On 05/11/2007 21:17, Gordon JC Pearce wrote:
I'm fairly sure the one with memory problems is
down to a faulty 4116,
but I can't work out how to find out which one it is. I'm guessing it's
in the lower part of memory, because although the system boots it hasn't
got enough marbles to run
DDT.COM - if I could get that working I could
prod around in memory until I found the stuck bit.
If it's really a stuck bit regardless of address (within the bank), a
logic probe might show you, though you might need to make up a really
sophisticated circuit with an LS TTL OR gate to only show you the data
when the chip select line for the bank you're testing is low. Or use a
'scope and set it to only trigger on the appropriate chip select.
I've got a fast Z80 memory test routine I once wrote for a project,
which tests every bit for address, stuck-at, coupling, and some pattern
faults, which runs from ROM. It tests 8K bytes on a 4MHz Z80 in about 6
seconds unless it finds faults (and then writing the information to the
display slows it down). It doesn't need any RAM at all, though it was
written to drive a 2-line 16-char LCD display mapped to an address in
the target system. You could probably modify it to write to whatever
you can use in -- or hung off -- the Superbrain.
Give me a shout if you want a source listing.
--
Pete Peter Turnbull
Network Manager
University of York