3B1 parts needed

[Philip Pemberton]

Hi there,

A few of you might be aware of the 3B1 emulator I released a while back.
A small fraction of you might also be aware of the issues I've had
(namely, it boots Diagnostics but booting the kernel throws up a ton of
"out of space" errors during the software install).

Through a contact in the USA, I've managed to acquire a 3B1 - it appears
to be a P5.1 unit with two 67MB Miniscribe hard drives (does this make it
one of the fabled "P6" units I've seen banded around here?).

Sadly, it's suffered quite badly in transit due to bad packing (a few
handfuls of foam peanuts which I found in the bottom of the box...
squashed flat).

* The keyboard is missing some keycaps: these appear to have broken off
(the white pegs are still in the switches, the caps are long gone) and
fallen out of a hole in the box... I'm either after some new keycaps from
a dead keyboard or another keyboard. Keycaps missing:
F8
HELP/up/dn (top right of the keyboard)
NEXT/3
ROLL DOWN/0
RIGHT ARROW / .

The top cover of the keyboard also has a broken screwpost... I'm not too
worried about this but it'd be nice to fix it for the sake of
completeness.

* The monitor tilt/swivel base is utterly destroyed. Smashed to bits. One
of these would be very nice to have! Sadly I think this means "entire
lower monitor casing and upper system unit case"...

* The power supply connector (the one at the end of the ribbon cable from
the motherboard) is burned - black in places. Does anyone know what type
of connector it is, or a compatible alternative?

* Ideally I'd like to replace the mouse, the retaining latch for the
connector has broken off. This is of considerably lower priority :)


Thanks,
--
Phil.
phi...@philpem.me.uk
http://www.philpem.me.uk/

[DoN. Nichols]

On 2013-05-15, Philip Pemberton <phi...@philpem.me.uk> wrote:
> Hi there,
>
> A few of you might be aware of the 3B1 emulator I released a while back.
> A small fraction of you might also be aware of the issues I've had
> (namely, it boots Diagnostics but booting the kernel throws up a ton of
> "out of space" errors during the software install).

Ouch! Maximum physical memory was 4 MB, and the same applies to
maximum virtual memory -- so you need at least 4 MB disk space allocated
to virtual memory.

> Through a contact in the USA, I've managed to acquire a 3B1 - it appears
> to be a P5.1 unit with two 67MB Miniscribe hard drives (does this make it
> one of the fabled "P6" units I've seen banded around here?).

I think just P5.1. Were both 65 MB Miniscribe drives mounted in
place of one full height one? In any case -- this is likely why the
power supply connector is so badly burned. When I did the modification,
I moved both drives to an external box, and had the 5V from the 3B1
switch on the power supply in the external box -- which was designed
for a 3B2's external drives, and thus had a similar color scheme.

> Sadly, it's suffered quite badly in transit due to bad packing (a few
> handfuls of foam peanuts which I found in the bottom of the box...
> squashed flat).

Ouch!

[ ... ]

> * The monitor tilt/swivel base is utterly destroyed. Smashed to bits. One
> of these would be very nice to have! Sadly I think this means "entire
> lower monitor casing and upper system unit case"...

Probably two separate pieces, and the parts needed to join them.
Otherwise, it could not tilt nor swivel.

> * The power supply connector (the one at the end of the ribbon cable from
> the motherboard) is burned - black in places. Does anyone know what type
> of connector it is, or a compatible alternative?

I found a couple of them at a hamfest (at the time, I remembered
the number of pins on the connector). What I mostly remember was the
orange plastic color, and the fact that the wires from the ribbon cable
connected by being pressed into slots in the top of the connector by a
narrow screwdriver blade. I never actually knew who made the connector.

It should be possible to put some alternative connector in place
of that -- both on the end of the cable, and on the power supply.
Actually, probably several smaller pin-count connectors to allow dealing
with the overly stiff ribbon cable and the constrained space. Hmm ...
perhaps one of those connectors used in industrial equipment like
machine tool controllers, where the connector solders down to the board,
and the individual wires are stuck into holes and tightened with clamp
screws.

In any case -- pull out the power supply and unscrew the board
to check the underside. I'll bet that the solder associated with the
burnt pins above it are at best now cold solder joints (grainy), and
likely solder even melted away from around the pins. Remove the old
solder with a solder sucker, and flow in fresh tin-lead electrical
solder, ideally 60/40 ratio or 63/37, so it will melt well and flow
well. *Don't* use the modern ROHS compliant lead-free solders. They
have problems, and also melt at a much higher temperature, thus damaging
the material of the printed circuit board.

> * Ideally I'd like to replace the mouse, the retaining latch for the
> connector has broken off. This is of considerably lower priority :)

Understood.

Besf of luck,
DoN.

--
Remove oil spill source from e-mail
Email: <BPdnic...@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

[Philip Pemberton]

On Thu, 16 May 2013 01:50:28 +0000, DoN. Nichols wrote:

> On 2013-05-15, Philip Pemberton <phi...@philpem.me.uk> wrote:
>> Hi there,
>>
>> A few of you might be aware of the 3B1 emulator I released a while
>> back. A small fraction of you might also be aware of the issues I've
>> had (namely, it boots Diagnostics but booting the kernel throws up a
>> ton of "out of space" errors during the software install).
>
> Ouch! Maximum physical memory was 4 MB, and the same applies to
> maximum virtual memory -- so you need at least 4 MB disk space allocated
> to virtual memory.

Hmm. In that case, I wonder if there's something wrong with the hard
drive controller emulation - possibly incorrectly reading or writing.
I'll have to check that...

> I think just P5.1. Were both 65 MB Miniscribe drives mounted in
> place of one full height one? In any case -- this is likely why the
> power supply connector is so badly burned.

Got it in one -- a previous owner appears to have fashioned four metal
strips which are drilled to hold the two drives together. The bottom
drive is then screwed into the shielding plate.

There's a Molex power connector (like a PC 5.25in one) running off the
PSU, which may well have been unused. There was a Y-splitter power cable
and Molex extension cable in the shipping box -- so I suspect a previous
owner had the HDDs running from an external power supply at some point.

It looks like I'll need the 3.51m install disks if I ever need to do a
reinstall... Sadly it looks like the ones on Bitsavers are standard
v3.51 :(

>> * The monitor tilt/swivel base is utterly destroyed. Smashed to bits.
>> One of these would be very nice to have! Sadly I think this means
>> "entire lower monitor casing and upper system unit case"...
>
> Probably two separate pieces, and the parts needed to join them.
> Otherwise, it could not tilt nor swivel.

Indeed. There's undoubtedly a way to detach the monitor from the base,
though I'm at a loss to figure out how...

>> * The power supply connector (the one at the end of the ribbon cable
>> from the motherboard) is burned - black in places. Does anyone know
>> what type of connector it is, or a compatible alternative?
>
> I found a couple of them at a hamfest (at the time, I remembered
> the number of pins on the connector). What I mostly remember was the
> orange plastic color, and the fact that the wires from the ribbon cable
> connected by being pressed into slots in the top of the connector by a
> narrow screwdriver blade. I never actually knew who made the connector.

It looks like they're ITW-Pancon MAS-CON connectors:

http://www.iobium.com/wpc_idc_connectors.htm
http://itwpancon.com/Products/mascon_index.htm

There's also a similar Molex KK connector, part number 09-06-5027:

http://www.molex.com/molex/products/datasheet.jsp?
part=active/0009065027_IDT_AND_SOLDER_CON.xml

Molex declare it as obsolete, but Farnell have a few left in stock.

After measuring with callipers, I'm left with two options for pin pitch:
3.96mm (0.156in) and 4mm. As 4mm isn't a valid option for a MAS-CON, it
must be 3.96mm :)

I also found a document from Thad Florian with the PSU pinout:

http://unixpc.taronga.com/documents/powsup.info

It's an 18-pin connector, pinout as follows:
Odd numbered pins (1,3,5,7,9,11,13,15,17) are ground
Pins 2,4,6,8,10 are +5V (5x5A = 25A max assuming even load spread)
Pins 12,14,16 are +12V (3x5A = 15A max assuming even load spread)
Pin 18 is -12V (300mA max limited by the PSU, 5A terminal limit)

> It should be possible to put some alternative connector in place
> of that -- both on the end of the cable, and on the power supply.

I'm looking into doing that - chances are I'll remove the ribbon cable
and replace it with a bunch of colour-coded high-current wires terminated
in a Molex 0.156mm KK connector. If I can get the phosphor-bronze KKs,
they're good to 7A max per contact. The brass ones are only good to 5A
per contact.

I'm looking at the Molex 26-48-1185 (KK 0.156in 18way), but Farnell don't
seem to stock the mating crimp housings for those (they have them up to
14way). The 18way + 0.156mm combination seems rather difficult to find on
this side of the pond :(

> perhaps one of those connectors used in industrial equipment like
> machine tool controllers, where the connector solders down to the board,
> and the individual wires are stuck into holes and tightened with clamp
> screws.

A terminal block? That's one option I suppose, but the ones I've seen
aren't available in 0.156 pitch.

> In any case -- pull out the power supply and unscrew the board
> to check the underside. I'll bet that the solder associated with the
> burnt pins above it are at best now cold solder joints (grainy), and
> likely solder even melted away from around the pins.

Yep, got it in one. The soldering is a mess. There's visible heat damage
around the connector too :(

> Remove the old
> solder with a solder sucker, and flow in fresh tin-lead electrical
> solder, ideally 60/40 ratio or 63/37, so it will melt well and flow
> well. *Don't* use the modern ROHS compliant lead-free solders. They
> have problems, and also melt at a much higher temperature, thus damaging
> the material of the printed circuit board.

Lucky me, I only use RoHS stuff if I have to. Most of the time, the Metcal
is set up for 60/40.

Thanks,
Phil.

[DoN. Nichols]

On 2013-05-16, Philip Pemberton <phi...@philpem.me.uk> wrote:
> On Thu, 16 May 2013 01:50:28 +0000, DoN. Nichols wrote:
>
>> On 2013-05-15, Philip Pemberton <phi...@philpem.me.uk> wrote:
>>> Hi there,
>>>
>>> A few of you might be aware of the 3B1 emulator I released a while
>>> back. A small fraction of you might also be aware of the issues I've
>>> had (namely, it boots Diagnostics but booting the kernel throws up a
>>> ton of "out of space" errors during the software install).
>>
>> Ouch! Maximum physical memory was 4 MB, and the same applies to
>> maximum virtual memory -- so you need at least 4 MB disk space allocated
>> to virtual memory.
>
> Hmm. In that case, I wonder if there's something wrong with the hard
> drive controller emulation - possibly incorrectly reading or writing.
> I'll have to check that...

You should be emulating a swap partition as well as the main
disk.

>> I think just P5.1. Were both 65 MB Miniscribe drives mounted in
>> place of one full height one? In any case -- this is likely why the
>> power supply connector is so badly burned.
>
> Got it in one -- a previous owner appears to have fashioned four metal
> strips which are drilled to hold the two drives together. The bottom
> drive is then screwed into the shielding plate.

Just what I would do if I wanted the disks to be in the box.
But I did not like that kind of load on the power supply, and while one
of my disks was the 67 MB MiniScribe, the other was a 190 MB one -- not
the Maxtor, but rather another brand -- Priam, IIRC -- and that was full
height, so two in the case would not work anyway.

And earlier experimentation showed that the magnetic fields from
the drive's servo motor caused the monitor image to twitch and shift,
unless you could get the original magnetic shield to fit -- which it
would not with the Priam drive.

> There's a Molex power connector (like a PC 5.25in one) running off the
> PSU, which may well have been unused.

That was for the full-height disk drive in the 3B1. The earlier
7300 machines, which did not have room for anything more than a
half-height drive.

> There was a Y-splitter power cable
> and Molex extension cable in the shipping box -- so I suspect a previous
> owner had the HDDs running from an external power supply at some point.

That -- or perhaps used it with the cable from the power supply
to power two drives -- until the PS self-immolated.

> It looks like I'll need the 3.51m install disks if I ever need to do a
> reinstall... Sadly it looks like the ones on Bitsavers are standard
> v3.51 :(

They used to be downloadable from the AT&T site, IIRC. And were
mirrored on other sites -- but many of those are now gone.

IIRC, the only thing different between 3.51 and 3.51m is the
kernel installed by the diagnostics disk. Do you have the enhanced
diagonstics disk? If not, send me an e-mail address at which you can
accept an attachment of about a 365k floppy image. Since this does not
have the encryption software, I can easily send it to you. (I could not
receive an attachment of that size, and would not send one unless you
told me to do so.

>>> * The monitor tilt/swivel base is utterly destroyed. Smashed to bits.
>>> One of these would be very nice to have! Sadly I think this means
>>> "entire lower monitor casing and upper system unit case"...
>>
>> Probably two separate pieces, and the parts needed to join them.
>> Otherwise, it could not tilt nor swivel.
>
> Indeed. There's undoubtedly a way to detach the monitor from the base,
> though I'm at a loss to figure out how...

Well -- first there are the screws to detach the back of the
monitor case (shell) from the monitor and fame. IIRC there is an edge
connector in there to connect the cable which feeds through the neck so
you can separate the monitor from the case.

Then the connection of the monitor shell to the top of the case
has a ball end, and a spring metal disc screwed to the bottom of the
case top. (This from long ago memory.)

>>> * The power supply connector (the one at the end of the ribbon cable
>>> from the motherboard) is burned - black in places. Does anyone know
>>> what type of connector it is, or a compatible alternative?
>>
>> I found a couple of them at a hamfest (at the time, I remembered
>> the number of pins on the connector). What I mostly remember was the
>> orange plastic color, and the fact that the wires from the ribbon cable
>> connected by being pressed into slots in the top of the connector by a
>> narrow screwdriver blade. I never actually knew who made the connector.
>
> It looks like they're ITW-Pancon MAS-CON connectors:
>
> http://www.iobium.com/wpc_idc_connectors.htm
> http://itwpancon.com/Products/mascon_index.htm

That sounds good, and looks like it.

> There's also a similar Molex KK connector, part number 09-06-5027:
>
> http://www.molex.com/molex/products/datasheet.jsp?
> part=active/0009065027_IDT_AND_SOLDER_CON.xml
>
> Molex declare it as obsolete, but Farnell have a few left in stock.

Similar Molex connectors were used as the bus connectors on the
SWTP 6800 and 6809 computers, and were very unreliable over time -- the
pins would oxidize, and I would have to slide the cards up and down
several times on the pins to make it work again.

> After measuring with callipers, I'm left with two options for pin pitch:
> 3.96mm (0.156in) and 4mm. As 4mm isn't a valid option for a MAS-CON, it
> must be 3.96mm :)

Digital, Dial, or Vernier calipers? But I agree, the 3.96mm
sounds much more likely. Over 18 pins (17 inter-pin spaces) that adds
up to 0.644mm, or 0.0254", which might work anway as those connectors
and the pins are at least that flexible.

> I also found a document from Thad Florian with the PSU pinout:
>
> http://unixpc.taronga.com/documents/powsup.info
>
> It's an 18-pin connector, pinout as follows:
> Odd numbered pins (1,3,5,7,9,11,13,15,17) are ground
> Pins 2,4,6,8,10 are +5V (5x5A = 25A max assuming even load spread)
> Pins 12,14,16 are +12V (3x5A = 15A max assuming even load spread)
> Pin 18 is -12V (300mA max limited by the PSU, 5A terminal limit)

Useful information. One of the things which was not available
from anywhere was a diagram of the power supply itself, to allow
troubleshooting and repair.

>> It should be possible to put some alternative connector in place
>> of that -- both on the end of the cable, and on the power supply.
>
> I'm looking into doing that - chances are I'll remove the ribbon cable
> and replace it with a bunch of colour-coded high-current wires terminated
> in a Molex 0.156mm KK connector. If I can get the phosphor-bronze KKs,
> they're good to 7A max per contact. The brass ones are only good to 5A
> per contact.

Good.

> I'm looking at the Molex 26-48-1185 (KK 0.156in 18way), but Farnell don't
> seem to stock the mating crimp housings for those (they have them up to
> 14way). The 18way + 0.156mm combination seems rather difficult to find on
> this side of the pond :(
>
>> perhaps one of those connectors used in industrial equipment like
>> machine tool controllers, where the connector solders down to the board,
>> and the individual wires are stuck into holes and tightened with clamp
>> screws.
>
> A terminal block? That's one option I suppose, but the ones I've seen
> aren't available in 0.156 pitch.

And those that are that small probably will not accept the gauge
of wire you need to use.

>> In any case -- pull out the power supply and unscrew the board
>> to check the underside. I'll bet that the solder associated with the
>> burnt pins above it are at best now cold solder joints (grainy), and
>> likely solder even melted away from around the pins.
>
> Yep, got it in one. The soldering is a mess. There's visible heat damage
> around the connector too :(

Yep. That is what I've seen before.

>> Remove the old
>> solder with a solder sucker, and flow in fresh tin-lead electrical
>> solder, ideally 60/40 ratio or 63/37, so it will melt well and flow
>> well. *Don't* use the modern ROHS compliant lead-free solders. They
>> have problems, and also melt at a much higher temperature, thus damaging
>> the material of the printed circuit board.
>
> Lucky me, I only use RoHS stuff if I have to. Most of the time, the Metcal
> is set up for 60/40.

Great! So I did not need to mention it -- but just being safe. :-)

Good Luck,

[Philip Pemberton]

On Fri, 17 May 2013 02:20:13 +0000, DoN. Nichols wrote:

>>> Ouch! Maximum physical memory was 4 MB, and the same applies to
>>> maximum virtual memory -- so you need at least 4 MB disk space
>>> allocated to virtual memory.
>>
>> Hmm. In that case, I wonder if there's something wrong with the hard
>> drive controller emulation - possibly incorrectly reading or writing.
>> I'll have to check that...
>
> You should be emulating a swap partition as well as the main
> disk.

I'm actually emulating the hard disc as a whole -- what the kernel sees
is up to it. All my emulator provides is a WD1010 chip with a virtual
hard disc of arbitrary size and with no bad blocks :)

Sadly I don't have a copy of the kernel built with debugging enabled,
otherwise the messages might give away what's going wrong :(

>> There was a Y-splitter power
>> cable
>> and Molex extension cable in the shipping box -- so I suspect a
>> previous owner had the HDDs running from an external power supply at
>> some point.
>
> That -- or perhaps used it with the cable from the power supply
> to power two drives -- until the PS self-immolated.

Quite possibly. I've just resoldered the PSU - quite a few of the solder
joints were utterly shot, especially on large components. Typical wave
soldering cockups.

I'd put good money that the failure mode for the connector is something
along the lines of:

* Solder joint on PSU fails
* Load is now spread across (n-1) pins instead of (n). Connector gets
warm.
* Another joint fails
* Load is now spread across (n-2) pins. Connector is seriously
overloaded and gets cooked.

For extra fun, add in oxidation of connector plating - the warm connector
pins oxidise, increasing contact resistance and heating them up further.
It wouldn't take much for this to cause something akin to a chain
reaction until one or both connectors were destroyed.

>> It looks like I'll need the 3.51m install disks if I ever need to do a
>> reinstall... Sadly it looks like the ones on Bitsavers are standard
>> v3.51 :(
>
> They used to be downloadable from the AT&T site, IIRC. And were
> mirrored on other sites -- but many of those are now gone.

Yep :(
Bitsavers only has 3.51 "proper"; the Taronga UNIX-PC archive has FIXDISK
2.0 for 3.51 but not the 3.51m patch or any install media.

> IIRC, the only thing different between 3.51 and 3.51m is the
> kernel installed by the diagnostics disk. Do you have the enhanced
> diagonstics disk? If not, send me an e-mail address at which you can
> accept an attachment of about a 365k floppy image. Since this does not
> have the encryption software, I can easily send it to you. (I could not
> receive an attachment of that size, and would not send one unless you
> told me to do so.

Nope - I don't have that (or any 3.51m media for that matter). This email
address (<phi...@philpem.me.uk>) should be able to accept attachments up
to 10MB. If that fails, I can email you the details for an FTP account I
keep around for exchanging files.

>> It looks like they're ITW-Pancon MAS-CON connectors:
>>
>> http://www.iobium.com/wpc_idc_connectors.htm
>> http://itwpancon.com/Products/mascon_index.htm
>
> That sounds good, and looks like it.
>
>> There's also a similar Molex KK connector, part number 09-06-5027:
>>
>> http://www.molex.com/molex/products/datasheet.jsp?
>> part=active/0009065027_IDT_AND_SOLDER_CON.xml
>>
>> Molex declare it as obsolete, but Farnell have a few left in stock.
>
> Similar Molex connectors were used as the bus connectors on the
> SWTP 6800 and 6809 computers, and were very unreliable over time -- the
> pins would oxidize, and I would have to slide the cards up and down
> several times on the pins to make it work again.

Ugh :(
Sadly either of these will involve a Mouser or Digikey order - none of my
usual UK suppliers stock 0.156in pin connectors :-/

Having done a bit of reading up, I get the impression that Molex Trifurcon
headers are the "gold standard" in the pinball community. Sadly they seem
to be a US-only item...

>> After measuring with callipers, I'm left with two options for pin
>> pitch: 3.96mm (0.156in) and 4mm. As 4mm isn't a valid option for a
>> MAS-CON, it must be 3.96mm :)
>
> Digital, Dial, or Vernier calipers?

Digital.

>> I also found a document from Thad Florian with the PSU pinout:
>>
>> http://unixpc.taronga.com/documents/powsup.info
>>
>> It's an 18-pin connector, pinout as follows:
>> Odd numbered pins (1,3,5,7,9,11,13,15,17) are ground Pins 2,4,6,8,10
>> are +5V (5x5A = 25A max assuming even load spread)
>> Pins 12,14,16 are +12V (3x5A = 15A max assuming even load spread)
>> Pin 18 is -12V (300mA max limited by the PSU, 5A terminal limit)
>
> Useful information. One of the things which was not available
> from anywhere was a diagram of the power supply itself, to allow
> troubleshooting and repair.

I've reverse engineered the input section and rectifier to see if there
was any possibility of modifying it for 230V operation. I don't fancy
trying to R/E the whole thing, the epoxy-potted hybrid would probably
frustrate any serious attempt to do so.

Frankly, I'm tempted to install a good quality ~350W PC power supply in
place of the stock UNIX PC power supply...

>>> Remove the old solder with a solder sucker, and flow in fresh tin-lead
>>> electrical solder, ideally 60/40 ratio or 63/37, so it will melt well
>>> and flow well. *Don't* use the modern ROHS compliant lead-free
>>> solders. They have problems, and also melt at a much higher
>>> temperature, thus damaging the material of the printed circuit board.
>>
>> Lucky me, I only use RoHS stuff if I have to. Most of the time, the
>> Metcal is set up for 60/40.
>
> Great! So I did not need to mention it -- but just being safe. :-
)

"Specify every single step. That way you don't have to deal with the one
guy who didn't realise you were supposed to unplug the thing before
trying to repair it. Or, for that matter, his incredibly litigious
family." :-)

I've resoldered the PSU connector - melted in some fresh solder, removed
the same, then resoldered with fresh Multicore 60-40. I'll be damned if
those joints fail in a hurry. The existing joints looked both flux and
solder starved... had I been the production line inspector, I'd have sent
it back for rework. Could well explain why Power Systems Inc. aren't
around these days.


Cheers,
Phil.

[DoN. Nichols]

On 2013-05-17, Philip Pemberton <phi...@philpem.me.uk> wrote:
> On Fri, 17 May 2013 02:20:13 +0000, DoN. Nichols wrote:

[ ... ]

>> You should be emulating a swap partition as well as the main
>> disk.
>
> I'm actually emulating the hard disc as a whole -- what the kernel sees
> is up to it. All my emulator provides is a WD1010 chip with a virtual
> hard disc of arbitrary size and with no bad blocks :)

O.K. But you need the initial install disk to partition the
hard disk (even the virtual one -- this is just writing information into
the superblock and other parts of the disk structure.

Actually -- that "no bad blocks" may be the problem. It can't
believe that such disks exist. :-)

But really -- you need the install floppy to partition the disk
before the actual installation happens.

> Sadly I don't have a copy of the kernel built with debugging enabled,
> otherwise the messages might give away what's going wrong :(

You may -- have you downloaded the FIXDISK 2.0 yet? (I'll get
to more of that later.)

>
>>> There was a Y-splitter power
>>> cable
>>> and Molex extension cable in the shipping box -- so I suspect a
>>> previous owner had the HDDs running from an external power supply at
>>> some point.
>>
>> That -- or perhaps used it with the cable from the power supply
>> to power two drives -- until the PS self-immolated.
>
> Quite possibly. I've just resoldered the PSU - quite a few of the solder
> joints were utterly shot, especially on large components. Typical wave
> soldering cockups.

I wonder whether it is that, or the result of the heat as the
pins were failing.

> I'd put good money that the failure mode for the connector is something
> along the lines of:
>
> * Solder joint on PSU fails
> * Load is now spread across (n-1) pins instead of (n). Connector gets
> warm.
> * Another joint fails
> * Load is now spread across (n-2) pins. Connector is seriously
> overloaded and gets cooked.
>
> For extra fun, add in oxidation of connector plating - the warm connector
> pins oxidise, increasing contact resistance and heating them up further.
> It wouldn't take much for this to cause something akin to a chain
> reaction until one or both connectors were destroyed.

Actually -- what I have seen indicates that the first step is
more often oxidation of a pin, which then casues it to heat up and
eventually melt out the solder at the PC board. If you can catch it
when the connctor is just barely starting to brown at one pin, and treat
it with a really good contact cleaner, you will get a lot more life out
of it. ("Really Good Contact Cleaner" -- back then it was two spray
cans from Craig Laboratories, IIRC) called "Cramolin". There was a Blue
and a Red version -- and you were supposed to spray one on and work
the mated connectors to scrub off oxidation, then wipe the connectors,
and spry the other color. (I think blue first, and then red, but it
might have been the other order. In reality, they were identical other
than color dyes in them -- and the idea was to make sure that you spray
scrub, clean and re-spray -- the first for cleaning, the second to leave
a clean protective film on the contact. Sort of like the help desk
asking "Is the plug from the computer (or terminal, depending on when)
plugged into the upper or lower outlet. It made no difference, but it
preempted the "Of course it is!" without looking when asked "Are you
sure it is plugged in?". :-)

Cramolin was kiled off in the interst of the ozone layer or
something way back when, and has been succeeded by at least two
replacements.

>>> It looks like I'll need the 3.51m install disks if I ever need to do a
>>> reinstall... Sadly it looks like the ones on Bitsavers are standard
>>> v3.51 :(

No -- all you need is the 3.51 disks, and the FIXDISK 2.0, and
to use that early on to get what you need.

>> They used to be downloadable from the AT&T site, IIRC. And were
>> mirrored on other sites -- but many of those are now gone.
>
> Yep :(
> Bitsavers only has 3.51 "proper"; the Taronga UNIX-PC archive has FIXDISK
> 2.0 for 3.51 but not the 3.51m patch or any install media.

FIXDISK 2.0 has the following (determined by:

cpio -icBt < FIXdISK-2.0+IN

from an image of the file contents. This gives me:

======================================================================
Size
Name
Install
Remove
Files
Fixes
Instructions
Readme.Fix
ate/Info
ate/Install
ate/Name
ate/async_main.z
date/Info
date/Install
date/Name
date/date.z
kernel/Info
kernel/Install
kernel/Name
kernel/UNIX3.51m.z
kernel/cmb.o.z
kernel/wind.o.z
kmap/Info
kmap/Install
kmap/Name
kmap/kmap.610
ksh/Info
ksh/Install
ksh/Name
ksh/ksh.z
magic/Info
magic/Install
magic/Name
magic/magic
modemcap/Info
modemcap/Install
modemcap/Name
modemcap/modemcap
ph/Info
ph/Install
ph/Name
ph/ph.z
remove/Info
remove/Install
remove/Name
remove/Uninstall.sh
scrset/Info
scrset/Install
scrset/Name
scrset/scrset.z
setgetty/Info
setgetty/Install
setgetty/Name
setgetty/setgetty.z
tam/Info
tam/Install
tam/Name
tam/libtam.a.z
tape/Info
tape/Install
tape/Name
tape/Tbackup.sh.z
utmp/Info
utmp/Install
utmp/Name
utmp/getty.z
utmp/getut.o.z
utmp/init.z
utmp/login.z
utmp/ttyslot.o.z
uucp/Info
uucp/Install
uucp/Name
uucp/uucico.z
======================================================================

The order of the files is critical, as the install script starts
by extracting the "Size" file and checking its contents against
available space in /tmp (where it normally is extracted.)

Note in particlar, part way down the list the file:

kernel/UNIX3.51m.z

A compressed version of the 3.51m kernel.

I made it a practice to keep a copy of the diag kernel, as well
as other kernels which implemented diagnostics for specific hardware
like the ethernet card, the floppy tape card, and others in /kernel, so
I could boot any one of them without needing to install the floppies.

Anyway, you can extract all of that into another directory by
doing a CD to that directory and typing:

cpio -idBdv < /Path-to-FIXDISK-image

And, looking though the files downloaded from various sites for
the 3B1, I find these two:

Note that the FIXDISK file above is the cpio for the contents
of two floppies, so to run it you need to either create a pair of
floppies using cpio and the "Lists" file as input to cpio in place of
running find in the directory to assure that things are in the right
order. Or -- you could extract the file in /tmp manually and then run
the install script -- but I forget whether that breaks anything. If
you were to find some 80 track 5-1/4" floppies which were not high
density ones -- or 3-1/2" ones which could be convinced to not go into
high density mode (some had jumpers for that), you could install it all
from a single floppy, as those would hold 800K instead of 400K. And
yes, I did use some 3-1/2" floppies on the system at one time.

======================================================================
-rw-r--r-- 1 dnichols family 86K Mar 15 23:41 s4diag.cpio.Z
-rw-r--r-- 1 dnichols family 5.0K Mar 15 23:41 s4test.info.Z
======================================================================

The second is a text file, while the first is:

======================================================================
# zcat s4diag.cpio.Z | cpio -iBctv
drwxrwxrwx 2 root sys 0 Jul 2 14:30 1989, .
-rw-r--r-- 1 102 _ldapd 2084 Jul 2 14:29 1989, Install
-rw-r--r-- 1 102 _ldapd 3328 Jul 2 14:29 1989, README
-rwxr-xr-x 1 102 _ldapd 150478 Jul 2 14:29 1989, s4diag
======================================================================

So be prepared for a bunch of files. I think that I'll send
them one at a time, so you don't hit a magic threshold for a single
e-mail.

It looks as though I downloaded these from unixpc.org some
unknown time ago. (I more recently moved them from one RAID array to
another using cpio -pdV (Sun version, not 3B1), and the dates all got
reset. :-)

>> IIRC, the only thing different between 3.51 and 3.51m is the
>> kernel installed by the diagnostics disk. Do you have the enhanced
>> diagonstics disk? If not, send me an e-mail address at which you can
>> accept an attachment of about a 365k floppy image. Since this does not
>> have the encryption software, I can easily send it to you. (I could not
>> receive an attachment of that size, and would not send one unless you
>> told me to do so.
>
> Nope - I don't have that (or any 3.51m media for that matter). This email
> address (<phi...@philpem.me.uk>) should be able to accept attachments up
> to 10MB. If that fails, I can email you the details for an FTP account I
> keep around for exchanging files.

All you need for 3.51m is a 3.51 set and the FIXDISK-2.0
floppies.

[ ... ]

> Having done a bit of reading up, I get the impression that Molex Trifurcon
> headers are the "gold standard" in the pinball community. Sadly they seem
> to be a US-only item...

How much current do pinball machines put through their connectors?

>>> After measuring with callipers, I'm left with two options for pin
>>> pitch: 3.96mm (0.156in) and 4mm. As 4mm isn't a valid option for a
>>> MAS-CON, it must be 3.96mm :)

BTW Is there room to mount a block connector with pigtails to the power
supply board instead of a rigid mount to the ps PC board? That way you
would not need to care about the spacing.

>> Digital, Dial, or Vernier calipers?
>
> Digital.

Easier to swap between metric and inch mode there. :-) I've seem
dial calipers which had two needles and two concentric dials, but really
hard to re-zero if they ever get bumped out of sync.

As for Vernier ones -- good ones were available in both systems,
but my eyes are not up to reading them these days without close-up
glasses and lots of light. :-)

[ ... ]

>> Useful information. One of the things which was not available
>> from anywhere was a diagram of the power supply itself, to allow
>> troubleshooting and repair.
>
> I've reverse engineered the input section and rectifier to see if there
> was any possibility of modifying it for 230V operation. I don't fancy
> trying to R/E the whole thing, the epoxy-potted hybrid would probably
> frustrate any serious attempt to do so.

Unless it is just a bridge rectifier? I don't have a PS in my
hands at the moment, so I don't remember what was on them.

> Frankly, I'm tempted to install a good quality ~350W PC power supply in
> place of the stock UNIX PC power supply...

That way, you have plenty of separate connectors to go to the
two hard drives and the floppy drive, to keep the total current in any
one connector from being too high.:-)

>>>> Remove the old solder with a solder sucker, and flow in fresh tin-lead
>>>> electrical solder, ideally 60/40 ratio or 63/37, so it will melt well
>>>> and flow well. *Don't* use the modern ROHS compliant lead-free
>>>> solders. They have problems, and also melt at a much higher
>>>> temperature, thus damaging the material of the printed circuit board.
>>>
>>> Lucky me, I only use RoHS stuff if I have to. Most of the time, the
>>> Metcal is set up for 60/40.
>>
>> Great! So I did not need to mention it -- but just being safe. :-
> )
>
> "Specify every single step. That way you don't have to deal with the one
> guy who didn't realise you were supposed to unplug the thing before
> trying to repair it. Or, for that matter, his incredibly litigious
> family." :-)

Indeed so. Or at least to keep someone from compounding the
damage to old hardware.

> I've resoldered the PSU connector - melted in some fresh solder, removed
> the same, then resoldered with fresh Multicore 60-40. I'll be damned if
> those joints fail in a hurry. The existing joints looked both flux and
> solder starved... had I been the production line inspector, I'd have sent
> it back for rework. Could well explain why Power Systems Inc. aren't
> around these days.

:-)

>
> Cheers,
> Phil.

[Philip Pemberton]

On Sat, 18 May 2013 01:25:56 +0000, DoN. Nichols wrote:

> O.K. But you need the initial install disk to partition the
> hard disk (even the virtual one -- this is just writing information into
> the superblock and other parts of the disk structure.

I used the Diagnostics disk to initialise it - I think I picked the
largest disk available -- Miniscribe 64MB, 1024/8/16 geometry. I'm
emulating a WD2010, by the way...

Booting from Diags, I get:

2097151 bytes On-board memory
2097148 bytes Expansion memory

Initialising:

Format completed.
Surface test
Volume name: WINCHE
Total space = 65536 blocks
Partition 0 = 64 blocks
Partition 1 = 4000 blocks ( SWAP ) Single-user
User space = 61472 blocks


From Expert mode, test 6,12:

WINCHESTER DISK TEST
Volume Name: WINCHE
Winchester disk
Volume Name: WINCHE
1024 Cylinders, 8 Heads per Cylinder
There are 17 Physical Sectors (of 512 bytes) per Track
136 Physical Sectors per Cylinder, 139264 Physical Sectors per Disk
There are 8 Logical Blocks (of 1024 bytes) per Track
64 Logical Blocks per Cylinder, 65536 Logical Blocks per Disk
Reduced write current is valid
The Step Rate supplied to the controller is 0
Partition 0: start Track=0, size (in Blocks)=64
Partition 1: start Track=8, size (in Blocks)=4000
Partition 2: start Track=508, size (in Blocks)=61472
Bad Block Table starts at Block 1 (size=1 Blocks)

The Bad Block Table contains 0 entries.


Now I boot from Disk 2 (FD BOOT)...

Version 3.31
Real memory = 4194304
Available memory = 3850240
Main board is P5.1
Floppy Unix

Now swap in Disk 3 (Floppy Filesystem)...

WARNING If you continue all files currently on the hard disk will be
erased!! (Yes/continue)

Making filesystem on the hard disk ....

Copying the loader onto the hard disk ....

-sh: no space

Copying initial 3.51 files onto hard disk filesystem ....

cannot write in </mnt>
mv: cannot access /mnt/etc/profile.hd



And then a prompt to insert the Hard Disk Boot floppy -- if I do, the
machine reboots, but locks up at the Loader.

> You may -- have you downloaded the FIXDISK 2.0 yet? (I'll get
> to more of that later.)

No, but I got the three e-mail attachments. Thanks for those, by the way.

> It looks as though I downloaded these from unixpc.org some
> unknown time ago. (I more recently moved them from one RAID array to
> another using cpio -pdV (Sun version, not 3B1), and the dates all got
> reset. :-)

I have a couple of files from unixpc.org -- sadly it seems the site
itself has gone. Unfortunately I didn't take a full mirror before it
died :(

>> Having done a bit of reading up, I get the impression that Molex
>> Trifurcon headers are the "gold standard" in the pinball community.
>> Sadly they seem to be a US-only item...
>
> How much current do pinball machines put through their connectors?

I guess it depends on the machines... could be five or six amps per pin
on the bigger machines. Lots of electromechanical parts and lamps drawing
fairly high currents.

> BTW Is there room to mount a block connector with pigtails to the power
> supply board instead of a rigid mount to the ps PC board? That way you
> would not need to care about the spacing.

I probably could... but eighteen pins? That's a long terminal block...

>>> Useful information. One of the things which was not available
>>> from anywhere was a diagram of the power supply itself, to allow
>>> troubleshooting and repair.
>>
>> I've reverse engineered the input section and rectifier to see if there
>> was any possibility of modifying it for 230V operation. I don't fancy
>> trying to R/E the whole thing, the epoxy-potted hybrid would probably
>> frustrate any serious attempt to do so.
>
> Unless it is just a bridge rectifier? I don't have a PS in my
> hands at the moment, so I don't remember what was on them.

The hybrid seems to be some form of switchmode power supply controller
built on a separate PCB (with pins to the main PSU PCB) and covered in
green epoxy. More likely an anti-cloning/reverse-engineering measure than
an actual hybrid chip.

>> Frankly, I'm tempted to install a good quality ~350W PC power supply in
>> place of the stock UNIX PC power supply...
>
> That way, you have plenty of separate connectors to go to the
> two hard drives and the floppy drive, to keep the total current in any
> one connector from being too high.:-)

And the "keep the machine original!!!" crowd will no doubt have me locked
in the stocks and pelted with rotten fruit... :P

Right... back to finding a usable power connector I guess... :-/

Cheers,
Phil.

[DoN. Nichols]

On 2013-05-21, Philip Pemberton <phi...@philpem.me.uk> wrote:
> On Sat, 18 May 2013 01:25:56 +0000, DoN. Nichols wrote:
>
>> O.K. But you need the initial install disk to partition the
>> hard disk (even the virtual one -- this is just writing information into
>> the superblock and other parts of the disk structure.
>
> I used the Diagnostics disk to initialise it - I think I picked the
> largest disk available -- Miniscribe 64MB, 1024/8/16 geometry. I'm
> emulating a WD2010, by the way...

Hmm ... with the WD2010, you can use larger disks. The limiting
factor with the original WD1010 was a maximum of 1024 cylinders. The
WD2010 bumps it up to a number of cylinders which is larger than any MFM
drive made.

Another (hardware) limiting factor is the number of heads (disk
surfaces). That was defined by the number of head-select wires going to
the drive. There were only three, thus the maximum of 8 heads. The P5.1
hardwware mod gives you an extra head select wire, so you can get 16
heads. The largest drive which was actually ever used with a system
with both upgrades was A Maxtor, and a Priam with the same specs. That
one was a 190 Mb drive, or something like 160 MB when the OS takes out a
sector per head per cylinder. (Yes, I know that you are emulating a
perfect drive -- but that is not something which the OS can be convinced
to believe in It *will* reserve that sector per head per cylinder no
matter what you do. (Maybe set up the emulator to give it a few bad
sectors to give it something to do? :-)

> Booting from Diags, I get:
>
> 2097151 bytes On-board memory
> 2097148 bytes Expansion memory
>
> Initialising:
>
> Format completed.
> Surface test
> Volume name: WINCHE
> Total space = 65536 blocks
> Partition 0 = 64 blocks
> Partition 1 = 4000 blocks ( SWAP ) Single-user
> User space = 61472 blocks

Hmm ... that looks weird to me There should be a named partition
for the user space.

O.K. The disk test below does show the user space *is* a third
partition.

> From Expert mode, test 6,12:
>
> WINCHESTER DISK TEST
> Volume Name: WINCHE
> Winchester disk
> Volume Name: WINCHE
> 1024 Cylinders, 8 Heads per Cylinder
> There are 17 Physical Sectors (of 512 bytes) per Track
> 136 Physical Sectors per Cylinder, 139264 Physical Sectors per Disk
> There are 8 Logical Blocks (of 1024 bytes) per Track
> 64 Logical Blocks per Cylinder, 65536 Logical Blocks per Disk
> Reduced write current is valid
> The Step Rate supplied to the controller is 0
> Partition 0: start Track=0, size (in Blocks)=64
> Partition 1: start Track=8, size (in Blocks)=4000
> Partition 2: start Track=508, size (in Blocks)=61472
> Bad Block Table starts at Block 1 (size=1 Blocks)

The swap space looks to be a full 4 MB -- the maximum memory
which the system can handle.

> The Bad Block Table contains 0 entries.
>
>
> Now I boot from Disk 2 (FD BOOT)...
>
> Version 3.31
> Real memory = 4194304
> Available memory = 3850240
> Main board is P5.1
> Floppy Unix
>
> Now swap in Disk 3 (Floppy Filesystem)...
>
> WARNING If you continue all files currently on the hard disk will be
> erased!! (Yes/continue)
>
> Making filesystem on the hard disk ....
>
> Copying the loader onto the hard disk ....
>
> -sh: no space
>
> Copying initial 3.51 files onto hard disk filesystem ....
>
> cannot write in </mnt>
> mv: cannot access /mnt/etc/profile.hd

Is your emulator perhaps write protected?

> And then a prompt to insert the Hard Disk Boot floppy -- if I do, the
> machine reboots, but locks up at the Loader.

Yes, since you were not able to put a loader onto the emulated disk.
Can you log what happens with the emulated disk to see why it is not
allowing writing?

>> You may -- have you downloaded the FIXDISK 2.0 yet? (I'll get
>> to more of that later.)
>
> No, but I got the three e-mail attachments. Thanks for those, by the way.

You're welcome.

>> It looks as though I downloaded these from unixpc.org some
>> unknown time ago. (I more recently moved them from one RAID array to
>> another using cpio -pdV (Sun version, not 3B1), and the dates all got
>> reset. :-)
>
> I have a couple of files from unixpc.org -- sadly it seems the site
> itself has gone. Unfortunately I didn't take a full mirror before it
> died :(

That is a pity. It had a *lot* of stuff there.

>>> Having done a bit of reading up, I get the impression that Molex
>>> Trifurcon headers are the "gold standard" in the pinball community.
>>> Sadly they seem to be a US-only item...
>>
>> How much current do pinball machines put through their connectors?
>
> I guess it depends on the machines... could be five or six amps per pin
> on the bigger machines. Lots of electromechanical parts and lamps drawing
> fairly high currents.

O.K.

>> BTW Is there room to mount a block connector with pigtails to the power
>> supply board instead of a rigid mount to the ps PC board? That way you
>> would not need to care about the spacing.
>
> I probably could... but eighteen pins? That's a long terminal block...

Who says:

1) -- that it has to be a single row of pins?

2) -- that you can't use two smaller grid pin connectors.

I would use two of different numbers of pins -- say a 10 pin and
an 8-pin, so you could not cross-connect them. Or a 12-pin and a
10-pin, with duplicate ground wires going to both, so you are *sure*
that the power supply has a ground reference even if one connector is
not in place.

>>>> Useful information. One of the things which was not available
>>>> from anywhere was a diagram of the power supply itself, to allow
>>>> troubleshooting and repair.
>>>
>>> I've reverse engineered the input section and rectifier to see if there
>>> was any possibility of modifying it for 230V operation. I don't fancy
>>> trying to R/E the whole thing, the epoxy-potted hybrid would probably
>>> frustrate any serious attempt to do so.

Likely so. It goes along with the absence of any repair
documentation, even in the _Hardware Reference Manual_, which if you
don't have it, you should try to find a copy. Maybe someone has scanned
it -- including the fold-out schematics pages.

>> Unless it is just a bridge rectifier? I don't have a PS in my
>> hands at the moment, so I don't remember what was on them.
>
> The hybrid seems to be some form of switchmode power supply controller
> built on a separate PCB (with pins to the main PSU PCB) and covered in
> green epoxy. More likely an anti-cloning/reverse-engineering measure than
> an actual hybrid chip.
>
>>> Frankly, I'm tempted to install a good quality ~350W PC power supply in
>>> place of the stock UNIX PC power supply...
>>
>> That way, you have plenty of separate connectors to go to the
>> two hard drives and the floppy drive, to keep the total current in any
>> one connector from being too high.:-)
>
> And the "keep the machine original!!!" crowd will no doubt have me locked
> in the stocks and pelted with rotten fruit... :P

Just avoid doing anything to it which can't be reversed if you
get the proper power supply. IIRC, only some of the systems had both
fan locations occupied, so route the wires in through the other.

> Right... back to finding a usable power connector I guess... :-/

Remember -- it does not *have* to look like what is in there
now. Any combination of connectors which will fit somewhere in there
and provide the total pin count needed should work.

Good Luck,
DoN.

[none Aharon Robbins]

In article <slrnkpoacf.d0...@Katana.d-and-d.com>,

DoN. Nichols <BPdnic...@d-and-d.com> wrote:
>> I have a couple of files from unixpc.org -- sadly it seems the site
>> itself has gone. Unfortunately I didn't take a full mirror before it
>> died :(
>
> That is a pity. It had a *lot* of stuff there.

Has anybody tried the Internet Archive at www.archive.org?

Arnold
--
Aharon (Arnold) Robbins arnold AT skeeve DOT com
P.O. Box 354 Home Phone: +972 8 979-0381
Nof Ayalon
D.N. Shimshon 9978500 ISRAEL

[Thad Floryan]

On 5/22/2013 11:39 PM, none Aharon Robbins wrote:
> In article <slrnkpoacf.d0...@Katana.d-and-d.com>,
> DoN. Nichols <BPdnic...@d-and-d.com> wrote:
>>> I have a couple of files from unixpc.org -- sadly it seems the site
>>> itself has gone. Unfortunately I didn't take a full mirror before it
>>> died :(
>> That is a pity. It had a *lot* of stuff there.
>
> Has anybody tried the Internet Archive at www.archive.org?

Excellent idea! I just did and found this:

http://web.archive.org/web/20111018205812/http://www.unixpc.org/

which I don't have time to examine at the moment.

Thad

[Aharon Robbins]

In article <519E02E6...@thadlabs.com>,

From the slider thingy at the top it looks like there may be snapshots as
far back as 2004.

It would be really good to recover that stuff, organize and make it available,
esp once the emulator gets up and working.

I for one, *can't wait*. I spent many happy hours coding and writing on
my 3B1 and it would be be fun to try to run some modern incarnations of
some things on it.

Thanks,

[Philip Pemberton]

On Thu, 23 May 2013 13:02:54 +0000, Aharon Robbins wrote:

> From the slider thingy at the top it looks like there may be snapshots
> as far back as 2004.
>
> It would be really good to recover that stuff, organize and make it
> available,
> esp once the emulator gets up and working.

I think the trick is going to be in figuring out why I'm getting that
blasted "no space" error...

Oh, I contacted Corey Lindsly, the guy behind unixpc.org. He's renewed
the domain and the site and all its files are now back online :)

As for the (physical) UNIX PC - I've procured some Molex "KK 0.156" high-
current connectors. Sadly I couldn't get the Trifurcon contacts, best I
could find were the 7-amp phosphor-bronze KKs.

The plan for "when I next have some spare time" (maybe this coming bank
holiday Monday) is to replace the motherboard power connector and cable,
then build up an ATX-to-UNIXPC cable so I can power the thing up without
having to make any permanent modifications. It won't look pretty, but it
should work until I can get my hands on a decently beefy 240-to-110V
transformer.

I may also try and borrow a FLIR camera to check for overheating on the
new connector, but that's a "maybe if I'm bored" task :)

Cheers,
Phil.

[Bill Gunshannon]

In article <519f9cc6$0$11933$c3e8da3$fe7f...@news.astraweb.com>,

Philip Pemberton <phi...@philpem.me.uk> writes:
> On Thu, 23 May 2013 13:02:54 +0000, Aharon Robbins wrote:
>
>> From the slider thingy at the top it looks like there may be snapshots
>> as far back as 2004.
>>
>> It would be really good to recover that stuff, organize and make it
>> available,
>> esp once the emulator gets up and working.
>
> I think the trick is going to be in figuring out why I'm getting that
> blasted "no space" error...
>
> Oh, I contacted Corey Lindsly, the guy behind unixpc.org. He's renewed
> the domain and the site and all its files are now back online :)
>
> As for the (physical) UNIX PC - I've procured some Molex "KK 0.156" high-
> current connectors. Sadly I couldn't get the Trifurcon contacts, best I
> could find were the 7-amp phosphor-bronze KKs.
>
> The plan for "when I next have some spare time" (maybe this coming bank
> holiday Monday) is to replace the motherboard power connector and cable,
> then build up an ATX-to-UNIXPC cable so I can power the thing up without
> having to make any permanent modifications. It won't look pretty, but it
> should work until I can get my hands on a decently beefy 240-to-110V
> transformer.

Do you live anywhere near a US Military Base? We had lots of them
and I would imagine now they still show up in thrift shops. Heck, I
still have two here. One I actually reversed so I could run some of
the European stuff I brought home with me (although that stuff is all
long gone at this point. I would happily give you the one I still
have (I think it is something like 1200 Watt) but the cost of shipping
it back over the pond is probably prohibitive.

bill

>
> I may also try and borrow a FLIR camera to check for overheating on the
> new connector, but that's a "maybe if I'm bored" task :)
>
> Cheers,
> Phil.

--
Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
bill...@cs.scranton.edu | and a sheep voting on what's for dinner.
University of Scranton |
Scranton, Pennsylvania | #include <std.disclaimer.h>

[tlvp]

On 24 May 2013 17:18:14 GMT, Bill Gunshannon wrote:

> ... the cost of shipping

> it back over the pond is probably prohibitive.

Via USPS, yes, certainly. But if you can dig up how to exploit their
services, DHL seems to have a trans-Atlantic surface mail service that
seriously undercuts USPS air rates -- Amazon's CreateSpace service, for
example, uses DHL to ship books they print in S. Carolina over the pond for
less than a third of what USPS charges.

HTH. Cheers, -- tlvp
--
Avant de repondre, jeter la poubelle, SVP.

[Bill Gunshannon]

In article <t4cybnopvtf$.9l8cgmsv...@40tude.net>,

Books almost always go at a cheaper rate. Even with USPS. I think they
could tell the difference. :-)

bill

[DoN. Nichols]

On 2013-05-24, Philip Pemberton <phi...@philpem.me.uk> wrote:

> As for the (physical) UNIX PC - I've procured some Molex "KK 0.156" high-
> current connectors. Sadly I couldn't get the Trifurcon contacts, best I
> could find were the 7-amp phosphor-bronze KKs.

O.K. One option at least.

> The plan for "when I next have some spare time" (maybe this coming bank
> holiday Monday) is to replace the motherboard power connector and cable,
> then build up an ATX-to-UNIXPC cable so I can power the thing up without
> having to make any permanent modifications. It won't look pretty, but it
> should work until I can get my hands on a decently beefy 240-to-110V
> transformer.

Just out of curiosity -- what does the nameplate say beside the
AC input connector? It *might* be that the power supply is a
self-switching one, which automatically switchs between 120 VAC and 240
VAC.

> I may also try and borrow a FLIR camera to check for overheating on the
> new connector, but that's a "maybe if I'm bored" task :)

It is not a matter of "if" but of "when" it will overheat, based
on previous experience. :-)