jwest at classiccmp.org
Fri Nov 17 20:08:29 CST 2006
Bob S. wrote...
> So what was the DSD440 issue?
The DSD440 is the only component not working. I hooked up an RX02 externally
today just long enough to get the software on it I wanted.
If you'll recall, you were going to help me troubleshoot the main logic
board on the DSD440 before you got busy on some other things. The DSD440 has
an 8085 master cpu that controls general housekeeping and the communication
between it and the host controller. It also has a bitslice cpu using three
2901's that does the actual bit stream decoding from the floppy.
I have diagnosed the problem far enough to know that the 8085 and the
bitslice processor are not communicating. The 8085 sends a command to the
bit slice cpu which is supposed to do a few tasks asynchronously and then
signal back to the 8085 saying "done". The 8085 is never getting that
completion signal back from the bit slice cpu (nor is it getting an error
signal back). The 8085 gets waken up by a watchdog timeout signal, so it is
able to present a "hey, the bitslice cpu isn't responding" message instead
of just waiting forever. The problem is, I have no idea if the 8085 isn't
getting the ready signal because the bit slice cpu is dead/nonrunning, or if
it's caught in a loop or error state (that prevents if from sending back the
ready signal), or if the bit slice cpu is actually doing what it's supposed
to be doing ok and the ready signal just isn't getting back to the 8085. It
is one of those three possibilities. Due to other tests, I'm fairly sure the
8085 is working correctly itself.
I have checked that all the major components of the bit slice cpu are
getting good Vcc, and all the pertinent chips are getting a good clock
signal as well. On the directly related chips I can see a good clock signal
being output too (there's a 12mhz crystal which is sliced up into several
clock signals forming a bus that drives a lot of stuff on the board)..
However, I lack the technical skills to go further and see if the bitslice
cpu is actually DOING something, or if it's stuck/not running, etc. The
handshake logic between the two is confusing me. Not to mention I really
would rather not have to learn the microinstruction set the 2901 lash-up
implements. So, the DSD440 has sat on the back burner. The clearence is such
that you can't really get probes on a running board, so I built a small
extension to let me run it mostly out of the chassis. If I ever got time or
direction/help, I was thinking the next step would be to hook a logic
analyzer up to the address lines on the bitslice cpu and see if it's moving
or repeating a specific set of address accesses ad-nauseum. But I'm getting
out of my league there.
More information about the cctech