31 Mar
1997
31 Mar
'97
8:16 p.m.
At 06:01 PM 3/31/97 -0800, you wrote:
I was talking to a friend this weekend who brought
up this idea that before long EPROMs in our old
computers are going to start to go bad. This is
something I've heard before but, to be honest, I
don't know enough to make a judgement on it. If
this is the case, I suppose I should be burning
backups.
While I'm at it... What are some other concerns
along these lines. What should we be planning
for as these machines grow 5, 10, 15 years older?
Good questions all, so some comments and some thoughts.
EPROMS: yes, they will forget. Being stored charge devices they will
eventually start to lose the stored information. As I recall the original
estimated life for stored data was 10-15 years. Punched tape (mylar
preferably or paper if stored properly) may be the closest thing we have to
a true "archival" medium.
Electrolytic Capacitors: They dry out, or will start to ooze the
electrolyte. (I address this to some extent on the 'Repair Tips & Tricks'
section on my web pages) Average lifetime (of large caps such as found in
power supplies) is probably 5-10 years. On machines that I restore they
are (almost) always replaced before I do anything else. The very small
caps found on logic boards for timing and similar circuits do not seem to
be a prone to failure.
Disks: depends highly on how they are stored. Should be treated similar
to records. (remember those?) Store upright, loosely packed, dry but cool.
Watch for oxide flaking on very old disks.
Tapes: similar to disks for storage. Should be retensioned (repacked?)
from time to time. (fast forward to end, then rewind)
As with disks,
watch for oxide flaking. Dampness will cause some tapes to adhere
to
itself, and strip off the oxide when run. Had this happen to my "Standard
Pack" tape for an HP-85 I obtained recently. Went to load the tape and got
a read error. Manual inspection of the tape showed a 3/4 inch strip where
there was no oxide at all. it was stripped away by the tape adhesion when
I rewound the tape for probably the first time in 4-5 years!
Card edge connectors: corrosion is common, expecially is stored in damp
places or contaminated by skin oils. Can be cleaned with the classic "Pink
Pearl" eraser. (Don't use the grey erasers - too abrasive!)
Hmmm... probably more, but it's late and I forget... (often it seems - bad
EPROM?)
More later...
-jim
---
jimw(a)agora.rdrop.com
The Computer Garage - http://www.rdrop.com/~jimw
Computer Garage Fax - (503) 646-0174
1 Apr
1 Apr
6:52 a.m.
Electrolytic Capacitors: They dry out, or will start
to ooze the
electrolyte. (I address this to some extent on the 'Repair Tips & Tricks'
section on my web pages) Average lifetime (of large caps such as found in
power supplies) is probably 5-10 years. On machines that I restore they
are (almost) always replaced before I do anything else. The very small
caps found on logic boards for timing and similar circuits do not seem to
be a prone to failure.
I think the average lifetime may be longer for electrolytics, especially
those made in the past thirty years. I have had decent luck getting 50
year old capacitors alive again - but only if they are treated right. The
best way to keep them happy is to reform them, treat them gently, and use
them.
Most of the following probably is better suited to guys with minis, as
often they have not been powered for 10 years and stored in a damp
warehouse...
Always disconnect the power supply(s) and check them out first! For
non-switching supplies, disconnect the output (no load condition) and use
a variac (variable transformer) on the input. Slowly ramp up the variac,
starting from 0. For really old electronics, I do this over a 30 minutes,
but that might be me being over-cautious. An ammeter on the input is also
handy, in case something starts to go (moisture in transformers - but
that is a different rant).
For switching supplies, connect a decent load, as some units _need_ a
load or they will break. Still ramp up the input voltage, but do it
quickly - perhaps over just a few seconds (the supplies will complain
about the low input voltage). I suppose the best thing to do is to power
the capacitors out of circuit - probably too much work for electonics of
this vintage.
The object of all of this is to prevent the capacitors from getting
whacked with full voltage from the rectifier. The longer the ramp-up on
the input, the better, as it allows the electroyte in the capacitor to
reform (I do not know the exact chemical reactions that take place during
decay and reformation, but I could probably dig them up).
Finally, the best way to keep the capacitors from decaying is to use them
from time to time - just like a vintage car engine.
Eventually they will
wear out, but I think that would be _far_ longer than 10 or 15
years.
Tantalum capacitors are pretty bulletproof, and do not need this treatment.
Card edge connectors: corrosion is common, especially
is stored in damp
places or contaminated by skin oils. Can be cleaned with the classic "Pink
Pearl" eraser. (Don't use the grey erasers - too abrasive!)
I would not do this, unless it is followed by a cleaning with a solvent
of some sort. Erasers do leave a residue, mainly in the microscopic pits
in the metal.
William Donzelli
william(a)ans.net
6:59 p.m.
At 12:52 PM 4/1/97 -0500, you wrote:
True enough in many cases, but check them before proceeding...
Agreed. The part I seem to have omitted that is part of my regular routine
after cleaning with the eraser is to remove all residue with a soft
(non-static) cloth. I've tended to avoid solvents on PC boards when possible.
-jim
---
jimw(a)agora.rdrop.com
The Computer Garage - http://www.rdrop.com/~jimw
Computer Garage Fax - (503) 646-0174
Electrolytic
Capacitors: They dry out, or will start to ooze the
electrolyte. (I address this to some extent on the 'Repair Tips & Tricks'
section on my web pages) Average lifetime (of large caps such as found in
power supplies) is probably 5-10 years. On machines that I restore they
are (almost) always replaced before I do anything else. The very small
caps found on logic boards for timing and similar circuits do not seem to
be a prone to failure.
I think the average lifetime may be longer for electrolytics, especially
those made in the past thirty years. I have had decent luck getting 50
year old capacitors alive again - but only if they are treated right. The
best way to keep them happy is to reform them, treat them gently, and use
them. Tantalum capacitors are pretty bulletproof, and do not
need this treatment.
Yep!
Card edge
connectors: corrosion is common, especially is stored in damp
places or contaminated by skin oils. Can be cleaned with the classic "Pink
Pearl" eraser. (Don't use the grey erasers - too abrasive!)
I would not do this, unless it is followed by a cleaning with a solvent
of some sort. Erasers do leave a residue, mainly in the microscopic pits
in the metal.
2 Apr
2 Apr
10:49 a.m.
New subject: PDP-8 boards?
Can anyone enlighten me on the following?
(1) M8330
(1) M8349B K-SP-MR8FE-1 Rev F M1 1k prom board
(1) M8349B K-SP-MR8FE-1 Rev F M2 1k prom board
(1) M8349B K-SP-MR8FE-1 Rev F M3 1k prom board
(1) M8349B K-SP-MR8FE-1 Rev F M4 1k prom board
The prom boards have Intel 1702A eproms on them, but they are soldered in.
I am pretty sure the first is one of the processor cards from a PDP-8/e,
but I have no idea what the other boards are for.
The 1702As, as noted, are soldered in, which makes me think that
this was more than just a generic, burn-your-own-ROM card set.
Also, if the M8330 is surplus to me (I think I have quite a few base 8/e
card sets, save the bus load card), does anyone want it?
William Donzelli
william(a)ans.net
5:14 p.m.
New subject: PDP-8 boards?
William Donzelli wrote:
Can anyone enlighten me on the following?
Ok, a PDP-8 question. Which version used the Harris
IML6100 processor?
My first bench job was fixing Transaction II compatable
charge card machines based on this processor. I was told
that it was the PDP-8 proc, all 12 multiplexed bus bits
of it. Cmos processor, no bus drivers, driving straigt
into the edge card connector. Spent a lot of time cleaning
edge connectors. So much that I had to "re-plate" them with
the soldering iron when the copper started to show through
too much. But that was a VERY long time ago!
--
| Allen Underdown - wbrco(a)valuenet.net |
| Amateur Radio Operator - N0GOM, computer geek, |
| homebrewer and outdoor enthusiast! |
| Try My BBS at 314.939.9445! |
(1) M8330
(1) M8349B K-SP-MR8FE-1 Rev F M1 1k prom board
(1) M8349B K-SP-MR8FE-1 Rev F M2 1k prom board
(1) M8349B K-SP-MR8FE-1 Rev F M3 1k prom board
(1) M8349B K-SP-MR8FE-1 Rev F M4 1k prom board
The prom boards have Intel 1702A eproms on them, but they are soldered in.
I am pretty sure the first is one of the processor cards from a PDP-8/e,
but I have no idea what the other boards are for.
The 1702As, as noted, are soldered in, which makes me think that
this was more than just a generic, burn-your-own-ROM card set.
Also, if the M8330 is surplus to me (I think I have quite a few base 8/e
card sets, save the bus load card), does anyone want it?
William Donzelli
william(a)ans.net 
6 p.m.
New subject: PDP-8 boards?
Can anyone enlighten me on the following?
>(1) M8330
PDP8/e Clock Generator/Timing logic. Part of the CPU board set (along with
M8300 and M8310, which are the data path and control in some order) as you
guessed
>(1) M8349B K-SP-MR8FE-1 Rev F M1 1k prom
board
>(1) M8349B K-SP-MR8FE-1 Rev F M2 1k prom board
>(1) M8349B K-SP-MR8FE-1 Rev F M3 1k prom board
>(1) M8349B K-SP-MR8FE-1 Rev F M4 1k prom board
>
>The prom boards have Intel 1702A eproms on them, but they are soldered in.
They are 1K * 13 bits (!) of EPROM (7 off 1702s), + an optional 256 12 bit
words of RAM. The idea is that if the 13th bit of a word is clear, then
the processor sees the other 12 bits of that EPROM location. If the 13th
bit is set then 8 bits of the EPROM data select one of the RAM locations,
which can be read/written by the CPU. It basically gives you 1K of EPROM
with the ability to have some changeable locations in the middle of it.
I am pretty sure the first is one of the processor cards from a PDP-8/e,
but I have no idea what the other boards are for.
The 1702As, as noted, are soldered in, which makes me think that
this was more than just a generic, burn-your-own-ROM card set.
You'll notice there are _8_ edge plugs on these boards. The bottom 4 (away
from the EPROMs) go into a normal Omnibus slot. The top
4 take 4 little
shorting blocks that link the pins on the solder side to the pins on
the
component side. The pins on the component side are linked to the EPROMs,
those on the solder side to the rest of the logic. To program it, you pull
off the shorting blocks and link up a special programmer to the top
connectors.
AFAIK it was just a generic ROM board. The PDP8/e Maintenance Manual
Volume 2 covers it, but doesn't AFAIK mention any versions for specific
applications, like bootstraps.
On the other hand, the set of 4 boards you mentioned just might contain
some standard code.
William Donzelli
--
-tony
ard12(a)eng.cam.ac.uk
The gates in my computer are AND,OR and NOT, not Bill
6:20 a.m.
In article <199704011752.AA23897(a)interlock.ans.net>et>, William Donzelli
<william(a)ans.net> writes
Tantalum capacitors are pretty bulletproof, and do not
need this treatment.
No - we're using computer-assisted equipment from the 60's (74
type logic) where the tanatalum's are going short-circuit or leaky, so
allow for a 20-30 year lifetime.
--
Martin Evans at
Whitfield Street Studio, London
5 Apr
5 Apr
1:05 p.m.
New subject: Timonium/The goods are out there...
I attended the Timonium hamfest/computer fest today near Baltimore, and I
must admit to classics are still out there...
I saw many of the old standbys (Atari, Amiga, etc.), but the real treat
for me was the higher end stuff. Sun-3s are starting to come out in
adundance (I purchased an empty box for RCS/RI), even some older SPARCs
and an SGI, and plenty of other older workstations.
Also evident was a bit of old DEC stuff. I went away with quite a bit of
a PDP-11/20, and a small VAX 11/730 system.
I had to leave two PDP-11/44s and a custom DEC thingie (PDP-16?), plus a
Lisa 2, because my car was stuffed.
Was is my point (other than boasting about my latest find)?
If you do not go to the fests, GO!!!!! Most likely those PDP-11/44s are
furnace fodder, unless another collector gets them. The only way these
systems will be saved is if we collectors save them.
William Donzelli
william(a)ans.net
6 Apr
6 Apr
7:40 a.m.
New subject: Timonium/The goods are out there...
At 07:05 PM 4/5/97 -0500, William Donzelli wrote:
I attended the Timonium hamfest/computer fest today
near Baltimore, and I
must admit to classics are still out there...
i agree it was pretty wild. but you really do have to scrounge to find the
pre-PC era stuff (translation: there were TONS of 286/386s, PS/2s, etc.)
On the other hand if you wanted to put together the pieces of an "original"
PC or PC/XT it would be a cinch (PC video and other cards going for as
little as 10 cents; hard disks for 50 cents! cabinets for a buck or two).
on the pre-PC front there was still a fair amount of commodore stuff (e.g.
i picked up a $3 C/16 plus a box of Vic 20 stuff) but you have to shop for
price. i didn't see any "cool" microcomputer stuff like any S/100,
Heathkit or other early era systems or SBC's like the KIM, but if you go to
enough of these something's bound to turn up.
there were handouts there for upcoming ham/computer fests, e.g. Luzerne Co,
(PA) on 7/6, Frederick Co. (MD) on father's day (6/15), and Manassas (VA)
on 6/1. There doesn't appear to be a web site for any of these hamfests.
then again there's always the upcoming Trenton Computer Festival on 4/26
and 4/27:
http://www.tcf.net/index1.html
- glenn
+=========================================================+
| Glenn F. Roberts, Falls Church, VA
| Comments are my own and not the opinion of my employer
| groberts(a)mitre.org
11:22 p.m.
On Wed, 2 Apr 1997, Martin Evans wrote:
In article
<199704011752.AA23897(a)interlock.ans.net>et>, William Donzelli
<william(a)ans.net> writes
Maybe my memory is a bit vague - but tantalums in the 1960's ??????????
Jim
Tantalum capacitors are pretty bulletproof, and do
not need this treatment.
No - we're using computer-assisted equipment from the 60's (74
type logic) where the tanatalum's are going short-circuit or leaky, so
allow for a 20-30 year lifetime.
7 Apr
7 Apr
7:56 a.m.
> No - we're using computer-assisted
equipment from the 60's (74
> type logic) where the tanatalum's are going short-circuit or leaky, so
> allow for a 20-30 year lifetime.
It sounds like you just may have a bunch of cruddy capacitors (or a power
supply problem that is not killing the chips). I have never found a bad
tantalum capacitor.
Maybe my memory is a bit vague - but tantalums in the
1960's ??????????
They were around in the 1960s, but basically for military applications.
William Donzelli
william(a)ans.net
2:06 p.m.
> No - we're using
computer-assisted equipment from the 60's (74
> type logic) where the tanatalum's are going short-circuit or leaky, so
> allow for a 20-30 year lifetime.
It sounds like you just may have a bunch of cruddy capacitors (or a power
supply problem that is not killing the chips). I have never found a bad
tantalum capacitor. William Donzelli
--
-tony
ard12(a)eng.cam.ac.uk
The gates in my computer are AND,OR and NOT, not Bill
2 Apr
2 Apr
4:18 p.m.
I think the average lifetime may be longer for
electrolytics, especially
those made in the past thirty years. I have had decent luck getting 50
year old capacitors alive again - but only if they are treated right. The
best way to keep them happy is to reform them, treat them gently, and use
them.
The other big point is that (in general) computer capacitors are standard
parts. A 5000uF capacitor in a DEC PSU is replaceable by a standard 4700uF
one, avaialbe from a dozen suppliers in the UK alone. OK, so it may not
_look_ original, but the machine still works.
[...]
Always disconnect the power supply(s) and check them
out first! For
non-switching supplies, disconnect the output (no load condition) and use
Absolutely. This is not to protect the PSU, but to protect the rest of the
machine. An overvoltage on the 5V line can ruin _every_ chip in the
machine. It's at times like those that you are glad of the ROM backups :-)
Seriously, I _never_ power up a machine without first checking the PSU on
a dummy load. Only once was the supply faulty (and the 5V line sat at
about 8.5V), but none the less I do it. I seems sensible to make sure
you're not about to ruin a piece of history.
[...]
For switching supplies, connect a decent load, as some
units _need_ a
load or they will break. Still ramp up the input voltage, but do it
Light bulbs make good dummy loads. I was suprised that my DEC PC04 tape
reader uses a 12V car bulb running off the 5V line - and it's still quite
bright. I guess car headlamp bulbs would be ideal for SMPS loads. You can
get 6V car bulbs in the UK, BTW, but I had to order some when I wanted
them
In general you only need to load the main output - the one it takes
regulation from. This is likely to be a 5V output, but check the
schematics to be sure. For example, the PSU in my DEC LA324 printer has a
36V main output.
quickly - perhaps over just a few seconds (the
supplies will complain
about the low input voltage). I suppose the best thing to do is to power
the capacitors out of circuit - probably too much work for electonics of
this vintage.
It's not necessary to run up _every_ small capacitor. It may be worth
doing the big ones, particularly the mains smoothing caps in an SMPS.
Tantalum capacitors are pretty bulletproof, and do not
need this treatment.
Odd, I've had a couple of tants go short circuit with impressive results
(pop, smoke, small flame). No other damage, though.
Card edge connectors: corrosion is common,
especially is stored in damp
places or contaminated by skin oils. Can be cleaned with the classic "Pink
Pearl" eraser. (Don't use the grey erasers - too abrasive!)
I would not do this, unless it is followed by a cleaning with a solvent
of some sort. Erasers do leave a residue, mainly in the microscopic pits
in the metal. William Donzelli
--
-tony
ard12(a)eng.cam.ac.uk
The gates in my computer are AND,OR and NOT, not Bill
9968
days inactive
9974
days old
12 comments
7 participants
participants (7)
-
ard12＠eng.cam.ac.uk -
groberts＠mitre.org -
jimw＠agora.rdrop.com -
jj＠scs.leeds.ac.uk -
MARTIN＠paladio.demon.co.uk -
wbrco＠valuenet.net -
william＠ans.net