P4R800-VM and Boxed Celeron D 330

  • Thread starter Thread starter Andy
  • Start date Start date
A

Andy

I have just fitted a Celeron D 330 CPU into a P4R800-VM revision 1.03 board.

The mobo functions OK with one exception. The CPU cooling fan generates an
oscillating noise when the CPU is under load. A second CPU and fan assembly
was connected to this mobo but also has the same problem. Therefore I
conclude its not a faulty fan.

The boxed Celeron D 330 comes as standard with a temp controlled fan which
is rated at 12v 0.44A (i.e. 5.3Watts). The mobo manual states that the fan
connectors can supply over 8Watts.

If a boxed P4 2.8G fan is connected, no oscillation occurs. However this fan
is only rated at 12V 0.22A (i.e. 2.6Watts).

If I connect the Celeron D 330 fan into a P4P800-E mobo, no oscillation
occcur.

I measured the voltage across the fan when it was ocillating. It was
approximattley 5V. I have checked to make sure Q Fan is disabled and even
reset the CMOS just to be on the safe side.

The conclusions I have drawn are: -

1) The mobo is fault
2) Q Fan is enabled even thought the bios setting has been disabled
3) The fan connector can't supply enough power to drive the new Celeron D
fan.

Has anyone else fitted a Celeron D processor into this mobo sucessfully.

Andy
 
"Andy" said:
I have just fitted a Celeron D 330 CPU into a P4R800-VM revision 1.03 board.

The mobo functions OK with one exception. The CPU cooling fan generates an
oscillating noise when the CPU is under load. A second CPU and fan assembly
was connected to this mobo but also has the same problem. Therefore I
conclude its not a faulty fan.

The boxed Celeron D 330 comes as standard with a temp controlled fan which
is rated at 12v 0.44A (i.e. 5.3Watts). The mobo manual states that the fan
connectors can supply over 8Watts.

If a boxed P4 2.8G fan is connected, no oscillation occurs. However this fan
is only rated at 12V 0.22A (i.e. 2.6Watts).

If I connect the Celeron D 330 fan into a P4P800-E mobo, no oscillation
occcur.

I measured the voltage across the fan when it was ocillating. It was
approximattley 5V. I have checked to make sure Q Fan is disabled and even
reset the CMOS just to be on the safe side.

The conclusions I have drawn are: -

1) The mobo is fault
2) Q Fan is enabled even thought the bios setting has been disabled
3) The fan connector can't supply enough power to drive the new Celeron D
fan.

Has anyone else fitted a Celeron D processor into this mobo sucessfully.

Andy
Click to expand...

First of all, that "spec" for the current the fan headers can provide,
is a copy/paste from another manual. There are two fan headers on your
board, and that spec is scaled as if there are three fan headers.
350ma * 3 = 1A. So, the following is a load of BS.

"350mA~740mA (8.88W max.) or a total of 1A~2.22A (26.64W max.) at +12V."

Put in a request to Asus Tech Support and see if they can figure out
what the real spec is. (Hint - you'll have to be very patient :-)

As a work around, find one of those adapters for powering a fan from
a disk drive power connector. You will want an adapter that has a
single wire with a three pin connector on the end, so that the CPU_FAN
connector on the motherboard can continue to get a tachometer signal.

The chip that controls the fans on your board is probably this one:

http://www.winbond-usa.com/products/winbond_products/pdfs/PCIC/627hf.pdf

On the second from last page of that document, are shown three ways to
interface from the Winbond 83627THF to a fan. None of the methods
has protection against overload. Now, the first solution that uses
a bipolar transistor, maybe if the beta of the power transistor
was low enough, the op amp could be operating in current limiting,
or the op amp is thermally protecting itself. That is about the
only mechanism I can think of for the oscillation. Note that the
Winbond control method is not PWM, but rather is linear control,
and this means the control device (MOSFET, bipolar transistor, or
adjustable regulator) will run hot. The PWM (pulse width modulation)
method of control, has as its main virtue, that the MOSFET operates
in saturation, either fully switched on or fully switched off,
and produces a variable duty cycle square wave, which is further
smoothed by a filter network to give the voltage fed to the fan.
The linear method is superior from the fan's perspective, but the
linear method also runs the risk of burning out the transistor
that controls the fan. On the PWM, that is only going to happen if
there is a really bad overload, whereas a moderate overload can
damage the linear method. (Asus could design in overload
protection, but every penny counts.)

The Winbond chip has the ability to mess with fan speed all by itself,
with no processor intervention. So, it is possible the oscillation
could be caused by whatever control mode is enabled in the Winbond
chip. That seems to be a trifle unlikely scenario, but it is the
only other mechanism I can see.

I would find an adapter for the fan, so it can run off a disk drive
connector, and move on with life :-)

HTH,
Paul
 
Paul,

Thanks for the suggestions

I reported this problem to Asus support over a week ago and we exchange
emails daily on the subject.

Great minds think alike. I have already attempted to power the fan directly
from a hard drive connector as a test. It makes no difference. The fan noise
still oscillates when the CPU is under load.

Just checked the the 12V supply with a digital multimeter at a hard drive
connector. The meter reads 11V with no CPU load and dipps below 10.5V when
the CPU is under load. Asus probe measure the 12V line as 11.8V

I have disconnected all the periferals i.e. DVDRW and the floppy. Repeated
the test and the 12V supply still dips below 10.5V during CPU load. Replaced
the PSU with another identially rated unit. Same result.

The PSU is a 310W unit which can supply 198W from its 12V line i.e. 16A as
specified. The only components the PSU is powering is the mobo, 1x256MB
PC3200, Celeron D 330 and its fan.

I can only conclude that the CPU is drawing more that 16A from the PSU to
drag down the volts. The rated power consumption of a Celeron D 330 is
approx 85W - no where near 198W!

I emailed Asus support with this info this morning and asked what capacity
of power supply they used to test the P4R800-VM with a Celeron D 330. No
response as of yet but I will keep you posted.

I can't belive that the Celeron D 330 is draging more that 16A from the 12V
line. I know the new Prescott core needs more power and generates more heat
that the Northwood but this is just getting silly. Just as a check I
refitted a Northwood P4 2.8G. The fan noise disappeared and the supply does
not drop below 11V.

Andy
 
"Andy" said:
Paul,

Thanks for the suggestions

I reported this problem to Asus support over a week ago and we exchange
emails daily on the subject.

Great minds think alike. I have already attempted to power the fan directly
from a hard drive connector as a test. It makes no difference. The fan noise
still oscillates when the CPU is under load.

Just checked the the 12V supply with a digital multimeter at a hard drive
connector. The meter reads 11V with no CPU load and dipps below 10.5V when
the CPU is under load. Asus probe measure the 12V line as 11.8V

I have disconnected all the periferals i.e. DVDRW and the floppy. Repeated
the test and the 12V supply still dips below 10.5V during CPU load. Replaced
the PSU with another identially rated unit. Same result.

The PSU is a 310W unit which can supply 198W from its 12V line i.e. 16A as
specified. The only components the PSU is powering is the mobo, 1x256MB
PC3200, Celeron D 330 and its fan.

I can only conclude that the CPU is drawing more that 16A from the PSU to
drag down the volts. The rated power consumption of a Celeron D 330 is
approx 85W - no where near 198W!

I emailed Asus support with this info this morning and asked what capacity
of power supply they used to test the P4R800-VM with a Celeron D 330. No
response as of yet but I will keep you posted.

I can't belive that the Celeron D 330 is draging more that 16A from the 12V
line. I know the new Prescott core needs more power and generates more heat
that the Northwood but this is just getting silly. Just as a check I
refitted a Northwood P4 2.8G. The fan noise disappeared and the supply does
not drop below 11V.

Andy
Click to expand...

Interesting. And, your report is the first user report of getting a
Celeron D to work. Two other users couldn't get their boards to POST
(on different boards).

So, you've confirmed with the voltmeter, that the PSU is at fault.
Or, at least it cannot supply the current the CeleronD is drawing.
My cutoff for PSUs is somewhere between 5% and 10% dip below 12V
nominal, and your measured value of 10.5 with a voltmeter is well
below that.

I use a clamp-on ammeter, to measure DC currents in wires like the
two +12V wires on the 2x2 power connector. That connector is the
one I'd be looking at, to determine if the processor draws current
in spec or not. While you could try to find some Molex Minifit Jr
plugs, and make an adapter cable for the amps input on the voltmeter,
there is a danger that the current could be higher than the
limited range of your meter. My meters here, have a limit of 10A
on the ammeter ranges, and that might not be enough. My clamp-on
ammeter can measure up to 400 amps DC or AC (because it is based
on a Hall probe that measured the magnetic field, and doesn't touch
the circuit).

When you enter the BIOS monitor, does the Vcore value make sense
for the VID value printed on the box the processor came in ?
The Intel site had been listing the processor, as using around
1 volt, which is most likely wrong. Being related to other Prescott
processors, it should be in the 1.25 to 1.4 range. I'm curious if
perhaps the VID isn't being read properly by the Asus board.

If the Vcore voltage is actually higher than it is supposed to be,
that could cause excess consumption on the CeleronD. But, the heatsink
would be overheating if that were the case. Does the CeleronD
run extremely hot ?

Hmmm. I wonder if the BIOS doesn't recognize the processor properly,
if it will program the VID to an inappropriate value. We know
that Asus BIOS don't like to undervolt. The BIOS generally will
refuse to accept manual Vcore settings known to be less than the
nominal VID values for a family of processors, so maybe the BIOS
is messing about with Vcore and running it higher than the VID
value being sent by the processor.

In any case, logical thinking would be this. You have a 16A supply
to work with at 12V. That is a total of 192 watts, and it is going
somewhere. I would be using my fingers, to figure out exactly where
the heat from that power ends up. Something has to get hot. Even
the air coming out of the PSU should be hot, due to the excess
current drain.

In the case of a simple (partial) short, most of the heat could
be dissipated in the PSU. But, because you've operated another
processor in there just fine, that rules out that kind of fault.

So, follow the heat...

Paul
 
Paul,

I contacted Asus before installing the Celeron D because nothing was posted
on their CPU compatibility table. They confirmed via email that the
P4R800-VM PCB revision 1.03 and later did support the Celeron D. However the
bios needs flashing to V1005, published the day before. As I had a P4 2.8G
already installed it was no problem to flash the bios.

I fitted the Celeron D , booted, installed XP Home SP1a and its worked
reliably ever since, apart from the oscillating fan syndrome.

Unfortunately my digital multimeter is quite old and cannot measure above
200mA so I can't measure the Amps being drawn by the CPU. The VCore on the
box is V1.40. The mobo reads it as between V1.36 to V1.40. No problem there.

The mobo and CPU have been stressed with SIS SandraSoft for a continuous
period of eight hours without a problem. CPU temp averaged 54 degrees and
the mobo temp averaged 38 degrees, cooled by the Intel supplied fan &
heatsink. The PSU temp was within its thermal limits even after eight hours.

Don't get me wrong, it generated more heat than a small electric fan heater
but the mobo, CPU and PSU did not overheat and I had no stability problems.
But I had to open the window in the small room to let all the heat out!

My current perception is that the PSU maybe marginally underrated or its
reaction time to a demand for high amps is a little too slow, therefore I
get a volt drop on the 12V line. The fan rotates at 3000rpm at 12V when hot,
which translates to say a 15% drop in revs when the 12V line is loaded.
Hence the oscillating fan noise syndrome. Its sounds like a hamster revving
a small electric car - very irritating after 10 minutes.

It strange that I have not heard anything from Asus support today since I
reported my findings this morning. Maybe I've given them something to think
about.

Andy
 
"Andy" said:
Paul,

I contacted Asus before installing the Celeron D because nothing was posted
on their CPU compatibility table. They confirmed via email that the
P4R800-VM PCB revision 1.03 and later did support the Celeron D. However the
bios needs flashing to V1005, published the day before. As I had a P4 2.8G
already installed it was no problem to flash the bios.

I fitted the Celeron D , booted, installed XP Home SP1a and its worked
reliably ever since, apart from the oscillating fan syndrome.

Unfortunately my digital multimeter is quite old and cannot measure above
200mA so I can't measure the Amps being drawn by the CPU. The VCore on the
box is V1.40. The mobo reads it as between V1.36 to V1.40. No problem there.

The mobo and CPU have been stressed with SIS SandraSoft for a continuous
period of eight hours without a problem. CPU temp averaged 54 degrees and
the mobo temp averaged 38 degrees, cooled by the Intel supplied fan &
heatsink. The PSU temp was within its thermal limits even after eight hours.

Don't get me wrong, it generated more heat than a small electric fan heater
but the mobo, CPU and PSU did not overheat and I had no stability problems.
But I had to open the window in the small room to let all the heat out!

My current perception is that the PSU maybe marginally underrated or its
reaction time to a demand for high amps is a little too slow, therefore I
get a volt drop on the 12V line. The fan rotates at 3000rpm at 12V when hot,
which translates to say a 15% drop in revs when the 12V line is loaded.
Hence the oscillating fan noise syndrome. Its sounds like a hamster revving
a small electric car - very irritating after 10 minutes.

It strange that I have not heard anything from Asus support today since I
reported my findings this morning. Maybe I've given them something to think
about.

Andy
Click to expand...
I hope the CeleronD doesn't shorten the life of your PSU. When a PSU
blows, you never know what it might take with it.

Maybe you could post the brand and model number of the PSU units in
question, so other people will know they don't live up to their
rating.

Paul
 
Paul,

I have seen quite a few bad PSU's destroy whole PCs in my time so I don't
intend to power the mobo and CPU until Asus responds. Nothing as of this
morning.

I'm not totally convinced that the PSU is at fault although it maybe under
rated for the Celeron D. We will have to wait and see. The PSU is Chieftec
unit fitted as standard to some of their cases. The PSU model number is
ATX-310-202. It has no problem powering the P4R800-VM fitted with a
Northwood P4 2.8G CPU.

If I get some time next week, I will perform a static load test on the
Cheiftec PSU. I have access to a testing rig but I will have to confirm the
maximum dummy loads it can handle before I proceed.

Andy.
 
Back
Top