A7V333 problems and power supplies

[John Wilson]

Ordered three A7V333 (rev. 1.02) boards back in December. (I am running
a Athlon 1700+ (0.13um) CPU in these boards.)
Didn't have time to set one of them up until recently, and the other was
only used a little. The one I setup in December had to have the power
supply replaced in Feb. The 3.3V portion of the supply (300W) was
putting out 3.9V. I lost a SCSI controller. Replaced the supply and all
worked well until recently. Then the power supply started to make high
pitched noises. Not the fan, but very high freq. like a fly-back
inductor in the switching power supply that had a large switching
current flowing through it causing magnetostriction of the magnetic
core.

Then....
I built another machine using one of my other boards and new CPU (same
speed and also 0.13um). The box for this machine is an Antec with a 350W
supply. It ran for a little less than a day and then started the make
the same (more or less) high pitched sound. Now that machine reboots and
gives trouble.

My conclusion is that ASUS screwed up the on-board buck converter (the
DC-DC converter used to step the 3.3V from the power supply down to the
CPU voltage ~ 1.5V to 1.8V depending on CPU) and that is producing large
ripple currents in the fly-back inductor in the 3.3V output, which is
probably causing the output filter capacitors of the 3.3V section to
fail since they are probably seeing large ripple current as well.

The interesting part is that we have a machine in the lab that I built
using a rev.1.02 A7V333, but it is using a 1900+ Athlon that was
fabricated in a 0.18um process. This machine has given no trouble. It
could be that the on-board buck converter produces large ripple currents
when stepping down to the lower voltage that is used by the 0.13um CPU.
I wouldn't draw such a conclusion, except that I have encountered it
twice with 0.13um CPUs, but when a 0.18um CPU was used I have had no
problems.

I have ordered a 0.18um CPU so that I can experiment at home (the lab
machine is not mine to experiment with) and I have a new and unused 350W
supply to experiment with too.


Has anyone else had similar woes?

Please reply to the newsgroup and my email address.


Thanks in advance.

John Wilson
Ph.D. Electrical Engineering
Visiting Research Scientist
North Carolina State University
Electrical and Computer Engineering Deptartment

 

[Paul]

Ordered three A7V333 (rev. 1.02) boards back in December. (I am running
a Athlon 1700+ (0.13um) CPU in these boards.)
Didn't have time to set one of them up until recently, and the other was
only used a little. The one I setup in December had to have the power
supply replaced in Feb. The 3.3V portion of the supply (300W) was
putting out 3.9V. I lost a SCSI controller. Replaced the supply and all
worked well until recently. Then the power supply started to make high
pitched noises. Not the fan, but very high freq. like a fly-back
inductor in the switching power supply that had a large switching
current flowing through it causing magnetostriction of the magnetic
core.

Then....
I built another machine using one of my other boards and new CPU (same
speed and also 0.13um). The box for this machine is an Antec with a 350W
supply. It ran for a little less than a day and then started the make
the same (more or less) high pitched sound. Now that machine reboots and
gives trouble.

My conclusion is that ASUS screwed up the on-board buck converter (the
DC-DC converter used to step the 3.3V from the power supply down to the
CPU voltage ~ 1.5V to 1.8V depending on CPU) and that is producing large
ripple currents in the fly-back inductor in the 3.3V output, which is
probably causing the output filter capacitors of the 3.3V section to
fail since they are probably seeing large ripple current as well.

The interesting part is that we have a machine in the lab that I built
using a rev.1.02 A7V333, but it is using a 1900+ Athlon that was
fabricated in a 0.18um process. This machine has given no trouble. It
could be that the on-board buck converter produces large ripple currents
when stepping down to the lower voltage that is used by the 0.13um CPU.
I wouldn't draw such a conclusion, except that I have encountered it
twice with 0.13um CPUs, but when a 0.18um CPU was used I have had no
problems.

I have ordered a 0.18um CPU so that I can experiment at home (the lab
machine is not mine to experiment with) and I have a new and unused 350W
supply to experiment with too.


Has anyone else had similar woes?

Please reply to the newsgroup and my email address.


Thanks in advance.

John Wilson
Ph.D. Electrical Engineering
Visiting Research Scientist
North Carolina State University
Electrical and Computer Engineering Deptartment

There is a table of processors here:

http://www.qdi.nl/support/CPUQDISocketA.htm

Have you set the Vcore voltage correctly for the processors you are
using ?

Has anything failed completely ? As in a burst capacitor or leaking
brown fluid onto the board ? Abit boards like to fail like that,
whereas on Asus boards, we get the occasional report of a failed
MOSFET (burned and cracked).

Usually when Asus screws up a Vcore design, the result is frequent
processor crashes due to insufficient current or wobbly output voltage.

I think with the A7V333 it is easy to apply too much Vcore. There
is an overvoltage jumper, that bumps Vcore 0.3 above nominal. You should
be able to use the Power Monitor BIOS page, to read the voltage,
or get your trusty voltmeter, to make sure it isn't overvoltage.

DDR DIMM voltage jumpers and mention of the overvoltage jumper:
http://www.spodesabode.com/content/article/a7v333/print

Overvoltage is mentioned here also:
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&[email protected]

Lostcircuits says it uses an Onsemi NCP5322A for Vcore (because I cannot
read the part number on the chip in the picture of the board in the PDF
user manual):
http://www.lostcircuits.com/motherboard/asus_a7v333/3.shtml

Onsemi says the NCP5322A regulates 12V down to 1.6V at up to 45Amps
(with the Onsemi reference circuit, presumably). Could the sound
you are hearing actually be coming from the computer PS as it
supplies the +12V ?

http://www.onsemi.com/pub/Collateral/NCP5322A-D.PDF

Now you can check the Asus designer's calculations

Paul

 

[... et al.]

Lostcircuits says it uses an Onsemi NCP5322A for Vcore (because I cannot
read the part number on the chip in the picture of the board in the PDF
user manual):
http://www.lostcircuits.com/motherboard/asus_a7v333/3.shtml

Both the A7V333 1.01 & the 1.04 i've looked at have location U8 occupied
by an Onsemi NCP5322A.