15 Feb
2014
15 Feb
'14
2:02 p.m.
My PDP-8/A with two 16K core boards has developed a flaky bit on one
of the boards. The failure is always in the LSB (11), but there is no
pattern to the addresses that I can discern.
The +20 volt supply appears to be ok (as is +/- 15 and +/- 5)... it
was 19.86v, but adjusting it to 20.00 didn't help.
DHKMAD.DG runs beautifully with only the lower 16k installed. But
while testing the upper 16k, eventually (30 sec. to several minutes) a
random word will fail with its LSB (Bit 11) in the wrong state.
Usually 1 when it should be 0 but *not* always!
Often this bombs the diagnostic program itself (I suppose if it's
self-relocated to the upper field). So far the failures have all been
in field 6 or 7.
Any tips for checking currents/timing? Or should I just replace the
G650, the H219B core stack, or both as a pair? They aren't easy to
find these days and my small chassis only has room for two core boards
that occupy two hex slots each - so they both have to be 16K boards,
as ADVENT requires 32K even though OS/8 runs in 12K :(
-Charles
16 Feb
16 Feb
7:59 p.m.
Hi, I agree with Jon, sounds like a sense, data latch or inhibit driver
problem on just bit 11, intermitants can be a pain.
What I would do, and have done, is to write a small test routine, write
7777, test bit 11 for 1, write 0000, agian test for bit 11 =0, the fail on
eitrher of these tests should go to a delay routine, the write and reads
must be to the failing field so will need to be indirect and you will need a
set data field to the suspect field at the start of the routine.
Then run this in a good field say 0, and "listen " to the
program using an AM radio, when a test fails you should here a change , you
can then try the old fraeser spray on likley components, with luck this may
induce the fault.
Iv used this sucsessfully in the past.
Due to the destructive read of core it can be a bit trickey to find out if
its a read or a write fault, what I do is to "break the loop" by opening the
drive to the inhibit and strapping this high /low, you can then sciope the
sense amp knowing what should be there.
All my experiance so far has been with an 8e , i dont know the actual 8a
boards but the theroy is the same,
Interesting you should mension 32k for adventure, I am ATM trying to get 4
x 8k core working so I can can dio the same im one of my 8m's
Cheers
DaveH
----- Original Message -----
From: "Charles" <charlesmorris800 at centurytel.net>
To: <cctalk at classiccmp.org>
Sent: Saturday, February 15, 2014 10:02 PM
Subject: PDP-8 core memory troubles
My PDP-8/A with two 16K core boards has developed a
flaky bit on one
of the boards. The failure is always in the LSB (11), but there is no
pattern to the addresses that I can discern.
The +20 volt supply appears to be ok (as is +/- 15 and +/- 5)... it
was 19.86v, but adjusting it to 20.00 didn't help.
DHKMAD.DG runs beautifully with only the lower 16k installed. But
while testing the upper 16k, eventually (30 sec. to several minutes) a
random word will fail with its LSB (Bit 11) in the wrong state.
Usually 1 when it should be 0 but *not* always!
Often this bombs the diagnostic program itself (I suppose if it's
self-relocated to the upper field). So far the failures have all been
in field 6 or 7.
Any tips for checking currents/timing? Or should I just replace the
G650, the H219B core stack, or both as a pair? They aren't easy to
find these days and my small chassis only has room for two core boards
that occupy two hex slots each - so they both have to be 16K boards,
as ADVENT requires 32K even though OS/8 runs in 12K :(
-Charles
-----
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2012.0.2247 / Virus Database: 3705/6599 - Release Date: 02/17/14
17 Feb
17 Feb
1:32 a.m.
In message <bsovf9l51pr6bp1098bii2o7nkgmkoiq94 at 4ax.com>
Charles <charlesmorris800 at centurytel.net> wrote:
My PDP-8/A with two 16K core boards has developed a
flaky bit on one
of the boards. The failure is always in the LSB (11), but there is no
pattern to the addresses that I can discern.
The +20 volt supply appears to be ok (as is +/- 15 and
+/- 5)... it
was 19.86v, but adjusting it to 20.00 didn't help.
DHKMAD.DG runs beautifully with only the lower 16k
installed. But
while testing the upper 16k, eventually (30 sec. to several minutes) a
random word will fail with its LSB (Bit 11) in the wrong state.
Usually 1 when it should be 0 but *not* always!
Often this bombs the diagnostic program itself (I
suppose if it's
self-relocated to the upper field). So far the failures have all been
in field 6 or 7.
Any tips for checking currents/timing? Or should I
just replace the
G650, the H219B core stack, or both as a pair? They aren't easy to
find these days and my small chassis only has room for two core boards
that occupy two hex slots each - so they both have to be 16K boards,
as ADVENT requires 32K even though OS/8 runs in 12K :(
-Charles
My first thought is -check the bus-driver-.
On my PDP-11/05 several boards had problems with warming up in both
ways, either operational till warming up or (most often) fail until
warmed up.
Rolf.
3965
days inactive
3967
days old
2 comments
3 participants
participants (3)
-
charlesmorris800＠centurytel.net -
david＠hheng.plus.com -
harten＠injectstar.de