4 Jul
2025
4 Jul
'25
3:32 p.m.
Jon Elson's take hits home. A 780 was delivered and VMS was running. We
installed 4.1BSD and it ran fine until it crashed. Field service insisted
we needed a full set of RS-232 wires in our cable. Still crashed
(surprise!). Switched to VMS, still crashed after a while. Local field
service couldn't find it. The big guns flew in from Maynard. First day:
Couldn't find it. Second day: "What, what's that wire doing there? Have a
wire-wrap tool?". Removed wire from backplane. Boots, runs. Engineer
flies home.
4 Jul
4 Jul
3:55 p.m.
On Jul 4, 2025, at 12:32 PM, Mark Kahrs via cctalk
<cctalk(a)classiccmp.org> wrote:
Jon Elson's take hits home. A 780 was delivered and VMS was running. We
installed 4.1BSD and it ran fine until it crashed. Field service insisted
we needed a full set of RS-232 wires in our cable. Still crashed
(surprise!). Switched to VMS, still crashed after a while. Local field
service couldn't find it. The big guns flew in from Maynard. First day:
Couldn't find it. Second day: "What, what's that wire doing there? Have a
wire-wrap tool?". Removed wire from backplane. Boots, runs. Engineer
flies home.
My story is about a 750. Under maintenance it ran for some time until the
tape drive failed. Tech installs new controller board, fixed.
Then the CPU fails. New CPU, fixed.
Then the tape drive failed again. New board, fixed.
And so on through various peripherals, disks, tty’s, etc.
Eventually they send out a guy who sits with it for a day, poking and prodding just
about everywhere for a while. At the end of the day says that he’ll be back tomorrow
with the correct part.
The correct part: A new backplane.
Took the entire day to remove/replace the hardware, and lo - the machine never failed
after that.
David
You can know the name of a bird in all the languages of the world, but when you're
finished,
you'll know absolutely nothing whatever about the bird... So let's look at the
bird and
see what it's doing -- that's what counts. I learned very early the difference
between
knowing the name of something and knowing something.
--Richard Feynman
David Barto
barto(a)kdbarto.org
4:21 p.m.
This is similar to an IBM mainframe story i have.
IBM 370 wouldn’t IMPL.
CE and FE couldn’t figure it out.
They called in a “Special Consultant.”
I
Came in with a briefcase that contained his “tools”, which was a big brush that you would
use for light plaster work and a wood mallet.
He proceeded to sweep the brush across the wirewrap backplane and a wire cam loose.
He replaced the wire and all was good again.
I never found out what the mallet was used for.
Sent from my iPhone
On Jul 4, 2025, at 12:40, Mark Kahrs via cctalk
<cctalk(a)classiccmp.org> wrote:
after
4:35 p.m.
On Fri, 4 Jul 2025, Wayne S via cctalk wrote:
This is similar to an IBM mainframe story i have.
IBM 370 wouldn’t IMPL.
CE and FE couldn’t figure it out.
They called in a “Special Consultant.”
I
Came in with a briefcase that contained his “tools”, which was a big brush that you would
use for light plaster work and a wood mallet.
He proceeded to sweep the brush across the wirewrap backplane and a wire cam loose.
He replaced the wire and all was good again.
I never found out what the mallet was used for.
. . . and a piece of chalk?
measure from top edge, measure from side edge.
Make a chalk mark.
Hit with mallet.
"$1 for hitting with mallet; $499 for knowing where to hit"
5:21 p.m.
I didn’t see chalk, but it was $1600/hr plus expenses to know where to sweep.
IBM was paying it!
Sent from my iPhone
On Jul 4, 2025, at 13:44, Fred Cisin via cctalk
<cctalk(a)classiccmp.org> wrote:
On Fri, 4 Jul 2025, Wayne S via cctalk wrote:
This is similar to an IBM mainframe story i
have.
IBM 370 wouldn’t IMPL.
CE and FE couldn’t figure it out.
They called in a “Special Consultant.”
I
Came in with a briefcase that contained his “tools”, which was a big brush that you would
use for light plaster work and a wood mallet.
He proceeded to sweep the brush across the wirewrap backplane and a wire cam loose.
He replaced the wire and all was good again.
I never found out what the mallet was used for.
. . . and a piece of chalk?
measure from top edge, measure from side edge.
Make a chalk mark.
Hit with mallet.
"$1 for hitting with mallet; $499 for knowing where to hit"
5:29 p.m.
Had another odd one with our 780. FE came in replaced board tge remote diagnostics called
out and everything was good again. When he looked at the bad board he started laughing.
One of the IC’s had a hole in it. Looked like it burned out.
He said it was the first time he’s ever seen that!
Sent from my iPhone
On Jul 4, 2025, at 14:21, Wayne S
<wayne.sudol(a)hotmail.com> wrote:
I didn’t see chalk, but it was $1600/hr plus expenses to know where to sweep.
IBM was paying it!
Sent from my iPhone
>> On Jul 4, 2025, at 13:44, Fred Cisin via cctalk <cctalk(a)classiccmp.org>
wrote:
>>
>> On Fri, 4 Jul 2025, Wayne S via cctalk wrote:
>> This is similar to an IBM mainframe story i have.
>> IBM 370 wouldn’t IMPL.
>> CE and FE couldn’t figure it out.
>> They called in a “Special Consultant.”
>> I
>> Came in with a briefcase that contained his “tools”, which was a big brush that
you would use for light plaster work and a wood mallet.
>> He proceeded to sweep the brush across the wirewrap backplane and a wire cam
loose.
>> He replaced the wire and all was good again.
>> I never found out what the mallet was used for.
>
> . . . and a piece of chalk?
> measure from top edge, measure from side edge.
> Make a chalk mark.
> Hit with mallet.
> "$1 for hitting with mallet; $499 for knowing where to hit"
7:54 p.m.
One of the IC's had a hole in it. Looked like it
burned out.
He said it was the first time he's ever seen that!
The Apple II floppy drives were notorious for this. The chip
that drove the stepper motor (I forget the part number) would
literally explode, leaving a quite visible crater in the top
of the IC.
It got to the point where, as a customer walked into the shop with
a broken drive, I could diagnose the problem from across the room
because those exploded chips emitted a unique odor.
We made a fair bit of cash swapping those ICs :-)
--lyndon
8:39 p.m.
On 2025-07-04 17:29, Wayne S via cctalk wrote:
Had another odd one with our 780. FE came in replaced
board tge remote diagnostics called out and everything was good again. When he looked at
the bad board he started laughing.
One of the IC’s had a hole in it. Looked like it burned out.
He said it was the first time he’s ever seen that!
Sent from my iPhone
> On Jul 4, 2025, at 14:21, Wayne S<wayne.sudol(a)hotmail.com> wrote:
>
> I didn’t see chalk, but it was $1600/hr plus expenses to know where to sweep.
> IBM was paying it!
>
> Sent from my iPhone
>
>>> On Jul 4, 2025, at 13:44, Fred Cisin via cctalk<cctalk(a)classiccmp.org>
wrote:
>>>
>>> On Fri, 4 Jul 2025, Wayne S via cctalk wrote:
>>> This is similar to an IBM mainframe story i have.
>>> IBM 370 wouldn’t IMPL.
>>> CE and FE couldn’t figure it out.
>>> They called in a “Special Consultant.”
>>> I
>>> Came in with a briefcase that contained his “tools”, which was a big brush
that you would use for light plaster work and a wood mallet.
>>> He proceeded to sweep the brush across the wirewrap backplane and a wire cam
loose.
>>> He replaced the wire and all was good again.
>>> I never found out what the mallet was used for.
>> . . . and a piece of chalk?
>> measure from top edge, measure from side edge.
>> Make a chalk mark.
>> Hit with mallet.
>> "$1 for hitting with mallet; $499 for knowing where to hit"
I had a bit of luck once with a burned out chip. Back in the day we
would change boards in the field and then fix them back at our office
with live equipment that was also being used for engineers to write
software for the Canadian Post Office. So service on live equipment was
rushed while they had lunch. One day, i used am extender card that had
the index plastic broken off, and, you guessed it, plugged it the wrong
way hanging down from a DEC RK05 drive.
Chip smoked on power-up, hastily cut power and removed smoking card.
Replaced chip to go back and further troubleshoot and the card worked
perfectly.
Only time I ever solved a problem by plugging a card in backwards!
cheers
Nigel
--
Nigel Johnson, MSc., MIEEE, VE3ID/G4AJQ/VA3MCU
Amateur Radio, the origin of the open-source concept!
5 Jul
5 Jul
2:11 a.m.
When I worked for DEC in field service, we had a customer almost next door
to us. We gathered outside to watch their new (maybe used) VAX 11/780 get
delivered. For whatever reason they used a ramp instead of the lift. The
780 got away from them, and when the wheels hit the pavement we watched as
it flipped over and landed on its side. One of the techs calmly said
"Someone should go order a new backplane".
On Fri, Jul 4, 2025 at 7:49 PM Nigel Johnson Ham via cctalk <
cctalk(a)classiccmp.org> wrote:
On 2025-07-04 17:29, Wayne S via cctalk wrote:
Had another odd one with our 780. FE came in
replaced board tge remote
diagnostics called out and everything was good again. When he looked at the
bad board he started laughing.
One of the IC’s had a hole in it. Looked like it
burned out.
He said it was the first time he’s ever seen that!
Sent from my iPhone
> On Jul 4, 2025, at 14:21, Wayne S<wayne.sudol(a)hotmail.com> wrote:
>
> I didn’t see chalk, but it was $1600/hr plus expenses to know where to
sweep.
> IBM was paying it!
>
> Sent from my iPhone
>
>>> On Jul 4, 2025, at 13:44, Fred Cisin via cctalk<cctalk(a)classiccmp.org>
wrote:
>>>
>>> On Fri, 4 Jul 2025, Wayne S via cctalk wrote:
>>> This is similar to an IBM mainframe story i have.
>>> IBM 370 wouldn’t IMPL.
>>> CE and FE couldn’t figure it out.
>>> They called in a “Special Consultant.”
>>> I
>>> Came in with a briefcase that contained his “tools”, which was a big
brush that you would use for light plaster work and a wood mallet.
>>> He proceeded to sweep the brush
across the wirewrap backplane and a
wire cam loose.
>>> He replaced the wire and all was good
again.
>>> I never found out what the mallet was used for.
>> . . . and a piece of chalk?
>> measure from top edge, measure from side edge.
>> Make a chalk mark.
>> Hit with mallet.
>> "$1 for hitting with mallet; $499 for knowing where to hit"
I had a bit of luck once with a burned out chip. Back in the day we
would change boards in the field and then fix them back at our office
with live equipment that was also being used for engineers to write
software for the Canadian Post Office. So service on live equipment was
rushed while they had lunch. One day, i used am extender card that had
the index plastic broken off, and, you guessed it, plugged it the wrong
way hanging down from a DEC RK05 drive.
Chip smoked on power-up, hastily cut power and removed smoking card.
Replaced chip to go back and further troubleshoot and the card worked
perfectly.
Only time I ever solved a problem by plugging a card in backwards!
cheers
Nigel
--
Nigel Johnson, MSc., MIEEE, VE3ID/G4AJQ/VA3MCU
Amateur Radio, the origin of the open-source concept!
11:24 a.m.
On 7/5/25 01:11, Paul Anderson via cctalk wrote:
When I worked for DEC in field service, we had a
customer almost next door
to us. We gathered outside to watch their new (maybe used) VAX 11/780 get
delivered. For whatever reason they used a ramp instead of the lift. The
780 got away from them, and when the wheels hit the pavement we watched as
it flipped over and landed on its side. One of the techs calmly said
"Someone should go order a new backplane".
This ALMOST happened to our first 780! The truck driver was
getting it off the truck with a pallet jack (I think) and
there was a slight misalignment between the truck bed and
the loading dock. The 780 started to tip, but he got his
shoulder against it and kept it from going over.
Another one I saw was our computer center was raising their
raised floor a foot, as it was getting TOO filled with
cables for airflow. They had the 360/65 disassembled
(processor and memory/power cabinet) and a SINGLE guy was
moving the CPU up a ramp to the higher portion of the
floor. He started to lose control of it, and it started
tipping. My first thought was that I was going to see a guy
get crushed to death right there! Incredibly, he get the
thing righted and pushed back up the ramp!
Jon
7:17 p.m.
I will never forget moving an RM03 down stairs into a friend's basement.
CZ on top, RM03 in the middle, RS on bottom. Halfway down the flight of
steps.
CZ: If we lose control of this it's going to crush you like a grape.
RS: Trust me CZ, there will be nothing grape like about it....
I don't miss high-mass hobbies as much as one might think....
CZ
On 7/5/2025 2:11 AM, Paul Anderson via cctalk wrote:
When I worked for DEC in field service, we had a
customer almost next door
to us. We gathered outside to watch their new (maybe used) VAX 11/780 get
delivered. For whatever reason they used a ramp instead of the lift. The
780 got away from them, and when the wheels hit the pavement we watched as
it flipped over and landed on its side. One of the techs calmly said
"Someone should go order a new backplane".
On Fri, Jul 4, 2025 at 7:49 PM Nigel Johnson Ham via cctalk <
cctalk(a)classiccmp.org> wrote:
> On 2025-07-04 17:29, Wayne S via cctalk wrote:
>> Had another odd one with our 780. FE came in replaced board tge remote
> diagnostics called out and everything was good again. When he looked at the
> bad board he started laughing.
>> One of the IC’s had a hole in it. Looked like it burned out.
>> He said it was the first time he’s ever seen that!
>>
>> Sent from my iPhone
>>
>>> On Jul 4, 2025, at 14:21, Wayne S<wayne.sudol(a)hotmail.com> wrote:
>>>
>>> I didn’t see chalk, but it was $1600/hr plus expenses to know where to
> sweep.
>>> IBM was paying it!
>>>
>>> Sent from my iPhone
>>>
>>>>> On Jul 4, 2025, at 13:44, Fred Cisin via
cctalk<cctalk(a)classiccmp.org>
> wrote:
>>>>>
>>>>> On Fri, 4 Jul 2025, Wayne S via cctalk wrote:
>>>>> This is similar to an IBM mainframe story i have.
>>>>> IBM 370 wouldn’t IMPL.
>>>>> CE and FE couldn’t figure it out.
>>>>> They called in a “Special Consultant.”
>>>>> I
>>>>> Came in with a briefcase that contained his “tools”, which was a big
> brush that you would use for light plaster work and a wood mallet.
>>>>> He proceeded to sweep the brush across the wirewrap backplane and a
> wire cam loose.
>>>>> He replaced the wire and all was good again.
>>>>> I never found out what the mallet was used for.
>>>> . . . and a piece of chalk?
>>>> measure from top edge, measure from side edge.
>>>> Make a chalk mark.
>>>> Hit with mallet.
>>>> "$1 for hitting with mallet; $499 for knowing where to hit"
>
> I had a bit of luck once with a burned out chip. Back in the day we
> would change boards in the field and then fix them back at our office
> with live equipment that was also being used for engineers to write
> software for the Canadian Post Office. So service on live equipment was
> rushed while they had lunch. One day, i used am extender card that had
> the index plastic broken off, and, you guessed it, plugged it the wrong
> way hanging down from a DEC RK05 drive.
>
> Chip smoked on power-up, hastily cut power and removed smoking card.
> Replaced chip to go back and further troubleshoot and the card worked
> perfectly.
>
> Only time I ever solved a problem by plugging a card in backwards!
>
> cheers
>
> Nigel
>
>
> --
> Nigel Johnson, MSc., MIEEE, VE3ID/G4AJQ/VA3MCU
> Amateur Radio, the origin of the open-source concept!
>
>
>
11:15 a.m.
On 7/4/25 16:29, Wayne S via cctalk wrote:
Had another odd one with our 780. FE came in replaced
board tge remote diagnostics called out and everything was good again. When he looked at
the bad board he started laughing.
One of the IC’s had a hole in it. Looked like it burned out.
He said it was the first time he’s ever seen that!
We had a group at the university that had a PDP-11 with a
fancy graphics frame buffer, they were doing planetary
imaging for NASA. One day the frame buffer went down, and
one of the boards had a big HOLE in it. I gather that a
Tantalum capacitor had shorted across the main 5 V power rails.
Jon
4 Jul
4 Jul
5:56 p.m.
>> Came in with a briefcase that contained his
“tools”, which was a big brush that you would use for light plaster work and a wood
mallet.
>> He proceeded to sweep the brush across the wirewrap backplane and a wire cam
loose.
>> He replaced the wire and all was good again.
>> I never found out what the mallet was used for.
> . . . and a piece of chalk?
> measure from top edge, measure from side edge.
> Make a chalk mark.
> Hit with mallet.
> "$1 for hitting with mallet; $499 for knowing where to hit"
On Fri, 4 Jul 2025, Wayne S wrote:
I didn’t see chalk, but it was $1600/hr plus expenses
to know where to sweep.
IBM was paying it!
True, there has been some inflation in the half century since I was around 360/370
6:15 p.m.
Yeah, this was 1978.
What we used to spend on IT stuff back then is really amazing.
So much!
Sent from my iPhone
On Jul 4, 2025, at 15:04, Fred Cisin via cctalk
<cctalk(a)classiccmp.org> wrote:
>> Came in with a briefcase that contained his “tools”, which was a big brush that
you would use for light plaster work and a wood mallet.
>> He proceeded to sweep the brush across the wirewrap backplane and a wire cam
loose.
>> He replaced the wire and all was good again.
>> I never found out what the mallet was used for.
> . . . and a piece of chalk?
> measure from top edge, measure from side edge.
> Make a chalk mark.
> Hit with mallet.
> "$1 for hitting with mallet; $499 for knowing where to hit"
On Fri, 4 Jul 2025, Wayne S wrote:
I didn’t see chalk, but it was $1600/hr plus expenses to know where to sweep.
IBM was paying it!
True, there has been some inflation in the half century since I was around 360/370
8:08 p.m.
On 2025-07-04 4:15 p.m., Wayne S via cctalk wrote:
Yeah, this was 1978.
What we used to spend on IT stuff back then is really amazing.
So much!
True, but you got field service,
not some AI telling you read web page xxx, with a dead computer.
Ben.
PS: serial port had some active input on boot, and windows thought that
was a mouse not the usb port.
5 Jul
5 Jul
2:13 a.m.
On 05/07/2025 01:08, ben via cctalk wrote:
On 2025-07-04 4:15 p.m., Wayne S via cctalk wrote:
But IBM field service had to follow the diagnostic MAPs which listed the
steps to debug a problem, sometimes in order of cost rather that
probability of failure.
Our CE once remarked they were written by accountants not engineers.
The result was that when a 3380 disk drive (DASD in IBM terms) failed
the Customer Engineer (CE) had to follow a tortuous set of tests...
.. the 3380 had two sets of heads, so every was duplicated and the
diagnostics swapped every component in turn between the data paths to
make sure nothing else was faulty...
.. invariably it was the "head and drive" assembly which on a 3380 was a
sealed unit and very expensive to make.
The CE usually brought one with him which was good, as it had to
acclimatise for a couple of hours before fitting, so running the other
tests didn't cost much time...
Yeah, this was 1978.
What we used to spend on IT stuff back then is really amazing.
So much!
True, but you got field service,
not some AI telling you read web page xxx, with a dead computer. Ben.
PS: serial port had some active input on boot, and windows thought
that was a mouse not the usb port.
Dave
7:48 a.m.
On Jul 4, 2025, at 8:08 PM, ben via cctalk
<cctalk(a)classiccmp.org> wrote:
True, but you got field service,
not some AI telling you read web page xxx, with a dead computer.
Ben.
PS: serial port had some active input on boot, and windows thought that was a mouse not
the usb port.
The Windows problem with misidentifying a serial port as a mouse still exists in Windows
10. I discovered that connecting our small ARM based embedded device to a Windows machine
via its USB serial port sometimes causes the Windows mouse cursor to go crazy. It took a
while to find the registry setting to disable serial mouse detection.
- Rod
4 Jul
4 Jul
8:19 p.m.
On Fri, Jul 4, 2025 at 3:29 PM Wayne S via cctalk <cctalk(a)classiccmp.org> wrote:
I never found out what the mallet was used for.
I heard people joke back in the day that CEs carried a rubber mallet
for remediating core memory problems.
No idea whether percussive maintenance to jostle the cores was
actually a thing or not.
5 Jul
5 Jul
11:12 a.m.
On 7/4/25 15:21, Wayne S via cctalk wrote:
This is similar to an IBM mainframe story i have.
IBM 370 wouldn’t IMPL.
CE and FE couldn’t figure it out.
They called in a “Special Consultant.”
I
Came in with a briefcase that contained his “tools”, which was a big brush that you would
use for light plaster work and a wood mallet.
He proceeded to sweep the brush across the wirewrap backplane and a wire cam loose.
He replaced the wire and all was good again.
I never found out what the mallet was used for.
We had a long-in-the-tooth 360/65 that was starting to get
flaky. IBM asked if they could be given the time for a
"field overhaul". They explained they would have the
machine down for about a week. They were given the
go-ahead. The brought in a 55 gallon barrel of some kind of
cleaning solvent, took all the circuit cards out of the
machine, and dunked them in the barrel. After cleaning, they
plugged EVERY board into some tester box and checked for
weak signal levels. They cleaned the "boards" which were
the wire-wrapped panels that took the SLT cards on the other
side. Finally, they put all the cards back in (after
replacing a few that failed on the tester box) and the
machine came up and ran until IBM sent the dreaded letter
that said you were welcome to keep running the machine, but
they wouldn't service it beyond a certain date.
Jon
4 Jul
4 Jul
6:35 p.m.
My backplane horror story isn't so fun.
The old TOPS-20 had an 'emergency power off' button. I'm told that it
would connect 5V to ground through
the backplane for when you really, really didn't want someone to bring
it back up easily. Seems like a really
stupid feature, but the university I went to had a bunch of military
connections, and it may have been for
a contract with them (or maybe we got the machine surplus from them, I
never could figure out the details).
Anyway, in March we announced that we'd be switching the TOPS-20
machine off just before school started
again in August. One of the tops-20 operators "accidentally" hit this
switch just before finals week because
they needed more time to finish their final... That didn't end well for them...
We had a field service person rebuild the backplane, replacing the
charred wires with fresh, and finding
all the ones that had been too stressed out by the surge. They started
in May and were done the last
week of July. So we were able to runt the TOPS-20 machine again for 3
weeks before we shut it off (well,
due to the failures, that extended to 5 weeks). So 10 weeks of field
service time to get 5 more weeks of
life from the machine... And even then, it was "brittle" and useful
only for data extraction and transfer to
the new Sun workstations...
The power of stupidity... None of this billed out at $1600/hr, though
it was the mid 80s.
Warner
On Fri, Jul 4, 2025 at 1:39 PM Mark Kahrs via cctalk
<cctalk(a)classiccmp.org> wrote:
Jon Elson's take hits home. A 780 was delivered and VMS was running. We
installed 4.1BSD and it ran fine until it crashed. Field service insisted
we needed a full set of RS-232 wires in our cable. Still crashed
(surprise!). Switched to VMS, still crashed after a while. Local field
service couldn't find it. The big guns flew in from Maynard. First day:
Couldn't find it. Second day: "What, what's that wire doing there? Have a
wire-wrap tool?". Removed wire from backplane. Boots, runs. Engineer
flies home.
5 Jul
5 Jul
11:05 a.m.
On 7/4/25 14:32, Mark Kahrs via cctalk wrote:
Jon Elson's take hits home. A 780 was delivered
and VMS was running. We
installed 4.1BSD and it ran fine until it crashed. Field service insisted
we needed a full set of RS-232 wires in our cable. Still crashed
(surprise!). Switched to VMS, still crashed after a while. Local field
service couldn't find it. The big guns flew in from Maynard. First day:
Couldn't find it. Second day: "What, what's that wire doing there? Have a
wire-wrap tool?". Removed wire from backplane. Boots, runs. Engineer
flies home.
Holy cow, HOW did he find it!!?? The KA780 backplane was a
HUGE mass of wires!
Jon
12:33 p.m.
On Jul 5, 2025, at 8:05 AM, Jon Elson via cctalk
<cctalk(a)classiccmp.org> wrote:
On 7/4/25 14:32, Mark Kahrs via cctalk wrote:
My story is about a 750. Under maintenance it ran for some time until the
tape drive failed. Tech installs new controller board, fixed.
Then the CPU fails. New CPU, fixed.
Then the tape drive failed again. New board, fixed.
And so on through various peripherals, disks, tty’s, etc.
Eventually they send out a guy who sits with it for a day, poking and prodding just
about everywhere for a while. At the end of the day says that he’ll be back tomorrow
with the correct part.
The correct part: A new backplane.
Took the entire day to remove/replace the hardware, and lo - the machine never failed
after that.
David
"Nature doesn't care how smart you are. You can still be wrong." - Richard
Feynman
David Barto
barto(a)kdbarto.org
Jon Elson's take hits home. A 780 was
delivered and VMS was running. We
installed 4.1BSD and it ran fine until it crashed. Field service insisted
we needed a full set of RS-232 wires in our cable. Still crashed
(surprise!). Switched to VMS, still crashed after a while. Local field
service couldn't find it. The big guns flew in from Maynard. First day:
Couldn't find it. Second day: "What, what's that wire doing there? Have a
wire-wrap tool?". Removed wire from backplane. Boots, runs. Engineer
flies home.
Holy cow, HOW did he find it!!?? The KA780 backplane was a HUGE mass of wires!
Jon
3:34 p.m.
On Jul 5, 2025, at 11:05 AM, Jon Elson via cctalk
<cctalk(a)classiccmp.org> wrote:
On 7/4/25 14:32, Mark Kahrs via cctalk wrote:
If you understand the machine well enough and trace the broken data path around the
machine, you can get there without too much trouble. That assumes the issue is reasonably
repeatable. If it only messes up once a day or so, it's harder, or at least more time
consuming.
Compared to CDC 6000 mainframes, the 780 (and other DEC computers) are marvels of
simplicity. Consider the 6600: 15 chassis each with 750-ish module slots, each with 28
signal pins. A bunch of slots were filled instead by memory modules (5 slots wide), but
still you're looking at maybe 5000 wires (or rather, twisted pairs) per chassis, plus
30 or so cable assemblies each with 19 coax inside, to run signals from one chassis to
another (or to I/O devices). I assume those were all done by hand; it's not obvious
how a robot could do that in the early 1960s, unlike wire wrap backplanes.
paul
Jon Elson's take hits home. A 780 was
delivered and VMS was running. We
installed 4.1BSD and it ran fine until it crashed. Field service insisted
we needed a full set of RS-232 wires in our cable. Still crashed
(surprise!). Switched to VMS, still crashed after a while. Local field
service couldn't find it. The big guns flew in from Maynard. First day:
Couldn't find it. Second day: "What, what's that wire doing there? Have a
wire-wrap tool?". Removed wire from backplane. Boots, runs. Engineer
flies home.
Holy cow, HOW did he find it!!?? The KA780 backplane was a HUGE mass of wires!
Jon
8:55 p.m.
On 2025-07-05 1:34 p.m., Paul Koning via cctalk wrote:
And all the wires tuned to have the same delay I bet.
On Jul 5, 2025, at 11:05 AM, Jon Elson via cctalk
<cctalk(a)classiccmp.org> wrote:
On 7/4/25 14:32, Mark Kahrs via cctalk wrote:
If you understand the machine well enough and trace the broken data path around the
machine, you can get there without too much trouble. That assumes the issue is reasonably
repeatable. If it only messes up once a day or so, it's harder, or at least more time
consuming.
Compared to CDC 6000 mainframes, the 780 (and other DEC computers) are marvels of
simplicity. Consider the 6600: 15 chassis each with 750-ish module slots, each with 28
signal pins. A bunch of slots were filled instead by memory modules (5 slots wide), but
still you're looking at maybe 5000 wires (or rather, twisted pairs) per chassis, plus
30 or so cable assemblies each with 19 coax inside, to run signals from one chassis to
another (or to I/O devices). I assume those were all done by hand; it's not obvious
how a robot could do that in the early 1960s, unlike wire wrap backplanes.
Jon Elson's take hits home. A 780 was
delivered and VMS was running. We
installed 4.1BSD and it ran fine until it crashed. Field service insisted
we needed a full set of RS-232 wires in our cable. Still crashed
(surprise!). Switched to VMS, still crashed after a while. Local field
service couldn't find it. The big guns flew in from Maynard. First day:
Couldn't find it. Second day: "What, what's that wire doing there? Have a
wire-wrap tool?". Removed wire from backplane. Boots, runs. Engineer
flies home.
Holy cow, HOW did he find it!!?? The KA780 backplane was a HUGE mass of wires!
Jon paul
Ben.
6 Jul
6 Jul
11:38 a.m.
On Jul 5, 2025, at 5:55 PM, ben via cctalk
<cctalk(a)classiccmp.org> wrote:
On 2025-07-05 1:34 p.m., Paul Koning via cctalk wrote:
And all the wires tuned to have the same delay I bet.
At Celerity Computing (later called Sun Supercomputing, long story) we started by making
small dishwasher sized deskside workstations.
We were using an NCR RISC chip as our CPU, getting about 2 MIPS with a custom FPU hand
crafted from the most interesting parts.
We wire-wrapped the first few boards and backplanes during R&D before we went to
production. We switched to a proper 6 layer PCB for the backplane.
The backplane had 3 traces that had long curved lines in them to get the delay right.
David
Without others thinking you are good, it doesn't matter how good you think you are.
David Barto
barto(a)kdbarto.org
On Jul 5,
2025, at 11:05 AM, Jon Elson via cctalk <cctalk(a)classiccmp.org> wrote:
On 7/4/25 14:32, Mark Kahrs via cctalk wrote:
If you understand the machine well enough and trace the broken data path around
the machine, you can get there without too much trouble. That assumes the issue is
reasonably repeatable. If it only messes up once a day or so, it's harder, or at
least more time consuming.
Compared to CDC 6000 mainframes, the 780 (and other DEC computers) are marvels of
simplicity. Consider the 6600: 15 chassis each with 750-ish module slots, each with 28
signal pins. A bunch of slots were filled instead by memory modules (5 slots wide), but
still you're looking at maybe 5000 wires (or rather, twisted pairs) per chassis, plus
30 or so cable assemblies each with 19 coax inside, to run signals from one chassis to
another (or to I/O devices). I assume those were all done by hand; it's not obvious
how a robot could do that in the early 1960s, unlike wire wrap backplanes.
Jon Elson's take hits home. A 780 was
delivered and VMS was running. We
installed 4.1BSD and it ran fine until it crashed. Field service insisted
we needed a full set of RS-232 wires in our cable. Still crashed
(surprise!). Switched to VMS, still crashed after a while. Local field
service couldn't find it. The big guns flew in from Maynard. First day:
Couldn't find it. Second day: "What, what's that wire doing there? Have a
wire-wrap tool?". Removed wire from backplane. Boots, runs. Engineer
flies home.
Holy cow, HOW did he find it!!?? The KA780 backplane was a HUGE mass of wires!
Jon
paul
Ben.
5 Jul
5 Jul
10:14 p.m.
On Sat, Jul 5, 2025 at 2:40 PM Paul Koning via cctalk
<cctalk(a)classiccmp.org> wrote:
I assume those were all done by hand; it's not
obvious how a robot could do that in the early 1960s, unlike wire wrap backplanes.
Nice page on the hand-assembly of early Cray systems with cool pictures:
https://cray-history.net/2022/08/28/sonjas-story-about-cray-1-fabrication-a…
0
days inactive
2
days old
25 comments
17 participants
participants (17)
-
ben -
cz -
David [CCTALK] Barto -
David [TUHS] Barto
-
David Barto -
David Wade -
Fred Cisin -
Gavin Scott -
Jon Elson -
Lyndon Nerenberg (VE7TFX/VE6BBM) -
Mark Kahrs -
Nigel Johnson Ham -
Paul Anderson -
Paul Koning -
Rod Bartlett -
Warner Losh -
Wayne S