New MOBO, no POST.....what's the deal with socket 775?

I've been building computers on and off for 10 years and I've never
had so much problems as I've had with socket 775 motherboards. I have
two of them right now that have the same problems and I'm stuck hard.
Here's what's happened:

1.) I have a MSI motherboard that I bought as a motherboard combo from
TigerDirect. I did everything by the book, kept the 'cap' on the
socket while I installed the processor, took extreme care grounding
myself to the case before I installed the processor, made sure there
were no bent pins, installed the processor in the socket with the
tabls lined up right and pin 1 int he 'clipped corner' of the socket,
locked the arm down on the processor, lifted the 'cap' off, put a thin
coat of heatsink compound on the processor case and installed the fan
to lock down hard to the processor and board - looked great and fan
was verified tight to the mobo. Installed the mobo, cabled everything
up and doubled checked everything (24 pin ATX, 12V connector for
processor, front panel connectors, USB, etc). Installed the memory,
cabled up a mouse, keyboard, and AC. Flipped the switch on the power
supply, hit the switch on the front panel watched the fan turn on on
the processor and...........NOTHING. No POST, no customary 'beeps'
that you normally get, no video, no lights on the keyboard.
Immediately checked the power supply - all the voltages (with the
exception of the -5V DC) were there - even the pin on the power supply
that says all the power supply voltages are tested ok (active high)
read fine. Read a post that -5V isn't needed with the new
motherboards, but just to be sure, I swapped power supplies with
another one........same thing. No keyboard lights, fan spins on the
processor, no beeps, no video, no POST, no joy.

2.) Did the same thing with an ECS 945GCT-M/1333 (V3.0) motherboard -
- and a NEW processor. Same power supply (the one in the new case).
Same drill, same thing. DEAD. Fan spins, no video, POST, no joy. I
did take out both sticks of memory just to **** off the board and it
did some kind of 3-beep alarm - then nothing. Fan spins, no video,
etc, etc. Took same precautions as I did with the first motherboard.
Went overboard to make sure this time, too.

What's happening here? I'm completely stymied by all this. My trusty
Intel 945 motherboard (not a socket 775) was NEVER this hard to set up
- - worked like a champ right out of the packaging and it's the
computer I'm using right now to post this message. But these two
socket 775-based motherboards are giving me the SAME issue and driving
me nuts. I want Intel Dual Core, but this is ridiculous that two
boards would be the same in terms of the failure. What gives?

Please let me know if you've run into this before - - this is really
an SOS here.......thanks!

Marty_KN0CK wrote:
> I've been building computers on and off for 10 years and I've never
> had so much problems as I've had with socket 775 motherboards. I have
> two of them right now that have the same problems and I'm stuck hard.
> Here's what's happened:
>
> 1.) I have a MSI motherboard that I bought as a motherboard combo from
> TigerDirect. I did everything by the book, kept the 'cap' on the
> socket while I installed the processor, took extreme care grounding
> myself to the case before I installed the processor, made sure there
> were no bent pins, installed the processor in the socket with the
> tabls lined up right and pin 1 int he 'clipped corner' of the socket,
> locked the arm down on the processor, lifted the 'cap' off, put a thin
> coat of heatsink compound on the processor case and installed the fan
> to lock down hard to the processor and board - looked great and fan
> was verified tight to the mobo. Installed the mobo, cabled everything
> up and doubled checked everything (24 pin ATX, 12V connector for
> processor, front panel connectors, USB, etc). Installed the memory,
> cabled up a mouse, keyboard, and AC. Flipped the switch on the power
> supply, hit the switch on the front panel watched the fan turn on on
> the processor and...........NOTHING. No POST, no customary 'beeps'
> that you normally get, no video, no lights on the keyboard.
> Immediately checked the power supply - all the voltages (with the
> exception of the -5V DC) were there - even the pin on the power supply
> that says all the power supply voltages are tested ok (active high)
> read fine. Read a post that -5V isn't needed with the new
> motherboards, but just to be sure, I swapped power supplies with
> another one........same thing. No keyboard lights, fan spins on the
> processor, no beeps, no video, no POST, no joy.
>
> 2.) Did the same thing with an ECS 945GCT-M/1333 (V3.0) motherboard -
> - and a NEW processor. Same power supply (the one in the new case).
> Same drill, same thing. DEAD. Fan spins, no video, POST, no joy. I
> did take out both sticks of memory just to **** off the board and it
> did some kind of 3-beep alarm - then nothing. Fan spins, no video,
> etc, etc. Took same precautions as I did with the first motherboard.
> Went overboard to make sure this time, too.
>
> What's happening here? I'm completely stymied by all this. My trusty
> Intel 945 motherboard (not a socket 775) was NEVER this hard to set up
> - - worked like a champ right out of the packaging and it's the
> computer I'm using right now to post this message. But these two
> socket 775-based motherboards are giving me the SAME issue and driving
> me nuts. I want Intel Dual Core, but this is ridiculous that two
> boards would be the same in terms of the failure. What gives?
>
> Please let me know if you've run into this before - - this is really
> an SOS here.......thanks!

Well, your two results are different. The second board, the
ECS 945GCT-M/1333 made more progress, because you were able to
get a beep code from it. That is one of the tests for a board,
the "removing the RAM" test. You should get a beep code that
says the RAM is missing. To generate the beep code, the
processor executes some BIOS code, then programs one of the
chips to make the beep. (The BIOS can use processor registers
for intermediate storage, until the RAM is set up and working.
That is how it manages to do this.)

You should try the missing RAM test on the first board, and
see if it gets as far.

If there is a separate video card involved, you can install
your RAM, and remove the video card, then listen for the
"missing video" beep. It should be a different pattern than
the RAM one. If you hear "missing video", that tells you the
processor is happy with the RAM. If the processor continues
to give the failed RAM beep pattern, then you know it still
isn't happy with the RAM. So the two possible responses of
a beep test, can help distinguish what is not working right.

Bad RAM can also cause no beeps at all. Sometimes a bad RAM
will prevent the processor from getting very far in the POST
sequence.

On modern boards -

1) The BIOS must be prepared to support the processor being used.
If Intel released a processor today, it might be a while before
a BIOS was available to support it. And a much longer while, before
that new BIOS, was actually shipping in retail products. The
seller already has a container full of the older motherboards,
so it takes a while to clear the older stock.

In some cases, you'd need to borrow an older LGA775 processor, to
use long enough to flash the BIOS to a later revision. Either
that, or make arrangements with the manufacturer, to get another
BIOS chip to snap into place. (Some boards now, have small SPI
BIOS chips soldered to the board, so swapping chips is not an
option.)

2) DDR2 or DDR3 RAM can be cranky, and may need a slight voltage boost
to work. Try with just one stick of RAM, as sometimes a board will
start with one stick.

-5V is no longer needed, so you don't have to worry about that one.

HTH,
Paul

Hey Paul,

Thanks for posting a reply back - much appreciated. Both motherboards
did the 'beep' test when I removed both sticks of memory, so I'm
assuming that memory may be an issue? Please let me know on that - -
it's good to know that the motherboard and processor are at least
healthy. I may try a new (different) stick of memory and see if that
makes a difference with POST.

I have integrated video on both motherboards - so not seeing any video
is making me queasy, too.

Let me know what you think about the new information I've provided and
in the meantime, I'm going to get a new stick of memory and give that
a try this evening.

Thanks for the help thus far and please reply back on the above when
you have a moment today - thanks,

Marty

Marty_KN0CK wrote:
> Hey Paul,
>
> Thanks for posting a reply back - much appreciated. Both motherboards
> did the 'beep' test when I removed both sticks of memory, so I'm
> assuming that memory may be an issue? Please let me know on that - -
> it's good to know that the motherboard and processor are at least
> healthy. I may try a new (different) stick of memory and see if that
> makes a difference with POST.
>
> I have integrated video on both motherboards - so not seeing any video
> is making me queasy, too.
>
> Let me know what you think about the new information I've provided and
> in the meantime, I'm going to get a new stick of memory and give that
> a try this evening.
>
> Thanks for the help thus far and please reply back on the above when
> you have a moment today - thanks,
>
> Marty

Memory could be the issue. Either the memory, or some problem with
the motherboard. And since both boards give the same symptoms,
a missing RAM beep pattern, that suggests the RAM is a common
ingredient. Make sure the RAM is fully seated, and both lock
latches are properly engaged. You shouldn't be able to see the
gold fingers on the RAM, if it is well down into the slot.

Say I had some DDR2 memory (standards say 1.8V nominal, and the
datasheet for the product says "5-5-5-15 at 2.0 volts". What that
is telling you is, the memory needs 2.0V to meet 5-5-5 timing. Now,
if the SPD (the tiny EEPROM on the DIMM module), happens to set the
timing to 5-5-5, and the BIOS uses Auto (which it should when the
motherboard is new), then it is going to try to run at 5-5-5
timing. But when the motherboard is new, it may not be applying
the 2.0V. There are a couple (non-standard) EEPROM formats, that
allow specifying the voltage value needed, and on select
motherboard models, the motherboard will apply the "boosted" values.
But most motherboards are too stupid, and the memory
module itself may not contain the optional voltage
information. So most of the time, no boost is applied
to Vdimm.

So that can create a problem for a home builder. On the one
hand, you can buy some "uber" DDR2-1066 RAM, assuming it'll
run at sky high speeds. But the part you're missing, is it might
need 2.1V to do it. And since the motherboard might start up the
first time, applying 1.8V, the memory may fail miserably.

In such a case, buy "ordinary" RAM, slow stuff, something that
is rated to run at 1.8V. If you insert one stick of that in the
board, then enter the BIOS and tweak Vdimm to the value the "uber"
memory wants, then you can save, exit, and shut down, and install
the new "uber" RAM.

Paul

Paul,

Well, heck.....YOU ARE A GENIUS! I did exactly what you said (use
slower memory - I just happened to have some Crucial 512M PC4200
hanging around that I probably paid a handsome sum for awhile back)
and danged if the ECS Motherboard didn't boot to the startup screen
and stop when it didn't find a HD (that I didn't install). Everything
looked great and I was able to navigate to ALL the BIOS setup screens
effortlessly. So, I really need to check this with the other MSI
motherboard and get it running since that's what my Son would rather
have running - and it should respond similarly with the slower memory
installed until I readjust the RAM speed. As it turns out, he bought
(almost) uber memory (PC6400, 800 MHz) and the ECS motherboard came up
with the slowest setting on - which meant that his uber memory was
getting the lower Vdimm value (probably as you mentioned 1.8V). So
I'll keep you tuned in tomorrow, but the ECS booted and there's no
reason not to do the same thing with the MSI and just readjust
everything to the faster memory speed, turn it off, and reboot with
the faster memory installed as you instructed earlier.

Anyway, Paul - MANY THANKS for the help with this. More to follow on
the MSI's status once I get everything up and running on slower RAM.

Stay tuned......

Marty

Paul,

Well, this is the final installment for now on the Dual Core mobo
journey - read on.....

Got the MSI motherboard running this evening with the slower Crucial
DDR2 RAM (PC4200) and actually got everything installed on the drives
with XP running and everything. Tried getting readjusting the memory
voltage in BIOS but it didn't take - - the BIOS apparently 'knows'
better about what the voltage needs to be adjusted to and pre-adjusted
the memory voltage to 1.90V for memory he has (PC6400 with Vdimm 2.1V)
that just wouldn't hack it at 1.90V. So when I was in Windows I tried
using MSI's Windows app (Dual Core Center 2.0) to adjust the memory
voltage and did a reboot - - same thing. BIOS took over and the PC
acted as dead as it was when I was looking for help the other night.
Well, I've told my Son to use the PC4200 memory that's in there now
(the machine still had awesome speed - I'm going to keep the other ECS
motherboard now based on what I saw) and I told him to buy some PC6400
Kingston that should take 1.9 - 2.0V with no issue (according to their
website), so he's off to exchange the 2.1V memory he has now for
Kingston......should work great after that exchange happens.

So, for now, all is right with the world again. My Son has his
machine working, I have a new motherboard, processor, fan, and memory
to play with and it looks like the dual boot tower with Linux and
Windows in my Ham Shack (geek room) is going to be a reality
now......go figure. But all that credit for making work goes to you,
Paul.

Take care and thanks again for all the help,

Marty

Marty_KN0CK wrote:
> Paul,
>
> Well, this is the final installment for now on the Dual Core mobo
> journey - read on.....
>
> Got the MSI motherboard running this evening with the slower Crucial
> DDR2 RAM (PC4200) and actually got everything installed on the drives
> with XP running and everything. Tried getting readjusting the memory
> voltage in BIOS but it didn't take - - the BIOS apparently 'knows'
> better about what the voltage needs to be adjusted to and pre-adjusted
> the memory voltage to 1.90V for memory he has (PC6400 with Vdimm 2.1V)
> that just wouldn't hack it at 1.90V. So when I was in Windows I tried
> using MSI's Windows app (Dual Core Center 2.0) to adjust the memory
> voltage and did a reboot - - same thing. BIOS took over and the PC
> acted as dead as it was when I was looking for help the other night.
> Well, I've told my Son to use the PC4200 memory that's in there now
> (the machine still had awesome speed - I'm going to keep the other ECS
> motherboard now based on what I saw) and I told him to buy some PC6400
> Kingston that should take 1.9 - 2.0V with no issue (according to their
> website), so he's off to exchange the 2.1V memory he has now for
> Kingston......should work great after that exchange happens.
>
> So, for now, all is right with the world again. My Son has his
> machine working, I have a new motherboard, processor, fan, and memory
> to play with and it looks like the dual boot tower with Linux and
> Windows in my Ham Shack (geek room) is going to be a reality
> now......go figure. But all that credit for making work goes to you,
> Paul.
>
> Take care and thanks again for all the help,
>
> Marty
>

Place both the slow stick and one fast stick in the computer. Put
the slow stick in the slot nearest the CPU. Put the fast stick
next to it. The idea is, the slow stick will be providing
low memory. The BIOS will select 1.9V, as it has before, on
noting a new hardware configuration. The slow stick will behave
like memory should. The BIOS should be able to display on the
screen, because the slow stick is holding its crucial information.
Now, change the voltage and do Save and Exit.

If that doesn't work, try reversing the order of the sticks.

It is a long shot, but perhaps it will hold the setting
when you swap out the slow stick and put in the fast stick.
Make sure, before you do that, that all the BIOS settings
regarding memory, are in manual mode. To reduce the temptation
of the BIOS, to screw around.

There is no guarantee it'll work, but it might be worth a
shot.

Or something like this -


CPU slow fast ---- fast (set the voltage with three sticks)
(should be single channel mode, with
slow stick at address 0x00000000)

Now, remove the slot stick, like this, and cross your fingers.

CPU ---- fast ---- fast

The reason I thought up stuff like this, is I recently had
a stick of RAM fail, and I could operate the computer and
run memtest86+, only if I changed the order of the sticks,
and moved the bad stick so it wasn't located at 0x00000000.
I also had to make sure the computer was not in dual channel
mode, so could not afford to use a "balanced" memory config.
Thus the three stick configuration - that is what worked
for me.

HTH,
Paul

I have seen a couple dozen or more socket 775 computers come into my shop (all major brands) that quit working, and "ALL" have -5v out on the power supply. The first one I put in a new PS and computer still dead and blown -5v on the new PS. So if -5v isn't needed why do these new computers keep blowing PS's. These have been factory machines and were working until they weren't