15 Jan
2012
15 Jan
'12
5:09 p.m.
We reformed the capacitors in all three power supplies in the PDP-8/I
at the RICM and powered on the CPU to see if it would show any signs
of life. It randomly lit some lights and would not react to the front
panel switches. Looks like we have some debugging to do.
The front panel power switch was corroded and burned. Any idea where
we can get a replacement?
Were can we get some Oshinos OL-1 bulbs to repair the front panel?
Does anyone have diagnostics for the 8/I on DECtape?
--
Michael Thompson
18 Jan
18 Jan
5:34 p.m.
On 1/15/12 9:09 AM, "Michael Thompson" <michael.99.thompson at gmail.com>
wrote:
We reformed the capacitors in all three power supplies
in the PDP-8/I
Don't "reform" them, replace them. See, e.g.,
http://www.cde.com/tech/reliability.pdf on the limitations of lifespan of
aluminum electrolytic capacitors. Keep in mind that the electrolytics in
a PDP-8/I are at least 40+ years old.
at the RICM and powered on the CPU to see if it would
show any signs
of life. It randomly lit some lights and would not react to the front
panel switches. Looks like we have some debugging to do.
Again, be sure your power supplies are in good shape, not only in terms of
static voltage but also ripple - put a scope on them under load. Some
folks don't have an oscilloscope (although I would assume you at RICM do):
for those reading along, keep in mind it's not necessary to have a 500MHz
dual-trace digital phosphor "video game" if you're working on PDP-8's
and
such. A simple 10MHz scope that should be available for under $50 (I sold
one a couple of years back for $20) is perfectly good for not only
measuring power supply ripple but for lots of other diagnostic purposes in
machines of this vintage. (I have a vacuum-tube Tektronix 561A that I
love to use for things like this, just because I can.)
The front panel power switch was corroded and burned. Any idea where
we can get a replacement?
Were can we get some Oshinos OL-1 bulbs to repair the front panel?
Does anyone have diagnostics for the 8/I on DECtape?
I'd recommend finding paper tape images (Bitsavers, pdp8online.com) and
feeding them in through a terminal emulator. The DECtape controller is
itself a complex piece of equipment - minimize your variables! :-) Once
you have the CPU running reliably, then you'll probably need to align the
(presumably) TC08 - it has a handful of trimmers that adjust various pulse
widths, and I can tell you from experience they aren't just for show.
(You'll need a scope for this, too.) -- Ian
6:41 p.m.
Ian King wrote of aluminum electrolytic capacitors:
Don't "reform" them, replace them. See,
e.g.,
http://www.cde.com/tech/reliability.pdf on the limitations of lifespan
of aluminum electrolytic capacitors. Keep in mind that the
electrolytics in a PDP-8/I are at least 40+ years old.
I respectfully disagree.
If you want to ensure that the artifact will be operable for the maximum
possible time with the minimum required maintenance (e.g., replacing the
capacitor again), then you are correct that the capacitors should be
replaced. However, this is not generally the primary objective of
museums and similar entities with regard to the artifacts.
For example, in the PDP-1 restoration, we went to great pains to reform
the original capacitors, using a grossly over-engineered technique, and
to document every aspect of this process (and of the entire
restoration). It is true that the reformed capacitors can't be expected
to have as much remaining operating life as a new capacitor, but that
isn't the point. We attempted to keep as much of the original artifact
as possible, and found that with careful reformation, nearly all of the
capacitors still met the original specifications. Given the relatively
small percentage of time that the machine is powered up, generally less
than 2% of the time, the capacitor lifetime will be predominantly
determined by the non-powered wearout mechanisms, which are quite slow.
The machine is normally powered up for demonstrations for at least a few
hours every two weeks, which is sufficient to prevent deterioration of
the oxide layer which would eventually necessitate another round of
reformation. We expect to check the power supply ripple on occasion to
see how the capacitors are holding up, and certainly we will replace
them if they fail, but we don't expect that to happen for some time. If
at some point the demonstrations cease and the machine goes unpowered
for a long interval, the capacitors will need to be rechecked thoroughly
and possibly reformed before the machine can be powered again.
Eric
19 Jan
19 Jan
2:13 a.m.
For example, in the PDP-1 restoration, we went to
great pains to reform the
original capacitors, using a grossly over-engineered technique, and to
document every aspect of this process (and of the entire restoration).
Has there been any thought about installing some sort of non-invasive,
reversible monitoring circuitry to the PDP-1 (or other operating
artifacts)? Perhaps a circuit to watch the ripple or leakage in the
capacitors, and shut down the machine if and danger is sensed?
--
Will
2:21 a.m.
William Donzelli wrote:
Has there been any thought about installing some sort
of non-invasive,
reversible monitoring circuitry to the PDP-1
I proposed that, but I don't think the idea got any serious consideration.
(or other operating artifacts)?
I may be involved in an effort to get one or two DECSYSTEM-2065 (KL10)
systems running in the near future, and I'm seriously considering adding
quite a few voltage/current/temperature monitors.
3:48 a.m.
On 1/18/12 6:21 PM, "Eric Smith" <eric at brouhaha.com> wrote:
William Donzelli wrote:
Instead of all of that, why not just replace the capacitors?
We've restored a 2065 that runs 24x7, and that's where we learned that you
can cross your fingers and pray, or recognize that even when they're
sitting on a shelf aluminum electrolytic capacitors (AlECs) degrade with
time.
I understand the concern about cosmetics - modern AlECs may be
significantly different in dimensions, and in situations where they are
visible it can impact the appearance of the artifact. On our PDP-7, we
elected not to replace the filter capacitors, which are the size of large
soup cans. Instead, we added small power supplies in a non-destructive
manner (no holes were drilled) and added cabling to provide their output
to the machine's circuits. The visual impact is small and the amendment
can be completely reversed.
It's a real problem that modern AlECs, while superior in construction and
density, are commonly far smaller than original components. But replacing
existing components with contemporaneous "new old stock" doesn't solve the
problem, because AlECs degrade simply sitting on the shelf, according to
research. We're going to have to get creative. I've heard of folks who
hollow out original AlECs and embed new, higher-density components. We
supplemented the existing system with modern switching supplies (which
also has the advantage of far higher efficiency). The tension between
access and preservation continues, with no simple or obvious answers. --
Ian
Has there been any thought about installing some
sort of non-invasive,
reversible monitoring circuitry to the PDP-1
I proposed that, but I don't think the idea got any serious consideration.
(or other operating artifacts)?
I may be involved in an effort to get one or two DECSYSTEM-2065 (KL10)
systems running in the near future, and I'm seriously considering adding
quite a few voltage/current/temperature monitors.
5:07 a.m.
On Jan 18, 2012, at 22:48, Ian King <IanK at vulcan.com> wrote:
On 1/18/12 6:21 PM, "Eric Smith" <eric at
brouhaha.com> wrote:
Why not do both? The monitors will help detect when the new capacitors inevitably start
failing.
- Dave
I may be involved in an effort to get one or two
DECSYSTEM-2065 (KL10)
systems running in the near future, and I'm seriously considering adding
quite a few voltage/current/temperature monitors.
Instead of all of that, why not just replace the capacitors? 
7:08 a.m.
Zitat von David Riley <fraveydank at gmail.com>:
Why not do both? The monitors will help detect when
the new
capacitors inevitably start failing.
The problem is, they are going bad, because you didn't use them.
Monitoring them after switching on the power supply isn't not really
neccessary, as you can see the hole in the pcb very easily.
And a monitoring circuit to monitor a CAP in circuit is not going to
be easy anyway. Electrolytic CAPS are changing their capacitance
drastically over temperature, and actually the worst is, that going
bad, they are getting hot during startup, where you have the worst
ripple on the power line.
I may be involved in an effort to get one or two
DECSYSTEM-2065 (KL10)
systems running in the near future, and I'm seriously considering adding
quite a few voltage/current/temperature monitors.
Instead of all of that, why not
just replace the capacitors?
1:11 p.m.
And a monitoring circuit to monitor a CAP in circuit
is not going to be easy
anyway. Electrolytic CAPS are changing their capacitance drastically over
temperature, and actually the worst is, that going bad, they are getting hot
during startup, where you have the worst ripple on the power line.
Such circuits are well documented and available.
Mainframe equipment (or maybe I should say *real* mainframe equipment
- zing!) tends to have such circuitry by default.
--
Will
10:08 p.m.
Why not do both? The monitors will help detect when
the new capacitors
inevitably start failing.
I am not sure I understnd the point of the monitor circuit.
If the capacitor fails dead-short, then the PSU fuses will blow (You do
hae fuses in series with the transfomer windings, I hope). No monitor
circuit is going to do much of use in that situation.
If the capacitor fails open, or high ESR, or loses capacitance, the
ripple on that supply line will increase. A monitor could detect that,
but equally if that happens, the machine stops working correctly. And the
first thing I'd do to any piece of vintage hardware that wasn't working
-- whatever the fault -- is to check all the supply lines for voltage and
ripple. At which point you home in on the faulty capacitor quite quickly.
A monitor would perhaps save you 2 or 3 'scope tests. Is that worth it?
-tony
11 p.m.
On Thu, 19 Jan 2012, Tony Duell wrote:
I am not sure I understnd the point of the monitor
circuit.
[snip]
If the capacitor fails open, or high ESR, or loses
capacitance, the
ripple on that supply line will increase. A monitor could detect that,
but equally if that happens, the machine stops working correctly.
But it is quite possible that a monitor circuit could detect an increase
in the ripple before it was large enough to stop the machine from working
correctly, and it might be desirable to shut the machine down before the
ripple becomes excessive.
Alexey
20 Jan
20 Jan
2:59 a.m.
On 01/19/2012 02:08 PM, Tony Duell wrote:
I am not sure I understnd the point of the monitor
circuit.
[...]
If the capacitor fails open, or high ESR, or loses
capacitance, the
ripple on that supply line will increase. A monitor could detect that,
but equally if that happens, the machine stops working correctly
.
The ripple can increase quite a bit before it causes failures. A
monitoring circuit can catch it before it gets that far.
3:21 a.m.
New subject: Straight-8 at VCF East
In a little bit of a friendly rivalry with RICM :) ..... David
Gesswein, who you all know, intends to bring his restored/operational
ORIGINAL pdp-8 to the VCF East. Our museum has one too, but it's not
(yet!) restored.
Now, I ask: maybe 10% of cctalker's have ever seen a Straight-8; here in
MARCH we're fortunate to have our own; but how many people have ever
seen one WORKING? I sure haven't! I fully expect all you DEC-heads who
are the foundation of this list to make your pilgrimage to New Jersey
this spring.
While you're here, you can heckle Dave McGuire, who'll lead our "intro
to minicomputer care and feeding" class. But be warned, he heckles
back. :-)
9:54 p.m.
On 01/19/2012 02:08 PM, Tony Duell wrote:
True, but given the complexity of the monitoring circuit [1], do you really
gain that much by having it. I think I'd just conenct all the supply
lines to a rotary switch [2] and fed them to a low-end 'scope. And then
once a day (or so) trun the swithc round and look at the ripple on each
line.
[1] When I add bits to a classic computer, I try to use compoennts and
designs that were avaialble when the machine was made. So something like
this woudl be done using analogue circuitry. Sampling the supply lien
withaAnd ADC and tryign to measure the ripple that way would not be
somethbing I'd want to do in a classic machine
[2] Make sure it's break-brfore-make for very obvious reasons!
-tony
I am not sure I understnd the point of the
monitor circuit.
[...]
If the capacitor fails open, or high ESR, or
loses capacitance, the
ripple on that supply line will increase. A monitor could detect that,
but equally if that happens, the machine stops working correctly
.
The ripple can increase quite a bit before it causes failures. A
monitoring circuit can catch it before it gets that far.
19 Jan
19 Jan
5:44 a.m.
Ian King wrote:
Instead of all of that, why not just replace the
capacitors?
Why not just verify that the capacitors meet specifications, and only
replace them if they've failed?
Anyhow, even if the capacitors are replaced, I'll *still* want to
monitor the voltage and ripple.
even when they're sitting on a shelf aluminum
electrolytic capacitors
(AlECs) degrade with time.
There is very little non-reversible degradation. Most of the
degradation is oxide breakdown, and is reversed by the reformation process.
Note that the reformation process isn't doing something bizarre to the
capacitor. It's basically doing the same thing the vendor does to it
during manufacturing to grow the oxide layer in the first place.
On our PDP-7, we elected not to replace the filter
capacitors, which
are the size of large soup cans.
Have you looked at the capacitors in the "capacitor box" of the ECL
power supply of the KL10, between the H760 raw supply and H761 regulated
supply? There are 13 300,000uF 15V capacitors. There are certainly
modern replacements that are smaller, but if they're not broken, I'm not
going to replace them. They're also fairly expensive, roughly $50 each.
20 Jan
20 Jan
1:58 a.m.
From: Eric Smith
Sent: Wednesday, January 18, 2012 9:44 PM
Ian King wrote:
> Instead of all of that, why not just replace the capacitors?
Why not just verify that the capacitors meet
specifications, and only
replace them if they've failed?
Ian was not at LCM yet when we did exactly this, after reading the CDE
white paper. More than 95% of the aluminum electrolytic caps in the 2065
had failed. We first tried "re-forming" a few. They failed again in
days.
At that point we made the decision never to waste our time on this. We
buy new capacitors (or higher-level components; cf. our PDP-7) and put
them in.
> On our PDP-7, we elected not to replace the filter
capacitors, which
> are the size of large soup cans.
Have you looked at the capacitors in the
"capacitor box" of the ECL
power supply of the KL10, between the H760 raw supply and H761 regulated
supply? There are 13 300,000uF 15V capacitors. There are certainly
modern replacements that are smaller, but if they're not broken, I'm not
going to replace them. They're also fairly expensive, roughly $50 each.
As I noted above, Ian was not on the 2065 project. (He did the same kind
of surgery on our VAX-11/785 systems.) Keith and I did the 2065, with
the aid of two friends of yours, Bruce Kennard and Bob Sellars.
The point is that we *run* our systems, unlike the PDP-1 at CHM. We can't
afford to mess around.
Rich Alderson
Vintage Computing Sr. Systems Engineer
Vulcan, Inc.
505 5th Avenue S, Suite 900
Seattle, WA 98104
mailto:RichA at vulcan.com
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
4:58 p.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
In article <4F19986B.1080209 at bitsavers.org>,
Al Kossow <aek at bitsavers.org> writes:
On 1/19/12 5:58 PM, Rich Alderson wrote:
I know that both the LCM and the CHM run their systems, but I think
you'll agree that running them 24/7 is a bit different than running
them a couple hours every two weeks.
Besides, LCM and CHM have different missions as I see it. CHM has a
primary mission of preservation and while it's cool to be able to run
things, that seems less important to CHM. Please correct me if I'm
wrong about that.
When I've been to the CHM, things were in static displays and behind
glass cases and turned off. I haven't been there when the PDP-1 was
running and the only stuff I've seen operating there was at the PLATO at 50
conference where the PLATO terminals that were running were brought in
by someone else and not part of the CHM's collection. Maybe some
things are different with the big revolution exhibit, I haven't been
there to see that yet.
LCM has a primary mission of operating the equipment 24/7. Preservation
is still important to them, but not as important as having operating
equipment. So they upgrade the power supplies and replace the original
capacitors. The PS upgrades are non-destructive and reversible and they
keep the old capacitors -- I don't know how you could non-destructively
replace capacitors since they have to be soldered into the circuits.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 version available for download
<http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/>
Legalize Adulthood! <http://legalizeadulthood.wordpress.com>
The point is that we *run* our systems, unlike
the PDP-1 at CHM.
What exactly was the point of posting this bit of disinformation? 
5:13 p.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
[...] -- I don't know how you could
non-destructively replace
capacitors since they have to be soldered into the circuits.
Soldered? Not necessarily. I have a number of large caps which have
screw terminals rather than solder terminals.
I'm also not entirely sure why you consider unsoldering one thing and
soldering in another a destructive operation.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
5:37 p.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
This debate is not unique to the computer world. This debate comes up
all the time in the rail and aviation preservation communities as
well. In general, groups that tend to be tied in with the museum and
academic communities tend to favor a "static display of relics"
approach while those from the various enthusiast communities tend to
favor the "keep your objects operational" approach. Sadly, many
foundations and funding sources tend to favor the "super-preservation"
approach which means that groups that want to keep old computers
computing, locomotives hauling and planes flying are FAR more reliant
on large volunteer groups and/or support from wealthy individuals than
more traditional museums are.
6:06 p.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
On 01/20/2012 12:13 PM, Mouse wrote:
In fact, nearly ALL "big" capacitors use screw terminals, or to a
lesser extent, spade lugs.
-Dave
--
Dave McGuire
New Kensington, PA
[...] -- I
don't know how you could non-destructively replace
capacitors since they have to be soldered into the circuits.
Soldered? Not necessarily. I have a number of large caps which have
screw terminals rather than solder terminals. 
6:45 p.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
I know that both the LCM and the CHM run their
systems, but I think
you'll agree that running them 24/7 is a bit different than running
them a couple hours every two weeks.
Running a paper tape based computer that requires operator training makes
no sense to run 24/7. The point plot CRT has problems with deflection amplifier
failures and the team that maintains the PDP-1 has never suggested that we run it
unattended.
24/7 operation makes sense for timesharing environments, which has been what
LCM has been offering access to for a while now.
Besides, LCM and CHM have different missions as I see
it. CHM has a
primary mission of preservation and while it's cool to be able to run
things, that seems less important to CHM. Please correct me if I'm
wrong about that.
the actual mission statement is:
http://www.computerhistory.org/about/
"The mission of the Computer History Museum is to preserve and present for posterity
the artifacts and stories of the information age."
So, how do you balance "present for posterity", keeping an artifact in no worse
condition that you received it
for future generations, with the desire of enthusiasts to have "hands on" access
to them.
My personal interpretation is that we make available all of the documentation and software
that can be located, which is of the
most use to people not in the San Fransisco Bay area, and not use CHM's space and
other resources for restorations. This view is
shared by the other curators at the museum.
In general, groups that tend to be tied in with the
museum and
academic communities tend to favor a "static display of relics"
approach while those from the various enthusiast communities tend to
favor the "keep your objects operational" approach.
We had around 65,000 visitors last year. With that many members of the general public
in an exhibit, theft and vandalism becomes a real concern. Along with that, many people
feeling the need to touch artifacts, climb on platforms, and try to do things like open
equipment drawers.
Hopefully, that explains the current thinking of the curatorial staff of the museum.
7:21 p.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
On Fri, Jan 20, 2012 at 1:45 PM, Al Kossow <aek at bitsavers.org> wrote:
We had around 65,000 visitors last year. With that many members of the
general public
in an exhibit, theft and vandalism becomes a real concern. Along with that,
many people
feeling the need to touch artifacts, climb on platforms, and try to do
things like open
equipment drawers.
Hopefully, that explains the current thinking of the curatorial staff of the
museum.
I'm not saying that the strict preservationist approach is necessarily
a bad thing, especially where you've got lots of visitors.
Additionally, I do feel that the CHM's software and documentation
preservation and sharing work is IMMENSELY beneficial to everyone. I
was merely pointing out that the tensions between a preservation-first
approach and an operations-oriented approach exist in many contexts.
Mike
8:41 p.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
In article <4F19B636.4050602 at bitsavers.org>,
Al Kossow <aek at bitsavers.org> writes:
I agree with everything you've written. I didn't mean to imply that
the difference between the way LCM does things and the way CHM does
things means that one is better than the other. I like that we have
different organizations so that we can pursue both the preservation
goal and the operational goal.
I know that
both the LCM and the CHM run their systems, but I think
you'll agree that running them 24/7 is a bit different than running
them a couple hours every two weeks.
Running a paper tape based computer that requires operator training makes
no sense to run 24/7. The point plot CRT has problems with deflection
amplifier failures and the team that maintains the PDP-1 has never
suggested that we run it unattended.
24/7 operation makes sense for timesharing environments, which has been what
LCM has been offering access to for a while now. My personal interpretation is that we make available
all of the
documentation and software that can be located, which is of the most
use to people not in the San Fransisco Bay area, and not use
CHM's space and other resources for restorations. This view is
shared by the other curators at the museum.
I have no problem with that either.
We had around 65,000 visitors last year. With that
many members
of the general public in an exhibit, theft and vandalism becomes
a real concern. Along with that, many people feeling the need to
touch artifacts, climb on platforms, and try to do things like open
equipment drawers.
Yes, proper treatment of the equipment is something that concerns me
in making interactive hands-on exhibits. Ultimately, it may boil down
to making sacrificial replicas for which no harm is ultimately done
should vandalism and monkeying about be a concern. I think my museum
will be small enough for a while that this won't be an immediate
concern (i.e. with 3 seats and an exhibit that's not being used
without someone overseeing it's use).
However, this is something that will most likely need to be addressed
in some fashion at some point. Remember, the cages in zoos protect
not only the visitors from the animals, but also protect the animals
from the visitors.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 version available for download
<http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/>
Legalize Adulthood! <http://legalizeadulthood.wordpress.com>
21 Jan
21 Jan
10:24 a.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
If it can run it should run.
Yes protect it from those who know not.
What I am agin is electronic taxidermy
?
Rod Smallwood
?
?
-----Original Message-----
From: cctech-bounces at classiccmp.org [mailto:cctech-bounces at classiccmp.org]
On Behalf Of Richard
Sent: 20 January 2012 20:41
To: cctalk
Subject: Re: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
In article <4F19B636.4050602 at bitsavers.org>,
Al Kossow <aek at bitsavers.org> writes:
what
I know that
both the LCM and the CHM run their systems, but I think
you'll agree that running them 24/7 is a bit different than running
them a couple hours every two weeks.
Running a paper tape based computer that requires operator training makes
no sense to run 24/7. The point plot CRT has problems with deflection
amplifier failures and the team that maintains the PDP-1 has never
suggested that we run it unattended.
24/7 operation makes sense for timesharing environments, which has been LCM has been offering access to for a while now.
I agree with everything you've written. I didn't mean to imply that
the difference between the way LCM does things and the way CHM does
things means that one is better than the other. I like that we have
different organizations so that we can pursue both the preservation
goal and the operational goal.
My personal interpretation is that we make available
all of the
documentation and software that can be located, which is of the most
use to people not in the San Fransisco Bay area, and not use
CHM's space and other resources for restorations. This view is
shared by the other curators at the museum.
I have no problem with that either.
We had around 65,000 visitors last year. With that
many members
of the general public in an exhibit, theft and vandalism becomes
a real concern. Along with that, many people feeling the need to
touch artifacts, climb on platforms, and try to do things like open
equipment drawers.
Yes, proper treatment of the equipment is something that concerns me
in making interactive hands-on exhibits. Ultimately, it may boil down
to making sacrificial replicas for which no harm is ultimately done
should vandalism and monkeying about be a concern. I think my museum
will be small enough for a while that this won't be an immediate
concern (i.e. with 3 seats and an exhibit that's not being used
without someone overseeing it's use).
However, this is something that will most likely need to be addressed
in some fashion at some point. Remember, the cages in zoos protect
not only the visitors from the animals, but also protect the animals
from the visitors.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 version available for download
<http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/>
Legalize Adulthood! <http://legalizeadulthood.wordpress.com>
2:36 a.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
We had around 65,000 visitors last year. With that
many members of the
general public
in an exhibit, theft and vandalism becomes a real concern.
Not just a concern, but a problem. The nautical museums I am involved
with have minor artifacts disappear all the time. One time, during a
Boy Scout overnight on Battleship MASSACHUSETTS, I was working in the
normally padlocked radar shop - a closet sized space - and some little
shit STOLE the lock WHILE I was inside.
People do these crimes not to make money, but for the simple thrill. I
am surer that padlock was dumped overboard.
--
Will
3:16 a.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
On Jan 20, 2012, at 9:36 PM, William Donzelli <wdonzelli at gmail.com> wrote:
Oh shit, that was you?!
(kidding...KIDDING!)
-Dave
--
Dave McGuire
New Kensington, PA
We had around
65,000 visitors last year. With that many members of the
general public
in an exhibit, theft and vandalism becomes a real concern.
Not just a concern, but a problem. The nautical museums I am involved
with have minor artifacts disappear all the time. One time, during a
Boy Scout overnight on Battleship MASSACHUSETTS, I was working in the
normally padlocked radar shop - a closet sized space - and some little
shit STOLE the lock WHILE I was inside.
People do these crimes not to make money, but for the simple thrill. I
am surer that padlock was dumped overboard. 
8:23 a.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
Im involved with the preservation of a working stationary steam engine
We run about 6 times a year and we replace wear parts as would have
been done throughout its working life.
http://www.middleton-leawood.org.uk/leawood/
I take guidance from 5.8 below and the fact that the maintenance
required to do that means we inspect it more so can deal with its
condition and any problems as they arise etc.
See section 5 of
http://www.collectionslink.org.uk/discover/site-information/487-standards-i…
5.7 Demonstrating an object in motion can add immeasurably to our
understanding of its
purpose and significance; riding on a transport exhibit designed
to carry passengers not
only shows us how it was made and controlled, but demonstrates
conditions of life for
those who worked on it or used it. A visitor who has never seen
textile machinery at work
will have difficulty understanding a static loom, let alone the
working conditions of
Victorian mill-hands. Running an early computer gives an unique
insight into early
computing practices.
5.8 Operating an object may also actually contribute to its
preservation, eg through distributing
lubricants or varying stress points.
5.9 Operating an object may also help to preserve or rediscover
appropriate skills.
5.10 However, operating most mechanical objects will inevitably result
in their wearing out. At
the very least surfaces will be worn and information thus be
lost; at the worst the object
will end up like the proverbial Grandfather's Axe, with little or
nothing preserved from the
days before it entered the museum.
Dave Caroline
7:41 p.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
[..]
I am not involved with a museum. but I do try to preserve the computers
(and other objects) that I own. However, I do keep them operational, I do
run them from time to time The reason is essentially your '5.7' -- A
computer was designed to run programs, that's what makes it ineresting.
It is not a piece of fine art, to be used for its appearance.
5.7 Demonstrating an object in motion can add
immeasurably to our
understanding of its
purpose and significance; riding on a transport exhibit designed
to carry passengers not
only shows us how it was made and controlled, but demonstrates
conditions of life for
those who worked on it or used it. A visitor who has never seen
textile machinery at work
will have difficulty understanding a static loom, let alone the
working conditions of
Victorian mill-hands. Running an early computer gives an unique
insight into early
computing practices.
5.8 Operating an object may also actually contribute to its
preservation, eg through distributing
lubricants or varying stress points.
5.9 Operating an object may also help to preserve or rediscover
appropriate skills.
5.10 However, operating most mechanical objects will inevitably result
in their wearing out. At
the very least surfaces will be worn and information thus be
lost; at the worst the object
This depends o nthe object IMHO. Let's take an example I can see from
where I am sitting -- the HP9810 calculator. If I run the card reader in
that then the motor brushes will wear. But what is lost if that happens?
It is obvious to anyone whp/s ever seen a DC permanent magnet motor that
there should be brushes on the spring metal arms. OK, the lenght of the
brush that was there when I got the amchine is lost _but since I didn't
have the machjine from new, that's not the length of a new brush_. It's
the length of a brush after it's been used to read an unknown number of
cards. So I don't think that's particularly improtant.
Of course it's entirey reasonable to measure and record information
about the machine when you got it -- diameters of spindles, thickness of
motor brushes, resistance of a themral printhed (this increases as the
head wears). inductance of magnetic heads (which will drop as the head
gap wears), and so on.
will end up like the proverbial Grandfather's
Axe, with little or
nothing preserved from the
days before it entered the museum.
I think that's unliekly. In most machines, and certainly in computers,
there are parts that are not going to wear out or fail just becuase the
machine is used. For example, it's very unlikely a computer PCB would
ever be changed in a museum exhibit. The actual board will be the one
that was in the machine when you got it. Components _on_ the board
might be changed, but not the board itself. Similarly case parts are
lilely to remain original.
As I said I run my machines. I don't run them 25/7 of course, but I do
use them. And I would estimate I've changed mush less that 1% of the
components in any machine I own.
-tony
7:52 p.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
On Fri, 20 Jan 2012, William Donzelli wrote:
normally padlocked radar shop - a closet sized space -
and some little
shit STOLE the lock WHILE I was inside.
You learned to never leave the lock hanging from the shackle!
Be glad that he didn't LOCK it, instead!
9:01 p.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
On 1/21/12 11:52 AM, "Fred Cisin" <cisin at xenosoft.com> wrote:
On Fri, 20 Jan 2012, William Donzelli wrote:
I wanted to check out a tablet at a local Worst Buy and its screen was
locked with a passcode. An employee unlocked it so I could play with it,
telling me that if they didn't lock them with a *known* passcode, random
sh*ts (to use your technical term) would set the passcode just to be
obnoxious. I've never quite understood vandalism of any sort?. -- Ian
normally padlocked radar shop - a closet sized
space - and some little
shit STOLE the lock WHILE I was inside.
You learned to never leave the lock hanging from the shackle!
Be glad that he didn't LOCK it, instead!
20 Jan
20 Jan
8:38 p.m.
New subject: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
From: Al Kossow
Sent: Friday, January 20, 2012 8:38 AM
On 1/19/12 5:58 PM, Rich Alderson wrote:
> The point is that we *run* our systems, unlike the
PDP-1 at CHM.
What exactly was the point of posting this bit of
disinformation?
My apologies, Al, if that came off a little more strident than I
intended it to. I'll plead cabin fever (I've been at home due to snow
storms in a region with no snow plans even though we're 7.5 degrees
north of Chicago), exacerbated by the power outage that woke me up when
my CPAP stopped working night before last and which had me running my
emergency generator until after 12:30AM this morning.
In the context of the discussion (which had gone on to the perennial
topic of replacing vs. "reforming" capacitors), our desire to keep those
systems which it makes sense to run 24/7 up and running means that we do
not use what we consider, based on experience, to be half measures.
No, it may not make sense to run the PDP-1 24/7, but the fact that a
potential visitor must know when the PDP-1 *will* run, rather than being
able to see whenever she has an opportunity to go to to CHM, makes for a
different experience. We have similar systems in our much smaller
collection, which we are happy to turn on for any visitor any time, if
they have been made runnable at all.
That's all that was meant by the telegraphic statement you quoted.
Rich Alderson
Vintage Computing Sr. Systems Engineer
Vulcan, Inc.
505 5th Avenue S, Suite 900
Seattle, WA 98104
mailto:RichA at vulcan.com
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
21 Jan
21 Jan
8:46 a.m.
New subject: OT: CPAP, was Re: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
On 01/20/2012 03:38 PM, Rich Alderson wrote:
Eeeeeek! I am on CPAP as well; power failures are VERY
disconcerting. I run mine on a small UPS now. I strongly recommend it.
In the event of a power failure, the beeping of the UPS will likely
wake you up, but without that oh-so-wonderful suffocating feeling.
-Dave
--
Dave McGuire
New Kensington, PA
What exactly
was the point of posting this bit of disinformation?
My apologies, Al, if that came off a little more strident than I
intended it to. I'll plead cabin fever (I've been at home due to snow
storms in a region with no snow plans even though we're 7.5 degrees
north of Chicago), exacerbated by the power outage that woke me up when
my CPAP stopped working night before last and which had me running my
emergency generator until after 12:30AM this morning. 
11:11 a.m.
New subject: CPAP, was Re: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
Eeeeeek! I am on CPAP as well; power failures are
VERY
disconcerting. I run mine on a small UPS now. I strongly recommend it.
In the event of a power failure, the beeping of the UPS will likely
wake you up, but without that oh-so-wonderful suffocating feeling.
Shoudn't these toys be battery-backed from factory? :oO
12:17 p.m.
New subject: CPAP, was Re: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
I've never seen one that is. I have two bipap machines, one which runs
a compression fan continuously, and one when you breath. Neither would
probably run long on batteries, they use a lot of power.
the newer machines are a lot like my second machine, but still probably
draw enough to need a pretty substantial supply to run for very long.
Jim
On 1/21/2012 3:11 AM, Alexandre Souza - Listas wrote:
Eeeeeek! I
am on CPAP as well; power failures are VERY
disconcerting. I run mine on a small UPS now. I strongly recommend
it. In the event of a power failure, the beeping of the UPS will
likely wake you up, but without that oh-so-wonderful suffocating
feeling.
Shoudn't these toys be battery-backed from factory? :oO
23 Jan
23 Jan
7:27 p.m.
New subject: CPAP, was Re: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
From: jim s
Sent: Saturday, January 21, 2012 4:17 AM
On 1/21/2012 3:11 AM, Alexandre Souza - Listas wrote:
[quoting Dave McGuire --rma]
>> Eeeeeek! I am on CPAP as well; power
failures are VERY
>> disconcerting. I run mine on a small UPS now. I strongly recommend
>> it. In the event of a power failure, the beeping of the UPS will
>> likely wake you up, but without that oh-so-wonderful suffocating
>> feeling.
> Shoudn't these toys be battery-backed from
factory? :oO
I've never seen one that is. I have two bipap
machines, one which runs
a compression fan continuously, and one when you breath. Neither would
probably run long on batteries, they use a lot of power.
the newer machines are a lot like my second machine,
but still probably
draw enough to need a pretty substantial supply to run for very long.
A good friend of mine built his own UPS for his CPAP using a deep-cycle
marine battery. Plenty of power, long run time on one battery, and he
keeps a spare battery for longer outages. He also has a generator at
his house, but this means he doesn't have to get up in the middle of the
night to kick it into life.
My wife has a blind spot WRT these things: "Why don't you just sleep
without it for a night?" She doesn't understand that breathing is worse
than rock cocaine as an addiction: Since I became able to sleep through
the night, I can't go back. :-)
Rich
7:51 p.m.
New subject: CPAP, was Re: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
On Jan 23, 2012, at 2:27 PM, Rich Alderson wrote:
My wife has a blind spot WRT these things: "Why
don't you just sleep
without it for a night?" She doesn't understand that breathing is worse
than rock cocaine as an addiction: Since I became able to sleep through
the night, I can't go back. :-)
Well, also, the withdrawal symptoms include death by asphyxiation, so... there's
that.
- Dave
8:25 p.m.
New subject: CPAP, was Re: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
On 01/23/2012 02:51 PM, David Riley wrote:
On Jan 23, 2012, at 2:27 PM, Rich Alderson wrote:
Perhaps Rich has a big life insurance policy. ;)
O2 is good to have while one is sleeping.
-Dave
--
Dave McGuire
New Kensington, PA
My wife has a blind spot WRT these things:
"Why don't you just sleep
without it for a night?" She doesn't understand that breathing is worse
than rock cocaine as an addiction: Since I became able to sleep through
the night, I can't go back. :-)
Well, also, the withdrawal symptoms include death by asphyxiation, so... there's
that. 
10:29 p.m.
New subject: CPAP, was Re: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
On 1/23/12 1:51 PM, David Riley wrote:
On Jan 23, 2012, at 2:27 PM, Rich Alderson wrote:
A couple of my wives would have called that a feature, not a bug...
Doc
My wife has a blind spot WRT these things:
"Why don't you just sleep
without it for a night?" She doesn't understand that breathing is worse
than rock cocaine as an addiction: Since I became able to sleep through
the night, I can't go back. :-)
Well, also, the withdrawal symptoms include death by asphyxiation, so... there's
that. 
21 Jan
21 Jan
5:40 p.m.
New subject: OT: CPAP, was Re: CHM's PDP-1 (was Re: PDP-8/I at the RICM)
Eeeeeek! I am on CPAP as well; power failures are
VERY
disconcerting. I run mine on a small UPS now. I strongly recommend it.
In the event of a power failure, the beeping of the UPS will likely
wake you up, but without that oh-so-wonderful suffocating feeling.
I've certainly written letters in my day job capacity for patients with CPAP
and oxygen concentrators to give them priority in an outage, also. Depending
on your state regulations the power company will honour these from a
physician or other qualified medical professional.
--
------------------------------------ personal: http://www.cameronkaiser.com/ --
Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckaiser at floodgap.com
-- The only thing to fear is fearlessness -- R. E. M. -------------------------
20 Jan
20 Jan
9:47 p.m.
Ian was not at LCM yet when we did exactly this, after
reading the CDE
white paper. More than 95% of the aluminum electrolytic caps in the 2065
had failed. We first tried "re-forming" a few. They failed again in
days.
I was under the impression tha the oxide dielectric film was originally
produced by a process that's much the same as the reforming process. So I
wonder why the reformed capactiros didn't last. Was theres somethign else
wrogn with them?
At that point we made the decision never to waste our time on this. We
buy new capacitors (or higher-level components; cf. our PDP-7) and put
them in.
AS somebody who. like you, wants to run their classic computers, I must
admit that I also replace defective capacitors. I use the neares new
equivalents that I can find (same or a little hight capacitance, same or
a bit hight working voltage). The appearance foesn't matter too much to
me. I simply think 'what would have been don e if this capacitor had
failed when the machine was in 'serious' use' -- and the answer is 'It
would have been replaced by a suitable new part'.
Perhaps I've been lucky, but I've not found that electrolytiics are the
problem that soem make out. As I said the other doay, I've replaced
perhaps 10 of them in all my old machines.
-tony
21 Jan
21 Jan
7 p.m.
On 01/20/2012 01:47 PM, Tony Duell wrote:
I was under the impression tha the oxide dielectric
film was
originally produced by a process that's much the same as the reforming
process.
It is.
So I wonder why the reformed capactiros didn't
last. Was theres
somethign else wrogn with them?
I would think so.
If you do the reforming by simply throwing the capacitor on a
current-limited supply and not measuring anything (as seems to be the
usual method), then a completely defective capacitor may improve a
little bit but still be fundamentally broken, and still fail either
immediately or shortly after being put into service.
The process we used on the PDP-1 project was to start with the power
supply set to 0.5V, watch the current until it dropped below specified
leakage, ramp to 1.0V, watch the leakage, and so on up to the rated
voltage. At each step the current limit of the power supply is set such
that the product of the voltage and the current limit were less than the
maximum rated power dissipation of the capacitor. (I think there may
have been an additional constraint to avoid having unreasonably high
current limit settings at low voltages, but I don't recall the details
of that.) Due to the current limit, for a capacitor that needs
reformation the power supply will be in current limiting, with the
voltage slowly ramping to the voltage setting as the reformation
proceeds. If the process takes more than a few hours at a voltage step,
the capacitor is probably not going to reform, and should be replaced.
If you don't do at least some monitoring of the process (not necessarily
as thorough as we did), you won't really know whether the reformation is
successful.
In the PDP-1 restoration, we encountered a few capacitors that had
obviously gone bad and did not successfully complete the restoration
process. Most of the capacitors completed the restoration process, and
at the completion, performed well within the factory specifications.
Although there are other wearout mechanisms, we expect that most of the
reformed capacitors will enjoy a reasonable remaining service life.
Eric
7:46 p.m.
-----Original Message-----
From: cctalk-bounces at classiccmp.org
[mailto:cctalk-bounces at classiccmp.org] On Behalf Of Eric Smith
Sent: 21 January 2012 19:01
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: PDP-8/I at the RICM
On 01/20/2012 01:47 PM, Tony Duell wrote:
This really suprises me. I guess that the caps Radio Hams use may run at
higher voltages than those in vintage PSUs but the process has as I was
taught it was always to ramp up the voltage whilst checking the leakage.
Some interesting circuits here:-
http://www.vmars.org.uk/capacitor_reforming.htm
I was under the impression tha the oxide
dielectric film was
originally produced by a process that's much the same as
the reforming
process.
It is.
So I wonder why the reformed capactiros
didn't last. Was theres
somethign else wrogn with them?
I would think so.
If you do the reforming by simply throwing the capacitor on a
current-limited supply and not measuring anything (as seems to be the
usual method), then a completely defective capacitor may improve a
little bit but still be fundamentally broken, and still fail either
immediately or shortly after being put into service.
The process we used on the PDP-1 project was to start
with the power
supply set to 0.5V, watch the current until it dropped below
specified
leakage, ramp to 1.0V, watch the leakage, and so on up to the rated
voltage. At each step the current limit of the power supply
is set such
that the product of the voltage and the current limit were
less than the
maximum rated power dissipation of the capacitor. (I think there may
have been an additional constraint to avoid having unreasonably high
current limit settings at low voltages, but I don't recall
the details
of that.) Due to the current limit, for a capacitor that needs
reformation the power supply will be in current limiting, with the
voltage slowly ramping to the voltage setting as the reformation
proceeds. If the process takes more than a few hours at a
voltage step,
the capacitor is probably not going to reform, and should be replaced.
If you don't do at least some monitoring of the process (not
necessarily
as thorough as we did), you won't really know whether the
reformation is
successful.
In the PDP-1 restoration, we encountered a few capacitors that had
obviously gone bad and did not successfully complete the restoration
process. Most of the capacitors completed the restoration
process, and
at the completion, performed well within the factory specifications.
Although there are other wearout mechanisms, we expect that
most of the
reformed capacitors will enjoy a reasonable remaining service life.
Eric
9:33 p.m.
On 01/21/2012 02:46 PM, Dave wrote:
For ham radio, as with anything else, it depends completely on the
specific application. A B+ (plate) power supply filter capacitor for a
2KW HF amplifier can easily be 3KV.
-Dave
--
Dave McGuire
New Kensington, PA
If you do the
reforming by simply throwing the capacitor on a
current-limited supply and not measuring anything (as seems to be the
usual method), then a completely defective capacitor may improve a
little bit but still be fundamentally broken, and still fail either
immediately or shortly after being put into service.
This really suprises me. I guess that the caps Radio Hams use may run at
higher voltages than those in vintage PSUs but the process has as I was
taught it was always to ramp up the voltage whilst checking the leakage.
7:51 p.m.
On 01/20/2012 01:47 PM, Tony Duell wrote:
I would seem that when the 'leakage current' gets very low then no more
oxide is goign to be depositied - -the film will not get any thicker.
After all, it's an electrolytic process.
There may be some trick the voltage profile (and/or controlling the
voltage as a fucntion of the leakage current). I susepct the official way
to do this is a trade secret, but there may be some details avaiable for
the older processes.
-tony
I was under the impression tha the oxide
dielectric film was
originally produced by a process that's much the same as the reforming
process.
It is.
19 Jan
19 Jan
7 a.m.
Zitat von Ian King <IanK at vulcan.com>:
We've restored a 2065 that runs 24x7, and that's where we learned that you
can cross your fingers and pray, or recognize that even when they're
sitting on a shelf aluminum electrolytic capacitors (AlECs) degrade with
time.
I was involved in a project where we had to guarantee long term
availability of some
electronic circuits. At the end of it, we decided not to populate the
electrolytics and
store them on the shelf. There was an adpater build to
rejuvenate/reform them every few
years and test before being used.
Just storing them on a shelf is the worst thing you can do to them ;-)
4:12 a.m.
On 1/18/2012 7:21 PM, Eric Smith wrote:
William Donzelli wrote:
It seems to me that you'd need to understand the failure
mechanism to build a monitor for it.
I just had a big Heathkit Audio Amp loose both of the big filter
capacitors... I'm not sure I understand the sequence of events.
It's obvious that the rectifier diodes were hot - the board is a
little discolored - but the diodes are still good. Both big
capacitors are now fully open. In fact the 'innards' rattle
inside the capacitor can.
It is as-if the capacitors initially shorted for a while,
and then opened... I'm not sure why both capacitors failed
at the same time - one is on the positive voltage, the other
is on the negative voltage.
What exactly is the physics of failure for big capacitors?
Rob.
Has there been any thought about installing some
sort of non-invasive,
reversible monitoring circuitry to the PDP-1
I proposed that, but I don't think the idea got any serious consideration.
(or other operating artifacts)?
I may be involved in an effort to get one or two DECSYSTEM-2065 (KL10)
systems running in the near future, and I'm seriously considering adding
quite a few voltage/current/temperature monitors.
5:13 a.m.
On Jan 18, 2012, at 23:12, Rob Doyle <radioengr at gmail.com> wrote:
It seems to me that you'd need to understand the
failure
mechanism to build a monitor for it.
In my experience, they tend to lose capacity with age, eventually becoming an open circuit
as the electrolyte dries out. This is often reflected as higher ripple.
It is as-if the capacitors initially shorted for a
while,
and then opened... I'm not sure why both capacitors failed
at the same time - one is on the positive voltage, the other
is on the negative voltage.
Could that not reflect a higher peak current due to high ripple? Or possibly inadequate
thermal design to begin with?
- Dave
5:50 a.m.
Rob Doyle wrote:
It seems to me that you'd need to understand the
failure
mechanism to build a monitor for it.
You don't need to understand the failure mechanism of a component to
monitor for the failure. You only need to understand the observable
behaviour of a failing component.
Nevertheless, the failure mechanisms of aluminum electrolytic capacitors
are well understood. Three of the most common are:
1) oxide breakdown - when not under bias for an extended period, the
oxide breaks down, which reduces the safe working voltage of the
capacitor. At lower voltages, the capacitor appears to be more leaky
than it should be. This can be fixed by reforming the capacitor.
2) catastrophic breakdown - when the applied voltage exceeds the safe
working voltage (which may be reduced from factory spec due to oxide
breakdown), the capacitor will short out in a spectacular manner.
Non-repairable.
3) electrolyte drying out - this happens if the seals are bad, or if
the capacitor is overstressed and ruptures the seals. Non-repairable.
It is as-if the capacitors initially shorted for a
while,
and then opened...
Sounds like it had oxide breakdown, which then resulted in catastrophic
breakdown.
7:41 a.m.
On 1/18/12 9:50 PM, "Eric Smith" <eric at brouhaha.com> wrote:
Rob Doyle wrote:
I've cited this research before in this thread, and it states that the
common failure mode is that the equivalent series resistance (ESR)
increases, which contributes to other failure modes. "The wearout process
is rarely driven by
evaporation and escape of the electrolyte unless the safety vent is
compromised due to high leakage current and pressure buildup."
Overvoltage is not likely to be a factor in an existing system unless
there has been some other failure in the power supply, in which case you
have other things to worry about, too.
http://www.cde.com/tech/reliability.pdf
These are the people who *make* these things. I think they probably have
some insight here.
It seems to me that you'd need to understand
the failure
mechanism to build a monitor for it.
You don't need to understand the failure mechanism of a component to
monitor for the failure. You only need to understand the observable
behaviour of a failing component.
Nevertheless, the failure mechanisms of aluminum electrolytic capacitors
are well understood. Three of the most common are:
1) oxide breakdown - when not under bias for an extended period, the
oxide breaks down, which reduces the safe working voltage of the
capacitor. At lower voltages, the capacitor appears to be more leaky
than it should be. This can be fixed by reforming the capacitor.
2) catastrophic breakdown - when the applied voltage exceeds the safe
working voltage (which may be reduced from factory spec due to oxide
breakdown), the capacitor will short out in a spectacular manner.
Non-repairable.
3) electrolyte drying out - this happens if the seals are bad, or if
the capacitor is overstressed and ruptures the seals. Non-repairable.
It is as-if the capacitors initially shorted for
a while,
and then opened...
Sounds like it had oxide breakdown, which then resulted in catastrophic
breakdown.
9:44 a.m.
Ian King wrote:
I've cited this research before in this thread,
and it states that the
common failure mode is that the equivalent series resistance (ESR)
increases, which contributes to other failure modes. "The wearout
process is rarely driven by evaporation and escape of the electrolyte
unless the safety vent is compromised due to high leakage current and
pressure buildup." Overvoltage is not likely to be a factor in an
existing system unless there has been some other failure in the power
supply,
Actually I don't see anything in that paper that contradicts my claims
that the oxide breaks down when the capacitor is not used under bias for
a long interval, nor that such an oxide breakdown reduces the effective
working voltage of the capacitor, leading to failure when used at the
original rated voltage.
I've seen other papers from capacitor vendors, possibly including CDE,
that discuss this problem. I don't have those papers handy at the
moment, so I'll have to dig around for them.
10:06 a.m.
I wrote:
Actually I don't see anything in that paper that
contradicts my claims
that the oxide breaks down when the capacitor is not used under bias
for a long interval, nor that such an oxide breakdown reduces the
effective working voltage of the capacitor, leading to failure when
used at the original rated voltage.
I've seen other papers from capacitor vendors, possibly including CDE,
that discuss this problem. I don't have those papers handy at the
moment, so I'll have to dig around for them.
Here's a copy of the CDE application guide:
http://www.cde.com/catalogs/AEappGUIDE.pdf
They explain how the oxide formation is done during manufacture (p. 3)
and mention that there is an aging step that does some additional
forming (p. 4). They discuss field-reforming capacitors with high DC
leakage in the "Shelf Life" section (p. 16). They don't directly state
that the reformation process should ramp the voltage relatively slowly,
but that's one effect of their recommended 1K series resistor.
I'm pretty sure I saw some applications information from a capacitor
vendor that had more details on reformation, but I don't recall which
vendor and still can't find it.
Eric
1:02 p.m.
3) ?electrolyte drying out - this happens if the seals
are bad, or if the
capacitor is overstressed and ruptures the seals. Non-repairable.
Seals going bad are the curse of every capacitor, not just
electroytics, and is the failure mode that they "all" will reach.
Right now, we are seeing many of the capacitors of the 1950s starting
to go (Bumblebees, we are looking at you!), with even some of the
bulletproof designs starting to fail. In the past, mica capacitors
were overlooked in repair jobs, simply because they "never" went bad,
and could be trusted. Not any more.
Just a tiny amount of moisture or contaminant getting by any capacitor
seal can really wreck your whole day.
--
Will
6:40 a.m.
New subject: Restoring a DEC KL10 (was: PDP-8/I at the RICM)
In article <4F177E37.8030204 at brouhaha.com>,
Eric Smith <eric at brouhaha.com> writes:
I may be involved in an effort to get one or two
DECSYSTEM-2065 (KL10)
systems running in the near future, and I'm seriously considering adding
quite a few voltage/current/temperature monitors.
I'd advise you to consult with the folks at the Living Computer
Museum. IMO, they did a wonderful job of replacing the original power
supply in their machines with a modern efficient switching design. It
was all done in such a way that the upgrade was fully reversible to
restore the machine to it's original condition.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 version available for download
<http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/>
Legalize Adulthood! <http://legalizeadulthood.wordpress.com>
7:05 a.m.
New subject: Restoring a DEC KL10
Richard wrote:
I'd advise you to consult with the folks at the
Living Computer
Museum. IMO, they did a wonderful job of replacing the original power
supply in their machines with a modern efficient switching design.
My
understanding is that they replaced the entire ECL power supply
system with modern switchers.
I plan to keep the original power supply front end (three-phase to -12V
DC), and only replace the linear regulator modules (nine each -12V to
-5.2V at 35A, and four each -12V to -2.0V at 35A) with switchers. That
preserves (and actually improves upon) the AC power-fail functionality
of the system, allowing the processor state to be written to RAM, to
allow resumption of execution on power restoration. (Assuming that the
RAM is nonvolatile, which in the original machines was accomplished by
either core memory, or battery-backed DRAM.)
However, I plan to get the machine working with the original regulators
before I change anything.
It was all done in such a way that the upgrade was
fully reversible to
restore the machine to it's original condition.
Naturally I plan to do that
also.
20 Jan
20 Jan
2:14 a.m.
New subject: Restoring a DEC KL10
From: Eric Smith
Sent: Wednesday, January 18, 2012 11:06 PM
Richard wrote:
> I'd advise you to consult with the folks at
the Living Computer
> Museum. IMO, they did a wonderful job of replacing the original power
> supply in their machines with a modern efficient switching design.
My understanding is that they replaced the entire ECL
power supply
system with modern switchers.
On our 1095, yes. The 2065 has not undergone surgery quite as extensive
as that. We re-built the DEC power supplies with new capacitors [thread
tie]. The 1095 was the second effort, as an experiment.
I plan to keep the original power supply front end
(three-phase to -12V
DC), and only replace the linear regulator modules (nine each -12V to
-5.2V at 35A, and four each -12V to -2.0V at 35A) with switchers. That
preserves (and actually improves upon) the AC power-fail functionality
of the system, allowing the processor state to be written to RAM, to
allow resumption of execution on power restoration. (Assuming that the
RAM is nonvolatile, which in the original machines was accomplished by
either core memory, or battery-backed DRAM.)
However, I plan to get the machine working with the
original regulators
before I change anything.
> It was all done in such a way that the upgrade was
fully reversible to
> restore the machine to it's original condition.
Naturally I plan to do that also.
Well worth it.
Rich Alderson
Vintage Computing Sr. Systems Engineer
Vulcan, Inc.
505 5th Avenue S, Suite 900
Seattle, WA 98104
mailto:RichA at vulcan.com
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
19 Jan
19 Jan
1:08 p.m.
On Wed, Jan 18, 2012 at 05:34:02PM +0000, Ian King wrote:
I'd recommend finding paper tape images (Bitsavers, pdp8online.com) and
feeding them in through a terminal emulator. The DECtape controller is
itself a complex piece of equipment - minimize your variables! :-)
To save time I changed the timing capactior on a spare M452 and put
that in my 8/I and used a current loop to 232 adapter so I can quickly
download the diags from a PC. I aimed for 9600 baud but found with the
capacitor I picked I could only adjust it to 4800 baud. Also I found none
of the cheap USB to serial boards will do 110 baud, only real serial ports
can.
My 8/I tape images are here
http://www.pdp8online.com/ftp/software/paper_tapes/
4724
days inactive
4732
days old
56 comments
25 participants
participants (25)
-
aek＠bitsavers.org -
alexeyt＠freeshell.org -
ard＠p850ug1.demon.co.uk -
cisin＠xenosoft.com -
dave.g4ugm＠gmail.com -
dave.thearchivist＠gmail.com -
djg＠pdp8online.com -
doc＠vaxen.net -
emu＠e-bbes.com -
eric＠brouhaha.com -
evan＠snarc.net -
fraveydank＠gmail.com -
IanK＠vulcan.com -
jws＠jwsss.com -
legalize＠xmission.com -
madcrow.maxwell＠gmail.com -
mcguire＠neurotica.com -
michael.99.thompson＠gmail.com -
mouse＠Rodents-Montreal.ORG -
pu1bzz.listas＠gmail.com -
radioengr＠gmail.com -
RichA＠vulcan.com -
rodsmallwood＠btconnect.com -
spectre＠floodgap.com -
wdonzelli＠gmail.com