AT/ATX PS question

[Pelysma]

The computer on my workbench right now has an AT form factor, but the
motherboard has connectors for both AT and ATX power supplies. I want to
replace the 230w Linksys AT PSU that's in there now with a newer L&S 250w
ATX PSU that I have sitting here.

What I know about the system is what I can see looking at it. Here's what I
can collect:

VA-503+ motherboard
VIA chipset; the next line on the Northbridge chip is VT82C598MVP
K6-2 400MHz on 100MHz FSB
128 MB RAM in two 64-MB 168-pin DIMMs; four slots for 72-pin EDO are empty
6.4 GB IBM ATA HDD
Windows 98SE upgraded from W95

I'm adding two more hard drives, 8 GB each, but I'm keeping the 6.4 as the
system drive because there is software on it I want to use for a while and
have no way to install on another OS (notably MS Access 97). I'll be using
the Reset button as a start button as mentioned in another thread.

The computer has a role to play in my collection but won't really be turned
on very often.

Is there anything I need to know, or anything I need to reconfigure, before
unplugging the AT PSU and plugging in the ATX one?

 

[Floyd L. Davidson]

Is there anything I need to know, or anything I need to reconfigure, before
unplugging the AT PSU and plugging in the ATX one?

How is the front panel ON/OFF power switch wired?

With an AT supply there will be a (probably black) AC cord coming
out of the supply going to the ON/OFF switch. That cord carries
120VAC, and it is switched at the front panel.

With an ATX supply, no such AC cord exists, and instead the
green wire on the connector (routed through pins on the
motherboard) is switched to ground by the front panel ON/OFF
switch, and that causes the AC to be remotely switched inside
the power supply.

Very different arrangements, which have more to do with the case
and the power supply than with the motherboard. Your
motherboard was apparently designed to work with either. Which
probably means you could switch it to an ATX case. But swapping
the AT supply in an AT case to an ATX supply isn't that easy.

 

[Pelysma]

How is the front panel ON/OFF power switch wired?

With an AT supply there will be a (probably black) AC cord coming
out of the supply going to the ON/OFF switch. That cord carries
120VAC, and it is switched at the front panel.

With an ATX supply, no such AC cord exists, and instead the
green wire on the connector (routed through pins on the
motherboard) is switched to ground by the front panel ON/OFF
switch, and that causes the AC to be remotely switched inside
the power supply.

Very different arrangements, which have more to do with the case
and the power supply than with the motherboard. Your
motherboard was apparently designed to work with either. Which
probably means you could switch it to an ATX case. But swapping
the AT supply in an AT case to an ATX supply isn't that easy.

Well, they both physically fit. I know about the AC cord to the front; I
also know that the power switch on the front of an ATX case is identical to
the reset switch on both, a momentary spst pushbutton in series with a small
resistor. In most ATX boards there is a pair of pins labeled RST for the
reset switch, and a pair labeled SOFT PWR for the power switch. I've
located the RST pins and the wire connected to them.

Usually if the pins aren't labeled I go in there, determine some likely
candidates, and briefly short each likely pair with a tool I made, a
1000-ohm resistor plugged into a little red jumper-style connector and
epoxied to a popsicle stick. (I used to use a Phillips screwdriver, but had
second thoughts about shorting the wrong thing without resistance.)

The pair that starts the computer gets the wire from what was the reset
switch.

So far on this board I haven't found a pair that starts the machine. But
there are some jumpers around the switch headers and around the power
connectors that I can't identify, and I don't have TFM for the board at the
moment.

Anyway, thanks for your input.

 

[kony]

The computer on my workbench right now has an AT form factor, but the
motherboard has connectors for both AT and ATX power supplies. I want to
replace the 230w Linksys AT PSU that's in there now with a newer L&S 250w
ATX PSU that I have sitting here.

What I know about the system is what I can see looking at it. Here's what I
can collect:

VA-503+ motherboard
VIA chipset; the next line on the Northbridge chip is VT82C598MVP
K6-2 400MHz on 100MHz FSB
128 MB RAM in two 64-MB 168-pin DIMMs; four slots for 72-pin EDO are empty
6.4 GB IBM ATA HDD
Windows 98SE upgraded from W95

I'm adding two more hard drives, 8 GB each, but I'm keeping the 6.4 as the
system drive because there is software on it I want to use for a while and
have no way to install on another OS (notably MS Access 97). I'll be using
the Reset button as a start button as mentioned in another thread.

The computer has a role to play in my collection but won't really be turned
on very often.

Is there anything I need to know, or anything I need to reconfigure, before
unplugging the AT PSU and plugging in the ATX one?

Download the motherboard manual and it will reveal the ATX
power-on pin header. It is likely there is a jumper you
change to switch from AT to ATX power supply and it could be
important to change that jumper before trying it with the
ATX PSU. Otherwise it should only be an issue of what
you'd mentioned, the need to use the reset switch for power
switch.

 

[ric]

With an ATX supply, no such AC cord exists, and instead the
green wire on the connector (routed through pins on the
motherboard) is switched to ground by the front panel ON/OFF
switch, and that causes the AC to be remotely switched inside
the power supply.

Not quite accurate. The PS_ON (green) wire goes to a logic circuit
on the MB. The on/off switch goes to this circuit as well. Engaging
the momentary on/off switch toggles the output of this logic circuit,
changing its state, and turning the PSU either on or off. (If the
logic output toggled low, the PSU turns on. If the logic output
toggled high, the PSU turns off.)

Beside, the OP said he was going to use the RESET switch for his
on/off switch. That will work just fine.

 

[Floyd L. Davidson]

Not quite accurate.

You've just repeated the same thing I said with more (and
unnecessary) detail. How is it not accurate?

The PS_ON (green) wire goes to a logic circuit
on the MB. The on/off switch goes to this circuit as well. Engaging
the momentary on/off switch toggles the output of this logic circuit,
changing its state, and turning the PSU either on or off. (If the
logic output toggled low, the PSU turns on. If the logic output
toggled high, the PSU turns off.)

That's correct detail, but it doesn't add anything at all to
understanding the point about the switch.

Beside, the OP said he was going to use the RESET switch for his
on/off switch. That will work just fine.

It will work. Of course then he doesn't have a RESET switch,
only an ON/OFF switch that is mislabeled, and an ON/OFF switch
that is labeled but doesn't work... which is a nice recipe for
confusion.

Seems to me that finding another AT power supply is a *much*
better idea. Or just switching the whole thing to an ATX case.

 

[Pelysma]

You've just repeated the same thing I said with more (and
unnecessary) detail. How is it not accurate?


That's correct detail, but it doesn't add anything at all to
understanding the point about the switch.


It will work. Of course then he doesn't have a RESET switch,
only an ON/OFF switch that is mislabeled, and an ON/OFF switch
that is labeled but doesn't work... which is a nice recipe for
confusion.

or a nice job for a sharpie.

Seems to me that finding another AT power supply is a *much*
better idea. Or just switching the whole thing to an ATX case.

or what I did, which is to replace the two damaged Molex connectors on the
AT PSU so I can install two more drives. It probably won't do what I
wanted to do with the extra drives, anyway. I do still need to transfer
something like 400 MB of images from it to a CD, and plan to keep it running
so I can use Access for a little while for another project, but it was given
to me to scrap and scrap it soon shall be.

The two pins that should have been the power-on header in the diagram are
missing from this board, so I'm not going to get anywhere with the ATX idea.

Thanks, both, for your input.

 

[Eric]

You've just repeated the same thing I said with more (and
unnecessary) detail. How is it not accurate?

You stated that the green wire is switched to ground by the front
panel switch. It is not. It is switched to ground by the OUTPUT
of a logic circuit. The front panel switch toggles the INPUT of
the logic circuit.

It will work. Of course then he doesn't have a RESET switch,

And a reset switch is used oh so often in an ATX computer.

only an ON/OFF switch that is mislabeled, and an ON/OFF switch
that is labeled but doesn't work... which is a nice recipe for
confusion.

For those who are easily confused. I was under the impression
that the OP would be operating this system. I think he would
get through the confusion.

BTW, a little simple surgery will turn a AT type ON/OFF switch
into a momentary ATX type ON/OFF switch.

 

[Floyd L. Davidson]

You stated that the green wire is switched to ground by the front
panel switch. It is not.

It *is*. (Please read what people *say*, not what you want them
to have said.)

It is switched to ground by the OUTPUT
of a logic circuit. The front panel switch toggles the INPUT of
the logic circuit.

The front panel switch is what causes it to be switched. I said
*nothing* to indicate that the physical circuit itself was
switched *through* the front panel switch. You are reading too
much into a one sentence statement of the effect.

And a reset switch is used oh so often in an ATX computer.

I've got one case that came without one. Talk about *annoying*!
I had to add one.

For those who are easily confused. I was under the impression
that the OP would be operating this system. I think he would
get through the confusion.

The original post did not make that clear at all. Certainly he
has since clarified that it is a short term use, where that
isn't terribly important. But cute little hacks that only the
perpetrator knows about, aren't conducive to great group
dynamics...

BTW, a little simple surgery will turn a AT type ON/OFF switch
into a momentary ATX type ON/OFF switch.

The OP seems to have it well in hand, and isn't bothered by
word games...

 

[VWWall]

It *is*. (Please read what people *say*, not what you want them
to have said.)

You did *say*: "causes the AC to be remotely switched...".
The AC is already on, which provides the +5 VSB. The grounding of the
green wire just turns on the DC switch circuit. Holding in the front
panel switch causes the MB logic to remove the ground and turn off the
ATX power supply.

 

[Floyd L. Davidson]

You did *say*: "causes the AC to be remotely switched...".
The AC is already on, which provides the +5 VSB. The grounding
of the green wire just turns on the DC switch circuit. Holding

"The DC switch circuit"??

in the front panel switch causes the MB logic to remove the
ground and turn off the ATX power supply.

I think you're trying a little too hard on that one.

 

[VWWall]

"The DC switch circuit"??

The pulse width modulator to the DC switching transistors. That's why
they're called "switching power supplies".

I think you're trying a little too hard on that one.

Not hard at all--that's how it works!

 

[Floyd L. Davidson]

The pulse width modulator to the DC switching
transistors. That's why they're called "switching power
supplies".

Ahem... technically that fits what I said.

Not hard at all--that's how it works!

Nobody was trying to describe the internal circuitry in the power
supply. It was intended to be a single phrase reference, not a
detailed description.

It was correct. You are trying awful hard, for nothing.

 

[ric]

The pulse width modulator to the DC switching transistors. That's why
they're called "switching power supplies".

Not hard at all--that's how it works!

Note to VWW: I have put Floyd in my "Don't wrestle with pigs" file.
[You just get dirty and the pig enjoys it.] No matter what he said
incorrectly, his response will be "That's what I said. You're trying
too hard on this one."

 

[VWWall]

Note to VWW: I have put Floyd in my "Don't wrestle with pigs" file.
[You just get dirty and the pig enjoys it.] No matter what he said
incorrectly, his response will be "That's what I said. You're trying
too hard on this one."

I found that out in alt.engineering.electrical, where he's constantly
doing the same. :-(

He must have a boring job in Alaska!

 

[Pelysma]

I think he would
get through the confusion.

Thanks for the vote of confidence.
;-)

BTW, I do understand all about these switches. I only wanted to know if I
needed to change a jumper on the motherboard before unplugging the AT PSU
and plugging in the ATX one! That's moot now, except for the pure academic
exercise of learning something new.

As I've come to expect, Kony gave the best information.