GeForce 2 MX400 PCI voltage regulator burned up, missing solder
 

I bought two GeForce 2 MX400 PCI cards at the same time over a year
ago and installed them in two identical machines with industrial PICMG
backplanes and industrial PICMG PIII 540MHz CPU boards. These
computers each have 2 chassis fans, one at front and one at rear and 1
power supply fan. The chosen power supply delivers more than enough
power for an entire system.

After about 3 months of use, one of these GeForce 2 MX400 PCI cards
began to cause 3D graphics intensive games to intermittently pause,
stray pixels and odd coloring to appear and then a few seconds later,
the entire machine would lock up. This occurred intermittently for
about two weeks until the computer finally locked up and refused to
power up.

Upon removing both identical GeForce cards from each computer, I
noticed a difference between them. On the failed board, the
manufacturing process had neglected to solder the thermal fin of one
of the regulators (lower left corner of the board) to the circuit
board, causing an eventual thermal runaway condition. The regulator
appeared lifted slightly from the board and had discolored the circuit
board from the soft yellow-beige to a burnt brown. The GPU also showed
burnt brown discoloration on the opposite side of the circuit board.

The regulator had not been properly installed in the first place
because there was no excess solder to be seen; the regulator did not
desolder itself under extreme heat conditions, rather, there was no
trace of any attempt at soldering at all. I believe that this stage of
manufacturing must be a hand-solder job and prone to many mistakes,
including poor quality assurance gates. The functioning GeForce
board's regulator was properly soldered.

I the remaining GeForce 2 card is still functioning to this day. I
bought these cards around March 2002.

Researching the frequency of this problem, through internet search
engines and newsgroups I have found that other people have experienced
similar thermal problems with many flavors of NVIDIA graphics cards
and conclude that NVIDIA has either a persistent thermal design
shortcoming or a manufacturing issue that needs resolving. I suspect
they are aware of these thermal problems, however their inability
(unwillingness?) to report or even admit these flaws on their web site
proves to me that this is yet another sociopathically irresponsible
corporation.

Beerden

On 17 Jul 2003 16:58:05 -0700, (Beerden) wrote:

I bought two GeForce 2 MX400 PCI cards at the same time over a year
ago and installed them in two identical machines with industrial PICMG
backplanes and industrial PICMG PIII 540MHz CPU boards. These
computers each have 2 chassis fans, one at front and one at rear and 1
power supply fan. The chosen power supply delivers more than enough
power for an entire system.

After about 3 months of use, one of these GeForce 2 MX400 PCI cards
began to cause 3D graphics intensive games to intermittently pause,
stray pixels and odd coloring to appear and then a few seconds later,
the entire machine would lock up. This occurred intermittently for
about two weeks until the computer finally locked up and refused to
power up.

Upon removing both identical GeForce cards from each computer, I
noticed a difference between them. On the failed board, the
manufacturing process had neglected to solder the thermal fin of one
of the regulators (lower left corner of the board) to the circuit
board, causing an eventual thermal runaway condition. The regulator
appeared lifted slightly from the board and had discolored the circuit
board from the soft yellow-beige to a burnt brown. The GPU also showed
burnt brown discoloration on the opposite side of the circuit board.

The regulator had not been properly installed in the first place
because there was no excess solder to be seen; the regulator did not
desolder itself under extreme heat conditions, rather, there was no
trace of any attempt at soldering at all. I believe that this stage of
manufacturing must be a hand-solder job and prone to many mistakes,
including poor quality assurance gates. The functioning GeForce
board's regulator was properly soldered.

I the remaining GeForce 2 card is still functioning to this day. I
bought these cards around March 2002.

Researching the frequency of this problem, through internet search
engines and newsgroups I have found that other people have experienced
similar thermal problems with many flavors of NVIDIA graphics cards
and conclude that NVIDIA has either a persistent thermal design
shortcoming or a manufacturing issue that needs resolving. I suspect
they are aware of these thermal problems, however their inability
(unwillingness?) to report or even admit these flaws on their web site
proves to me that this is yet another sociopathically irresponsible
corporation.


errrrr................sorry...........wrong target............

Who was the manufacturer of the faulty card ? I suggest that you
narrow your fault search to cards made by the same board
manufacturer.

NVidia just makes the GPU chips, provides reference designs and
commissions some boards for QC, design reliabilty-testing, demo and
design-samples for 3rd-party manufacturers. Quality control of the
final product is in the hands of the 3rd-parties. That is why some
card manufacturers have excellent reputations and others questionable.

I have had 6 video cards in my home PCs using Nvidia technology
over the past couple of years, all from various manufacturers, all
pretty reputable.

No functional or thermal-related problems other than a couple of
noisy fans which a spot of teflon lubricant cured. AFAIK nVidia does
not design fans and the choice of fan is distinctly a third-party
manufacturer exercise.

By the way, I build computers as a hobby activity for family and
friends. I always inspect boards for any visible manufacturing defects
before installation. Stitch in time saves nine............

John Lewis

Beerden

On 17 Jul 2003 16:58:05 -0700, (Beerden) wrote:

I bought two GeForce 2 MX400 PCI cards at the same time over a year
ago and installed them in two identical machines with industrial PICMG
backplanes and industrial PICMG PIII 540MHz CPU boards. These
computers each have 2 chassis fans, one at front and one at rear and 1
power supply fan. The chosen power supply delivers more than enough
power for an entire system.

After about 3 months of use, one of these GeForce 2 MX400 PCI cards
began to cause 3D graphics intensive games to intermittently pause,
stray pixels and odd coloring to appear and then a few seconds later,
the entire machine would lock up. This occurred intermittently for
about two weeks until the computer finally locked up and refused to
power up.

Upon removing both identical GeForce cards from each computer, I
noticed a difference between them. On the failed board, the
manufacturing process had neglected to solder the thermal fin of one
of the regulators (lower left corner of the board) to the circuit
board, causing an eventual thermal runaway condition. The regulator
appeared lifted slightly from the board and had discolored the circuit
board from the soft yellow-beige to a burnt brown. The GPU also showed
burnt brown discoloration on the opposite side of the circuit board.

The regulator had not been properly installed in the first place
because there was no excess solder to be seen; the regulator did not
desolder itself under extreme heat conditions, rather, there was no
trace of any attempt at soldering at all. I believe that this stage of
manufacturing must be a hand-solder job and prone to many mistakes,
including poor quality assurance gates. The functioning GeForce
board's regulator was properly soldered.

I the remaining GeForce 2 card is still functioning to this day. I
bought these cards around March 2002.

Researching the frequency of this problem, through internet search
engines and newsgroups I have found that other people have experienced
similar thermal problems with many flavors of NVIDIA graphics cards
and conclude that NVIDIA has either a persistent thermal design
shortcoming or a manufacturing issue that needs resolving. I suspect
they are aware of these thermal problems, however their inability
(unwillingness?) to report or even admit these flaws on their web site
proves to me that this is yet another sociopathically irresponsible
corporation.

Beerden

Nvidia just makes chips, not the actual cards. You need to yell at
whomever actually manufactured the card as they are the ones at fault
here. Nvidia has zero control over how well (or cheaply) the cards
for their chips are made.

On a small related note, I had an old, reliable GF2MX card that gave
me intermittent problems which turns out was caused by that infamous
defective capicitor fiasco. Three of the large caps were bulging and
one had evidently leaked out some of it's electrolyte. It sucks that
my card died, but it wasn't nVidia's doing, that's for sure.

Greetings...

On 17 Jul 2003 16:58:05 -0700, (Beerden) wrote:

Researching the frequency of this problem, through internet search
engines and newsgroups I have found that other people have experienced
similar thermal problems with many flavors of NVIDIA graphics cards
and conclude that NVIDIA has either a persistent thermal design
shortcoming or a manufacturing issue that needs resolving. I suspect
they are aware of these thermal problems, however their inability
(unwillingness?) to report or even admit these flaws on their web site
proves to me that this is yet another sociopathically irresponsible
corporation.

Beerden


The one major flaw of your theory is that Nvidia doesn't make or sell Graphics cards...

They make chip sets and drivers for those chipsets.

Perhaps you should concentrate your efforts on the manufacturer of the card, and quit
blaming Nvidia for something that is NOT their problem...