24 Aug
2016
24 Aug
'16
10:41 a.m.
I have a MicroVAX II which has started garbling and losing characters output
to the console. It had seemed that re-seating the processor board would fix
it, but that no longer seems to be the case.
I was just wondering if anyone else has ever come across this failure mode?
Regards
Rob
24 Aug
24 Aug
10:53 a.m.
On Aug 24, 2016, at 9:41 AM, Rob Jarratt <robert.jarratt at ntlworld.com> wrote:
I have a MicroVAX II which has started garbling and losing characters output
to the console. It had seemed that re-seating the processor board would fix
it, but that no longer seems to be the case.
I was just wondering if anyone else has ever come across this failure mode?
I?ve seen similar with under two occasions.
1) Non-compliant USB adaptor - Doesn?t fully support RS232C levels
2) Poor grounding, noise in the data lines - long cable runs or devices on different sides
of the (US) power main.
I have not dealt with component problems on the KA630 itself.
Jerry
11:34 a.m.
I have a
MicroVAX II which has started garbling and losing
characters output to the console. [...]
I???ve seen similar with under two
occasions. [...] > It had seemed that re-seating the processor board
would fix it, but
> that no longer seems to be the case.
It occurs to me that it might not have been the re-seating that was
responsible, but something else incidental to that. For example, if
the problem is thermal, powering it off briefly might have helped; if
(to continue that theory) if the environment has been getting hotter,
it may be that leaving it off briefly now doesn't cool it enough.
Without more details, we can't really do much but take stabs in the
dark.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
11:55 a.m.
On 08/24/2016 10:34 AM, Mouse wrote:
Furthermore, "garbling" is horribly imprecise.
And, of course, you should make sure it is not the terminal,
cabinet kit or cable that has gone bad, instead of the CPU
board. (I can see where a bad cabinet kit or internal cable
could cause the serial clock to change speed, that would
cause the same result.)
Jon
I have a MicroVAX II which has started garbling and
losing
characters output to the console. [...]
I???ve seen similar with under two
occasions. [...]
12:05 p.m.
-----Original Message-----
From: cctalk [mailto:cctalk-bounces at classiccmp.org] On Behalf Of Mouse
Sent: 24 August 2016 16:35
To: cctalk at classiccmp.org
Subject: Re: MicroVAX II Console Garbling Characters
Furthermore, "garbling" is horribly imprecise.
Fair enough, let me see if I can answer more precisely.
I have a
MicroVAX II which has started garbling and losing characters
output to the console. [...]
I???ve seen similar with under two occasions.
[...]
If certain characters always get corrupted, and a given character always
gets
corrupted to the same thing?
No, the corruption appears to be random. Unfortunately I don't have a
protocol analyser to help, but I could try to get a log using Putty
connecting via a DECserver 90M. I will try that later.
If it's always the same characters getting corrupted, but they get
corrupted
to different things on different occasions?
If there's no uniformity on which characters get corrupted, but a given
character, if corrupted, always gets corrupted to the same thing?
If both which characters get corrupted and what they get converted into
show no consistency?
Regarding all the above, it looks pretty random to me, but I will try to get
a dump of the characters and compare with what they should be.
Each of those scenarios points towards a different constellation of
plausible
causes. (Furthermore "same characters" can
be taken either of two ways,
either "characters with the same bit pattern, regardless of where they
occur"
or "characters at the same place in the output,
regardless of bit
pattern". It
could even be a cross between those.)
Similar remarks apply to "losing".
Well, it does "lose" characters. It seems to be that after a short burst of
characters it just gives up and I don't see anymore characters for a while,
then it will show a few more after a bit of a pause (but not always).
The post also said
continue
> It had seemed that re-seating the processor
board would fix it, but
> that no longer seems to be the case.
It occurs to me that it might not have been the re-seating that was
responsible, but something else incidental to that. For example, if the
problem is thermal, powering it off briefly might have helped; if (to that theory) if the environment has been getting
hotter, it may be that
leaving it off briefly now doesn't cool it enough.
Unfortunately this happens both with the machine cold and warm.
Without more details, we can't really do much but take stabs in the dark.
I will try to get more info.
Incidentally, the terminal and the machine *were* plugged into different
sockets, but now they are connected to the same wall socket and the problem
persists.
Regard
Rob
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
9:59 p.m.
On 08/24/2016 11:05 AM, Rob Jarratt wrote:
Well, it does "lose" characters. It seems to
be that after
a short burst of characters it just gives up and I don't
see anymore characters for a while, then it will show a
few more after a bit of a pause (but not always).
Hmmm, that sounds like possibly a
buffer overflow in the
terminal. Some terminals had really slow micros running
them, and needed to stop transmission from the host for a
while. If the host is not obeying the stop-sending command
(could be the terminal is using the CTS line that is not
wired to the VAX, or the VAX is not set to obey Xoff/Xon
characters) then this would happen.
Jon
12:21 p.m.
Incidentally, the terminal and the machine *were*
plugged into different
sockets, but now they are connected to the same wall socket and the
problem persists.
Actually, since connecting terminal and machine to the same wall socket, the
garbling seems to have much reduced, but I do get lots of dropped characters
all the same.
For example the prompt to enter date and time appeared without the final
colon. I then pressed CTRL-R to repeat it and it just typed "PLEASE", with
no further characters, but it then echoed the date and time as I entered it.
Like I said, it seems to be able to send a certain number of characters and
then loses the rest.
I saw the cab kit comment, and I also wonder about that too. I am pretty
sure the terminal is OK though, because it seems to work OK on other
machines and I had exactly the same problems when using a DECserver 90M and
Putty (so no terminal at all). I can try swapping some parts around to see
if or when the problem goes away.
I suppose I was just hoping someone might have seen similar problems, but I
can do some deeper analysis to narrow it down more.
Regards
Rob
12:59 p.m.
On Aug 24, 2016, at 12:21 PM, Rob Jarratt
<robert.jarratt at ntlworld.com> wrote:
Check that the shield ground and signal ground are both connected. If they are, then the
fact that it depends on power suggests your RS232 interface is inadequately designed.
paul
Incidentally, the terminal and the machine *were*
plugged into different
sockets, but now they are connected to the same wall socket and the
problem persists.
Actually, since connecting terminal and machine to the same wall socket, the
garbling seems to have much reduced, but I do get lots of dropped characters
all the same.
25 Aug
25 Aug
4:08 a.m.
-----Original Message-----
From: Paul Koning [mailto:paulkoning at comcast.net]
Sent: 24 August 2016 17:59
To: rob at jarratt.me.uk; General Discussion: On-Topic and Off-Topic Posts
<cctalk at classiccmp.org>
Subject: Re: MicroVAX II Console Garbling Characters
Check that the shield ground and signal ground are both connected. If
they
On Aug 24, 2016, at 12:21 PM, Rob Jarratt
<robert.jarratt at ntlworld.com>
wrote:
Incidentally, the terminal and the machine *were*
plugged into
different sockets, but now they are connected to the same wall socket
and the problem persists.
Actually, since connecting terminal and machine to the same wall
socket, the garbling seems to have much reduced, but I do get lots of
dropped characters all the same. are, then the fact that it depends on power suggests
your RS232 interface
is
inadequately designed.
It turns out the problem was the cab kit. I replaced it with a spare and the
machine worked great. On closer inspection I found some corrosion and
possible evidence of a previous battery leak, so I am going to take it
apart, check it and clean it up.
The other thing that made me suspect the cab kit is that the act of pulling
the serial cable out (not unplugging the terminal) would cause the machine
to reset. MicroVAX II does this if you switch off the console terminal while
the terminal is connected.
Thanks
Rob
12:53 p.m.
On Thu, Aug 25, 2016 at 4:08 AM, Rob Jarratt
<robert.jarratt at ntlworld.com> wrote:
It turns out the problem was the cab kit. I replaced
it with a spare and the
machine worked great. On closer inspection I found some corrosion and
possible evidence of a previous battery leak, so I am going to take it
apart, check it and clean it up.
Ah. Good to hear. I'm in the middle of cleaning up battery damage on
a cab kit myself. I have it apart and the obvious corrosion
neutralized and cleaned, but I need to go over the traces where the
damage is before I replace the DE9. It's been a month of cleaning and
inspecting and repair (last night's project was a PDP-11/04)
The other thing that made me suspect the cab kit is
that the act of pulling
the serial cable out (not unplugging the terminal) would cause the machine
to reset. MicroVAX II does this if you switch off the console terminal while
the terminal is connected.
Hmm... that is a peculiar symptom. Glad it led to a solution.
Cheers,
-ethan
1:05 p.m.
It turns out the problem was the cab kit. I replaced
it with a spare and
the
machine worked great. On closer inspection I found
some corrosion and
possible evidence of a previous battery leak, so I am going to take it
apart,
check it and clean it up.
The other thing that made me suspect the cab kit is that the act of
pulling the
serial cable out (not unplugging the terminal) would
cause the machine to
reset. MicroVAX II does this if you switch off the console terminal while
the
terminal is connected.
Just a bit more info on this.
I tried cleaning up the corrosion, and that briefly seemed to make a
difference, but then I realised the problem persisted. I looked again and
realised that the corrosion around the DE-9 pins was a bit bad. Initially I
thought to solder some jumper wires around pins 2 and 3, but when the solder
would not melt very easily, I found that it kind of flaked off under a bit
of heat and physical pressure. This made me think that although I was
getting continuity with my DMM, the joint was not that good. So I desoldered
the whole DE-9 connector, cleaned the board with PCB cleaner and then
resoldered it back.
This seems to have fixed it.
Regards
Rob
1:11 p.m.
On Thu, Aug 25, 2016 at 1:05 PM, Rob Jarratt
<robert.jarratt at ntlworld.com> wrote:
Just a bit more info on this.
I tried cleaning up the corrosion, and that briefly seemed to make a
difference, but then I realised the problem persisted. I looked again and
realised that the corrosion around the DE-9 pins was a bit bad. Initially I
thought to solder some jumper wires around pins 2 and 3, but when the solder
would not melt very easily, I found that it kind of flaked off under a bit
of heat and physical pressure. This made me think that although I was
getting continuity with my DMM, the joint was not that good. So I desoldered
the whole DE-9 connector, cleaned the board with PCB cleaner and then
resoldered it back.
This seems to have fixed it.
My own cab kit was also exhibiting gray solder that did not flow, so
without even testing it, I presumed it was faulty and I removed the
DE-9 and cleaned the PCB. That's as far as I've gotten this week. I
still plan to test the resistance of the traces because, sometimes,
those are consumed or heavily damaged (I've been reading a lot of
Amiga Varta repair threads lately - that's another platform I'm
cleaning damage on). The question I have for the MicroVAX is if I
trust the integrity of the DE-9 itself. I can solder in a fresh
replacement, but I _think_ the old one is probably still good.
-ethan
1:52 p.m.
-----Original Message-----
From: cctalk [mailto:cctalk-bounces at classiccmp.org] On Behalf Of Ethan
Dicks
Sent: 25 August 2016 18:12
To: rob at jarratt.me.uk; General Discussion: On-Topic and Off-Topic Posts
<cctalk at classiccmp.org>
Subject: Re: MicroVAX II Console Garbling Characters
On Thu, Aug 25, 2016 at 1:05 PM, Rob Jarratt <robert.jarratt at ntlworld.com>
wrote:
Unfortunately I spoke too soon. When I tested it again a bit more extensively, it suddenly
stopped working. I am going to have to solder in some wires to pins 2 and 3 and see if
that helps.
Regards
Rob
Just a bit more info on this.
I tried cleaning up the corrosion, and that briefly seemed to make a
difference, but then I realised the problem persisted. I looked again
and realised that the corrosion around the DE-9 pins was a bit bad.
Initially I thought to solder some jumper wires around pins 2 and 3,
but when the solder would not melt very easily, I found that it kind
of flaked off under a bit of heat and physical pressure. This made me
think that although I was getting continuity with my DMM, the joint
was not that good. So I desoldered the whole DE-9 connector, cleaned
the board with PCB cleaner and then resoldered it back.
This seems to have fixed it.
My own cab kit was also exhibiting gray solder that did not flow, so without
even testing it, I presumed it was faulty and I removed the
DE-9 and cleaned the PCB. That's as far as I've gotten this week. I still plan
to
test the resistance of the traces because, sometimes, those are consumed or
heavily damaged (I've been reading a lot of Amiga Varta repair threads lately
- that's another platform I'm cleaning damage on). The question I have for
the MicroVAX is if I trust the integrity of the DE-9 itself. I can solder in a fresh
replacement, but I _think_ the old one is probably still good.
6:54 p.m.
Unfortunately I spoke too soon. When I tested it again a bit more
extensively, it suddenly stopped working. I am going to have to solder in
some wires to pins 2 and 3 and see if that helps.
Removing and then resoldering J3 (the smaller of the two ribbon cable connectors) seems to
have fixed it this time. It is managing to continue to operate for rather longer than
before. Fingers crossed I have resolved it this time. I will test it for longer tomorrow.
Regards
Rob
24 Aug
24 Aug
11:51 a.m.
On 08/24/2016 09:53 AM, Jerry Weiss wrote:
The console interface is a set of 9636, 9643 and 9639 chips
(8-pin DIP). I had mine blown out about 1988 or so when a
power supply failed. There are a couple capacitors and
resistors related to that circuitry, too.
Jon
On Aug 24,
2016, at 9:41 AM, Rob Jarratt <robert.jarratt at ntlworld.com> wrote:
I have a MicroVAX II which has started garbling and losing characters output
to the console. It had seemed that re-seating the processor board would fix
it, but that no longer seems to be the case.
I was just wondering if anyone else has ever come across this failure mode?
I?ve seen similar with under two occasions.
1) Non-compliant USB adaptor - Doesn?t fully support RS232C levels
2) Poor grounding, noise in the data lines - long cable runs or devices on different
sides of the (US) power main.
I have not dealt with component problems on the KA630 itself.
2888
days inactive
2889
days old
14 comments
6 participants
participants (6)
-
elson＠pico-systems.com -
ethan.dicks＠gmail.com -
jsw＠ieee.org -
mouse＠Rodents-Montreal.ORG -
paulkoning＠comcast.net -
robert.jarratt＠ntlworld.com