Norm Why wrote:
This is a continuation of an old thread:

Paul suggested I try an Intel E4700 CPU. So far I have only acquired a
couple of Intel E4800 CPU(s), TDP 65 W. I cannot feel heat emitted, like
with Intel Q9650 CPU. With it, PC would try to boot 3 times then give up.
Now it tries forever, but gives beep code; Continuous short beeps: Power
error. No BIOS screen.

This is progress. I must try harder to find Intel E4700 CPU. I'll re-solder
new 470uF 25V Aluminum Electrolytic Capacitor



Those are probably E8400 processors, which are fine
as well. The E8400 could have a higher FSB value.

One way to find "friends" of a CPU, is via cpu-world.
The table at the bottom of the page, shows processors
for the same socket (LGA775). The E4700 should be
"cheaper" as it is gutless, but likely draws
the least power of the lot. Back when I had the e4700
set up on a VIA board, the whole PC drew 65W. That's
because everything in the PC was gutless.

http://www.cpu-world.com/CPUs/Core_2...7E4700%29.html

You look the individuals up on "ark".

https://ark.intel.com/content/www/us...0-mhz-fsb.html

Processor Base Frequency 2.60 GHz
Bus Speed 800 MHz 200MHz (four txfr per clk)

https://ark.intel.com/content/www/us...3-mhz-fsb.html

Processor Base Frequency 3.00 GHz
Bus Speed 1333 MHz 333MHz (four txfr per clk)

The memory clock is related to the CPU clock choice.

These processors don't need a lot of VCore power.
36W to 43W should do, at VCore (12V @ 3A or 12V @ 4A approx).

As long as the processors in question are in the CPU Support
chart, everything is great from that perspective.

The power will beep, if the HWmonitor spots a
rail which is out of spec. Some of the rails are
like "VCore". When you change CPUs, the BIOS should
be checking the VID values for the processor and so on,
and doing the right things.

To give an example of "doing the wrong thing", take
my P4B with 1.8 GHz Northwood in it. When the BIOS
battery is dead, it re-computes start conditions
and so on. It applies *1.75V* to the CPU, because it
is mis-interpreting something. I have to hurry and
enter the BIOS, and set VCore to 1.5V again, save
and exit settings, and now it's back to 1.5V, like
it says on the box. This is *dangerous* because
one of the processors in that era, had "instant death
syndrome" at that wrong voltage.

The Q9650 should not be like that, or doing that.
This occurred on some previous VRM/VID scenario.

This has also happened on some AMD motherboards,
where the idiotic BIOS uses a previous-family voltage
and it's way way off. Fortunately, some of those
could take 2.5V or so, and the error merely made
the CPU get hot. The CPU itself was in no danger.

When Intel releases info on processes at ISSCC,
sometimes the headroom isn't that high. This is
why on systems later than yours, memory rated as
1.5V nominal, is not to be run over 1.65V. There's
some reliability issue if you ignore that warning.
And it's not always the interface itself. Sometimes
if you raise memory higher than VCore, a phantom diode
forms in the substrate, between power distribution
"rings", and power flows where it should not.
A lot of silicon devices have "power sequence"
requirements and when devices like the Southbridge
(five rails) come up, it's a bitch to keep all those
things in the right sequence. That was one of my
most hated jobs at work, finding two chips I
wanted to use, had different sequence requirements.
And you couldn't satisfy both chip requirements
at the same time.

*******

Summary:

The nature of your problem hasn't changed.

I could find one thread, where they cleared power
error by "clearing CMOS". Do that with *PSU unplugged*.

For the person here, it was "dogged persistence". You can
see the power beep might have been a false positive here
(maybe it should have beeped something else).

https://forums.tweaktown.com/gigabyt...ast-beeps.html

Paul