Re: Black screen on boot: MB or PS defect?
- From: Paul <nospam@xxxxxxxxxx>
- Date: Sun, 16 Oct 2011 00:17:20 -0400
Tom McCreadie wrote:
Thanks for the typically thorough reply, Paul. You bring up a lot of
considerations that I still have to work my way through.
My 'Clear CMOS' procedure was: disconnect MB power > move a MB jumper from 1-2
to 2-3 for a few secs, then back to 1-2 > restore power. But I did not remove or
replace the CMOS battery. OK?
I get no Winbond Voice audible error message and also get no PC case speaker
beeps when I boot with no memory installed. Ditto when booting with all memory
re-installed. (<- all these tests were run with video card installedI). My
Winbond Voice is presumably still active, as I recall using it during my system
build; but I guess I need to dredge up the handbook of my Antec P160 case to
check if my case speaker is still properly connected :-) All the caps on the
MB appear fine. I still have to inspect the innards of the Antec or measure the
voltage rails...but at least the MB and case fans, drive bay lights etc. all
FWIW, on a few occasions (generally unconnected to my aforementioned fuzzy
screen issue ) when my system has hung and needed a power-off reset, I often
have seen an errror message on reboot: "system overclocking failure". I found
that mystifying as I never overclock. But maybe that was linked to my
impatient, multiple-pressing of the case 'reset' button. (I recall a thread back
in June where you pointed out the consequences to the bios of using the 'reset')
If you've verified the case speaker is connected (to listen for a beep pattern),
you can try again, to remove both memory and video, and listen for beeps.
If it won't beep, one possible reason, might be if it was "stuck in reset".
The power supply has a "power good" type signal, which must be asserted
by the power supply, before the system will start. You might hear the
power supply fan start, the case cooling fans start and so on, and yet
no other reaction. The power good signal is tied into the reset logic,
and the system comes out of reset, at some time after that signal is
asserted. A similar thing might happen, if you pushed on the reset button
too hard, the reset button was cheap, and got crushed in the "ON" state.
You could check the power good signal with a multimeter, on the main ATX
cable, while the system is running. (You can place a probe into the nylon
shell and touch the wire, with some effort.) You could also try that,
with the power supply removed and sitting on a table top, but then that
isn't as good as an "in-system" test.
I'd probably also want to check a few of the supply pins. You can access
+5V and +12V on a Molex disk connector. The +3.3V would be on the main
connector, or if your supply has it, it might be one of the pins on the
1x6 aux connector (seldom used). Again, I'd do that in-system, to save
a little time. (I don't recommend in-system testing, if the power supply
has made it obvious it's sick. If the power supply has emitted any
"magic smoke", then all testing should be done with the power supply
removed from the system. But since you don't report any symptoms pointing
at the supply, then testing it in-system should be OK.)
The Winbond, as far as I know, would be enabled after you've reset the
CMOS. You may be able to verify that, by looking at the user manual,
in the BIOS section. The Winbond output, is sometimes capacitively coupled
to the motherboard sound Line_Out lime green colored connector, and may be
coupled into one of the stereo sound channels. If you're using a separate
SoundBlaster, and have speakers hooked up to that, then you'd miss it.
Move your computer amplified speakers over to the motherboard sound,
and listen for some of that "illegible" voice audio. I use the tempo and
intonation, comparing to the list of errors in the manual, to kinda guess
at what she's saying :-)
The Winbond works, even when there is no CPU present. If you pull the CPU
from the motherboard socket, the Winbond has one message that detects
there is nothing plugged into the socket. Two of the messages are timer
based (Winbond timer). If the CPU doesn't clear the timer, within the
timeout period, then those messages will come from the speakers. The
rest of the messages, require proactive BIOS activity, writing something
into the Winbond chip, to get the audio message delivered. The Winbond
has its own processor (state machine) inside, which is why it can clock
out an audio message, when everything else is dead. But if the reset
signal is stuck ON, it's possible even the Winbond will be silent.
I can't remember whether it's connected to reset or not...
Between either the Winbond feature (and your computer amplified speakers),
or the beeping on the case speakers, you should be able to get some kinda
feedback as you remove stuff and test.
- Re: Black screen on boot: MB or PS defect?
- From: Tom McCreadie
- Re: Black screen on boot: MB or PS defect?
- Prev by Date: Re: Building your own computer
- Next by Date: Re: Black screen on boot: MB or PS defect?
- Previous by thread: Re: Black screen on boot: MB or PS defect?
- Next by thread: Re: Black screen on boot: MB or PS defect?