Norm Why wrote:
I need more info on what you mean by "Continuing CPU reboot cycle"
PC knows what it is doing, it recycles in short time. Power comes on, fans
turn, RAM phase lights glow. The PC shuts down. A moment later, PC comes
back to life and cycles ad finitum.
This is an improvement over two RAM DIMMs inserted when cycles ends after ~3
tries. These tries take long because CPU is confused.
Normally, when a PC shuts down like that, it's caused by THERMTRIP.
That's a signal from the CPU, that it's past 100C or so. It's
supposed to cause the PC to go off and stay off. Clearance of
such a fault requires removal of +5VSB (switch off at the back
of the PC).
You could also cause a shutoff, by overloading +5VSB. The +5VSB
runs the USB ports and it's only rated at 2.5A or 3A or so. But
I would expect that fault to latch off too. Only if the BIOS
was set to a particular "recovery policy when power comes back"
in the BIOS, might such a flaw result in a loop.
I don't have a good explanation that precisely fits the
symptoms. There are things that can cause it, but generally
the loop should be broken.
Discovery that the processor is not one the BIOS supports,
might do that, but I don't recollect any recent results
to refer to. My P2B-S for example (year 2000), didn't know what a
Tualatin was, but I fixed that by loading a different
BIOS and doing a microcode patch. The "ID string" on the
BIOS screen was still wrong (called it a Pentium II or
something), but at least the CPU was "accepted" and boot
would finish.
On page 48 of manual it says PCI graphics is the default. Without PCI
graphics one will never see the BIOS screen to change the default.
Wrote to Cougar.
That setting allows you to select which bus to examine first
for a video card. That motherboard has only PCI bus and PCI Express
bus, and has no AGP bus. And since it's a P45, at a guess
there are no graphics. P45, the P likely stands for Performance,
and a gamer would use their own video card. Other letters
they might use, would include G45 or Q45, and SKUs like
that would have a GPU in the Northbridge. Q45 is a designation
used for boards including the Intel Management Engine (ME)
for remote control of the PC via NIC packets. (The refurb Optiplex
I've got, has a Q45.)
Even without using the setting, if there is no PCI Express card,
it should still examine the PCI bus when it gets around to it.
At the current time, with the looping behavior you've got, I
don't think it is getting near enough to the conditional beep
codes, to issue any beep codes. Something is wrong in the
(register-only based) BIOS code. And whether a hardware signal
is taking it out of action, or the BIOS code itself is having
a wobble for itself and resetting, we're just not getting to
the beep code section.
Even if you had a PCI Port 80 card, the Port 80 display might not
update with new digit values before the looping hits it. Being
able to monitor the voltages, might give a hint if the problem
is power related, but you need the right electronics trash
to catch stuff like that.
*******
You can see there is something that isn't quite right
about these boards. But, what is it ? For one guy, it was
the switch to a PCI graphics card that allowed him to see
the screen and get on with it. Any old PCI graphics card
would do for this - there is no need for anything fancy.
like a 6200 PCI or a FX5200 PCI would be candidates.
The 6200 and the FX5200 PCI were available for quite
a while before they stopped making them. (The 6200 was
notable, in that it had native support for more
than one bus standard, which made it cheaper to make
for the manufacturers. It didn't have to be as expensive
as a bridged board.)
https://forums.tweaktown.com/gigabyt...ot-loop-2.html
Does yours loop with both RAM and video removed ?
Don't forget to remove all power, before adding
or removing those...
Paul