(Note crossposting)

Any PC power supply experts out there? I'm just trying to repair a
friend's PC with a dead PSU. Well I assume it's dead - it does produce
the +5V standby rail OK and when the PC's front panel power button is
pressed the PS_ON signal line is pulled low, but nothing else happens -
none of the main power rails come up. IOW the PSU is getting the
command to turn on, but appears to be failing to do so.

Anyway, a nice simple job I thought, trotting down to the local Chinese
computer take-away (MKC) to buy a new PSU. "Oh they all come with a
24-pin motherboard connector now, not 20-pin like that" (pointing at the
old one I'd taken along). "Don't worry, it'll be compatible, you'll
just have 4 unconnected pins on the connector. Line the latching tabs
up and it'll be OK."

Tending not to trust anything anyone in a shop says, ever, I have
carefully verified that if the new PSU connector is inserted into the
mobo connector pin 1 - pin 1 then all rails and signals do correspond,
and the four new pins only duplicate the +3.3, +5 and 0V rails. The
latching tabs don't align in this position, but that's immaterial, other
than confirming my mistrust...

But there is one possibly significant difference. The old PSU provides
a -5V rail on pin 12, but the new one doesn't - there's no connection to
the corresponding connector position (now pin 14 in 24-pin land).

Hmm, time to RTFM. Does the motherboard - an Asus A7V8X - actually need
a -5V rail? The manual (user guide) from their web site shows the
20-pin connector pin-outs, including -5V, but says nothing about what
it's for or whether it's needed. Only slightly more helpfully, the new
PSU's installation guide - it's a CoolerMaster RS-380, by the way - says
this:

"This power supply does not support –5V for the motherboard for new ATX
form factor operation. Please update the BIOS setting of the motherboard."

Now the questions:

1. DoeS aNyBody know whether the Asus A7V8X motherboard needs a
-5V rail? Will it be harmed if powered-up with this rail
missing (o/c). What is/was this rail used for anyway?

2. What have BIOS settings got to do with this? How can changing a
software (firmware) setting compensate for the lack of a power rail?

3. What is the line labelled PG (pin 8 on both version) for? (This
question is for interest only.)

4. Can you still get ATX PSUs with 20-pin mobo connectors and -5V
outputs, if this turns out to be needed?

5. Is there anything else I should know, before powering-up.

The Joy of Standards? - There are just sooo many to choose from...

TIA
--
Andy

In uk.d-i-y Andy Wade wrote:
(Note crossposting)

Any PC power supply experts out there? I'm just trying to repair a
friend's PC with a dead PSU. Well I assume it's dead - it does produce
the +5V standby rail OK and when the PC's front panel power button is
pressed the PS_ON signal line is pulled low, but nothing else happens -
none of the main power rails come up. IOW the PSU is getting the
command to turn on, but appears to be failing to do so.
snip
But there is one possibly significant difference. The old PSU provides
a -5V rail on pin 12, but the new one doesn't - there's no connection to
the corresponding connector position (now pin 14 in 24-pin land).

Essentially nothing uses -5V AIUI.
It's very, very unlikely not to work.

Andy Wade wrote:
(Note crossposting)

Any PC power supply experts out there? I'm just trying to repair a
friend's PC with a dead PSU. Well I assume it's dead - it does
produce the +5V standby rail OK and when the PC's front panel power
button is pressed the PS_ON signal line is pulled low, but nothing
else happens - none of the main power rails come up. IOW the PSU is
getting the
command to turn on, but appears to be failing to do so.

Anyway, a nice simple job I thought, trotting down to the local
Chinese computer take-away (MKC) to buy a new PSU. "Oh they all come
with a 24-pin motherboard connector now, not 20-pin like that"
(pointing at the old one I'd taken along). "Don't worry, it'll be
compatible, you'll
just have 4 unconnected pins on the connector. Line the latching tabs
up and it'll be OK."


So that should mean you will have two unused pins either side of the block
if its lined up with the latching?

Personally I'd of try to get my hands on a 20pin PS.

Don't always assume its a power supply fault. :-)

ps who pays for the PSU if you find out its not that?

--
Sir Benjamin Middlethwaite

On Fri, 19 May 2006 16:05:26 +0000, Ian Stirling wrote:

In uk.d-i-y Andy Wade wrote:
(Note crossposting)

Any PC power supply experts out there? I'm just trying to repair a
friend's PC with a dead PSU. Well I assume it's dead - it does produce
the +5V standby rail OK and when the PC's front panel power button is
pressed the PS_ON signal line is pulled low, but nothing else happens -
none of the main power rails come up. IOW the PSU is getting the
command to turn on, but appears to be failing to do so.
snip
But there is one possibly significant difference. The old PSU provides
a -5V rail on pin 12, but the new one doesn't - there's no connection to
the corresponding connector position (now pin 14 in 24-pin land).

Essentially nothing uses -5V AIUI.
It's very, very unlikely not to work.

Seconded. I'm struggling to come up with anything that's ever used -5V
TBH; even things like soundcards and serial ports would have used +/-12V
lines and not +/-5V.

It's probably a hang-up from the original IBM PC days, but even then I'm
not sure what would have used it - memory chips were all single-rail
devices by then. I expect its inclusion goes back further than that, to
machines in the 70's that did need -5V for the memory, and IBM included it
in their machine "just in case".

Hmm, I'm going to have to go away and ask now, though.

cheers

Jules

Andy Wade wrote:

3. What is the line labelled PG (pin 8 on both version) for? (This
question is for interest only.)

IIRC, this stands for "Power Good", and is an indication that the power
supply has started up and stabilised so the motherboard can now get on
with booting. No idea whether it's actually used (as opposed to an
arbitrary delay or whatever) by most motherboards.


Kim.

Andy Wade wrote:

(Note crossposting)

Any PC power supply experts out there? I'm just trying to repair a
friend's PC with a dead PSU. Well I assume it's dead - it does produce
the +5V standby rail OK and when the PC's front panel power button is
pressed the PS_ON signal line is pulled low, but nothing else happens -
none of the main power rails come up. IOW the PSU is getting the
command to turn on, but appears to be failing to do so.

Anyway, a nice simple job I thought, trotting down to the local Chinese
computer take-away (MKC) to buy a new PSU. "Oh they all come with a
24-pin motherboard connector now, not 20-pin like that" (pointing at the
old one I'd taken along). "Don't worry, it'll be compatible, you'll
just have 4 unconnected pins on the connector. Line the latching tabs
up and it'll be OK."

Tending not to trust anything anyone in a shop says, ever, I have
carefully verified that if the new PSU connector is inserted into the
mobo connector pin 1 - pin 1 then all rails and signals do correspond,
and the four new pins only duplicate the +3.3, +5 and 0V rails. The
latching tabs don't align in this position, but that's immaterial, other
than confirming my mistrust...

When I bought a new ATX power supply recently I disdcovered that the
24-pin connector was actually a 20-pin job connected to a 4-pin affair
and it was easy enought to separate the two.

Regards

Neil

Ian Stirling wrote:
In uk.d-i-y Andy Wade wrote:
(Note crossposting)

Any PC power supply experts out there? I'm just trying to repair a
friend's PC with a dead PSU. Well I assume it's dead - it does produce
the +5V standby rail OK and when the PC's front panel power button is
pressed the PS_ON signal line is pulled low, but nothing else happens -
none of the main power rails come up. IOW the PSU is getting the
command to turn on, but appears to be failing to do so.
snip
But there is one possibly significant difference. The old PSU provides
a -5V rail on pin 12, but the new one doesn't - there's no connection to
the corresponding connector position (now pin 14 in 24-pin land).

Essentially nothing uses -5V AIUI.
It's very, very unlikely not to work.

The Old 20-pin ATX supplies were built to the following spec.
http://www.formfactors.org/developer...X12V_1_3dg.pdf

This spec on release 1.3 (in 2002), deleted the -5V rail as little
industry interest in continued use.

The New 24-pin ATX supplies are built to
http://www.formfactors.org/developer...20PSDG2.01.pdf
http://www.formfactors.org/developer...public_br2.pdf

-5V rail extinct!

Pin 8 is the power good (PWR_OK) signal, goes to 'high' if the main 12V,
3.3V and 5V supply lines are in spec.

--
Adrian C

Andy Wade wrote:

4. Can you still get ATX PSUs with 20-pin mobo connectors and -5V
outputs, if this turns out to be needed?

most recent couple of PSUs I've used come with a 20pin connector, plus a
"slide on" extra 4pin connector that makes it into a 24pin, I used one
on a 20pin mobo, and the other on a 24pin mobo, both worked fine.

"The3rd Earl Of Derby" wrote in message
. uk...
Andy Wade wrote:
(Note crossposting)

Any PC power supply experts out there? I'm just trying to repair a
friend's PC with a dead PSU. Well I assume it's dead - it does
produce the +5V standby rail OK and when the PC's front panel power
button is pressed the PS_ON signal line is pulled low, but nothing
else happens - none of the main power rails come up. IOW the PSU is
getting the
command to turn on, but appears to be failing to do so.

Anyway, a nice simple job I thought, trotting down to the local
Chinese computer take-away (MKC) to buy a new PSU. "Oh they all come
with a 24-pin motherboard connector now, not 20-pin like that"
(pointing at the old one I'd taken along). "Don't worry, it'll be
compatible, you'll
just have 4 unconnected pins on the connector. Line the latching tabs
up and it'll be OK."


So that should mean you will have two unused pins either side of the block
if its lined up with the latching?


I fitted an ATX 2.0 power supply recently. It has a 24 pin power plug. The
_4 right-hand side_ pins are designed to be separated from the 20-pin l/h
side of the plug, but I was able to fit the the plug leaving the 4
unconnected pins still attached.
_Only_ leave the r/h side 4 pins unconnected, and _not_, I repeat _not_,
have 2 pins on the left and 2 pins on the right disconnected.
I would investigate that plug/socket latching before doing anything else.
Is your computer a Dell, Packard Bell (and a few others)? I understand that
some of their bits can only fit these particular computers.

Sylvain.

Personally I'd of try to get my hands on a 20pin PS.


Not good for possible future upgrading.

Sylvain.

Don't always assume its a power supply fault. :-)

ps who pays for the PSU if you find out its not that?

--
Sir Benjamin Middlethwaite

Sylvain VAN DER WALDE wrote:

_Only_ leave the r/h side 4 pins unconnected, and _not_, I repeat _not_,
have 2 pins on the left and 2 pins on the right disconnected.

I thought they had a mixture of "O" and "D" shaped polarisation of the
pins to prevent that

Is your computer a Dell, Packard Bell (and a few others)? I understand that
some of their bits can only fit these particular computers.

Certainly true that 5+ year old Dells use non-standard 20pin connector :-(

Andy Burns wrote:

Certainly true that 5+ year old Dells use non-standard 20pin connector :-(

Or, more accurately, a standard connector with a non-standard pinout.
Recipe for disaster that it is.


Kim.

Ian Stirling wrote:
In uk.d-i-y Andy Wade wrote:
(Note crossposting)

Any PC power supply experts out there? I'm just trying to repair a
friend's PC with a dead PSU. Well I assume it's dead - it does
produce the +5V standby rail OK and when the PC's front panel power
button is pressed the PS_ON signal line is pulled low, but nothing
else happens - none of the main power rails come up. IOW the PSU is
getting the command to turn on, but appears to be failing to do so.
snip
But there is one possibly significant difference. The old PSU
provides a -5V rail on pin 12, but the new one doesn't - there's no
connection to the corresponding connector position (now pin 14 in
24-pin land).

Essentially nothing uses -5V AIUI.
It's very, very unlikely not to work.

A bunch of new boards from MSI use -5V

eg:
http://www.msi.com.tw/program/produc...il.php?UID=641

--
ThePunisher
Latitude: 54.67N
Longitude: 5.96W

Neil Jones wrote:

When I bought a new ATX power supply recently I disdcovered that the
24-pin connector was actually a 20-pin job connected to a 4-pin affair
and it was easy enought to separate the two.

The PSU in my Dell at work went south the other day; we wheeled in
a new PSU, plugged in the 20 pin connector - not a sausage. Then
we realized that the old PSU, and indeed the new one, had an extra
connector that went to another place on the motherboard (hadn't really
noticed that when we took the old one out in the midst of unplugging all
the disk/SATA power as well). Plugged the second connector in, and hey
presto. So I guess this corresponds to the detachable bit you mention
above, only it was already detached on the new PSU. What it is useful
for, I have no idea - but obviously it does something!

Andy Wade wrote:
(Note crossposting)

Noted, but since you didn't set follow-ups to either group... :-)

[snip]

But there is one possibly significant difference. The old PSU provides
a -5V rail on pin 12, but the new one doesn't - there's no connection to
the corresponding connector position (now pin 14 in 24-pin land).

Hmm, time to RTFM. Does the motherboard - an Asus A7V8X - actually need
a -5V rail?

I'm currently running on an A7V8X. AFAIK, nothing requires the -5V,
though I can't say I've checked carefully as I'm using a good Enermax
PSU with the correct connector anyway.

HOWEVER, I did have horrible problems with the original graphics card I
bought for the system. According to various on-line sources (after a few
months and several incidents of the problem cropping up) was some minor
deviation from spec of the motherboard and/or graphics card power
systems, which resulted in the system effectively hanging occasionally.

The original graphics card was an ATI Radeon 9700 Pro, which used only a
small-style disk drive power connector for auxiliary power. The
replacement, a 9800 Pro that uses a full-size connector, has been solid
as a rock for nearly three years.

This cautionary tale may or may not be of any relevance, but I guess I'd
be wary of any non-standard power issues with the A7V8X.

Hope that helps,
Chris

Andrew Haylett wrote:

Neil Jones wrote:
When I bought a new ATX power supply recently I disdcovered that the
24-pin connector was actually a 20-pin job connected to a 4-pin affair
and it was easy enought to separate the two.

Not all are like that though, including the one I bought. On mine the
24-pin PSU connector mated quite happily with the 20-pin one on the
mobo, with 4 legs out of bed, so to speak.

Anyway, a big thanks to everyone who replied. The machine is now back
together with new PSU and tidied-up wiring and appears to work quite
happily with no -5V (and no BIOS tweaking either).

The PSU in my Dell at work went south the other day; we wheeled in
a new PSU, plugged in the 20 pin connector - not a sausage. Then
we realized that the old PSU, and indeed the new one, had an extra
connector that went to another place on the motherboard (hadn't really
noticed that when we took the old one out in the midst of unplugging all
the disk/SATA power as well). Plugged the second connector in, and hey
presto. So I guess this corresponds to the detachable bit you mention
above, only it was already detached on the new PSU. What it is useful
for, I have no idea - but obviously it does something!

Maybe not. There are separate outputs carrying only +12V & 0V on square
4-pin connectors (four pins, two rows) on both the old and new PSUs.
They're labelled "12V CPU connector." These have no use with the Asus
motherboard in question; I think they're for later boards where the CPU
supply is via a local on-board DC-DC converter - a much more sensible
way of doing things when you want very few volts at an awful lot of amps.

--
Andy

In article , Andy Wade
wrote:


But there is one possibly significant difference. The old PSU provides
a -5V rail on pin 12, but the new one doesn't - there's no connection to
the corresponding connector position (now pin 14 in 24-pin land).

I believe the -5v rail is no longer used. It mat have been RS423 which uses
+-5v rather than RS232 +-12V.

The RISCOS motherboard does use it though just for interest. I think it
feeds the sound processor.


--
AJL

"Andy Wade" wrote in message
...
(Note crossposting)

Any PC power supply experts out there? I'm just trying to repair a
friend's PC with a dead PSU. Well I assume it's dead - it does produce
the +5V standby rail OK and when the PC's front panel power button is
pressed the PS_ON signal line is pulled low, but nothing else happens -
none of the main power rails come up. IOW the PSU is getting the
command to turn on, but appears to be failing to do so.

Some power supplies are unstable without a load on them and will shut down.
Not sure if this is true of PC power supplies.

Anyway, a nice simple job I thought, trotting down to the local Chinese
computer take-away (MKC) to buy a new PSU. "Oh they all come with a
24-pin motherboard connector now, not 20-pin like that" (pointing at the
old one I'd taken along). "Don't worry, it'll be compatible, you'll
just have 4 unconnected pins on the connector. Line the latching tabs
up and it'll be OK."

Sometimes the bit with the 4 extra pins is detachable. The two parts can be
slid apart.

In article ,
"Malcolm Stewart" writes:

I remember -5V being quite important back in 1960 when I was selecting core
store driver transistors, germanium, for a Ferranti Argus computer.

We had fun with cores too at GEC -- they also used +18V and -18V.
The PSU's were one of the more interesting sides. In order to
guarantee no data loss during power outage (core is non-volatile),
you have to be very careful how the power rails are shut down so
that no spikes are delivered to the core store which might corrupt
the data, or activate any of the driving logic as the power rails
decay below their operating spec. In order to do this, the power
supplies were switched off by shorting their outputs in a timed
shutdown sequence, which took around 50ms to shutdown all rails.
This meant the PSU's were effectively switched off by activating
their over-current protection, which always made me smile.

--
Andrew Gabriel

On Sat, 20 May 2006 12:37:20 UTC, (Andrew
Gabriel) wrote:

We had fun with cores too at GEC -- they also used +18V and -18V.

I was given some core this week! Looks like 512 bytes (except it
wouldn't have been bytes)...64 x 64 AFACS. From an Atlas, no
less....anyone have any tech info???

--
The information contained in this post is copyright the
poster, and specifically may not be published in, or used by
Avenue Supplies, http://avenuesupplies.co.uk

"Sylvain VAN DER WALDE" wrote in message
...

So that should mean you will have two unused pins either side of the
block
if its lined up with the latching?


I fitted an ATX 2.0 power supply recently. It has a 24 pin power plug. The
_4 right-hand side_ pins are designed to be separated from the 20-pin l/h
side of the plug, but I was able to fit the the plug leaving the 4
unconnected pins still attached.
_Only_ leave the r/h side 4 pins unconnected, and _not_, I repeat _not_,
have 2 pins on the left and 2 pins on the right disconnected.
I would investigate that plug/socket latching before doing anything else.
Is your computer a Dell, Packard Bell (and a few others)? I understand
that some of their bits can only fit these particular computers.

I agree.
The new "universal" ones have a removable portion that you don't use on
older motherboards.
The latching mechanism should align.

"Bob Eager" wrote in message
...
On Sat, 20 May 2006 12:37:20 UTC, (Andrew
Gabriel) wrote:

We had fun with cores too at GEC -- they also used +18V and -18V.

I was given some core this week! Looks like 512 bytes (except it
wouldn't have been bytes)...64 x 64 AFACS. From an Atlas, no
less....anyone have any tech info???

The Ladybird guide to core store (1971)....

http://davidguy.brinkster.net/computer/016.html

Andy Luckman (AJL Electronics) wrote:

I believe the -5v rail is no longer used. It mat have been RS423 which uses
+-5v rather than RS232 +-12V.

Actually some laptops now use +/- 5V (presumably using a DC-DC
converter) instead of +/- 12V on their serial ports, most equipment is
happy with the reduced voltage swing but some isn't, this makes those
laptops less useful as a roaming terminal for diagnosing problems on
"random" routers/switches :-(

On Sun, 21 May 2006 08:25:10 UTC, "CWatters"
wrote:


"Bob Eager" wrote in message
...
On Sat, 20 May 2006 12:37:20 UTC, (Andrew
Gabriel) wrote:

We had fun with cores too at GEC -- they also used +18V and -18V.

I was given some core this week! Looks like 512 bytes (except it
wouldn't have been bytes)...64 x 64 AFACS. From an Atlas, no
less....anyone have any tech info???

The Ladybird guide to core store (1971)....

http://davidguy.brinkster.net/computer/016.html

Don't think so...I used to teach that stuff! I meant the Atlas
specifics...

(later..)

'S OK....this is *exactly* what I have...

http://www.chilton-computing.org.uk/...atlas/p012.htm
--
The information contained in this post is copyright the
poster, and specifically may not be published in, or used by
Avenue Supplies, http://avenuesupplies.co.uk

"Andrew Gabriel" wrote in message
...
In article ,
"Malcolm Stewart" writes:

I remember -5V being quite important back in 1960 when I was selecting
core
store driver transistors, germanium, for a Ferranti Argus computer.

We had fun with cores too at GEC -- they also used +18V and -18V.
The PSU's were one of the more interesting sides. In order to
guarantee no data loss during power outage (core is non-volatile),
you have to be very careful how the power rails are shut down so
that no spikes are delivered to the core store which might corrupt
the data, or activate any of the driving logic as the power rails
decay below their operating spec. In order to do this, the power
supplies were switched off by shorting their outputs in a timed
shutdown sequence, which took around 50ms to shutdown all rails.
This meant the PSU's were effectively switched off by activating
their over-current protection, which always made me smile.


Which bit of GEC were you with Andrew?
I spent one summer around the late sixties in Coventry Telephone works car
park, building mobile telephone exchanges destined for Iraq. Even in those
days exchanges were blown up and a mobile would be driven in and cabled up
so minimising disruption.

"Bob Eager" wrote in message
...
On Sun, 21 May 2006 08:25:10 UTC, "CWatters"
wrote:
The Ladybird guide to core store (1971)....

http://davidguy.brinkster.net/computer/016.html

Don't think so...I used to teach that stuff! I meant the Atlas
specifics...

Well yes that was kind of tounge in cheek.

I worked at GEC Computers in Elstree from 1981-1984. I can't really remember
for sure but I think they were still supporting (possibly even making?) core
store memory cards for military computers of the day.

In uk.d-i-y Andy Burns wrote:
Andy Luckman (AJL Electronics) wrote:

I believe the -5v rail is no longer used. It mat have been RS423 which uses
+-5v rather than RS232 +-12V.

Actually some laptops now use +/- 5V (presumably using a DC-DC
converter) instead of +/- 12V on their serial ports, most equipment is

There are many, many chips that provide RS232 ports from one +3V or one
+5V rail.

kimble wrote:
Andy Burns wrote:

Certainly true that 5+ year old Dells use non-standard 20pin connector
:-(


Or, more accurately, a standard connector with a non-standard pinout.
Recipe for disaster that it is.

A decent PSU doesn't worked plugged into a dell mobo. A cheap PSU never
works again after being plugged into a dell mobo. Guess how I know!
Pinouts are on the web if you don't mind a bit of soldering (splice the
cables from the new PSU to the old connector)-
http://www.formfactors.org/ will have the ATX specs.


--
Spamtrap in use
To email replace 127.0.0.1 with blueyonder dot co dot uk

On Sun, 21 May 2006 16:17:23 UTC, "CWatters"
wrote:


"Bob Eager" wrote in message
...
On Sun, 21 May 2006 08:25:10 UTC, "CWatters"
wrote:
The Ladybird guide to core store (1971)....

http://davidguy.brinkster.net/computer/016.html

Don't think so...I used to teach that stuff! I meant the Atlas
specifics...

Well yes that was kind of tounge in cheek.

Sorry!

I understand that every IBM employee (no matter who) was given a copy of
that book at some point in the past...

I have used machines with core, but not for a while. Nice to have an
example, and a fairly historic one at that...!

--
The information contained in this post is copyright the
poster, and specifically may not be published in, or used by
Avenue Supplies, http://avenuesupplies.co.uk

Andy Burns wrote:
Andy Luckman (AJL Electronics) wrote:

I believe the -5v rail is no longer used. It mat have been RS423 which
uses
+-5v rather than RS232 +-12V.


Actually some laptops now use +/- 5V (presumably using a DC-DC
converter) instead of +/- 12V on their serial ports, most equipment is
happy with the reduced voltage swing but some isn't, this makes those
laptops less useful as a roaming terminal for diagnosing problems on
"random" routers/switches :-(


You're lucky to get serial on a laptop at all these days (but at least
the keyspan USB-serial adaptors seem to work).

--
Spamtrap in use
To email replace 127.0.0.1 with blueyonder dot co dot uk

The message
from "CWatters" contains these
words:

The Ladybird guide to core store (1971)....

http://davidguy.brinkster.net/computer/016.html

If only they wrote kids' books as well as that now.

I downloaded a copy of this
http://www.mininova.org/tor/116968
a few days ago. Allegedly some of it is "highly dangerous". Well, I
suppose some of it is, but with a bit of common sense (avoiding the
carbon tetrachloride experiments, for a start) it's mostly OK. Certainly
more fun than the stuff you get these days.

--
Skipweasel
Pay no attention to that man behind the curtain.

On Fri, 19 May 2006 17:55:57 +0100, Jules wrote:
It's probably a hang-up from the original IBM PC days, but even then I'm
not sure what would have used it - memory chips were all single-rail
devices by then. I expect its inclusion goes back further than that, to
machines in the 70's that did need -5V for the memory, and IBM included it
in their machine "just in case".

Hmm, I'm going to have to go away and ask now, though.

OK, so I asked...

Apparently the original 16-64K IBM PC motherboards used 4116 memory chips
which *are* 3-rail devices, so I was having a brain fart... (I'll have to
check my original PC when I'm back in the UK to see what it has)

Anyway, so that's why -5V is there...

cheers

Jules

On Sat, 20 May 2006 14:24:36 +0000, Bob Eager wrote:

On Sat, 20 May 2006 12:37:20 UTC, (Andrew
Gabriel) wrote:

We had fun with cores too at GEC -- they also used +18V and -18V.

I was given some core this week! Looks like 512 bytes (except it
wouldn't have been bytes)...64 x 64 AFACS. From an Atlas, no
less....anyone have any tech info???

Don't the Science Museum have an Atlas? We've got a whole stack of
Pegasus manuals, but AFAIK no Atlas documentation.

Word lengths were often *really* oddball on the older machines - usually
somewhere up in the 30-80 bits range, and often odd numbers too. Our
Elliott's 39 bits IIRC, and I think the Marconi is 40. 64 bits is
certainly possible.

cheers

Jules

Ah I remember when companies had component stores....and a little man in
brown overall to run them. However did we manage without RS Components and
overnight delivery? One of my first projects after leaving university was to
build a memory card for a small data logger. The store only had 7 one bit
wide memory chips and a jobsworth in stock. By the time I'd done all the
necessary paperwork to order the 8th bit and then waited for it to be
delivered .... I'd left the company..

"Jules" wrote in message
news

Word lengths were often *really* oddball on the older machines - usually
somewhere up in the 30-80 bits range, and often odd numbers too. Our
Elliott's 39 bits IIRC, and I think the Marconi is 40. 64 bits is
certainly possible.

A lot of the early memory cards had error correction on them so it was
frequently wider than the processor.

"Ian Stirling" wrote in message
...

Actually some laptops now use +/- 5V (presumably using a DC-DC
converter) instead of +/- 12V on their serial ports, most equipment is

There are many, many chips that provide RS232 ports from one +3V or one
+5V rail.

Actually it's quite hard to find a laptop with a serial port built in now.

CWatters wrote:
Ah I remember when companies had component stores....and a little man
in brown overall to run them. However did we manage without RS
Components and overnight delivery? One of my first projects after
leaving university was to build a memory card for a small data
logger. The store only had 7 one bit wide memory chips and a
jobsworth in stock. By the time I'd done all the necessary paperwork
to order the 8th bit and then waited for it to be delivered .... I'd
left the company..

Why didn't you just use two nibbles?

Jules wrote:

Word lengths were often *really* oddball on the older machines - usually
somewhere up in the 30-80 bits range, and often odd numbers too. Our
Elliott's 39 bits IIRC,

Interesting... which Elliott have you got? The 803 was the first
computer I ever used. 39 bit words indeed, with two instructions per
word: octal op codes, 16-bit absolute addresses and a modifier bit in
the middle for indirect addressing (referred to as the B digit). 5-hole
paper tape, or the keyboard push-buttons, for input and tape punches or
a Creed 75 teleprinter for output. 8K words of core, mercury delay line
registers and its infamous Algol 60 compiler. A fine machine :~)

--
Andy

Jules wrote:

Apparently the original 16-64K IBM PC motherboards used 4116 memory chips
which *are* 3-rail devices, so I was having a brain fart... (I'll have to
check my original PC when I'm back in the UK to see what it has)

Anyway, so that's why -5V is there...

.... or *was* there, it now seems. Thanks for that enlightenment.

--
Andy

On Mon, 22 May 2006 23:22:55 UTC, Andy Wade
wrote:

Jules wrote:

Word lengths were often *really* oddball on the older machines - usually
somewhere up in the 30-80 bits range, and often odd numbers too. Our
Elliott's 39 bits IIRC,

Interesting... which Elliott have you got? The 803 was the first
computer I ever used. 39 bit words indeed, with two instructions per
word: octal op codes, 16-bit absolute addresses and a modifier bit in
the middle for indirect addressing (referred to as the B digit). 5-hole
paper tape, or the keyboard push-buttons, for input and tape punches or
a Creed 75 teleprinter for output. 8K words of core, mercury delay line
registers and its infamous Algol 60 compiler. A fine machine :~)

Not long ago, there was a working 803 at Bletchley Park. If they get the
project off the ground, I'm supposed to provide software for a 2900 they
have been promised.

(I just missed the 803 at work; I arrived a year too late and started on
the Elliott 4130).

--
The information contained in this post is copyright the
poster, and specifically may not be published in, or used by
Avenue Supplies, http://avenuesupplies.co.uk

In message , at
22:47:57 on Mon, 22 May 2006, CWatters
remarked:

Actually it's quite hard to find a laptop with a serial port built in now.

Mine has two: USB and 1394 (Firewire).
--
Roland Perry