Re: OT?: Major PC hardware failure

Puddin' Man wrote:
On Sat, 09 Apr 2011 16:25:19 -0400, Paul <nospam@xxxxxxxxxx> wrote:

"Without cpu/mem (or anything else), I power up, nothing happens."

Do you mean, you press the button on the front, the PSU fan spins,
the system is powered ? Or it won't start at all (no PSU fan) ?

The latter.

I work these tests from the PSU on/off. If it doesn't start, I
work the soft-switch. The PSU-on bought me the green LED.
I got nothing more after working the soft-switch.

As far as I know, all my motherboards here, start when the CPU socket
is empty. But they could always set it up, I suppose, to not come
on. This is Intel after all, and they could do something different.

'Twere my understanding that they usually -set- the standards, and
don't break 'em unless setting a new one. But, anything's
possible.

This is an example. How many other brands of motherboard
do it this way ?

"BIOS Recovery Mode jumper"
http://www.intel.com/support/motherboards/desktop/sb/CS-023360.htm


To your knowledge, if one or more of the 478 socket pins were welded
to another during the "thermal event", we could conceivably see the results I've gotten?

Thx,
P

The S478 is a ZIF (zero insertion force) socket. The contacts get
some minor protection from the plastic around the pins. VCore can
probably provide up to 100 amps. Would the touching of two power
pins together, result in the pin being burned off ?

When there is no processor installed in the socket, the VID
signals go to all 1's, the Vcore regulator is asked to put out
zero volts. And that effectively means Vcore is switched off,
because there'd be no gate drive on the MOSFETs. If Vcore is
switched off, a short circuit wouldn't even be detectable.

If an Intel designer wanted to, they could sense a single pin
in the socket, and decide if no processor is inserted, to gate
off PS_ON#. I don't know why you'd want to do that, but it's an
option. All I can say, is the couple brands of motherboards
I've got here, if you do the bare motherboard test, you can
turn the PSU on and off. (Which is great, as a basic verification
that the PS_ON# signal is working, coming from the motherboard.)

At this point, you'd have to reach for the multimeter. Your first
task, would be to set the multimeter to volts, connect the black
lead to the chassis (I clamp onto an I/O screw with an alligator
clip). Then, with the red lead, you'd touch PS_ON# on the main
connector and make a voltage reading. You can jam the probe tip,
in where the wire goes into the nylon connector body (as there
is exposed metal in there you can touch with the probe tip).
What you'd be looking for, is whether the PS_ON# signal is making
good logic levels. Like, when it switches on the PSU, is PS_ON#
dropping below 0.8V like it should ? Or does it hover around
1.5 to 2.0 volts ? When the motherboard wants the PSU to be off,
the PS_ON# signal should go all the way up to 5.0 volts. It could
be that the motherboard driver of PS_ON#, isn't making good levels,
in which case you might want to try another motherboard.

Looking at the Intel reference schematic, they drive PS_ON#
directly from the SuperI/O. So there really isn't much more
probing you can do. Older motherboards (like the Intel reference
schematic for 440BX), have a separate external chip for driving
PS_ON#, so you can use your multimeter to probe both the
input and output on that chip, to verify good logic signals
are present.

Logic levels for PS_ON# can be found in the ATX spec, an example
of which is Figure 8 on page 27. And because the ATX power supply
doesn't use an actual logic gate, it doesn't necessarily respond
as shown. (If you have a signal in the undefined region, at
least some power supplies end up in the "half-on" state, where the
outputs work, but weakly.)

http://www.formfactors.org/developer/specs/ATX12V_PSDG_2_2_public_br2.pdf

(In the sample ATX schematic, transistor Q10 senses PS_ON#, so the
signal isn't sensed here directly, by a logic gate. So this
supply might do something goofy, if the PS_ON# is in the undefined
region.)

http://www.pavouk.org/hw/en_atxps.html

*******

The thing is, we can't know, just by looking at the motherboard, how many
safety systems are wired into PS_ON#. An obvious choice is THERMTRIP,
and all motherboards should be using that now. If the processor
sends THERMTRIP, the power should go off immediately. But many other
things could potentially be included, such as an overcurrent signal
coming from VCore.

In some cases, the inclusion of some of these safety features, results
in unnecessary product returns to retailers. You really only want
the safety feature to trigger, when a serious fault is happening.
Nuisance trips, won't win you any friends.

Paul
.

Relevant Pages

  • Re: OT?: Major PC hardware failure
    ... Motherboard circuit runs from +5VSB. ... VID signals prevent Vcore from running (it's asked to produce a zero ... If it can't stay running then, it's either motherboard or PSU. ... and if that were the case, it would be the fault of the power supply. ...
    (alt.comp.periphs.mainboard.asus)
  • Re: OT?: Major PC hardware failure
    ... Motherboard circuit runs from +5VSB. ... VID signals prevent Vcore from running (it's asked to produce a zero ... If it can't stay running then, it's either motherboard or PSU. ... and if that were the case, it would be the fault of the power supply. ...
    (alt.comp.periphs.mainboard.asus)
  • Re: OT?: Major PC hardware failure
    ... With cpu/mem, I power up, fans spin for a second, then it shuts down. ... Motherboard circuit runs from +5VSB. ... VID signals prevent Vcore from running (it's asked to produce a zero ... If it can't stay running then, it's either motherboard or PSU. ...
    (alt.comp.periphs.mainboard.asus)
  • Re: Najor Problem Please Help
    ... ASUS M2N-E Motherboard ... HIPER M-SERIES 530w PSU ... (apparantly the asus geforce 7600 GS silenced card doesnt require a power ... the Power Switch etc. ...
    (uk.comp.homebuilt)
  • Re: Dead Mobo or PSU?
    ... cleared out a bunch of dust and decided to buy a new heatsink and PSU ... was searching for a new driver to the new 8600 GTS, ... download and install that driver while the system was still shutting ... motherboard wasn't even going through post anymore. ...
    (alt.comp.hardware.pc-homebuilt)