24 Mar
2010
24 Mar
'10
1:42 p.m.
I am working with an IBM Monochrome "Black and White" Parallel card that
is showing a double image. I noticed that the 106 / 16+ tant cap at the
top left near where C1 / U47 is printed was broken off the board so I
put in a new one. There was no change. I tried another cap, no
change. I removed the cap entirely, no change. I am looking for a
schematic/docs/suggestions.
It looks something like this:
-------------------------------------------------
-128K 128K -
- -
- -
- -
- -
- -
- -
- -
-------------------------------------------------
Thanks.
24 Mar
24 Mar
2:08 p.m.
B Degnan wrote:
It looks something like this:
-------------------------------------------------
-128K 128K -
- -
- -
- -
- -
- -
- -
- -
-------------------------------------------------
I'd be looking for a stuck address bit, or possibly a failed RAM chip
(it's theoretically possible for an ESD-protection diode in a CMOS chip
to fail shorted, which would effectively pull that signal line high or low).
If I had a circuit diagram I could probably tell you which parts / pins
to test, but either way you'll need a good logic probe (one that has a
pulse detector in addition to the hi/lo comparators) or an oscilloscope
to pin down the fault.
My money's on a 74xx chip (possibly a counter, buffer or multiplexer)
that's had an output fail open.
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/
5:11 p.m.
B Degnan wrote:
I am working with an IBM Monochrome "Black and White" Parallel card that
is showing a double image. I noticed that the 106 / 16+ tant cap at the
top left near where C1 / U47 is printed was broken off the board so I
put in a new one. There was no change. I tried another cap, no
change. I removed the cap entirely, no change. I am looking for a
schematic/docs/suggestions.
That would be a 10,000,000 pF = 10 uF / 16V cap, in all likelihood just a local
power supply filter cap.
It looks something like this:
-------------------------------------------------
-128K 128K -
- -
- -
- -
- -
- -
- -
- -
-------------------------------------------------
Along with Phil's suggestion of memory addressing issues resulting in a
double-scan of the memory, the staggering of the images suggests interlacing
might be occurring when it shouldn't be. Sometimes it is possible to tweak the
V/H-hold controls enough on monitors to end up with interlacing occurring when
it shouldn't. Can you discern whether each of the scan images contains a full
set of lines vs half the number of lines?
I hate to ask, but what sort of monitor is this being displayed on?
IIRC the scan rate for MDA was higher than NTSC and I'm not not sure whether an
NTSC monitor would sync up or sync down to half the scan rate.
8:35 p.m.
In ~1990, I was given one IBM MDP card that had intermittent problems.
On close examination, I found that IBM (in original maufacturing) had bent
a pin under one of the chips, instead of it being through the hole. The
bent pin was making intermittent contact with the pad around the hole that
the pin was supposed to go through. Replacing that chip fixed it.
Was I too wasteful to replace the whole chip, rather than just one pin?
--
Grumpy Ol' Fred cisin at xenosoft.com
9:12 p.m.
In ~1990, I was given one IBM MDP card that had intermittent problems.
On close examination, I found that IBM (in original maufacturing) had bent
a pin under one of the chips, instead of it being through the hole. The
bent pin was making intermittent contact with the pad around the hole that
the pin was supposed to go through. Replacing that chip fixed it.
Was I too wasteful to replace the whole chip, rather than just one pin?
I suspect I would have desoldered the chip, cleaned out the hole,
straightened the pin and soldered it back in again.
-tony
11:51 p.m.
Date: Wed, 24 Mar 2010 13:35:48 -0700
From: cisin at xenosoft.com
To: cctalk at classiccmp.org
Subject: Re: IBM Monochrome "Black and White" Parallel card double image
In ~1990, I was given one IBM MDP card that had intermittent problems.
On close examination, I found that IBM (in original maufacturing) had bent
a pin under one of the chips, instead of it being through the hole. The
bent pin was making intermittent contact with the pad around the hole that
the pin was supposed to go through. Replacing that chip fixed it.
Was I too wasteful to replace the whole chip, rather than just one pin?
--
Hi Fred
I might have just ran a wire through the hole and tacked it onto the
pin.
This really depends on who I was doing this for. If for a customer,
I'd have replaced the chip. If for me, as I stated, I'd most like not
even bothered to remove the part.
Dwight
_________________________________________________________________
The New Busy is not the old busy. Search, chat and e-mail from your inbox.
http://www.windowslive.com/campaign/thenewbusy?ocid=PID27925::T:WLMTAGL:ON:…
25 Mar
25 Mar
7:02 p.m.
I might have just ran a wire through the hole and
tacked it onto the
pin.
This really depends on who I was doing this for. If for a customer=2C
I'd have replaced the chip. If for me=2C as I stated=2C I'd most like not
even bothered to remove the part.
It is my experience that desoldering and resoldering a chip (DIL, SOIC,
PQFP, etc) does not significantly reduce reliability. Many times I've
desodlerd a chip for some reason, and then soldered it back, and only
very rarely does that part fail first (or second or....).
-tony
24 Mar
24 Mar
8:55 p.m.
I am working with an IBM Monochrome "Black and White" Parallel card that
is showing a double image. I noticed that the 106 / 16+ tant cap at the
top left near where C1 / U47 is printed was broken off the board so I
put in a new one. There was no change. I tried another cap, no
change. I removed the cap entirely, no change. I am looking for a
schematic/docs/suggestions.
This is certainly not a capacitor problem...
I am working from memory (pardon the pun) and if necessary I can dig out
the schematics (they're in the Options and Adapters Techref).But I think
this board uses 2144s as the video memory (look for 18 pin DIL chips). My
expeirence is that that particular RAM has a record of unreliability, and
a missing adress line is just the sort of problem you get (and it would
cause the repeated image).
So my first guess is to replace the RAMs.
-tony
5463
days inactive
5464
days old
7 comments
6 participants
participants (6)
-
ard＠p850ug1.demon.co.uk -
billdeg＠degnanco.com -
cisin＠xenosoft.com -
classiccmp＠philpem.me.uk -
dkelvey＠hotmail.com -
hilpert＠cs.ubc.ca