6 Sep
2014
6 Sep
'14
5:59 a.m.
After the successful restore of the HP9830B I continued with the HP9810A
that I was able to get from the Swedish Maritime Administration. (
http://www.datormuseum.se/computers/hewlett-packard/hp9810a )
The machine has been sitting in a storage container for many years until I
was able to rescue it. And of course it was dead. Testing the CPU boards in
the working 9830 gave that three out of four boards were faulty.
This far I have replaced four TTL chips. Three on the clock board and one
on the ALU board and all of them are made by National Semiconductor, date
codes are mid 1972. All are plastic DIP. The failure mode seems to be that
the outputs are floating. I guess that the bonding wires are broken.
Can it be that the moisture in the storage container that has made it into
the chips corroding the wires?
What is the experience when it comes to different manufacturer and plastic
DIP TTL? Which are better, which are worse after 40 years?
6 Sep
6 Sep
6:18 a.m.
There was a problem with some plastics used in the early days, it was
mentioned in the press at the time (EDN etc), I dont remember it being
company specific
Dave Caroline
On 06/09/2014, Mattis Lind <mattislind at gmail.com> wrote:
After the successful restore of the HP9830B I
continued with the HP9810A
that I was able to get from the Swedish Maritime Administration. (
http://www.datormuseum.se/computers/hewlett-packard/hp9810a )
The machine has been sitting in a storage container for many years until I
was able to rescue it. And of course it was dead. Testing the CPU boards in
the working 9830 gave that three out of four boards were faulty.
This far I have replaced four TTL chips. Three on the clock board and one
on the ALU board and all of them are made by National Semiconductor, date
codes are mid 1972. All are plastic DIP. The failure mode seems to be that
the outputs are floating. I guess that the bonding wires are broken.
Can it be that the moisture in the storage container that has made it into
the chips corroding the wires?
What is the experience when it comes to different manufacturer and plastic
DIP TTL? Which are better, which are worse after 40 years?
7:31 a.m.
Can I hijack this thread for another IC quality related question?
I'm in the early stages of trying to restore a ROLM CBX 8000 and I'm
finding that on certain cards; mostly 8550 shelf expanders and 8551
16-channel coders; there are noticeable amounts of TI 7400 series logic
produced in the following facilities:
EL SALVADOR
MALAYSIA
TAIWAN
where the IC leads have experienced, in my opinion, "excessive corrosion".
Date of manufacture on these parts ranges from mid-1970s to early 1980s.
ICs in the switch produced by other vendors e.g. AMD, Mostek, Zilog, Intel,
Rockwell, GI, Signetics, Motorola, OKI, etc as well as passive components
don't seem to really have this problem; leads are free of corrosion; it's
really just 7400 series "little logic" from TI on these two card types,
where the corrosion is occurring.
Has anyone ever seen this sort of thing before? Before I got the unit,
AFAIK, it was stored in a nominally dry-and-out-of-the-elements but not
really climate-controlled storage locker for probably at least a decade or
so. I've had other bits of equipment spend some time (years) in lockers and
never seen any internal corrosion... the corrosion on 8550 and 8551 cards
occurs across all six shelves of the PBX from top to bottom so it doesn't
appear that, say, one part of the PBX was any more exposed to the elements,
corrosive atmosphere, etc.
No meaningful corrosion on the PBX frame, card shelves, terminal strips,
connectors, fuse panel, ... any other metal component, really.
Was TI cheaping out on the metal used for package leads for a little while
there, or ...?
Best,
Sean
On Sat, Sep 6, 2014 at 10:18 AM, Dave Caroline <dave.thearchivist at gmail.com>
wrote:
There was a problem with some plastics used in the
early days, it was
mentioned in the press at the time (EDN etc), I dont remember it being
company specific
Dave Caroline
On 06/09/2014, Mattis Lind <mattislind at gmail.com> wrote:
After the successful restore of the HP9830B I
continued with the HP9810A
that I was able to get from the Swedish Maritime Administration. (
http://www.datormuseum.se/computers/hewlett-packard/hp9810a )
The machine has been sitting in a storage container for many years until
I
was able to rescue it. And of course it was dead.
Testing the CPU boards
in
the working 9830 gave that three out of four
boards were faulty.
This far I have replaced four TTL chips. Three on the clock board and one
on the ALU board and all of them are made by National Semiconductor, date
codes are mid 1972. All are plastic DIP. The failure mode seems to be
that
the outputs are floating. I guess that the
bonding wires are broken.
Can it be that the moisture in the storage container that has made it
into
the chips corroding the wires?
What is the experience when it comes to different manufacturer and
plastic
DIP TTL? Which are better, which are worse after
40 years?
11:14 a.m.
> On 06/09/2014, Mattis Lind <mattislind at
gmail.com> wrote:
>> What is the experience when it comes to different manufacturer and
> plastic
>> DIP TTL? Which are better, which are worse after 40 years?
On 2014-Sep-06, at 8:31 AM, Sean Caron wrote:
Can I hijack this thread for another IC quality
related question?
I'm in the early stages of trying to restore a ROLM CBX 8000 and I'm
finding that on certain cards; mostly 8550 shelf expanders and 8551
16-channel coders; there are noticeable amounts of TI 7400 series logic
produced in the following facilities:
EL SALVADOR
MALAYSIA
TAIWAN
where the IC leads have experienced, in my opinion, "excessive corrosion".
Date of manufacture on these parts ranges from mid-1970s to early 1980s.
ICs in the switch produced by other vendors e.g. AMD, Mostek, Zilog, Intel,
Rockwell, GI, Signetics, Motorola, OKI, etc as well as passive components
don't seem to really have this problem; leads are free of corrosion; it's
really just 7400 series "little logic" from TI on these two card types,
where the corrosion is occurring.
Has anyone ever seen this sort of thing before? Before I got the unit,
AFAIK, it was stored in a nominally dry-and-out-of-the-elements but not
really climate-controlled storage locker for probably at least a decade or
so. I've had other bits of equipment spend some time (years) in lockers and
never seen any internal corrosion... the corrosion on 8550 and 8551 cards
occurs across all six shelves of the PBX from top to bottom so it doesn't
appear that, say, one part of the PBX was any more exposed to the elements,
corrosive atmosphere, etc.
No meaningful corrosion on the PBX frame, card shelves, terminal strips,
connectors, fuse panel, ... any other metal component, really.
Was TI cheaping out on the metal used for package leads for a little while
there, or ??
To my observation, the answer there is yes and no.
Some plastic TI TTL ICs from a period somewhere around the mid/late-70s were manufactured
with steel pins that corrode underneath the plating. They might look fine but simply fall
off with the slightest stress, or develop intermittent conductivity.
But with that said, a lot of TI ICs from around then show discolouring on the pins that
looks horrible and suspect, but the discolouring doesn't necessarily imply the
corrosion underneath. I believe the visible discolouring was (often) just oxide of the
silver final plating of the pins. It shows up on TI because they were the only ones doing
that plating.
I think the TI steel-pin rot is the only wide-spread IC fault I've really correlated
to a particular manufacturer/period, and I don't know whether it was company-wide or
production that came out of maybe just one of their plants. (It can occur with other
manufacturer's steel pins but then it seems to be more a result of particularly bad
environmental/storage conditions of the machine and not really a surprise.)
When it comes to plastic, well, there were reasons ceramic was an option.
1:24 p.m.
Hi Brett,
Ah, very interesting; I did not know that. Yes, these are all TI ICs in
plastic DIP package produced in any of three facilities that I mention in
my prior message. I haven't inspected the "corrosion" that closely or
actively tried to "remove" it yet so now I'm not 100% clear if what I am
seeing is oxide residue from the plating, or if it is actual corrosion... I
just eyeballed it and assumed it was rust; the "residue" on the pins is a
dark red-brown sort of color; maybe a little darker than I'd expect iron
oxide to look, actually, but my examination wasn't under the best light.
I haven't gotten to the point of testing anything under power but the ICs
seem to be fairly securely attached to the board; I don't see any broken
pins and I can't easily wiggle anything around. Obviously these will need a
lot of cleanup but nothing unsalvageable. Even if I do have to replace the
odd part here and there, most of the affected components should be easy to
source; nothing too obscure.
Thanks for the response! It certainly explains why the phenomenon was
limited to TI ICs!
Best,
Sean
On Sat, Sep 6, 2014 at 3:14 PM, Brent Hilpert <hilpert at cs.ubc.ca> wrote:
> On
06/09/2014, Mattis Lind <mattislind at gmail.com> wrote:
>> What is the experience when it comes to different manufacturer and
> plastic
>> DIP TTL? Which are better, which are worse after 40 years?
On 2014-Sep-06, at 8:31 AM, Sean Caron wrote:
Can I hijack this thread for another IC quality
related question?
I'm in the early stages of trying to restore a ROLM CBX 8000 and I'm
finding that on certain cards; mostly 8550 shelf expanders and 8551
16-channel coders; there are noticeable amounts of TI 7400 series logic
produced in the following facilities:
EL SALVADOR
MALAYSIA
TAIWAN
where the IC leads have experienced, in my opinion, "excessive
corrosion".
Date of manufacture on these parts ranges from
mid-1970s to early 1980s.
ICs in the switch produced by other vendors e.g. AMD, Mostek, Zilog,
Intel,
Rockwell, GI, Signetics, Motorola, OKI, etc as
well as passive components
don't seem to really have this problem; leads are free of corrosion; it's
really just 7400 series "little logic" from TI on these two card types,
where the corrosion is occurring.
Has anyone ever seen this sort of thing before? Before I got the unit,
AFAIK, it was stored in a nominally dry-and-out-of-the-elements but not
really climate-controlled storage locker for probably at least a decade
or
so. I've had other bits of equipment spend
some time (years) in lockers
and
never seen any internal corrosion... the
corrosion on 8550 and 8551 cards
occurs across all six shelves of the PBX from top to bottom so it doesn't
appear that, say, one part of the PBX was any more exposed to the
elements,
corrosive atmosphere, etc.
No meaningful corrosion on the PBX frame, card shelves, terminal strips,
connectors, fuse panel, ... any other metal component, really.
Was TI cheaping out on the metal used for package leads for a little
while
there, or ??
To my observation, the answer there is yes and no.
Some plastic TI TTL ICs from a period somewhere around the mid/late-70s
were manufactured with steel pins that corrode underneath the plating. They
might look fine but simply fall off with the slightest stress, or develop
intermittent conductivity.
But with that said, a lot of TI ICs from around then show discolouring on
the pins that looks horrible and suspect, but the discolouring doesn't
necessarily imply the corrosion underneath. I believe the visible
discolouring was (often) just oxide of the silver final plating of the
pins. It shows up on TI because they were the only ones doing that plating.
I think the TI steel-pin rot is the only wide-spread IC fault I've really
correlated to a particular manufacturer/period, and I don't know whether it
was company-wide or production that came out of maybe just one of their
plants. (It can occur with other manufacturer's steel pins but then it
seems to be more a result of particularly bad environmental/storage
conditions of the machine and not really a surprise.)
When it comes to plastic, well, there were reasons ceramic was an option.
8:33 a.m.
Hello Mattis,
On Sat, Sep 6, 2014 at 8:59 AM, Mattis Lind <mattislind at gmail.com> wrote:
What is the experience when it comes to different manufacturer and plastic
DIP TTL? Which are better, which are worse after 40 years?
For what it's worth, the Altair 8800 that I recently repaired had around
12-13 bad 74XX series chips on the front panel and one on the CPU board.
Among the group were (5) Nat Semi 1st logo DM/SN7405N and (1) Nat Semi 2nd
logo DM74LS30N. All parts had a 1980 date.
All chips failed with one or more dead / floating pins. At least one 7405
had all output pins dead, with most having two or more dead outputs. The
7430 had a single dead input.
11:46 p.m.
Mattis Lind wrote:
After the successful restore of the HP9830B I
continued with the HP9810A
that I was able to get from the Swedish Maritime Administration. (
http://www.datormuseum.se/computers/hewlett-packard/hp9810a )
The machine has been sitting in a storage container for many years until I
was able to rescue it. And of course it was dead. Testing the CPU boards in
the working 9830 gave that three out of four boards were faulty.
This far I have replaced four TTL chips. Three on the clock board and one
on the ALU board and all of them are made by National Semiconductor, date
codes are mid 1972. All are plastic DIP. The failure mode seems to be that
the outputs are floating. I guess that the bonding wires are broken.
Can it be that the moisture in the storage container that has made it into
the chips corroding the wires?
What is the experience when it comes to different manufacturer and plastic
DIP TTL? Which are better, which are worse after 40 years?
Not exactly an answer to your question but I have similar experiences with
russian chips made before 1990.
I've reapaired an GDR made Logic Analyzer calles MC80-LA in the past,
(http://www.robotrontechnik.de/index.htm?/html/computer/la.htm)
it wasn't muchg effort to repair the first fault, an russian K155LA3 (7400)
on the video board was broken. After this the machine was working...for a
short time. In the next following 2 weeks I had to replace approx. 60 (!)
broken russian TTL chips with exactly the same fault, floating outputs.
(relativley easy to find if you can just test the output Levels with an
scope or a TTL-Level probe).
Russian chips in ceramics don't had this problem, they where and still are
a live, only the chips in dark green or brown plasitcs where affected.
the chips died like popcorn...., I think that has todo with some moisture
in the plastic enclosure of the chips. Maybe they where still alive whn I
had dryed them in an oven before applying power to them?
Regards,
Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Stra?e 42, 09600 Obersch?na, USt-Id: DE253710583
www.tsht.de, info at tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
7 Sep
7 Sep
6:46 a.m.
Wow.. yeah... that's what I'd like to avoid; replacing hundreds of 7400
series TTL ICs... If I really had to replace them, I could probably get the
job down into the double-digit numbers of IC replacements by just reducing
the switch configuration a little bit...
Thanks for posting the link to the page about the logic analyzer from the
DDR. I'm very much enjoying reading it now. I always get a huge kick out of
seeing how they did things behind the iron curtain (though I'll keep my HP
1662A, thanks!)
I wonder if there is much of a computer collecting community in the former
Eastern Bloc and, if so, how successful they have been in preserving
historical items from the Cold War era. Eastern Bloc engineers tried some
very interesting stuff over there in isolation from the "mainstream" in the
West... ternary logic, hybrid analog/digital CPUs, VLIW - before we ever
saw that "over here", etc.
I think some of this was an attempt to make up for trailing in process
technology and assembly quality with architectural cleverness (although as
Intel well demonstrates, he with the best process technology always wins in
the long run, architecture be damned, LOL), some was maybe just straight-up
old time engineering creativity at work. It's neat to see what people come
up with when they're a bit removed from the "prevailing knowledge"
(although there was certainly PLENTY of straight cloning of Western designs
being done as well).
Best,
Sean
On Sun, Sep 7, 2014 at 3:46 AM, Holm Tiffe <holm at freibergnet.de> wrote:
Mattis Lind wrote:
After the successful restore of the HP9830B I
continued with the HP9810A
that I was able to get from the Swedish Maritime Administration. (
http://www.datormuseum.se/computers/hewlett-packard/hp9810a )
The machine has been sitting in a storage container for many years until
I
was able to rescue it. And of course it was dead.
Testing the CPU boards
in
the working 9830 gave that three out of four
boards were faulty.
This far I have replaced four TTL chips. Three on the clock board and one
on the ALU board and all of them are made by National Semiconductor, date
codes are mid 1972. All are plastic DIP. The failure mode seems to be
that
the outputs are floating. I guess that the
bonding wires are broken.
Can it be that the moisture in the storage container that has made it
into
the chips corroding the wires?
What is the experience when it comes to different manufacturer and
plastic
DIP TTL? Which are better, which are worse after
40 years?
Not exactly an answer to your question but I have similar experiences with
russian chips made before 1990.
I've reapaired an GDR made Logic Analyzer calles MC80-LA in the past,
(http://www.robotrontechnik.de/index.htm?/html/computer/la.htm)
it wasn't muchg effort to repair the first fault, an russian K155LA3 (7400)
on the video board was broken. After this the machine was working...for a
short time. In the next following 2 weeks I had to replace approx. 60 (!)
broken russian TTL chips with exactly the same fault, floating outputs.
(relativley easy to find if you can just test the output Levels with an
scope or a TTL-Level probe).
Russian chips in ceramics don't had this problem, they where and still are
a live, only the chips in dark green or brown plasitcs where affected.
the chips died like popcorn...., I think that has todo with some moisture
in the plastic enclosure of the chips. Maybe they where still alive whn I
had dryed them in an oven before applying power to them?
Regards,
Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Stra?e 42, 09600 Obersch?na, USt-Id: DE253710583
www.tsht.de, info at tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
9:59 a.m.
Related to the issue on corrosion on the pins I can not see any such
problem in this HP9810 I am working with. There is very little black oxide
on the TI chips with silver plated pins and nothing on the other pins. So
it is not obvious to me that it has been exposed to a bad climate. But
still, chips have failed. The PDP-11/04 dated around 1976 and the HP9830,
same vintage, I worked with earlier had only a few failed chips in the
entire machine.
Now the count is 6 NS chips and one Signetics chip. A very annoying N8885
chip has failed. A quad input NOR which unfortunately doesn't have the same
pinout as the 7402. Maybe it is similar to the 7436 which is supposed be
the same as 7402 but different pin config. Didn't find the data sheet
though.
From now on every NS chip is a suspect...
2014-09-07 16:46 GMT+02:00 Sean Caron <scaron at umich.edu>:
Wow.. yeah... that's what I'd like to avoid;
replacing hundreds of 7400
series TTL ICs... If I really had to replace them, I could probably get the
job down into the double-digit numbers of IC replacements by just reducing
the switch configuration a little bit...
Thanks for posting the link to the page about the logic analyzer from the
DDR. I'm very much enjoying reading it now. I always get a huge kick out of
seeing how they did things behind the iron curtain (though I'll keep my HP
1662A, thanks!)
I wonder if there is much of a computer collecting community in the former
Eastern Bloc and, if so, how successful they have been in preserving
historical items from the Cold War era. Eastern Bloc engineers tried some
very interesting stuff over there in isolation from the "mainstream" in the
West... ternary logic, hybrid analog/digital CPUs, VLIW - before we ever
saw that "over here", etc.
I think some of this was an attempt to make up for trailing in process
technology and assembly quality with architectural cleverness (although as
Intel well demonstrates, he with the best process technology always wins in
the long run, architecture be damned, LOL), some was maybe just straight-up
old time engineering creativity at work. It's neat to see what people come
up with when they're a bit removed from the "prevailing knowledge"
(although there was certainly PLENTY of straight cloning of Western designs
being done as well).
Best,
Sean
On Sun, Sep 7, 2014 at 3:46 AM, Holm Tiffe <holm at freibergnet.de> wrote:
with
Mattis Lind wrote:
> After the successful restore of the HP9830B I continued with the
HP9810A
> that I was able to get from the Swedish
Maritime Administration. (
> http://www.datormuseum.se/computers/hewlett-packard/hp9810a )
>
> The machine has been sitting in a storage container for many years
until
I
> was able to rescue it. And of course it was dead. Testing the CPU
boards
in
> the working 9830 gave that three out of four boards were faulty.
>
> This far I have replaced four TTL chips. Three on the clock board and
one
> on the ALU board and all of them are made by
National Semiconductor,
date
codes are
mid 1972. All are plastic DIP. The failure mode seems to be
that
the outputs are floating. I guess that the
bonding wires are broken.
Can it be that the moisture in the storage container that has made it
into
the chips corroding the wires?
What is the experience when it comes to different manufacturer and
plastic
DIP TTL? Which are better, which are worse after
40 years?
Not exactly an answer to your question but I have similar experiences russian chips made before 1990.
I've reapaired an GDR made Logic Analyzer calles MC80-LA in the past,
(http://www.robotrontechnik.de/index.htm?/html/computer/la.htm)
it wasn't muchg effort to repair the first fault, an russian K155LA3
(7400)
on the video board was broken. After this the
machine was working...for a
short time. In the next following 2 weeks I had to replace approx. 60 (!)
broken russian TTL chips with exactly the same fault, floating outputs.
(relativley easy to find if you can just test the output Levels with an
scope or a TTL-Level probe).
Russian chips in ceramics don't had this problem, they where and still
are
a live, only the chips in dark green or brown
plasitcs where affected.
the chips died like popcorn...., I think that has todo with some moisture
in the plastic enclosure of the chips. Maybe they where still alive whn I
had dryed them in an oven before applying power to them?
Regards,
Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Stra?e 42, 09600 Obersch?na, USt-Id: DE253710583
www.tsht.de, info at tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
12:38 p.m.
On 2014-Sep-07, at 10:59 AM, Mattis Lind wrote:
Related to the issue on corrosion on the pins I can
not see any such
problem in this HP9810 I am working with. There is very little black oxide
on the TI chips with silver plated pins and nothing on the other pins. So
it is not obvious to me that it has been exposed to a bad climate. But
still, chips have failed. The PDP-11/04 dated around 1976 and the HP9830,
same vintage, I worked with earlier had only a few failed chips in the
entire machine.
Now the count is 6 NS chips and one Signetics chip. A very annoying N8885
chip has failed. A quad input NOR which unfortunately doesn't have the same
pinout as the 7402. Maybe it is similar to the 7436 which is supposed be
the same as 7402 but different pin config. Didn't find the data sheet
though.
I remember that solitary N8885A when reverse-engineering the 9830, and checking for a 7400
equivalent.
The 7436 seems to be pretty much non-existent. There's an obscure reference here:
http://www.datasheetarchive.com/dl/Scans-005/Scans-00104143.pdf
The strange thing is there are half-a-dozen 7402s in the machine.
Looking at the specs they may have used the 8885 because it's slightly faster than the
7402 and it doesn't look like a 74H02 was being made, but to a cursory examination I
don't see where the few nS benefit would be needed in the circuit.
Oh well .. if you use a 74LS02 or 74S02 at least it's only two pins that have to be
swapped.
10:31 p.m.
2014-09-07 22:38 GMT+02:00 Brent Hilpert <hilpert at cs.ubc.ca>:
I remember that solitary N8885A when reverse-engineering the 9830, and
checking for a 7400 equivalent.
The 7436 seems to be pretty much non-existent. There's an obscure
reference here:
http://www.datasheetarchive.com/dl/Scans-005/Scans-00104143.pdf
The strange thing is there are half-a-dozen 7402s in the machine.
Looking at the specs they may have used the 8885 because it's slightly
faster than the 7402 and it doesn't look like a 74H02 was being made, but
to a cursory examination I don't see where the few nS benefit would be
needed in the circuit.
Oh well .. if you use a 74LS02 or 74S02 at least it's only two pins that
have to be swapped.
Yes. I could install a 7402 type chip 180 degrees rotated and just swap
the power
pins, I guess. But I happened to find used N8885 pulls at $0.50
each, so I bought a few. Hope that they are better.
I checked the rest of the boards in the machine. Other than already faulty
ones I found that there are no more than maybe five or six NS chips in it.
Two of them seems to be working and three or four are just standard 74xx
chips. There is one more N8885 chip in the memory section.
3:32 p.m.
N8885 is available on Ebay ..
Best,
Thomas
On Sun, Sep 7, 2014 at 7:59 AM, Mattis Lind <mattislind at gmail.com> wrote:
Related to the issue on corrosion on the pins I can
not see any such
problem in this HP9810 I am working with. There is very little black oxide
on the TI chips with silver plated pins and nothing on the other pins. So
it is not obvious to me that it has been exposed to a bad climate. But
still, chips have failed. The PDP-11/04 dated around 1976 and the HP9830,
same vintage, I worked with earlier had only a few failed chips in the
entire machine.
Now the count is 6 NS chips and one Signetics chip. A very annoying N8885
chip has failed. A quad input NOR which unfortunately doesn't have the same
pinout as the 7402. Maybe it is similar to the 7436 which is supposed be
the same as 7402 but different pin config. Didn't find the data sheet
though.
From now on every NS chip is a suspect...
2014-09-07 16:46 GMT+02:00 Sean Caron <scaron at umich.edu>:
moisture
Wow.. yeah... that's what I'd like to
avoid; replacing hundreds of 7400
series TTL ICs... If I really had to replace them, I could probably get
the
job down into the double-digit numbers of IC
replacements by just
reducing
the switch configuration a little bit...
Thanks for posting the link to the page about the logic analyzer from the
DDR. I'm very much enjoying reading it now. I always get a huge kick out
of
seeing how they did things behind the iron
curtain (though I'll keep my
HP
1662A, thanks!)
I wonder if there is much of a computer collecting community in the
former
Eastern Bloc and, if so, how successful they have
been in preserving
historical items from the Cold War era. Eastern Bloc engineers tried some
very interesting stuff over there in isolation from the "mainstream" in
the
West... ternary logic, hybrid analog/digital
CPUs, VLIW - before we ever
saw that "over here", etc.
I think some of this was an attempt to make up for trailing in process
technology and assembly quality with architectural cleverness (although
as
Intel well demonstrates, he with the best process
technology always wins
in
the long run, architecture be damned, LOL), some
was maybe just
straight-up
old time engineering creativity at work. It's
neat to see what people
come
up with when they're a bit removed from the
"prevailing knowledge"
(although there was certainly PLENTY of straight cloning of Western
designs
being done as well).
Best,
Sean
On Sun, Sep 7, 2014 at 3:46 AM, Holm Tiffe <holm at freibergnet.de> wrote:
with
working...for a
Mattis Lind wrote:
> After the successful restore of the HP9830B I continued with the
HP9810A
> that I was able to get from the Swedish
Maritime Administration. (
> http://www.datormuseum.se/computers/hewlett-packard/hp9810a )
>
> The machine has been sitting in a storage container for many years
until
I
> was able to rescue it. And of course it was dead. Testing the CPU
boards
in
> the working 9830 gave that three out of four boards were faulty.
>
> This far I have replaced four TTL chips. Three on the clock board and
one
> on the ALU board and all of them are made by
National Semiconductor,
date
codes are
mid 1972. All are plastic DIP. The failure mode seems to be
that
the outputs are floating. I guess that the
bonding wires are broken.
Can it be that the moisture in the storage container that has made it
into
the chips corroding the wires?
What is the experience when it comes to different manufacturer and
plastic
DIP TTL? Which are better, which are worse after
40 years?
Not exactly an answer to your question but I have similar experiences russian chips made before 1990.
I've reapaired an GDR made Logic Analyzer calles MC80-LA in the past,
(http://www.robotrontechnik.de/index.htm?/html/computer/la.htm)
it wasn't muchg effort to repair the first fault, an russian K155LA3
(7400)
> on the video board was broken. After this the machine was > short time. In the next following 2 weeks I
had to replace approx. 60
(!)
broken
russian TTL chips with exactly the same fault, floating outputs.
(relativley easy to find if you can just test the output Levels with an
scope or a TTL-Level probe).
Russian chips in ceramics don't had this problem, they where and still
are
> a live, only the chips in dark green or brown plasitcs where affected.
>
> the chips died like popcorn...., I think that has todo with some > in the plastic enclosure of the chips. Maybe
they where still alive
whn I
had dryed
them in an oven before applying power to them?
Regards,
Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Stra?e 42, 09600 Obersch?na, USt-Id: DE253710583
www.tsht.de, info at tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
10:29 p.m.
Sean Caron wrote:
Wow.. yeah... that's what I'd like to avoid;
replacing hundreds of 7400
series TTL ICs... If I really had to replace them, I could probably get the
job down into the double-digit numbers of IC replacements by just reducing
the switch configuration a little bit...
Who doesn't want to avoid that?
The problem is, nobody tells you how much of chips are dying after ysour
replaced the last faulty one. Every time you think that this thing is ready
now for the next years.
Thanks for posting the link to the page about the logic analyzer from the
DDR. I'm very much enjoying reading it now. I always get a huge kick out of
seeing how they did things behind the iron curtain (though I'll keep my HP
1662A, thanks!)
I have a 1631D, 2 Tek 7D01 and this MC80-LA.
The MC80-LA is the biggest one because they used acase from some kind of
universal development system for the LA, including this 12" CRT. The Card
cage was also pretty much Standard (Robotron, GDR) with 170x215 mm cards.
Most of Robotron Z80 Systems used this configuration with the same bus
system. I don't have space on the deskt for really using this LA.
I wonder if there is much of a computer collecting community in the former
Eastern Bloc and, if so, how successful they have been in preserving
historical items from the Cold War era. Eastern Bloc engineers tried some
very interesting stuff over there in isolation from the "mainstream" in the
West... ternary logic, hybrid analog/digital CPUs, VLIW - before we ever
saw that "over here", etc.
The linked site is one from such a community, www.robotrontechnik.de.
They people there collect Items from robotron as the name already tells,
but not only that stuff. If you get in eastern PDP11's you find many
interresting russian sites too.
Russians are obscure from time to time :-).
They cloned entire machine designs, the stuff is looking from the outside
similar to the western original and is calles a clone.
But if you are looking at the inside most things are made totally
different. As an example there was much rumour about the cloned VAXes,
but the russians built the first VAX CPU on its own only from the processor
hanbdbook of the 11/750 .. but as a single IC! That integrated CPU never
existed in the US, so this is'nt really a clone at all.
Or this thing:
http://www.youtube.com/watch?v=BxRSEyj6twY
A Casio Clone? I've own a such looking casio calculator, but the russian
one has a PDP11 CPU.
Things you mentioned above (analog/digital, VLIW etc..) are currently
unknown to me. I don't know of saved machines like this.
I think some of this was an attempt to make up for trailing in process
technology and assembly quality with architectural cleverness (although as
Intel well demonstrates, he with the best process technology always wins in
the long run, architecture be damned, LOL), some was maybe just straight-up
old time engineering creativity at work. It's neat to see what people come
up with when they're a bit removed from the "prevailing knowledge"
(although there was certainly PLENTY of straight cloning of Western designs
being done as well).
Best,
Sean
Hmm, yes. The war of people wanting the Z8000 and others wanting the 8086
line of CPUs existed in the GDR (DDR) before the fall of the iron curtain
too. There where companies like EAW that build process control systems
around the Zilog processors (for sure the better architecture)
(http://www.robotrontechnik.de/index.htm?/html/computer/p8000.htm)
and wanted that line do be developed further and the "Big Blue" of the GDR,
robotron that wanted the 8086 Series. But the small country hasn't had enough
recources to do both of them..
I have some Kind of Zilog S8000 "compatible" System running Unix (Zeus
clone) on a 4Mhz Z8000.
That was'nbt really a S8000 Clone, the system was build out of 3 SBC
connected thogether (Z80 for floppy I/O and serail interfaces, the Z8000
CPU Board with Memory Cards in slots and another z80 card for the
Winchester drive 50MByte. The System was capable to run CP/M and RIO
Compatible OS'es on the Z80 part too).
http://www.robotrontechnik.de/index.htm?/html/computer/p8000.htm
http://www.robotrontechnik.de/index.htm?/html/computer/p8000compact.htm
There existed a subsystem with an 80286 and extra memory for that machine
too, so this thig was able to run DOS on the 80286 over the serial
connected terminal. So the available OSes for that thing where OS/M (CP/M),
UDOS (RIO), WEGA (Zeus Unixi, V7 with Extensions) and finally MSDOS.
Regards,
Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Stra?e 42, 09600 Obersch?na, USt-Id: DE253710583
www.tsht.de, info at tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
24 Sep
24 Sep
8:11 a.m.
An update on the topic failing NS chips.
Now the calculator is running again:
http://www.datormuseum.se/computers/hewlett-packard/hp9810a
But it took replacing in total 13 chips to get it running. 12 out those
were NS chips and one SIgnetics chip. Mostly standard, 7400, 7402, 7404,
7410 and 7474. But also more annoying ones like a 74H53 and a N8885. The
74H53 I got from Bulgaria at a very good price! Two 74H53 and five 74H21
for five US dollars including shipping.
There are still around ten NS chip in the machine. I hope they will keep
working as they should.
2014-09-08 8:29 GMT+02:00 Holm Tiffe <holm at freibergnet.de>:
Sean Caron wrote:
Wow.. yeah... that's what I'd like to
avoid; replacing hundreds of 7400
series TTL ICs... If I really had to replace them, I could probably get
the
job down into the double-digit numbers of IC
replacements by just
reducing
the switch configuration a little bit...
Who doesn't want to avoid that?
The problem is, nobody tells you how much of chips are dying after ysour
replaced the last faulty one. Every time you think that this thing is ready
now for the next years.
Thanks for posting the link to the page about the logic analyzer from the
DDR. I'm very much enjoying reading it now. I always get a huge kick out
of
seeing how they did things behind the iron
curtain (though I'll keep my
HP
1662A, thanks!)
I have a 1631D, 2 Tek 7D01 and this MC80-LA.
The MC80-LA is the biggest one because they used acase from some kind of
universal development system for the LA, including this 12" CRT. The Card
cage was also pretty much Standard (Robotron, GDR) with 170x215 mm cards.
Most of Robotron Z80 Systems used this configuration with the same bus
system. I don't have space on the deskt for really using this LA.
I wonder if there is much of a computer collecting community in the
former
Eastern Bloc and, if so, how successful they have
been in preserving
historical items from the Cold War era. Eastern Bloc engineers tried some
very interesting stuff over there in isolation from the "mainstream" in
the
West... ternary logic, hybrid analog/digital
CPUs, VLIW - before we ever
saw that "over here", etc.
The linked site is one from such a community, www.robotrontechnik.de.
They people there collect Items from robotron as the name already tells,
but not only that stuff. If you get in eastern PDP11's you find many
interresting russian sites too.
Russians are obscure from time to time :-).
They cloned entire machine designs, the stuff is looking from the outside
similar to the western original and is calles a clone.
But if you are looking at the inside most things are made totally
different. As an example there was much rumour about the cloned VAXes,
but the russians built the first VAX CPU on its own only from the processor
hanbdbook of the 11/750 .. but as a single IC! That integrated CPU never
existed in the US, so this is'nt really a clone at all.
Or this thing:
http://www.youtube.com/watch?v=BxRSEyj6twY
A Casio Clone? I've own a such looking casio calculator, but the russian
one has a PDP11 CPU.
Things you mentioned above (analog/digital, VLIW etc..) are currently
unknown to me. I don't know of saved machines like this.
I think some of this was an attempt to make up for trailing in process
technology and assembly quality with architectural cleverness (although
as
Intel well demonstrates, he with the best process
technology always wins
in
the long run, architecture be damned, LOL), some
was maybe just
straight-up
old time engineering creativity at work. It's
neat to see what people
come
up with when they're a bit removed from the
"prevailing knowledge"
(although there was certainly PLENTY of straight cloning of Western
designs
being done as well).
Best,
Sean
Hmm, yes. The war of people wanting the Z8000 and others wanting the 8086
line of CPUs existed in the GDR (DDR) before the fall of the iron curtain
too. There where companies like EAW that build process control systems
around the Zilog processors (for sure the better architecture)
(http://www.robotrontechnik.de/index.htm?/html/computer/p8000.htm)
and wanted that line do be developed further and the "Big Blue" of the GDR,
robotron that wanted the 8086 Series. But the small country hasn't had
enough
recources to do both of them..
I have some Kind of Zilog S8000 "compatible" System running Unix (Zeus
clone) on a 4Mhz Z8000.
That was'nbt really a S8000 Clone, the system was build out of 3 SBC
connected thogether (Z80 for floppy I/O and serail interfaces, the Z8000
CPU Board with Memory Cards in slots and another z80 card for the
Winchester drive 50MByte. The System was capable to run CP/M and RIO
Compatible OS'es on the Z80 part too).
http://www.robotrontechnik.de/index.htm?/html/computer/p8000.htm
http://www.robotrontechnik.de/index.htm?/html/computer/p8000compact.htm
There existed a subsystem with an 80286 and extra memory for that machine
too, so this thig was able to run DOS on the 80286 over the serial
connected terminal. So the available OSes for that thing where OS/M (CP/M),
UDOS (RIO), WEGA (Zeus Unixi, V7 with Extensions) and finally MSDOS.
Regards,
Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Stra?e 42, 09600 Obersch?na, USt-Id: DE253710583
www.tsht.de, info at tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
3572
days inactive
3590
days old
13 comments
7 participants
participants (7)
-
dave.thearchivist＠gmail.com -
drlegendre＠gmail.com -
hilpert＠cs.ubc.ca -
holm＠freibergnet.de -
mattislind＠gmail.com -
scaron＠umich.edu -
thomasmernst＠gmail.com