8 Nov
2014
8 Nov
'14
10:58 a.m.
From: Noel Chiappa
I also discovered and repaired at least two
cold-solder joints (one was
one of the pins for W1, the master clock control - the pin was actually
loose when I wiggled it).
So it turns out that loose wire-wrap pins seem to be a systemic problem on
11/23's. I just discovered _3_ more on one single 11/23 board - two of them
the pins for the master clock! I put my finger-tip on them and wiggled, and
they were loose.
I'm not sure why it's the wire-wrap pins especially - maybe their physical
size meant that they have larger thermal inertia, and so need a longer heat
than the manufacturing process gave them, to bring them up to temp to prevent
a cold solder joint?
Anyway, if one has a flaky 11/23, that's definitely something to check.
Noel
8 Nov
8 Nov
1:26 p.m.
On 08/11/2014 15:58, Noel Chiappa wrote:
So it turns out that loose wire-wrap pins seem to be a
systemic problem on
11/23's. I just discovered _3_ more on one single 11/23 board - two of them
the pins for the master clock! I put my finger-tip on them and wiggled, and
they were loose.
I've had dozens of those pass through my hands or at least across my
workbench and I don't recall ever seeing that. But given your
experience, it's clearly something to watch for nowadays. I wonder if
it's exacerbated by the wire-wrapping being done.
--
Pete
Pete Turnbull
4:22 p.m.
On 11/08/2014 12:26 PM, Pete Turnbull wrote:
On 08/11/2014 15:58, Noel Chiappa wrote:
I'm guessing the problem was that the wire wrap pin had
broken free of the solder
joint to the etched PC board? My guess is more along the
lines of frequent
board insertion/extraction fatigued the solder joint, as
pushing the board in/out
does wiggle the pin just a little. I haven't seen this
problem either on DEC
stuff, but I can imagine it happening. it might have been a
batch of backplane
that was not soldered quite properly, either temp profile
wrong or some kind
of contamination of the solder bath.
Jon
So it turns out that loose wire-wrap pins seem to
be a
systemic problem on
11/23's. I just discovered _3_ more on one single 11/23
board - two of them
the pins for the master clock! I put my finger-tip on
them and wiggled, and
they were loose.
I've had dozens of those pass through my hands or at least
across my workbench and I don't recall ever seeing that.
But given your experience, it's clearly something to watch
for nowadays. I wonder if it's exacerbated by the
wire-wrapping being done.
6:02 p.m.
On Sat, 8 Nov 2014, Jon Elson wrote:
On 11/08/2014 12:26 PM, Pete Turnbull wrote:
It would have been hand soldered. Wirewrap pins can't be wave soldered
since that would contaminate the gold or silver plated pins with solder
(the solder would dissolve and strips the gold or silver from the pin).
On 08/11/2014 15:58, Noel Chiappa wrote:
I'm guessing the problem was that the wire wrap pin had broken free of
the solder joint to the etched PC board? My guess is more along the
lines of frequent board insertion/extraction fatigued the solder joint,
as pushing the board in/out does wiggle the pin just a little. I
haven't seen this problem either on DEC stuff, but I can imagine it
happening. it might have been a batch of backplane that was not
soldered quite properly, either temp profile wrong or some kind of
contamination of the solder bath. So it turns out that loose wire-wrap pins seem to
be a systemic
problem on 11/23's. I just discovered _3_ more on one single 11/23
board - two of them the pins for the master clock! I put my finger-tip
on them and wiggled, and they were loose.
I've had dozens of those pass through my hands or at least across my
workbench and I don't recall ever seeing that. But given your
experience, it's clearly something to watch for nowadays. I wonder if
it's exacerbated by the wire-wrapping being done.
7:21 p.m.
On 08/11/2014 23:02, Tothwolf wrote:
It would have been hand soldered. Wirewrap pins
can't be wave soldered
since that would contaminate the gold or silver plated pins with solder
(the solder would dissolve and strips the gold or silver from the pin).
The "wirewrap" pins on the 11/23 and many other DEC boards are short
pins that are accessed on the /component/ side of the board and used for
jumper links, eg to select baud rates, or in the case of an 11/23 to set
the boot configuration and other things. See for example the images at
http://store.reuseum.com/DEC_M8186_F11_Fonz_CPU_Processor_Board_Q_Bus_p/000…
There's a group of pins with their jumpers to the right of the middle
40-pin IC. Look carefully and you'll see 5 (IIRC) other pairs,
including one near the top left of the crystal oscillator with a jumper
on it, and one with no jumper to the right of the smaller IC two rows
above the main jumper group. Other DEC boards such as DLV11J and the
notorious (because of gazillions of jumper options) MXV11B do actually
use wirewrap to such pins rather than jumpers, in a few cases 2-level wrap.
--
Pete
10:05 p.m.
On 11/08/2014 06:21 PM, Pete Turnbull wrote:
The "wirewrap" pins on the 11/23 and many other DEC boards
are short pins that are accessed on the /component/ side
of the board and used for jumper links, eg to select baud
rates, or in the case of an 11/23 to set the boot
configuration and other things.
OHH! THOSE pins! I had no idea you were referring
to the
pins on the individual boards!
Jon
9 Nov
9 Nov
4:33 a.m.
The "wirewrap" pins on the 11/23 and many
other DEC boards are short pins
that are accessed on the /component/ side of the board and used for jumper
links, eg to select baud rates, or in the case of an 11/23 to set the boot
configuration and other things. See for example the images at
http://store.reuseum.com/DEC_M8186_F11_Fonz_CPU_Processor_Board_
Q_Bus_p/000021514002g14store.htm
There's a group of pins with their jumpers to the right of the middle
40-pin
IC.
On that pic those links look like a DIL socket, not wire wrap pins. But if
wire wrap pins are used like that its not surprising they work loose.
I would expect them to be plated steel and therefore have characteristics
that are bad for soldering. That is:-
1. Have a relatively large mass as they need to be stiff to wire wrap
2. Therefore have a large thermal mass or heat capacity
3. So they are slow to heat up -> may not get hot enough to solder
4. Slow to cool -> likely hood of dry joints if they move while cooling
5. Poor solder whetting as they can be hard to clean....
So not surprising they sometimes work loose.
Look carefully and you'll see 5 (IIRC) other
pairs, including one near the
top
left of the crystal oscillator with a jumper on it,
and one with no jumper
to
the right of the smaller IC two rows above the main
jumper group. Other
DEC boards such as DLV11J and the notorious (because of gazillions of
jumper
options) MXV11B do actually use wirewrap to such pins
rather than jumpers,
in a few cases 2-level wrap.
--
Pete
8:19 a.m.
On 09/11/2014 09:33, Dave G4UGM wrote:
On that pic those links look like a DIL socket, not
wire wrap pins.
I suppose they do look a bit like a socket in that shot, but they are
wirewrap -- or more accurately, jumper -- pins, though. And normally
pretty reliable, as much as any other jumper pins. They're quite short,
about the size of the jumper pins you'd see on older 3.5" or 5.25" disk
drives or a little shorter than Molex KK pins, much shorter than than
ordinary wirewrap pins, and don't have much thermal inertia.
--
Pete
Pete Turnbull
8 Nov
8 Nov
10:04 p.m.
On 11/08/2014 05:02 PM, Tothwolf wrote:
It would have been hand soldered. Wirewrap pins can't be
wave soldered since that would contaminate the gold or
silver plated pins with solder (the solder would dissolve
and strips the gold or silver from the pin).
No, I don't think so. I think the PDP-11 backplanes were
mass-soldered in some way.
I have some real DEC and also some clone backplanes. The
pins appear to be
pure tin plated, and the soldering is such that there's no
way it was done by
hand. Possibly some very early backplanes might have been
hand soldered.
The early ones had no PC board, just the Winchester
connectors and wire-wrap
wires. I've NEVER seen a DEC backplane with gold or
silver-plated pins.
Jon
6:04 p.m.
On 08/11/2014 21:22, Jon Elson wrote:
On 11/08/2014 12:26 PM, Pete Turnbull wrote:
> On 08/11/2014 15:58, Noel Chiappa wrote:
>
>> So it turns out that loose wire-wrap pins seem to be a systemic
>> problem on 11/23's.
> I wonder if it's exacerbated by the
wire-wrapping being done.
I'm guessing the problem was that the wire wrap
pin had broken free
of the solder joint to the etched PC board? My guess is more along
the lines of frequent board insertion/extraction fatigued the solder
joint, as pushing the board in/out does wiggle the pin just a
little. I haven't seen this problem either on DEC stuff, but I can
imagine it happening. it might have been a batch of backplane that
was not soldered quite properly, either temp profile wrong or some
kind of contamination of the solder bath.
I didn't think Noel was referring to backplane pins - they're not
wirewrap pins nor do they carry the master clock - but to the various
jumper pins on the 11/23 PCB itself. There's a pair there to disconnect
the master clock for factory testing. Noel?
--
Pete
3543
days inactive
3544
days old
9 comments
5 participants
participants (5)
-
dave.g4ugm＠gmail.com -
elson＠pico-systems.com -
jnc＠mercury.lcs.mit.edu -
pete＠dunnington.plus.com -
tothwolf＠concentric.net