Just reflowed a Samsung R70 board which wouldn't boot.

Stripped the board out.
Shielded the surrounding stuff with ali foil.
Started to heat it with a craft hot air gun (only 300w, bought to puff up
ink according to my daughter).
Couldn't get my digital temp probe to work above 150C (maybe bust) so I put
a piece of solder on the chip where there was only solder resist.
Heated it until the bit of solder melted and then for about 10 seconds more
to make sure it had penetrated the chip's substrate.
Allowed it to cool slowly by moving the gun away a bit at a time.
Reassembled and now have a working laptop.

I did manage to drop the damn keyboard and broke the ESC key.


Now I need to make sure the Nvidia drivers underclock it so it doesn't get
too hot.

"dennis@home" wrote in message
...
Just reflowed a Samsung R70 board which wouldn't boot.

Stripped the board out.
Shielded the surrounding stuff with ali foil.
Started to heat it with a craft hot air gun (only 300w, bought to puff up
ink according to my daughter).
Couldn't get my digital temp probe to work above 150C (maybe bust) so I
put a piece of solder on the chip where there was only solder resist.
Heated it until the bit of solder melted and then for about 10 seconds
more to make sure it had penetrated the chip's substrate.
Allowed it to cool slowly by moving the gun away a bit at a time.
Reassembled and now have a working laptop.

I did manage to drop the damn keyboard and broke the ESC key.


Now I need to make sure the Nvidia drivers underclock it so it doesn't get
too hot.


This technique is much quoted and demonstrated on the 'net for 'restoring'
connections that have given way on the graphics engine chip in a PS3. A
while back, I did a lot of experimenting with it, and every single time, it
worked. For about 2 weeks ...

As far as I can work out, it is not actually possible to 'reflow' the
existing solder balls by applying direct heat to the chip. The general
concensus of what is actually happening with this technique, is that the
heat produces uneven expansion between the chip and the board, and this
causes microscopic movement across the crack(s) on the failed joint(s). This
in turn causes the insulating oxide layer that has formed within the crack
from air exposure, to break up enough to get a connection back across the
crack - temporarily.

You might get away with it for quite some time, if the chip is not being
worked too hard. Underclocking may be an assistance here. With the ones that
I experimented with, I kept one running myself as just a media streamer, for
six months or more. However, just about all of the ones that came bouncing
back, were from "Call of Duty" players. This game works the graphics engine
extremely hard, so the chip gets very hot, and it seems to be this which
causes the failure in the first place, and then the re-failure after the
'repair' ...

Arfa

"Arfa Daily" wrote in message
...


"dennis@home" wrote in message
...
Just reflowed a Samsung R70 board which wouldn't boot.

Stripped the board out.
Shielded the surrounding stuff with ali foil.
Started to heat it with a craft hot air gun (only 300w, bought to puff up
ink according to my daughter).
Couldn't get my digital temp probe to work above 150C (maybe bust) so I
put a piece of solder on the chip where there was only solder resist.
Heated it until the bit of solder melted and then for about 10 seconds
more to make sure it had penetrated the chip's substrate.
Allowed it to cool slowly by moving the gun away a bit at a time.
Reassembled and now have a working laptop.

I did manage to drop the damn keyboard and broke the ESC key.


Now I need to make sure the Nvidia drivers underclock it so it doesn't
get too hot.


This technique is much quoted and demonstrated on the 'net for 'restoring'
connections that have given way on the graphics engine chip in a PS3. A
while back, I did a lot of experimenting with it, and every single time,
it worked. For about 2 weeks ...

As far as I can work out, it is not actually possible to 'reflow' the
existing solder balls by applying direct heat to the chip. The general
concensus of what is actually happening with this technique, is that the
heat produces uneven expansion between the chip and the board, and this
causes microscopic movement across the crack(s) on the failed joint(s).
This in turn causes the insulating oxide layer that has formed within the
crack from air exposure, to break up enough to get a connection back
across the crack - temporarily.

You might get away with it for quite some time, if the chip is not being
worked too hard. Underclocking may be an assistance here. With the ones
that I experimented with, I kept one running myself as just a media
streamer, for six months or more. However, just about all of the ones that
came bouncing back, were from "Call of Duty" players. This game works the
graphics engine extremely hard, so the chip gets very hot, and it seems to
be this which causes the failure in the first place, and then the
re-failure after the 'repair' ...

You can't actually fix it by re-balling either.
Its an internal design/manufacturing fault within the GPU package that is
the root cause of the problem.
I would have taken it back under the SOGA but I didn't buy it so I have no
cover under the SOGA.

I know that there is less chance it will fail if it doesn't get worked hard,
Nvidia also know this as they made the performance lower in their drivers
after the fault was found (but they didn't tell anyone).

I have set maximum power saving on the graphics chip and I am hoping it will
last longer this time, or if it does fail it may fail in a way that still
allows the machine to boot into a headless server. Before the "repair" it
was a door stop.

On 15/08/2012 08:02, dennis@home wrote:


"Arfa Daily" wrote in message
...


"dennis@home" wrote in message
...
Just reflowed a Samsung R70 board which wouldn't boot.

Stripped the board out.
Shielded the surrounding stuff with ali foil.
Started to heat it with a craft hot air gun (only 300w, bought to
puff up ink according to my daughter).
Couldn't get my digital temp probe to work above 150C (maybe bust) so
I put a piece of solder on the chip where there was only solder resist.
Heated it until the bit of solder melted and then for about 10
seconds more to make sure it had penetrated the chip's substrate.
Allowed it to cool slowly by moving the gun away a bit at a time.
Reassembled and now have a working laptop.

I did manage to drop the damn keyboard and broke the ESC key.


Now I need to make sure the Nvidia drivers underclock it so it
doesn't get too hot.


This technique is much quoted and demonstrated on the 'net for
'restoring' connections that have given way on the graphics engine
chip in a PS3. A while back, I did a lot of experimenting with it, and
every single time, it worked. For about 2 weeks ...

As far as I can work out, it is not actually possible to 'reflow' the
existing solder balls by applying direct heat to the chip. The general
concensus of what is actually happening with this technique, is that
the heat produces uneven expansion between the chip and the board, and
this causes microscopic movement across the crack(s) on the failed
joint(s). This in turn causes the insulating oxide layer that has
formed within the crack from air exposure, to break up enough to get a
connection back across the crack - temporarily.

You might get away with it for quite some time, if the chip is not
being worked too hard. Underclocking may be an assistance here. With
the ones that I experimented with, I kept one running myself as just a
media streamer, for six months or more. However, just about all of the
ones that came bouncing back, were from "Call of Duty" players. This
game works the graphics engine extremely hard, so the chip gets very
hot, and it seems to be this which causes the failure in the first
place, and then the re-failure after the 'repair' ...

You can't actually fix it by re-balling either.

Yes you can...

reball with leaded solder rather than lead free.

Also fix the heat ladder / sink so that it makes proper contact with the
device - on many laptops there is a poor contact between the metal and
the top of the device.

Its an internal design/manufacturing fault within the GPU package that
is the root cause of the problem.

BGA is not a particularly unusual form factor. Its just that in this
application there is more heat than with many devices.

I have set maximum power saving on the graphics chip and I am hoping it
will last longer this time, or if it does fail it may fail in a way that
still allows the machine to boot into a headless server. Before the
"repair" it was a door stop.

Note that on many laptops with nVidea GPUs, the GPU is used for running
the LCD, but not always the external VGA (this is sometimes embedded in
the south bridge chipset). Hence with laptops, you can sometimes get
normal operation by connecting to an external monitor.

--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

"John Rumm" wrote in message
o.uk...

You can't actually fix it by re-balling either.

Yes you can...

reball with leaded solder rather than lead free.

The fault is inside the package.
Re-balling it doesn't fix the problem, it just fixes the symptoms.


Also fix the heat ladder / sink so that it makes proper contact with the
device - on many laptops there is a poor contact between the metal and the
top of the device.

Its the internal heat sink that is the problem, not the external one.
You can fit a perfect heat sink to the outside and it still wont be correct.


Its an internal design/manufacturing fault within the GPU package that
is the root cause of the problem.

BGA is not a particularly unusual form factor. Its just that in this
application there is more heat than with many devices.

In this case the thermal bits inside the package not working correctly is
the problem.


I have set maximum power saving on the graphics chip and I am hoping it
will last longer this time, or if it does fail it may fail in a way that
still allows the machine to boot into a headless server. Before the
"repair" it was a door stop.

Note that on many laptops with nVidea GPUs, the GPU is used for running
the LCD, but not always the external VGA (this is sometimes embedded in
the south bridge chipset). Hence with laptops, you can sometimes get
normal operation by connecting to an external monitor.

Not on this one, it wouldn't even boot.

On 15/08/2012 19:31, dennis@home wrote:


"John Rumm" wrote in message
o.uk...

You can't actually fix it by re-balling either.

Yes you can...

reball with leaded solder rather than lead free.

The fault is inside the package.

Not sure what gives you that idea.

Re-balling it doesn't fix the problem, it just fixes the symptoms.

You mean it takes the symptom of it "not working" and fixes it?

Also fix the heat ladder / sink so that it makes proper contact with
the device - on many laptops there is a poor contact between the metal
and the top of the device.

Its the internal heat sink that is the problem, not the external one.
You can fit a perfect heat sink to the outside and it still wont be
correct.

So that is how you explain how all these "internally broken" BGA
packages that can be reballed on the outside, and then work perfectly
for years after.

Its an internal design/manufacturing fault within the GPU package that
is the root cause of the problem.

BGA is not a particularly unusual form factor. Its just that in this
application there is more heat than with many devices.

In this case the thermal bits inside the package not working correctly
is the problem.

ok den whatever you say...



--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

"John Rumm" wrote in message
o.uk...
On 15/08/2012 19:31, dennis@home wrote:


"John Rumm" wrote in message
o.uk...

You can't actually fix it by re-balling either.

Yes you can...

reball with leaded solder rather than lead free.

The fault is inside the package.

Not sure what gives you that idea.

Re-balling it doesn't fix the problem, it just fixes the symptoms.

You mean it takes the symptom of it "not working" and fixes it?

It not working correctly is just a symptom of the problem.


Also fix the heat ladder / sink so that it makes proper contact with
the device - on many laptops there is a poor contact between the metal
and the top of the device.

Its the internal heat sink that is the problem, not the external one.
You can fit a perfect heat sink to the outside and it still wont be
correct.

So that is how you explain how all these "internally broken" BGA packages
that can be reballed on the outside, and then work perfectly for years
after.

Well as I already said nvidia have changed the driver so they aren't
stressed as much to start with.
Before the new driver I could easily get GPU temps over 120C now they sit at
around 60C even though the CPU is only 46C.
That is the die temps BTW not the package which is a lot lower due to the
heat sink.


Its an internal design/manufacturing fault within the GPU package that
is the root cause of the problem.

BGA is not a particularly unusual form factor. Its just that in this
application there is more heat than with many devices.

In this case the thermal bits inside the package not working correctly
is the problem.

ok den whatever you say...

Well you *could* check with nvidia and then understand the truth rather than
claiming I am wrong.
Go and argue with
http://apcmag.com/nvidia_disaster_th...pus_faulty.htm or someone
that actually has any belief that you don't just argue to be a PITA.

On 16/08/2012 15:43, dennis@home wrote:


"John Rumm" wrote in message
o.uk...
On 15/08/2012 19:31, dennis@home wrote:


"John Rumm" wrote in message
o.uk...

You can't actually fix it by re-balling either.

Yes you can...

reball with leaded solder rather than lead free.

The fault is inside the package.

Not sure what gives you that idea.

Re-balling it doesn't fix the problem, it just fixes the symptoms.

You mean it takes the symptom of it "not working" and fixes it?

It not working correctly is just a symptom of the problem.

Interesting semantic angle...

Also fix the heat ladder / sink so that it makes proper contact with
the device - on many laptops there is a poor contact between the metal
and the top of the device.

Its the internal heat sink that is the problem, not the external one.
You can fit a perfect heat sink to the outside and it still wont be
correct.

So that is how you explain how all these "internally broken" BGA
packages that can be reballed on the outside, and then work perfectly
for years after.

Well as I already said nvidia have changed the driver so they aren't
stressed as much to start with.
Before the new driver I could easily get GPU temps over 120C now they
sit at around 60C even though the CPU is only 46C.
That is the die temps BTW not the package which is a lot lower due to
the heat sink.


Its an internal design/manufacturing fault within the GPU package that
is the root cause of the problem.

BGA is not a particularly unusual form factor. Its just that in this
application there is more heat than with many devices.

In this case the thermal bits inside the package not working correctly
is the problem.

ok den whatever you say...

Well you *could* check with nvidia and then understand the truth rather
than claiming I am wrong.
Go and argue with
http://apcmag.com/nvidia_disaster_th...pus_faulty.htm or
someone that actually has any belief that you don't just argue to be a
PITA.

The (non nVidea) article you cite claims problems with two specific GPUs.

You appear to have extended that to all BGA GPUs.



--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

"John Rumm" wrote in message
o.uk...


The (non nVidea) article you cite claims problems with two specific GPUs.

You appear to have extended that to all BGA GPUs.

Where?

Unlike you I have stated it was an R70, this has an 8400m GPU.

It is you that is claiming all faults can be fixed by re-balling and
claiming there are no internal faults on the nvidia chips.

Its not the only GPU that nvidia have screwed up, but there are different
reasons for some of them.
Indeed some of them fail due to useless heat sinks in particular machines,
however not in this case as they fail prematurely even in machines with good
heat sinks.

Do you intend to continue with your argument even though you are totally
wrong?

John Rumm wrote:
On 16/08/2012 15:43, dennis@home wrote:


"John Rumm" wrote in message
o.uk...
On 15/08/2012 19:31, dennis@home wrote:


"John Rumm" wrote in message
o.uk...

You can't actually fix it by re-balling either.

Yes you can...

reball with leaded solder rather than lead free.

The fault is inside the package.

Not sure what gives you that idea.

Re-balling it doesn't fix the problem, it just fixes the symptoms.

You mean it takes the symptom of it "not working" and fixes it?

It not working correctly is just a symptom of the problem.

Interesting semantic angle...

A direct translation from the original Esperanto?


--
Ineptocracy

(in-ep-toc-ra-cy) €“ a system of government where the least capable to
lead are elected by the least capable of producing, and where the
members of society least likely to sustain themselves or succeed, are
rewarded with goods and services paid for by the confiscated wealth of a
diminishing number of producers.

On 16/08/2012 16:50, dennis@home wrote:


"John Rumm" wrote in message
o.uk...


The (non nVidea) article you cite claims problems with two specific GPUs.

You appear to have extended that to all BGA GPUs.

Where?

Unlike you I have stated it was an R70, this has an 8400m GPU.

Have you considered a trip to your GP to discuss your memory problems?

A day ago Arfa was discussing repairs to the PS3 GPU, effected by hot
air reflows. He expressed his doubt that reflowing would effect a long
term repair. You then claimed:

"You can't actually fix it by re-balling either."

Which seems an odd claim since in the case of the PS3, this is the
standard fix and generally seems to be an effective long term fix.

You also claimed "Its an internal design/manufacturing fault within the
GPU package" supporting this with an article discussing a GPU which is
not even used in the PS3.

It is you that is claiming all faults can be fixed by re-balling and
claiming there are no internal faults on the nvidia chips.

Again this would seem to be your failure to comprehend.

Its not the only GPU that nvidia have screwed up, but there are
different reasons for some of them.
Indeed some of them fail due to useless heat sinks in particular
machines, however not in this case as they fail prematurely even in
machines with good heat sinks.

PS3s den?

Do you intend to continue with your argument even though you are totally
wrong?

Sorry your question is a non sequitur.


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

"John Rumm" wrote in message
...
On 16/08/2012 16:50, dennis@home wrote:


"John Rumm" wrote in message
o.uk...


The (non nVidea) article you cite claims problems with two specific
GPUs.

You appear to have extended that to all BGA GPUs.

Where?

Unlike you I have stated it was an R70, this has an 8400m GPU.

Have you considered a trip to your GP to discuss your memory problems?

have you?


A day ago Arfa was discussing repairs to the PS3 GPU, effected by hot air
reflows. He expressed his doubt that reflowing would effect a long term
repair. You then claimed:



"You can't actually fix it by re-balling either."

In this thread which has nothing to do with PS3.
Are you confused?
You sure appear to be.


Which seems an odd claim since in the case of the PS3, this is the
standard fix and generally seems to be an effective long term fix.

You also claimed "Its an internal design/manufacturing fault within the
GPU package" supporting this with an article discussing a GPU which is not
even used in the PS3.

This thread doesn't talk about PS3s at all, are you confused?
Let me remind you that its about a R70 laptop as stated in the original
post.
This has the 8400M chip as talked about in the article as was previously
stated.


It is you that is claiming all faults can be fixed by re-balling and
claiming there are no internal faults on the nvidia chips.

Again this would seem to be your failure to comprehend.

Are you sure you aren't confused?


Its not the only GPU that nvidia have screwed up, but there are
different reasons for some of them.
Indeed some of them fail due to useless heat sinks in particular
machines, however not in this case as they fail prematurely even in
machines with good heat sinks.

PS3s den?

PS3 probably has heat sink problems but this thread isn't about PS3.
The fix is to reflow it and make the heat sink better in case you wonder.


Do you intend to continue with your argument even though you are totally
wrong?

Sorry your question is a non sequitur.


I suggest you see your GP.
Your argument is still totally wrong and isn't even anything to do with this
thread.
Please try to remember this as it will make life easier, maybe you can write
it down somewhere obvious?

On 16/08/2012 22:58, dennis@home wrote:


"John Rumm" wrote in message
...
On 16/08/2012 16:50, dennis@home wrote:


"John Rumm" wrote in message
o.uk...


The (non nVidea) article you cite claims problems with two specific
GPUs.

You appear to have extended that to all BGA GPUs.

Where?

Unlike you I have stated it was an R70, this has an 8400m GPU.

Have you considered a trip to your GP to discuss your memory problems?

have you?

I have little interest in your memory problems, and certainly not enough
to draw them to the attention of my GP.


A day ago Arfa was discussing repairs to the PS3 GPU, effected by hot
air reflows. He expressed his doubt that reflowing would effect a long
term repair. You then claimed:



"You can't actually fix it by re-balling either."

In this thread which has nothing to do with PS3.

Arfa said:

"This technique is much quoted and demonstrated on the 'net for
'restoring' connections that have given way on the graphics engine chip
in a PS3. A while back, I did a lot of experimenting with it, and every
single time, it worked. For about 2 weeks ... "

Are you confused?
You sure appear to be.


Which seems an odd claim since in the case of the PS3, this is the
standard fix and generally seems to be an effective long term fix.

You also claimed "Its an internal design/manufacturing fault within
the GPU package" supporting this with an article discussing a GPU
which is not even used in the PS3.

This thread doesn't talk about PS3s at all, are you confused?

see above.

Let me remind you that its about a R70 laptop as stated in the original
post.

I recall. However you were replying to comments about PS3 GPUs.

Time for bed den, you need your beauty sleep.


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

"John Rumm" wrote in message
.. .


Let me remind you that its about a R70 laptop as stated in the original
post.

I recall. However you were replying to comments about PS3 GPUs.

No I wasn't.
I have even gone to the extent of stating that only some nvidia chips suffer
from the problem.
I was replying to the comment about reflowing only working for two weeks.

Then you claimed that re-balling always fixed things, which it doesn't, not
even on PS3.
Since then you have been ducking and diving to avoid other people knowing
you were wrong again.

If you don't have anything useful to add, don't reply, it does you no good.

On 17/08/2012 07:57, dennis@home wrote:


"John Rumm" wrote in message
.. .


Let me remind you that its about a R70 laptop as stated in the original
post.

I recall. However you were replying to comments about PS3 GPUs.

No I wasn't.

whatever, the evidence is there, but I doubt anyone cares.

--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

In message , John
Rumm writes
On 17/08/2012 07:57, dennis@home wrote:


"John Rumm" wrote in message
.. .


Let me remind you that its about a R70 laptop as stated in the original
post.

I recall. However you were replying to comments about PS3 GPUs.

No I wasn't.

whatever, the evidence is there, but I doubt anyone cares.

Other than watching dense squirm, no not really


--
geoff