@sailfish (Tom Horton) (06/07/91)
In article <1429@actisb.UUCP> federico@actisb.UUCP (Federico Heinz) writes: > I'm puzzled. Furthermore, I'm scared I just threw the money for these > 8MB out of the window. Any ideas about what may be going wrong here? This last remark leads me to think that you bought two 4Mx9 SIMMs to go with the eight 1Mx8 SIMMs originally in your cube. If this is so, you've got a problem: you must have four identical SIMMs in each bank, so just having two 4 Mb SIMMS will cause problems. If this is what you've done and it is a surprise for you, I'm almost certain it is in the FAQ. It's certainly in NeXTAnswers. Tom Horton tom@cs.fau.edu