P55-UD6 socket burn

[SF3D]

Stop speculating like that Sascha It sounds like me and Pieter are some first timers with our motherboards.
No one noticed Pt1t reply in the first page? Biostar have got the same issues with their early boards and with these sockets. They installed cooler a bit too tight and zzzzip!

Two different CPU's. Different voltages and different configurations. Both sockets are burned from VCC and VSS pin area. It is not coincidence.
I think I know the reason. Too heavy compression to socket and CPU. My LN2 pot is usually tightened very well, so this might be as simple as that.

My own fault

Only advice is that do not bend the board and do not use too much force when installine LN2 pot or some other cooler.

[[XC] gomeler]

Sounds like the amount of current being pushed through the pins caused them to heat up excessively which further increases resistance which further increases heat, repeat ad infinitum. Combine this with the clamping pressure of your heatsinks pressing the CPU and pins so that the pins are making contact with the plastic backing(like SF3D just mentioned) of the socket and voila, melted plastic socket with a side of roasted CPU. I'd say you need thicker pins in order to successfully transfer these amounts of current without causing this to happen. The pins shouldn't get that hot..

[Raja@ASUS]

or dust or lint in the air landing in the open socket..

i've never liked LGA sockets.

The increased lint in the air must be from all the tissues people are using due to swine flu. Seems to be 'intelli-lint', in that it's attracted to everywhere but the top left corner of the socket on P55 boards.

[massman]

My Vcore was not even that high. 1.57V maximum. Maybe the issue is load calibration or something reltated to that.

I had the melting down the only time I set loadline to "level2". Relatively low voltage as well (1.58-ish) and cooler wasn't attached particularly tight. Board wasn't bent.

[kiwi]

OK, board and cpu is working but will it clock and be benchable to similar speeds as before?

5cl Some genuine Vodka

Mate, it is not just genuine vodka It's one of the best vodkas and my favorite
Do they sell this in Finland too?

[Raja@ASUS]

Seems around 120-130 amps is enough to cause this according to what I saw on the clamp meter on the CPU I've got with the burnt pad.

[Drwho?]

Just a heads up for anyone running the UD6 board: both SF3D and me have had an situation where the socket seems to be burned. My combo is still alive, but SF3D's might be dead.

I'm not pointing fingers or trying to make anyone look bad. Just a heads up; this might be an unfortunate coincidence.

Think about it ... at 1.65V, you are around 250watts ... this is 150AMPS!
gigabyte gave what the Overclockers asked ... now, use it with moderation ... I tested on a mecanical thermal dummy ... the UD6 can deliver up to 160AMPS without occilation ... at this point, you are more than 2x the specs for currents ... and 60% over the voltage ...
and 263% of the based power ...
OC is like alcool, use it with moderation ..

Gigabyte gave what you guys ask ... you are now like F1 drivers ... if you try to corner at 280km/h, you ll end up in the wall

May the Core be with you!

Francois

[Raja@ASUS]

I don't even think you need 160 amps Francois - I think 120-130a can cause it. That being said, i5 never was designed to upstage i7 so I guess you get what you pay for (not that I'm complaining).

[proxis]

Gigabyte gave what you guys ask ...

when was the last time you heard an overclocker praying "16 phases is not enough , please give me more" ?
okay, we got the "more" anyway and what we ended up with?

[HDCHOPPER]

Sounds like the amount of current being pushed through the pins caused them to heat up excessively which further increases resistance which further increases heat, repeat ad infinitum. Combine this with the clamping pressure of your heatsinks pressing the CPU and pins so that the pins are making contact with the plastic backing(like SF3D just mentioned) of the socket and voila, melted plastic socket with a side of roasted CPU. I'd say you need thicker pins in order to successfully transfer these amounts of current without causing this to happen. The pins shouldn't get that hot..

You got it man ! resistance makes heat , clamping pressure so high the cpu rest on plastic and the "pins" all have there own touching pressure & multi phased 16+ the shape of the pins remind me of sharpening tungsten to weld moly
only the points melt after a while with a tiny tiny ball on the end after done,

Things are getting mighty precision now and need a clamp & socket to match that pwr ( need more flatplate contact at the pad per pin )...

the old saying carefull what you ask for u just might get it ....just my 2 cent

[Drwho?]

I don't even think you need 160 amps Francois - I think 120-130a can cause it. That being said, i5 never was designed to upstage i7 so I guess you get what you pay for (not that I'm complaining).

I did some study ... just because I wanted to see where it can go ...

So, oxidation on the pin or the contacts of the processor increase their resistance by a little ... 0.1 Ohm ... but at 120 or 160amps ... this is a lot of heat to discipate ... think about it:

... (ohm law)
(seconds formula on the left up corner ... P= IxI x R )

When the socket guys design all of this, they actually pay a lot of attention to this.

my guess in those cases is "finger print" or oxidation due to Liquide nitro condensation plus an incredible UD6 voltage supply that does not bend under load. (thing requested by the OC community)

so, make sure you don't condensate inside the socket, and use dry hands on the pins and contacts. (when I OC with low temps, I use a non conductive silicon grease to fill the socket, it avoid condensation 100% ... a little messy , but save for hours)

Francois

[kiwi]

Or just some vaseline

[Hondacity]

i'm NOT quite entirely sure ..but vcc is only one pin on this i5 series? how many vcc pins does core i7 have? one or more?

[Drwho?]

i'm quite entirely sure ..but vcc is only one pin on this i5 series? how many vcc pins does core i7 have? one or more?

http://download.intel.com/design/pro...hts/322164.pdf

page 88 to 91 !

-----------------------------------------------------------

http://download.intel.com/design/pro...nex/322167.pdf
you can figure out with this diagram what burned.
G38/39 or H38/39 are the pins that burned, and it looks like it is VCC

read the documentation before doing funny claims like this ...
one pin ... i almost swallowed the wrong way my coke when I read this.

Francois

[Hondacity]

so it was more than one..i can't imagine 100A on one pin LOL

thanks Francois



***i just noticed "not" should have been in my first statement, "i'm not quite entirely sure"

[chew*]

Yeah just checked - A38 line.

Mine did not happen on the GB board - but board is ok (no burned socket pins though) - only 1 CPU pad is gone. Did not use any more than 1.50VID.

I did some study ... just because I wanted to see where it can go ...

So, oxidation on the pin or the contacts of the processor increase their resistance by a little ... 0.1 Ohm ... but at 120 or 160amps ... this is a lot of heat to discipate ... think about it:

... (ohm law)
(seconds formula on the left up corner ... P= IxI x R )

When the socket guys design all of this, they actually pay a lot of attention to this.

my guess in those cases is "finger print" or oxidation due to Liquide nitro condensation plus an incredible UD6 voltage supply that does not bend under load. (thing requested by the OC community)

so, make sure you don't condensate inside the socket, and use dry hands on the pins and contacts. (when I OC with low temps, I use a non conductive silicon grease to fill the socket, it avoid condensation 100% ... a little messy , but save for hours)

Francois

Seems it's not just Gigabyte boards Francios......

So if its not giga and gigas "phase"

It's either the same socket used by 2 companies ( which is specced by intel ) or the other common denominator......the cpu...

[saaya]

or dust or lint in the air landing in the open socket..

i've never liked LGA sockets.

well... if you open any oldschool socket, it actually looks about the same as an lga socket... its under a plastic mesh, but i doubt that prevents it from getting dirty... on the contrary, the dust and dirt will find a way in, always, but it wont be able to be brushed off like in an lga socket... i think when it comes to dust and dirt, an lga socket is actually better cause you CAN clean in... on a traditional socket you have to take the socket apart and even then its tricky to clean... once you got vaseline or anything else inside a traditional socket, forget about ever getting rid of it again

Stop speculating like that Sascha

hm? speculating like what?

Seems around 120-130 amps is enough to cause this according to what I saw on the clamp meter on the CPU I've got with the burnt pad.

hmmm that chip suck a lot? current wise i mean

i'm quite entirely sure ..but vcc is only one pin on this i5 series? how many vcc pins does core i7 have? one or more?

vcc is power, and its a whole lot more than 1
vcc and vss are usually 30% of the pins iirc... quite a lot...
for 1366 there are around 200 vcc and 200 vss pins... well there are more vss pins but those are the vss pins for vcore...
1366 cpus are rated for up to 145A which means each pin is close to 1A... and it literally is just a tip of a needle... looking at the socket pins is actually quite deceiving, since they will bend down and left compared to where they are on the pic... the ones that were molten into the plastic of the socket are still in their original position further down and left...

looking at the cpu pads, there are both burned ground and vcore pads... and interesting is that away from the main spot, lets say ground zero, there is a diagonal row of 3 pins that shows slight burn marks as well... those pins are ground! yet no vcc pin close to them shows any burn signs... so where did the current come from?

basically most of this corner is vcore pads, with diagonal rows of ground pads seperating them. 1 diagonal row of ground pads and then 2 rows of vcore pads... ground zero is exactly between 2 vcore rows and one ground row... since there is a mesh inside the cpus package connecting all ground and vcore pins to even out fluctuations and pins making bad contact, it would explain how the further away ground pads got burned...

if you think of how the socket closes, the top is definately where the pins experience the most pressure... and the left is where all the vcore pins are located, so thats probably why it happened on the top left corner.

those 3 singed ground pins are definately sticking out... why would they be slightly burned? it means there must have been a lot of current flowing through them... why did it flow through those 3 pins? cause they are grounded slightly better than other vss pins i guess...

it looks like the other ground pins close to it couldnt ground well enough, so a lot of current went through those 3 pins instead, and not the other pins... why did the other ground pins not ground well enough? cause they were hot and had a higher resistance? or they were making bad contact?

[hipro5]

I did some study ... just because I wanted to see where it can go ...

So, oxidation on the pin or the contacts of the processor increase their resistance by a little ... 0.1 Ohm ... but at 120 or 160amps ... this is a lot of heat to discipate ... think about it:

... (ohm law)
(seconds formula on the left up corner ... P= IxI x R )

When the socket guys design all of this, they actually pay a lot of attention to this.

my guess in those cases is "finger print" or oxidation due to Liquide nitro condensation plus an incredible UD6 voltage supply that does not bend under load. (thing requested by the OC community)

so, make sure you don't condensate inside the socket, and use dry hands on the pins and contacts. (when I OC with low temps, I use a non conductive silicon grease to fill the socket, it avoid condensation 100% ... a little messy , but save for hours)

Francois

IF you take a CLOSER look at the photos the guys have uploaded, you will see that there was NO oxidation at ALL.....
Why?
Because those pins would have taken a white/green colour and the plastic around them would have been somehow white....
I can see a VERY CLEAN socket and golden pins BUT melted plastic....
PLUS the pins where "down" NOT only from pressure but they just lost their "flexibility" and shape due to 2 reasons:
a. BAD metal material (for flexibility) and
b. TOO much current for their "shape" and their material so when the current was rolling through them, the pins started to get hitted much and - maybe - they got RED from the temperature and they have melted the plastic of the socket...
So for me it's not Gigabyte or any Gigabyte’s fault. It's the company that has made them and shells these sockets to mobo manufacturers....

Guys what make is that socket?...

P.S. @Drwho: Would Intel RMA such a CPU now or not?....and IF NOT, whom these guys would blame if they don't get their money back that have paid (CPU and mobo) WORKING for a month to pay for them? (regular guys that are from "poor countries") - if you can understand what I'm talking about?
For SURE an employee in the computer shop, will NOT RMA this CPU and this Mobo with bended/melted socket/pins...

[webbyboy]

IF you take a CLOSER look at the photos the guys have uploaded, you will see that there was NO oxidation at ALL.....
Why?
Because those pins would have taken a white/green colour and the plastic around them would have been somehow white....
I can see a VERY CLEAN socket and golden pins BUT melted plastic....
PLUS the pins where "down" NOT only from pressure but they just lost their "flexibility" and shape due to 2 reasons:
a. BAD metal material (for flexibility) and
b. TOO much current for their "shape" and their material so when the current was rolling through them, the pins started to get hitted much and - maybe - they got RED from the temperature and they have melted the plastic of the socket...
So for me it's not Gigabyte or any Gigabyte’s fault. It's the company that has made them and shells these sockets to mobo manufacturers....

Guys what make is that socket?...

P.S. @Drwho: Would Intel RMA such a CPU now or not?....and IF NOT, whom these guys would blame if they don't get their money back that have paid (CPU and mobo) WORKING for a month to pay for them? (regular guys that are from "poor countries") - if you can understand what I'm talking about?
For SURE an employee in the computer shop, will NOT RMA this CPU and this Mobo with bended/melted socket/pins...

foxconn?

[hipro5]

when was the last time you heard an overclocker praying "16 phases is not enough , please give me more" ?
okay, we got the "more" anyway and what we ended up with?

Actually there is NO NEED for 164 phases that companies put in nowadays mobos.... FOUR (4) are enough with nowadays materials and maybe SIX (6) with the upcoming Gulftowns....
The 1024 phases is a "marketing trick" nowadays....
You will buy MY mobo because I put 2048 phases in my motherboard....

Someone please calculate the Wattage of a P4 Gallatin at 2.23VCore at 4600MHz with ONLY FOUR phases and tell us what you see.....

foxconn?

No, I don't believe so....It's not "their style"...

[Raja@ASUS]

The new Volterra slaves are 40~45amps each, can switch over 1MHz - 6 to 8 phase of those and you can more than max out CPU's even on LN2.

[Drwho?]

IF you take a CLOSER look at the photos the guys have uploaded, you will see that there was NO oxidation at ALL.....
Why?
Because those pins would have taken a white/green colour and the plastic around them would have been somehow white....
I can see a VERY CLEAN socket and golden pins BUT melted plastic....
PLUS the pins where "down" NOT only from pressure but they just lost their "flexibility" and shape due to 2 reasons:
a. BAD metal material (for flexibility) and
b. TOO much current for their "shape" and their material so when the current was rolling through them, the pins started to get hitted much and - maybe - they got RED from the temperature and they have melted the plastic of the socket...
So for me it's not Gigabyte or any Gigabyte’s fault. It's the company that has made them and shells these sockets to mobo manufacturers....

Guys what make is that socket?...

P.S. @Drwho: Would Intel RMA such a CPU now or not?....and IF NOT, whom these guys would blame if they don't get their money back that have paid (CPU and mobo) WORKING for a month to pay for them? (regular guys that are from "poor countries") - if you can understand what I'm talking about?
For SURE an employee in the computer shop, will NOT RMA this CPU and this Mobo with bended/melted socket/pins...

your opinion is a title too definitive for my taste, and based on no science ... figuring out the cause of this with the color of the plastic is "funny" at most.
you like looking for "fight", as your PS try to, I gave my opinion, you are free to dissagree ... you should try a less caustic style.

Look at the specs, do some experientation, and then conclude, this is the way it works in computer science.
FYI :Oxidation is rarely visible to the eyes on the processor connection pads, at least, when it is around few 1/10 of Ohms ... if it was "green" due to Copper oxidation:
1) those metal mix don't oxidize because copper usually
2) If it was visible to the eyes, you ll be in the 1 Ohm area...

I tested those thing on Skulltrail, then I did my Dual Chiller configuration , and it went it smoke after 2 months of running. I usually spend the time to understand my mistake scientifically.

lol ...

Francois

[massman]

Well, would Intel RMA this CPU? I reckon not?

[Drwho?]

Actually there is NO NEED for 164 phases that companies put in nowadays mobos.... FOUR (4) are enough with nowadays materials and maybe SIX (6) with the upcoming Gulftowns....
The 1024 phases is a "marketing trick" nowadays....
You will buy MY mobo because I put 2048 phases in my motherboard....

Someone please calculate the Wattage of a P4 Gallatin at 2.23VCore at 4600MHz with ONLY FOUR phases and tell us what you see.....



No, I don't believe so....It's not "their style"...

I trend to agree that over a certain number of VREG, there are not much difference, never the less, in cornet case , adding voltage source and current source in parrallel (as explain in Norton theorem ( http://en.wikipedia.org/wiki/Norton's_theorem or Thévenin for the hell of it ) , there is a benefit when the work load is heavier ... In the case of overclocking, it will cause more stable current when you are pulling hard on it.
It can cause some current rush too, so, it is an interesting trade between cost and perfection.

When you look at what overclocking is about, it is about making hundreds millions of transistors pulsing faster than their specs, having a more stable VCC is a plus ... it is hard to do the science behind this, but a variation of few mV during a massive pull of 200W can cause transistors to stop reconizing the same switching point over the dice, due to electric propagation. This is how multiphase actually help.
If you want to check the minimum Phare required, it is fairly simple, remove them one per one, and run at 4.4Ghz ... you will discover than under a certain number, the VREG warm up, start ocillating, and cause crashed ... I did this study on Skulltrail ... I havr no data for P55.

[Drwho?]

Well, would Intel RMA this CPU? I reckon not?

it did obviously run out of its specs ... so no RMA. Fair!