Getting there

Thanks. I should mention that at this point I have one 500 MB RAM
stick installed with the Intel heat sink cooler on the CPU. Maybe I
will try a no RAM boot. If I can get into BIOS I can tweak it. I have
BARTPE (preinstallation edition based on WinXP). On it I have the
BIOS F2 upgrade. If I can boot BARTPE and install BIOS v2.0 I'm home
free.
I removed the one RAM stick to see if I got different beep codes.
Nope. 'Continuous short beeps: Power error' takes priority. This means
the annoying beep speaker can be removed and the system can be allowed
to recycle. Paul gave a link where GA-EP45-DS3L MOBO eventually
POSTed. With Q9650, I could feel heat radiated. E4700 is cold. I can
still feel heat coming from somewhere but face is not directional.
Touching does not reveal hot spot. Going through MOBO I found a
ten-pin keyed connector not plugged in. I removed all power to
peripherals, like DVD R/W drive.

Like Louis Rossmann, I think I need to find a component that is
"shorting", (an active short?).

Where to find such tool?
The claim here is the continuous short beeps are a RAM error.
GA-EP45-DS3L ? I reseated my 500 MB RAM.

https://forums.tweaktown.com/gigabyt...t-beeps-4.html

It almost seems like a half-finished BIOS design or something.

Some RAMs working, other RAMs not working, participants going crazy
with
the custom settings.

I had less trouble with my VIA chipset board than that :-)
(That was the board where the chipset actually works with 2GB
RAM DIMMs, but the BIOS was never "tuned" to use the right
Tsu and Th and the like. When I used the 1GB sticks I had, that
board was flawless. Long Prime95 runs, no problem at all. BIOS
issues can make an awful mess when they happen.)

Paul
Thanks Paul. I got a non-contact infrared thermometer and found heat was
generated in all the usual suspects: components with heat sink
radiators. I keep the room thermostat at 15 C and allow heat from xBox
360 and Samsung plasma TV to keep it warm. Tonight outside temperature
should get down to 8 C. I'll turn off thermostat and electronics and let
the room get cold. With recycle the GA-EP45-DS3L MOBO might POST.


On the other hand, "Semiconductor materials (carbon, silicon, germanium)
typically have negative temperature coefficients of resistance." From:

https://www.allaboutcircuits.com/tex...nt-resistance/

So maybe the room is too cold for the GA-EP45-DS3L to POST?

In the fossil record, I see no sign of such a temperature related problem.

There may have been one or two chipsets long ago, where something
about the circuit was temperature sensitive. But that is by no means
a common situation. It's an outlier.

Saturating logic circuits can successfully run from -55C to way way
above the boiling point of water. They're not snowflakes, but you have
to select the right items for the job. I found a processor the other
day, for automotive usage, which goes way outside the range of anything
I've ever used.

I have a P45 here, and it's never had a problem starting. That's
the refurb Optiplex with a dual core in it.

Room temperature is the temperature it's *supposed* to work at.
Sure, some electronics have trouble at -20C (maybe LED lightbulbs
or some pedestal box for ATT), but that's considered one extreme
of the temperature range. On the high end, it's a function of
what simulation temperature the circuit was verified at. Like at work,
maybe you'd set the simulation temperature to 105C for margining,
even though you had no intention of ever running an actual chip
at that temperature. CMOS gets slower at high temperature, and you're
checking for timing failures by using a simulation temperature that
high. Then the idea is, operation at temperatures below that is
just fine and dandy.

Where circuits have problems, is in the ad-hoc circuit used for
backfeed protection and parts of the reset circuit. Some of these
rely on analog voltages and capacitors, and occasionally some
booboo in there causes a motherboard to have a temperature issue.
It's not normal for pure-digital circuits (Northbridge/Southbridge/CPU)
to become "wobbly" with temp.

On the Southbridge, the only problem I've heard of, is the RTC
and CMOS RAM, the RAM may not function at reduced voltage (CMOS
battery getting weak) and at some non-room temperature. Some
of these conditions when they happened, were considered to be
functional failures and the chipset should never have been
shipped that way. Again, this is not a common condition, and there
hasn't been a problem like that in yonks.

Have you tried your one-stick-of-RAM test, in each of the
four slots individually ? Don't forget to remove all power
from the system, before moving the stick of RAM. That means
switch off at the back or unplug the PC power cable, wait
*at least* 30 seconds for 5VSB to drain. Asus motherboards
all have a green LED onboard, that monitors +5VSB and when that
LED extinguishes, then it's safe to move RAM. I don't know
how many other brands have that.

On some motherboards, there is the chicken-versus-egg problem,
where the user needs to change BIOS versions, the board won't
come up, and you can't flash the BIOS. Your board has a Dual BIOS,
and by now, probably both sides have the same BIOS version. That
would be another variable at this point. That's what the hope of
using some other processor was about - getting the board to start
by using an alternate processor, so the BIOS version could get
changed when you wanted to change it.

For removable BIOS chips, you could flash them using a lab programmer.
But I no longer have access to stuff like that, and I don't know of
any computer stores in town who I would expect to own such equipment.

My newest motherboard here, has its own flasher onboard. You can
change the BIOS version, *without* a CPU being in the CPU socket.
There is a special USB port, you plug in a flash stick with a BIOS
image on it, there's a pushbutton on the back of the PC, you
press that, and a few minutes later, the machine has a new BIOS
version. I've never used the feature, but the feature sure was
attractive when I bought it. Because, it meant I "couldn't be
held hostage by a motherboard refusal-to-start". There is a
microcontroller next to the USB port, that reads the flash stick
and writes the BIOS chip with what it finds. What used to cost $150,
now costs a buck to do.

Paul

Thanks Paul,

I messed around with temperature enough. The MOBO has been recycling in a
hot room. The infrared thermometer allows me to do controlled but pointless
experiments with a hairdryer. I have now taken some voltages. Where I
removed a 470 uF 16V capacitor allows me to measure 12V rail. It goes up to
12V steady each cycle. The vacant PCI slot allows me to measure 5V again.
Now it shows 0V, no 5V. No 5V explains the beep codes:

Q: What do the beeps emitted during the POST mean?
A: The following Award BIOS beep code descriptions may help you identify
possible computer problems.
(For reference only.)
1 short: System boots successfully
2 short: CMOS setting error
1 long, 1 short: Memory or motherboard error
1 long, 2 short: Monitor or graphics card error
1 long, 3 short: Keyboard error
1 long, 9 short: BIOS ROM error
Continuous long beeps: Graphics card not inserted properly
Continuous short beeps: Power error

No 5V is a major power error. But why now?