23 Jan
2014
23 Jan
'14
11:45 a.m.
One step forward, two steps back...
I got a few minutes to test last night and now I can't even boot the RK05
from 5000, let alone 0030.
No data is being xfered. Something flaky here. What a shock that 40 year
old hardware has intermittent problems.
I think I'm going to pull the TC08 interface cards (M8350/60) just to rule
out interference. Then I'll crank up the logic analyzer. Luckily I have
an extender card that has 68 of the more interesting omnibus signals
hard-wired to analyzer pods. Debugging the disk interface itself would be
a nightmare due to it's multi-board stackup.
Marc
On Wed, Jan 22, 2014 at 7:44 PM, Toby Thain <toby at telegraphics.com.au>wrote:
On 22/01/14 10:23 PM, Rob Doyle wrote:
...
The RK8E bootstrap is a little strange....
The bootstrap program /assumes/ that the disk is faster than the CPU.
When I sped up my PDP8, I exposed this race condition.
If the code is still executing the instruction at address 0031 after the
disk read operation, the disk wasn't read...
Thanks for sharing that little bit of knowledge. Awesome.
--Toby
Rob.