9NPA+ Boot Problem
jmbo...@gmail.com
My friend has an EP-9NPA+ motherboard. He is currently running with
an nVidia 6600GT video card. When I went to upgrade the card to a
9600GT, I got a C1 error on the board. My first thought was not
enough power. He currently has an Antec SL-350S power supply. I have
an Antec SP-400 power supply in my PC, so I swapped that out with his
PS to test. Still a no go. I put his old 6600GT card back in, and
try to boot again with my SP-400 power supply. I get the blue LED on
the board to light up, but when I hit the PC power switch nothing
happens. When I put his SL-350S power supply back in, it powers on
fine.
Now here's the weird part. When I plug in the old SL-350S power
supply and flip the power supply switch on, the fans, blue LED, and
LED displays on the board jump for a second which gives me the
impression they have juice. When I plug in the newer SP-400 power
supply and flip the switch, that doesn't happen (and the board won't
power on). The SP-400 still works with my PC, so it isn't dead.
My first problem is figuring out why my power supply won't power the
board. I don't want to upgrade his power supply only to find that the
new one won't turn the PC on for some strange reason. Does anyone
have any idea what might be going on?
jmbo...@gmail.com
my friend has only has a 20-pin connector. The SP-400 I'm swapping
from my PC has a 24-pin connector. Is there any BIOS setting or other
board issue that could keep the board from even trying to boot because
of the power supply change?
Paul
The BIOS doesn't monitor the ATX power connector as such. Logic
circuits are checking for some conditions to be met though, before
the BIOS is allowed to run.
These specs discuss the two generations of power supplies.
http://www.formfactors.org/developer/specs/atx/ATX12V_1_3dg.pdf (page 30)
http://www.formfactors.org/developer/specs/ATX12V_PSDG_2_2_public_br2.pdf (page 37)
Note that pin 1 of the 24 pin, goes to pin 1 of the 20 pin connector.
Some 24 pin connectors make this easy, by allowing the end four
pins to be hinged out of the way. The "extra" four pins are
detachable. If the connector happens to be the non-detachable
version, the four left over pins on one end of the connector
can be left dangling in space. They're redundant and aren't adding
any new voltage values. They're just more pins to carry more
current (amperes).
When the motherboard wants to turn on the power supply, it grounds
the PS_ON# signal. The PS_ON# is held at 5V normally, via a pullup
resistor. A driver chip or transistor on the motherboard, is used
to pull the signal to ground level (logic 0), which tells the
power supply to turn on.
The PWR_OK goes to a 5V level, when the power supply has stabilized.
That is how the motherboard knows the power supply is ready.
The PWR_OK signal then goes through some logic chips, which
check the levels of any locally regulated voltages. When
the motherboard finds all voltages stable (external and internal),
and the reset interval is completed, the motherboard can start.
But PWK_OK only happens when the PS_ON# thing kicks
things off. (See figure 7 of the second document above,
for a timing diagram of the signals.)
What should happen, when the switch on the back of the
supply is turned on, is the motherboard should not be
glitching the PS_ON# and causing other rails on the
supply to glitch in response. The motherboard should
quietly wait until the user presses the front POWER
button, before the PS_ON# signal is fed to the power
supply. And the processor plus BIOS cannot boot the
board, unless "PWR_OK and friends" are happy.
I resisted answering your post when it first showed up,
because I was going to present a theory that the issue
had to do with -5V. But then I discovered that the
SL350S doesn't have -5V on it, so I could not explain
the issue as being something to do with a missing
-5V on the new supply. The old supply is missing
-5V anyway. You'll notice on page 30 of the first
document above, pin 18 is marked "reserved" and
that is the place where the -5V was removed from
the supply spec. The corresponding pin on the
24 pin connector will also be marked as reserved,
for the same reason.
There are a few old motherboards that won't run
unless -5V is present. And thus, when a user
has an old power supply fail, and then they
substitute a new supply, the missing -5V pin
trips them up. It isn't supposed to happen that
way, but apparently there are some motherboards
that have a dependency on -5V.
Anyway, I don't have anything really to add about
your original problem description. (No strong
suspicions.) It would be nice if the Epox documentation
had a table of POST codes, so you could
find out what "C1" is.
The 9600GT is a PCI Express revision 2 card, while
the motherboard is likely revision 1. In the case
of the 8800GT (G92), another Nvidia product, there
was an issue with the card failing to negotiate
PCI Express speed properly with revision 1 motherboards.
A workaround was to get a special version of the video
BIOS and reflash the card so that it would only run in
revision 1 mode. I'm not aware of that mistake
being made by Nvidia, other than on the 8800GT.
Perhaps you can try some Google searches, and
see if the 9600GT offered by some video card
manufacturers, may have a similar video BIOS
issue. (Like, maybe some Newegg reviews might
have evidence of a problem.)
Oh. Another dumb question. Does the 9600GT have
a 2x3 power connector on the end of the card ?
Is that plugged in ? If it isn't powered, you
may need a Molex 1x4 to PCI Express 2x3 adapter
cable, to give the card some power.
This picture highlights a typical 9600GT power connector.
http://c1.neweggimages.com/NeweggImage/productimage/14-127-333-07.jpg
This is the adapter you'd try, if the power supply
doesn't have a 2x3.
http://c1.neweggimages.com/NeweggImage/productimage/12-200-106-01.jpg
http://www.newegg.com/Product/Product.aspx?Item=N82E16812200106
Paul