25 Aug
2010
25 Aug
'10
5:59 p.m.
When you are cleaning up and restoring a piece of classic computer
hardware, would you correct any minor mistakes made in the original
production?
I have a particular exmaple in mind. In bits on myu bench at the moment
is an HP2623 graphics terminal This was HP's response to the well-knwon
Tektronix terminals, it is, however a raster scanned device. But it has a
little state machine (or parhapes you could call it a very simple
procesosr) to speed up line drawing.
But wanyway...
On the video monitor PCB (HP call it the 'Sweep' PCB), there is a 74LS221
dual monostable chip in the common HP horixontal drive circuit, The first
half of the chip delays the sync pulse (and the delay time is controlled
by the horizontal centering control), the second half is triggered from
that and provides a pulse of known width to the horizontal drive transistor.
But I digress. You don't need to know any of that...
WHile cleaning the dirt off this PCB (HV attracts dust electrostatcally,
of couse), I noticed that the ground pin of this IC never made it through
the board. It's folded under the package, but has been effectively
surface-mount soldered to the pad when the board was wave-soldered. It
works, but I can't beeleive it's as good as a proper through-hole
connection, and it's certian;y not what was intended.
I could trivially desolder the IC, straighten the pin, and solder it back
properly. But should I? What would others do?
-tony
25 Aug
25 Aug
6:10 p.m.
If it's working, I'd probably leave well enough alone and make a mental note of
it in case something happened that it could have had an effect on.
________________________________
From: Tony Duell <ard at p850ug1.demon.co.uk>
To: cctalk at classiccmp.org
Sent: Wed, August 25, 2010 4:59:31 PM
Subject: Should you correct production mistakes?
When you are cleaning up and restoring a piece of classic computer
hardware, would you correct any minor mistakes made in the original
production?
I have a particular exmaple in mind. In bits on myu bench at the moment
is an HP2623 graphics terminal This was HP's response to the well-knwon
Tektronix terminals, it is, however a raster scanned device. But it has a
little state machine (or parhapes you could call it a very simple
procesosr) to speed up line drawing.
But wanyway...
On the video monitor PCB (HP call it the 'Sweep' PCB), there is a 74LS221
dual monostable chip in the common HP horixontal drive circuit, The first
half of the chip delays the sync pulse (and the delay time is controlled
by the horizontal centering control), the second half is triggered from
that and provides a pulse of known width to the horizontal drive transistor.
But I digress. You don't need to know any of that...
WHile cleaning the dirt off this PCB (HV attracts dust electrostatcally,
of couse), I noticed that the ground pin of this IC never made it through
the board. It's folded under the package, but has been effectively
surface-mount soldered to the pad when the board was wave-soldered. It
works, but I can't beeleive it's as good as a proper through-hole
connection, and it's certian;y not what was intended.
I could trivially desolder the IC, straighten the pin, and solder it back
properly. But should I? What would others do?
-tony
26 Aug
26 Aug
2:15 p.m.
If it's working, I'd probably leave well enough alone and make a
mental note of
it in case something happened that it could have had an effect on.
It was wroking when I lat used the terminal some 20 years ago (and it was
an old unit then). But I am not the sort of person to apply power to
somethign that's not been run in 20 years without checking it over first.
So I can't be sure it's still working. I can be sure (and am sure -- I've
tested it ) that this pin is is correctly grounded even though its' not
going through the hole.
If the conenction failed, then the output of this circuit would become
a TTL high, I guess. Dure to an iverter on the board, this would turn off
the horizontla drive transistor. Sinec the output of that is transfoemr
coupled to the horizontal output transistor, that would be turend off (no
change in drive form the driver transsitor -> no base current in the
output stage). At which point I'd loose all the CRT voltages, but there
would be no damage. One thing that had worreid me was that in some video
boards if there was a failure of the horizotnal drive sircuit then the
ouput transitor could be turned on, leading to damage to it and/or the
flyback transfotmer. The former would be a pain, the latter would be a
very large pain! But this doesn't apply here.
On the other hand, I know it's an erorr. And getting to it involves
dismantling most of the terminal, including disconnecting the EHT
connector form the CRT. So I'd rather not have to keep on removing this
PCB.
-tony
25 Aug
25 Aug
6:19 p.m.
Tony Duell wrote:
When you are cleaning up and restoring a piece of
classic computer
hardware, would you correct any minor mistakes made in the original
production?
.....
Personally I'd correct it. I've generally tried to restore the few
machines I have owned in my "rolling collection" to a working state. Now
in my case I'm not so competent at the circuit level so for example I've
been known to replace the power supply in an Indigo with a modern PC
unit but hidden in the original casing. Asthetically you wouldn't know
and the machine has several more useful years as a result.
Like you, I think these machines should be shown to work, and ideally
used for some purpose. The heat and stress inside a CRT casing is no
place for shoddy construction!
On the other hand there's a strong case for not tinkering with "quirks"
that made it to production and affect all machines of a given type since
they define the character.
Regards,
James
26 Aug
26 Aug
2:22 p.m.
Tony Duell wrote:
I think it's fairly well-known here that I like my classics to be
working. I don't want a shelf of ronaments :-). However, AFAIK this
terminal will work even with the problem (it had been used for some years
before I got it, I suspect the fault was there from new, there is no
evidence that this video monitor PCB has ever been swapped.
When you are cleaning up and restoring a piece of
classic computer
hardware, would you correct any minor mistakes made in the original
production?
.....
Personally I'd correct it. I've generally tried to restore the few
machines I have owned in my "rolling collection" to a working state. Now
in my case I'm not so competent at the circuit
level so for example I've
been known to replace the power supply in an Indigo with a modern PC
unit but hidden in the original casing. Asthetically you wouldn't know
I bet I'd know :-). I gerneally take machines apart...
and the machine has several more useful years as a
result.
Like you, I think these machines should be shown to work, and ideally
used for some purpose. The heat and stress inside a CRT casing is no
place for shoddy construction!
This particualr PCB is right hehind the CRT. It's one of the HP 'ET Head'
units, and inside the casing are 3 PCBs. A large logic board on the
underside of the unit (accesible by hinging the case up on the stand and
releasing 4 Nylatch cips) containing some 200 ICs, the monitor PCB mouted
behind the CRT, and which is got at by removingthe logic board,
disconencing cablesm ad fiddlign it out downared, avoiding damaging the
CRT. and the PSU (along with a mains transformer [1]) which fits at the
very back is is accessed by removing the top cover. So the monitor PCB is
not trivial to get to.
[1] Mine is not the Opto 050 version with a built-in thermal printer.
That has a full SMPSU. The normal one like mine has a mains trasformer
giving about 24V AC, and then a linear regulator for the -12V line and
switching regulators for the +5V and +12V outputs.
On the other hand there's a strong case for not tinkering with "quirks"
that made it to production and affect all machines of a given type since
they define the character.
I don;t think this fault has 'chracter'. If it was an open connection, or
intermittant, the most likely there would be nothing on the screeen at all.
On the otehr hand, I suspect that prodsuction erorrs from 'classic HP'
that have made it into the field are rare, so perhaps it's worth keeping
as an example of that ;-)
-tony
25 Aug
25 Aug
6:29 p.m.
Tony Duell wrote:
I could trivially desolder the IC, straighten the pin,
and solder it
back
properly. But should I? What would others do?
Had we found any problem analogous to that in the PDP-1 restoration,
where the machine operated correctly despite a manufacturing defect, I'm
sure we would have had a debate on whether to fix it. I think my own
opinion in that case would be that we should leave it alone, but tag the
module (paper tag attached with a short loop of string) and document the
issue in the system logbook.
For something that isn't considered a museum artifact, I'd be more
inclined to fix it.
When we do fix PDP-1 modules, we mark the replaced components and log
that as well. Since the component marking is intended to indicate that
the component is not original, if we were just correcting an improperly
installed component, we probably wouldn't mark it, but would still log it.
Eric
(not speaking for the PDP-1 Restoration Team or the Computer History Museum)
7:13 p.m.
On Wed, 25 Aug 2010, Eric Smith wrote:
Had we found any problem analogous to that in the
PDP-1 restoration,
where the machine operated correctly despite a manufacturing defect, I'm
sure we would have had a debate on whether to fix it. I think my own
opinion in that case would be that we should leave it alone, but tag the
module (paper tag attached with a short loop of string) and document the
issue in the system logbook.
In the example that I mentioned of an IBM MDA board, it worked for about
10 years, and then became intermittent. Even that, I would NOT consider
to be "operates correctly", as the particular defect IS a defect, and may
eventually bring the system down.
7:20 p.m.
I wrote:
Had we found any problem analogous to that in the
PDP-1 restoration,
where the machine operated correctly despite a manufacturing defect, I'm
sure we would have had a debate on whether to fix it. I think my own
opinion in that case would be that we should leave it alone, but tag the
module (paper tag attached with a short loop of string) and document the
issue in the system logbook.
Fred Cisin wrote:
In the example that I mentioned of an IBM MDA board,
it worked for about
10 years, and then became intermittent. Even that, I would NOT consider
to be "operates correctly", as the particular defect IS a defect, and may
eventually bring the system down.
I think the way that I would state it is that just because the system is
working correctly (at the moment) doesn't mean that it isn't broken.
Part of the reason I don't think we'd fix such a problem in the PDP-1 is
that the PDP-1 isn't doing anything criticial. We can afford to have
downtime if a latent problem eventual causes a failure. If we've
properly documented that latent problem, we can check for it when the
system does fail, and fix it if necessary at that time.
Eric
7:34 p.m.
On Wed, 25 Aug 2010, Eric Smith wrote:
I think the way that I would state it is that just
because the system is
working correctly (at the moment) doesn't mean that it isn't broken.
Part of the reason I don't think we'd fix such a problem in the PDP-1 is
that the PDP-1 isn't doing anything criticial. We can afford to have
downtime if a latent problem eventual causes a failure. If we've
properly documented that latent problem, we can check for it when the
system does fail, and fix it if necessary at that time.
Your ability to tolerate downtime is significant, as also is your thorough
documentation, so that the repair won't end up in the hands of board
swappers.
Y'know, it would be fun to see an entire set of exhibits explicitly about
failure modes. Remember the HUH S100 boards for the TRS80, where one of
the entire early production runs was reversed, but could be used by
soldering all of the components to the back side of the board?
--
Grumpy Ol' Fred cisin at xenosoft.com
8:10 p.m.
On 08/25/2010 07:34 PM, Fred Cisin wrote:
On Wed, 25 Aug 2010, Eric Smith wrote:
I have a different take on it. What is the breakage?
If it's failure is intermittent and CANNOT result in damage
I'd class that as minor and leave it. But document it!
If the failure could do minor damage but easily repaired
It would required evaluation and a decision. Document
the problem and actions taken or not taken. It's a
intermediate concern
If the failure can result if significant damage or
unrepairable damage or the possibly of severe damage.
Repair or mitigation should be applied to prevent
that. Document the mitigations/repairs and all the
supporting decisions. This is a major concern as it
can render the machine useless inert mass or possibly
worse.
The minor could be a gate with a transistor that has
compromised characteristics. This has historical value.
Intermediate could be a core driver that threatens to
burn up a sense inhibit line making for a difficult and
time consuming core repair.
Major could be a power supply that fails and goes
over-voltage enough to destroy whole sections.
Or a situation like PDP-15 had where there was
a fan failure that could result in un-contained fire.
At the extreme this is a hazard to people and other
artifacts. At best it will result in the population of
working machines to very few or zero. A great loss.
I have not specified how the modification or repair is done
only the process leading to it. Something rare and hard
to find spares for repair should have all efforts made to
do so in a reasonable way. Safety should never be
compromised.
Allison
I think the way that I would state it is that
just because the system is
working correctly (at the moment) doesn't mean that it isn't broken.
Part of the reason I don't think we'd fix such a problem in the PDP-1 is
that the PDP-1 isn't doing anything criticial. We can afford to have
downtime if a latent problem eventual causes a failure. If we've
properly documented that latent problem, we can check for it when the
system does fail, and fix it if necessary at that time.
Your ability to tolerate downtime is significant, as also is your thorough
documentation, so that the repair won't end up in the hands of board
swappers.
Y'know, it would be fun to see an entire set of exhibits explicitly about
failure modes. Remember the HUH S100 boards for the TRS80, where one of
the entire early production runs was reversed, but could be used by
soldering all of the components to the back side of the board?
--
Grumpy Ol' Fred cisin at xenosoft.com
26 Aug
26 Aug
2:45 p.m.
On Wed, 25 Aug 2010, Eric Smith wrote:
That would apply to me too i nthe case of this HP terminal. I am not
depending on it working. Yes, I want to be able to use it, and I want to
play about with it, but if it fails I am quite happy to open it up and
grab the 'scope and logic analyser...
I think the way that I would state it is that
just because the system is
working correctly (at the moment) doesn't mean that it isn't broken.
Part of the reason I don't think we'd fix such a problem in the PDP-1 is
that the PDP-1 isn't doing anything criticial. We can afford to have
downtime if a latent problem eventual causes a failure. If we've
properly documented that latent problem, we can check for it when the
system does fail, and fix it if necessary at that time.
Your ability to tolerate downtime is significant, as also is your thorough documentation, so that the repair won't end up in
the hands of board
swappers.
Y'know, it would be fun to see an entire set of exhibits explicitly about
failure modes. Remember the HUH S100 boards for the TRS80, where one of
the entire early production runs was reversed, but could be used by
soldering all of the components to the back side of the board?
:-).
I am pretty sure there ws a DMM sold in the UK which sued as expensiv
'custom' 40 pin chip. Turns out it was a 7107 or something mirror-flippe.
You could use a normal one if you soldered it to the otehr side of the
PCN (and there was enough space to do that.
Sinclar machines are knwon for such bodges -- ICs and/or transistors
stuck on top of other comonents with kludgewires everywhere. Although
wether those machines ever worked correctly is debatable :-)
-tony
25 Aug
25 Aug
8:20 p.m.
On Wednesday 25 August 2010, Eric Smith wrote:
I wrote:
To be specific, during the original restoration, we checked every
solder joint in the PDP-1. We found four bad connections. We then
reviewed the implications of the bad connections in terms of function
and reliability. The Team decided to only repair two of the connections
because they could affect normal operation and demoing of the system.
We left the other connections "loose". As Eric already stated,
we documented everything and labeled all changes (fixes) we made.
BTW: We have never experienced a problem with the "unfixed" solder
joints ;-)
Cheers,
Lyle
--
Lyle Bickley, AF6WS
Bickley Consulting West Inc.
http://bickleywest.com
"Black holes are where God is dividing by zero"
Had we found any problem analogous to that in
the PDP-1 restoration,
where the machine operated correctly despite a manufacturing defect, I'm
sure we would have had a debate on whether to fix it. I think my own
opinion in that case would be that we should leave it alone, but tag the
module (paper tag attached with a short loop of string) and document the
issue in the system logbook.
Fred Cisin wrote:
In the example that I mentioned of an IBM MDA
board, it worked for about
10 years, and then became intermittent. Even that, I would NOT consider
to be "operates correctly", as the particular defect IS a defect, and may
eventually bring the system down.
I think the way that I would state it is that just because the system is
working correctly (at the moment) doesn't mean that it isn't broken.
Part of the reason I don't think we'd fix such a problem in the PDP-1 is
that the PDP-1 isn't doing anything criticial. We can afford to have
downtime if a latent problem eventual causes a failure. If we've
properly documented that latent problem, we can check for it when the
system does fail, and fix it if necessary at that time.
26 Aug
26 Aug
2:24 p.m.
Tony Duell wrote:
Seeems reasonable for such a rare machine.
I could trivially desolder the IC, straighten the
pin, and solder it
back
properly. But should I? What would others do?
Had we found any problem analogous to that in the PDP-1 restoration,
where the machine operated correctly despite a manufacturing defect, I'm
sure we would have had a debate on whether to fix it. I think my own
opinion in that case would be that we should leave it alone, but tag the
module (paper tag attached with a short loop of string) and document the
issue in the system logbook.
For something that isn't considered a museum artifact, I'd be more
inclined to fix it.
While HP262x terminals are not exactly common (this is the only one Iv'e
ever seen...), I don't put it in the same class as the PDP1 :-)
-tony
25 Aug
25 Aug
7:09 p.m.
On Wed, 25 Aug 2010, Tony Duell wrote:
I could trivially desolder the IC, straighten the pin,
and solder it back
properly. But should I? What would others do?
If you care about the machine, of course!
I did on an IBM MDA board.
If you don't care about using the machine, and are collecting based on
resale value, rarity, provenance, then maybe not. Although such things do
NOT have the artificial rarity and inflated value that a misprinted stamp
does, if the computer has a particularly unique history, then it might be
worth saving for that - "Woz mis-soldered this chip" kind of crap. Since
"all science is either physics or stamp-collecting", which of those YOU
are doing, could determine whether repairs, or even modifications, are
appropriate.
--
Grumpy Ol' Fred cisin at xenosoft.com
26 Aug
26 Aug
2:36 p.m.
On Wed, 25 Aug 2010, Tony Duell wrote:
I care about having it working (I am not a 'stamp collector' :-)), and I
care about any failures that I could anticipate not causing more damage.
But actualyl, having looked at the circuit, I dont' see how an
open-circuit on the ground pin of this IC would damage any of the
imposible-to-replace components in the terminal
I could trivially desolder the IC, straighten the
pin, and solder it back
properly. But should I? What would others do?
If you care about the machine, of course! I did on an IBM MDA board.
If you don't care about using the machine, and are collecting based on
resale value, rarity, provenance, then maybe not. Although such things do
WOuld I do that???
NOT have the artificial rarity and inflated value that
a misprinted stamp
does, if the computer has a particularly unique history, then it might be
worth saving for that - "Woz mis-soldered this chip" kind of crap. Since
I am darn sure this PCB was wave-soldered, and the parts were probably
automatically fitted. So there is no chance of me being able to say
$person made this error. Not that such athing would really interest me
anyhow.
"all science is either physics or
stamp-collecting", which of those YOU
are doing, could determine whether repairs, or even modifications, are
appropriate.
As a physicist, I think I know which side I am on here :-)
-tony
25 Aug
25 Aug
9:37 p.m.
On Wednesday (08/25/2010 at 10:59PM +0100), Tony Duell wrote:
I could trivially desolder the IC, straighten the pin, and solder it back
properly. But should I? What would others do?
_I_ would fix it. It's clearly not what HP intended, sadly it slipped
through some quality inspection and certainly isn't supposed to be
that way. I'd take the position of fixing it now, while the unit is
apart for other reasons, so that you don't have to take it apart again
later when it does fail and then risk damaging other aged things then.
I had a similar decision which you folks offered input on-- which was
the apparently failed transformer in my SWTPC 6800 system (c. 1976).
I failed to close the loop on that story but in the end, I discovered
that the real problem was the lousy MOLEX connectors that were used
between the power supply PCB and the end of the power harness going to
the motherboard. The pins in this connector were not mating well, were
corroded and had been heating up for some time (a long time ago!). The
white MOLEX shell was turning brown and was actually warm to the touch
when I would briefly power the machine on. I decided to replace this
connector with something similar but better and all the issues went away.
Chris
--
Chris Elmquist
26 Aug
26 Aug
1:34 p.m.
From: Chris Elmquist
Sent: Wednesday, August 25, 2010 6:37 PM
... I discovered that the real problem was the lousy
MOLEX connectors
^^^^^ ^^^^^
Redundant, isn't it?
Rich Alderson
Vintage Computing Sr. Server Engineer
Vulcan, Inc.
505 5th Avenue S, Suite 900
Seattle, WA 98104
mailto:RichA at vulcan.com
mailto:RichA at LivingComputerMuseum.org
http://www.PDPplanet.org/
http://www.LivingComputerMuseum.org/
3:27 p.m.
On Wednesday (08/25/2010 at 10:59PM +0100), Tony Duell wrote:
That much is certain :-). There is a hole and pad on the PCB for this pin.
I could trivially desolder the IC, straighten the pin, and solder it back
properly. But should I? What would others do?
_I_ would fix it. It's clearly not what HP intended, sadly it slipped through some quality inspection and certainly
isn't supposed to be
that way. I'd take the position of fixing it now, while the unit is
apart for other reasons, so that you don't have to take it apart again
later when it does fail and then risk damaging other aged things then.
The things most likely to be damaged if I have to keep on removing the
PCB are the Nylatch clips thst hold it in plave (those are those fixings
that look like an over-size LED clip with a central plunger where pushing
the plunger home spreds the 'clip', thus locking the thing into a hole in
the chassis). or the CRT. Now Nylatch clips I have 'in stock', but
finding a CRT coupdl be more of a problem. And I;d tather not have to
replace it for no good reason.
I had a similar decision which you folks offered input on-- which was
the apparently failed transformer in my SWTPC 6800 system (c. 1976).
I failed to close the loop on that story but in the end, I discovered
that the real problem was the lousy MOLEX connectors that were used
Seen that many times... Commodore PETs suffer from this, for example.
between the power supply PCB and the end of the power
harness going to
the motherboard. The pins in this connector were not mating well, were
corroded and had been heating up for some time (a long time ago!). The
white MOLEX shell was turning brown and was actually warm to the touch
when I would briefly power the machine on. I decided to replace this
connector with something similar but better and all the issues went away.
I think I've mentioend befroe that my PERQ 3A had disk problems when I
got it. It turned out that there's an in-line fuseholder in the 12V feed
ot the head disk (Maxtor XT1140 type of thing) and one of the terminal
screws in the holder was loose and there was thus a high-resistnace
connection there. It was a particularly nasty fault because the voltage
only driopped far enough to cause problems when the drive was seeking
(due to the extra current needed to drive the positioner coil). So it ws
not easy to spot....
-tony
25 Aug
25 Aug
9:46 p.m.
From: ard at p850ug1.demon.co.uk
---snip---
WHile cleaning the dirt off this PCB (HV attracts dust electrostatcally,
of couse), I noticed that the ground pin of this IC never made it through
the board. It's folded under the package, but has been effectively
surface-mount soldered to the pad when the board was wave-soldered. It
works, but I can't beeleive it's as good as a proper through-hole
connection, and it's certian;y not what was intended.
I could trivially desolder the IC, straighten the pin, and solder it back
properly. But should I? What would others do?
-tony
Hi
First, can you explain how it works as is?
Being one shots, such a fix might effect the timing. Is it adjustable?
At a place I once worked, we had a board where the +5 was missing from
a CMOS part. It was running from the inputs that luckily always had
at least one high.
Dwight
9:58 p.m.
Actually, on many boards there was effectively the copper went all the way thru the hole
("thru hole plating"), thus providing continuity from one side to the other.
Only on really cheap / very early boards did soldering provide that thru-the-hole
connection -- it caused reliability problems (surprise surprise).
Even without that, grounds specifically tend to be run as busses, and so missing one
thru-hole connection on the ground buss would probably be no big deal, as there are
probably tons of other connections where ground is passed thru from one side of the board
to the other.
So, as long as that pin is soldered to the pad on top, and so long as that pad runs to
ground somewhere else *OR* so long as the thru-hole plating is intact -- no biggy.
Either way, it is almost certainly fine just the way it is.
JRJ
---- dwight elvey <dkelvey at hotmail.com> wrote:
From: ard at p850ug1.demon.co.uk
---snip---
WHile cleaning the dirt off this PCB (HV attracts dust electrostatcally,
of couse), I noticed that the ground pin of this IC never made it through
the board. It's folded under the package, but has been effectively
surface-mount soldered to the pad when the board was wave-soldered. It
works, but I can't beeleive it's as good as a proper through-hole
connection, and it's certian;y not what was intended.
I could trivially desolder the IC, straighten the pin, and solder it back
properly. But should I? What would others do?
-tony
Hi
First, can you explain how it works as is?
Being one shots, such a fix might effect the timing. Is it adjustable?
At a place I once worked, we had a board where the +5 was missing from
a CMOS part. It was running from the inputs that luckily always had
at least one high.
Dwight
10:16 p.m.
On Wed, 25 Aug 2010 cube1 at charter.net wrote:
So, as long as that pin is soldered to the pad on top,
and that is normally a really shitty solder connection.
and so long as
that pad runs to ground somewhere else *OR* so long as the thru-hole
plating is intact -- no biggy.
Either way, it is almost certainly fine just the way it is.
10 years OK, then gradually worsening intermittent connection.
Even DURING those first 10 years, I do NOT consider that "fine".
The Board Swappers declared the board to be "NFG" and wanted to discard
it. One of my colleagues is about to get fired - of the fifteen
complaints against him for "unprofessional behavior", one of the most
serious is "took discarded stuff out of the college dumpster and
repaired it".
All reasons to repair it not to "good as new", but to "good as it was
supposed to be". Unless there was some historical rarity like "Gary
Kildall did this bad soldering", WHY would you want to keep it in a
defective state? Are you doing stamp collecting, or physics?
To bring this throughly OFF-topic, should you clean Ansel Adams'
thumbprint off of one of his negatives?
(which is more important? - the historical aspect, or the quality of it?)
26 Aug
26 Aug
3:43 p.m.
On Wed, 25 Aug 2010 cube1 at charter.net wrote:
That's what worries me...
So, as long as that pin is soldered to the pad on
top,
and that is normally a really shitty solder connection. and so long as
that pad runs to ground somewhere else *OR* so long as the thru-hole
plating is intact -- no biggy.
Either way, it is almost certainly fine just the way it is.
10 years OK, then gradually worsening intermittent connection.
Even DURING those first 10 years, I do NOT consider that "fine". The Board Swappers declared the board to be
"NFG" and wanted to discard
it. One of my colleagues is about to get fired - of the fifteen
complaints against him for "unprofessional behavior", one of the most
serious is "took discarded stuff out of the college dumpster and
repaired it".
I thought we were supposoied to be reducing the amount of waste sent ot
landfill.... I susepc the next bit of 'unprofessional behaviour' should
involve a testla coil and some bits of the management :-)
To bring this throughly OFF-topic, should you clean
Ansel Adams'
thumbprint off of one of his negatives?
(which is more important? - the historical aspect, or the quality of it?)
I thinki I would print it as-is to keep a record of the thumbprint (and
to preseve the image as far as I could if cleaning it ruined it), and
then clean off the thumbrint and print it 'properly' :-)
-tony
3:38 p.m.
Actually, on many boards there was effectively the copper went all the
way thru the hole ("thru hole plating"), thus providing continuity from
Yes, this PCB us through-hole plated.
one side to the other. Only on really cheap / very
early boards did
soldering provide that thru-the-hole connection -- it caused reliability
problems (surprise surprise).
I cna;t think why there are reliability problems if the sodlering is done
properly. I've built a number of prototpyes where I've had to sodler
components on both sides or fint pins to connect traces through the board
and haev never had problems from doing that.
Even without that, grounds specifically tend to be run
as busses, and
Actualyl, this PCB is somehwt odd there. Almost all the signal and power
traces are on the _component_ side of the board. The solder side is
essentailly a ground plane witha few signal traces in it (to 'hop over'
other traces on the top side). So the through hole plating is essential
to provide a gorund to this pin.
But the plating is there. The hole is full of solder, and in fact the pin
is soldered to the pad. I have checked the ground conenciton, it's find.
-tony
3:34 p.m.
From: ard at p850ug1.demon.co.uk
---snip---
=20
WHile cleaning the dirt off this PCB (HV attracts dust electrostatcally=
=2C=20
of couse)=2C I noticed that the ground pin of
this IC never made it throu=
gh=20
the board. It's folded under the package=2C
but has been effectively=20
surface-mount soldered to the pad when the board was wave-soldered. It=20
works=2C but I can't beeleive it's as good as a proper through-hole=20
connection=2C and it's certian=3By not what was intended.
=20
I could trivially desolder the IC=2C straighten the pin=2C and solder it =
back=20
properly. But should I? What would others do?
=20
-tony
=20
Hi
First=2C can you explain how it works as is? Being one shots=2C such a fix might effect the
timing. Is it adjustable?
The 2 one-shots are cascaded. The first one is driven by the Hsync signal
from the logic board and delays it. The output of that
triggers the scond
one which probides a constant-width pulse to the horizontal
drive transistor.
The first one-shot is adjustabnle. The adjustment is a 'customer preset'
control -- that is a preset that can be turned witha screwdriver through
a hole in the tope of the terminal casing. It's used to centre the
picture horizontally. The second one-shot is not adjustable. But I can
see no evidence of there being any select-on-text compoennts here.
Much the same circuit -- with the same component values is used in the
HP120 and HP150 comptuers, for example. Totally different layourt,
though. And the adjustments (or lack of them) are the smae there.
At a place I once worked=2C we had a board where the
+5 was missing from
a CMOS part. It was running from the inputs that luckily always had
at least one high.
More ofen the circuit works until all the ipuuts go high at the same
time. A nasty fault to find if you've not seen it before.
-tony
1:49 p.m.
Tony Duell wrote:
When you are cleaning up and restoring a piece of
classic computer
hardware, would you correct any minor mistakes made in the original
production?
"It depends" (doesn't it always?)
I love finding stuff like that when investigating a machine - it's part of the
machine's "story", I suppose. On the other hand, the perfectionist in me
would
likely want to fix it.
Maybe the solution is to document your change somehow and fix it, but only if
it's possible to back out of the change at a later date - that way you're not
harming the "story" but still fixing the fault while the machine is part of
your own collection for your own enjoyment.
I suppose the magnitude of the problem comes into play, too; for something
small like a solder joint that was currently working, I might be tempted just
to leave it as-is until it did break properly (which might never happen).
cheers
Jules
2:14 p.m.
On 2010 Aug 26, at 10:49 AM, Jules Richardson wrote:
Tony Duell wrote:
Maybe HP's QC isn't quite what we like to believe it to be. I've
mentioned in the past an HP9815 calculator (1970's) in which a power
supply filter electrolytic capacitor was factory installed with reverse
polarity. It has never presented a problem however, so in that instance
I didn't correct it, just noted it in my log for the machine.
When you are cleaning up and restoring a piece of
classic computer
hardware, would you correct any minor mistakes made in the original
production?
"It depends" (doesn't it always?)
I love finding stuff like that when investigating a machine - it's
part of the machine's "story", I suppose. On the other hand, the
perfectionist in me would likely want to fix it.
Maybe the solution is to document your change somehow and fix it, but
only if it's possible to back out of the change at a later date - that
way you're not harming the "story" but still fixing the fault while
the machine is part of your own collection for your own enjoyment.
I suppose the magnitude of the problem comes into play, too; for
something small like a solder joint that was currently working, I
might be tempted just to leave it as-is until it did break properly
(which might never happen).
4:45 p.m.
Maybe HP's QC isn't quite what we like to
believe it to be. I've
MAybe not... I think I've had the odd reveres small electrolytic too...
mentioned in the past an HP9815 calculator
(1970's) in which a power
supply filter electrolytic capacitor was factory installed with reverse
polarity. It has never presented a problem however, so in that instance
I didn't correct it, just noted it in my log for the machine.
Now that I would _certainly_ replace. The reason is that the 9815 (and
9825 for thatr matteer) do not have corwbar cirucits in the PSU. If the
PSU malfucntions in the 'right' way -- e.g. the chopepr transistors
shorts (on the heatsink on the side of the transforemr box), the 5V line
ca jum pt abotu 30V. Needless to say that's expensive...
So if there was anything remotely suspect about that PSU. I'd fix it.
It's a difficult PSU to test because it's all on the lower PCN that
contains the printer and tape interface circuits. But that's not a big
enough loard for it. What I do is remvoe the kyeboard/procesosr board
assembly, take uout the expansion slot unit and the tape drive, and
disconnect the printer (particualty the DIL header fo the prinhead which
contains driver circuity). Then connect a 6V bulb across the 5V lone on
the lower PCB and power up. If the PSU oges crazy, all it will ruin are a
few common logic parts on that PCB, no custom parts.
-tony
5:17 p.m.
On 2010 Aug 26, at 1:45 PM, Tony Duell wrote:
The capacitor in question is the 680uF unit on the output side of the
+15V regulator. The regulator (723) is configured with current
limiting.
Someday I'd like to remove the cap and measure whether it is actually
providing any C to the circuit after years of reverse polarity
operation. The cap is rated for 25V, I wonder whether there is a
threshold effect on 'lytic caps, where -Vapplied < fraction*Vrated is
not damaging to the cap.
Maybe HP's
QC isn't quite what we like to believe it to be. I've
MAybe not... I think I've had the odd reveres small electrolytic too...
mentioned in the past an HP9815 calculator
(1970's) in which a power
supply filter electrolytic capacitor was factory installed with
reverse
polarity. It has never presented a problem however, so in that
instance
I didn't correct it, just noted it in my log for the machine.
Now that I would _certainly_ replace. The reason is that the 9815 (and
9825 for thatr matteer) do not have corwbar cirucits in the PSU. If the
PSU malfucntions in the 'right' way -- e.g. the chopepr transistors
shorts (on the heatsink on the side of the transforemr box), the 5V
line
ca jum pt abotu 30V. Needless to say that's expensive...
27 Aug
27 Aug
4:07 p.m.
[HP9815 PSU]
The capacitor in question is the 680uF unit on the
output side of the
+15V regulator. The regulator (723) is configured with current
limiting.
I have not analyssed the circuit in detail, but it appears that that +15V
rail (oputput of the 723) supplies the +5V chopper control circuit. It's
not clear what happens if that supply is missing, but since the chopper
transsitors is PNP, I could see it being turned hard on, which would put
30V or so on the 5V line.
No, I am not goign to try this practically.
-tony
28 Aug
28 Aug
8:22 p.m.
On 2010 Aug 27, at 1:07 PM, Tony Duell wrote:
[HP9815 PSU]
Some further circuit analysis indicates that if +15 were lost the +5
chopper would end up biased off: there are intervening driver
transistors which need + drive V to turn on the chopper and the chopper
has a low-value R between B-E, biasing it off by default.
Regardless, the thing has been fine since it was manufactured over 30
years ago, including all the use it received during it's heyday.
Strikes me as a rather low-risk situation. To each his own. While one
part of me wants to correct it to the design intention, I'm not
worrying about it.
The capacitor in question is the 680uF unit on
the output side of the
+15V regulator. The regulator (723) is configured with current
limiting.
I have not analyssed the circuit in detail, but it appears that that
+15V
rail (oputput of the 723) supplies the +5V chopper control circuit.
It's
not clear what happens if that supply is missing, but since the chopper
transsitors is PNP, I could see it being turned hard on, which would
put
30V or so on the 5V line.
No, I am not goign to try this practically.
5026
days inactive
5030
days old
30 comments
13 participants
participants (13)
-
ajp166＠verizon.net -
ard＠p850ug1.demon.co.uk -
chrise＠pobox.com -
cisin＠xenosoft.com -
cube1＠charter.net -
dkelvey＠hotmail.com -
eric＠brouhaha.com -
hilpert＠cs.ubc.ca -
james＠machineroom.info -
jules.richardson99＠gmail.com -
lbickley＠bickleywest.com -
oltmansg＠bellsouth.net -
RichA＠vulcan.com