Cirtech GS 8 plusRAM woes...
amotyka
I have recently come into posession of a Cirtech GS 8 plusRAM board,
having given up on ever receiving the Sirius ram card I ordered a year
ago... Cirtech was a British company, which is probably why I can't
find much info about it on the net, only photo's, no manuals, and I
believe this was the first 8mb board available for the iigs. The
trouble is, ever since installing it I have been plagued by random
crashes. Replacing the card with my old 4mb AE GS-RAM III makes the
system (almost) impossible to crash. I have read old posts about 8mb
cards causing trouble with DMA and rom 3 machines, but I'm using a rom
1 and the cards installed are reactive micro's Microdrive (not the
turbo version so no DMA), the apple scsi card (not the high-speed
version so again no DMA) with cd drive attached, uthernet card (no DMA
to my knowledge) and a Transwarp clocked at 16mhz. The card checks out
fine with mega memory tester.
I'm going to run some more tests e.g. with the transwarp at a lower
clock and reducing the ram on the board, but would be interested to
hear if anyone else has encountered similar problems with 8mb boards,
or if anyone can give me some more info on this particular card. I
read elsewhere that this card also has a rom disk feature, and in fact
has two rows of header pins which I presume would be used to program
it.
Thanks,
Anthony
ict@ccess
I have a Sirius 8 mb card and I was getting quite a few crashes as
well. I did the power supply mod in which you change the 18 guage
wires to 12 guage. This helped quite a bit and cut my crashes down to
1 a day. I would liked to have done one more mod. The ground trace
is supposedly not heavy enough since it goes all the way to the
opposite side of the board where the RAM slot is. A ground wire could
be soldered at the power connector and go to the ground terminal of
the RAM slot. I was lazy and ok with 1 crash per day, so I didn't
test this one out. But I would recommend trying a larger power
supply you can get from www.reactivemicro.com and also do the ground
wire to RAM slot mod.
I use a ZIPGS rated at 12.5 Mhz in slot 3, ramfactor in slot 5 and
CFFA in slot 7.
Rob
amotyka
> wire to RAM slot mod.
> I use a ZIPGS rated at 12.5 Mhz in slot 3, ramfactor in slot 5 and
> CFFA in slot 7.
>
> Rob
Hi Rob,
Thanks for the feedback. Now you mention it, this card must draw a lot
of power as it uses 8 ram chips per megabyte, so 64 in total, plus the
glue chips. The GS-RAM III by comparison is tiny and has only 4 ram
chips plus a few others. I am using a PC power supply with a
littlejohn adapter but I know the power traces to the slots can be
flakey in rom 1 machines. I will have to try and pluck up the courage
to have a go at my iigs with a soldering iron and see if that helps.
The weird thing is I left mega memory tester running for four
iterations with no errors reported, though i did note during the test
the transwarp was inactive so not contributing to any possible power
issues.
I don't necessarily need the full 8mb, so if I can get it to work with
5 or 6mb instead then I would be happy; I'm just keen to break the 4mb
barrier.
Anthony
Michael J. Mahon
That suggests to me that putting several 0.1uF capacitors between +5v and
ground, either right at the slot or, preferably, on the memory card itself.
It shows the symptoms of being insufficiently decoupled.
Ground bounce could also be an issue, but it would take a scope to tell for
sure.
-michael - NadaNet 3.1: http://home.comcast.net/~mjmahon
rich12345
> ago...
Did you order it directly from 16sector.com ? You can reach Tony D
via 16sector.com, or chat with him in IRC (a2central chatroom, can be
accessed via reactivemicro.com).. I've ordered several things from
him in the past and have never had any trouble.
Rich
amotyka
Hi Rich,
Yes it was bought and paid for through the site last february,
delivery has been promised several times over the past year but sadly
it has never materialized.
I don't want to cause any problems for anyone, I know we all have a
lot going on in our lives; I think I might just have to write it off
as a bad investment.
Anthony
amotyka
> > wire to RAM slot mod.
> > I use a ZIPGS rated at 12.5 Mhz in slot 3, ramfactor in slot 5 and
> > CFFA in slot 7.
>
> > Rob
>
> That suggests to me that putting several 0.1uF capacitors between +5v and
> ground, either right at the slot or, preferably, on the memory card itself.
>
> It shows the symptoms of being insufficiently decoupled.
>
> Ground bounce could also be an issue, but it would take a scope to tell for
> sure.
>
> -michael - NadaNet 3.1: http://home.comcast.net/~mjmahon
Hi Michael,
My card actually has around 40 capacitors soldered in between the ic
sockets, so I don't think decoupling is a problem.
I'm beginning to suspect that maybe the card can't keep up with the
transwarp running so fast. My guess is the GAL(s) on the ram card are
only supposed to go so fast, especially since this was the first 8mb
board produced, its creators may never have expected the clock to go
above 3mhz.
I'm going to try running at at stock iigs speed, which will bring back
painful memories, and see if it works. Then I may have to face a
difficult choice between speed or memory...
Anthony
Michael J. Mahon
I understand, but I thought that accelerator cards did not, in fact,
run the memory any faster than "regulation" speed. They have caches
that run faster, but their interface with the rest of the system is
at standard speed.
Computer busses cannot, in general, be sped up by aftermarket
gear. They have a timing design and that's that. That applies
to the external memory bus of the IIgs as well.
Of course, as you increase the accelerator speed, certain timing
margins may change slightly in the wrong direction, and there may
be small changes in power dissipation in the memory due to a changing
proportion of reads and writes. These effects would usually have no
practical effect on a machine with good timing and voltage margins,
but if your machine is "close to the line", they could certainly
affect stability.
It might be worth checking all the decoupling capacitors on the
board carefully. I once had a 0.1uF cap decoupling the 6502 on
my ][+ develop a subtle break in one of its "legs", with the
result that the 6502 became quite unstable. Resoldering the cap
fixed the problem completely.
You could have a similar problem with one or more DRAM chips,
resulting in stability issues only when certain address ranges
are used.
-michael
NadaNet 3.1 for Apple II parallel computing!
Home page: http://home.comcast.net/~mjmahon/
"The wastebasket is our most important design
tool--and it's seriously underused."
amotyka
Thanks Michael, you were right, running the iigs at stock speed made
no difference. As an experiment I set up a 4mb ram disk; the system
crashed as soon as GS/OS started loading with fatal error $0308,
damaged heartbeat queue. Also, when trying to mount more than 3 disk
images with mount it, I get an error saying the disk is damaged. I've
also suffered the occasional $0201 unable to allocate memory block
error.
I'll thoroughly examine the board to check for dodgy contacts and also
try swapping out the socketed chips, but first I'll leave mega memory
tester running overnight to see if it catches any problems.
Anthony