P55-UD6 socket burn

Quote:

---

Originally Posted by Gautam

/resists making sauna joke

---

:rofl::ROTF:

VERY CHEAPY socket from what I see through the photos.....The pins are just AWFULL in this socket (I mean the new pins)....
What make is it..... :(

a cheap greek drained in Metaxa Foxconn derative probably....

Quote:

---

Originally Posted by Leeghoofd

a cheap greek drained in Metaxa Foxconn derative probably....

---

What exactly do you mean by that because I didn't understand it?... :)

I cant understand that either Leeghoofd..


My motherboard is not usable anymore. Pins are melted down to plastic, so they will not make contact to the CPU anymore.

Quote:

---

Originally Posted by raju

yeah mine was clamped properly. I had thought about non contacting pins but that can't be right as only one pin has gone on mine while the surrounding ones are blemish free (no visibly bent pins either). I will try to take a macro pic if I can, there is very little pad left. Looks like some i5 CPU's can have issues when pushed. Mine was not even pushed that hard, I just about got it to 4.8. No moisture anywhere as the whole socket and CPU were completely sealed from the air with eraser. Just one of those things.

---

did you use vaseline inside the cpu socket around the pins? if no then it could still have been insulation... or too much pressure maybe?

the socket pins are bad hipro? what exactly is wrong with them? bad shape or cheap metal?

Quote:

---

Originally Posted by saaya

did you use vaseline inside the cpu socket around the pins? if no then it could still have been insulation... or too much pressure maybe?

the socket pins are bad hipro? what exactly is wrong with them? bad shape or cheap metal?

---

In this UD6 case, insulation and condensation is not the problem. I believe, that these pins will get too hot under heavy conditions and they will just burn.

Well.. it can't be the whole truth, cause I did not even push my board yet. 1.57V was the max voltage in tests :shrug:

Quote:

---

Originally Posted by saaya

did you use vaseline inside the cpu socket around the pins? if no then it could still have been insulation... or too much pressure maybe?

the socket pins are bad hipro? what exactly is wrong with them? bad shape or cheap metal?

---

It was selaed - the whole socket with eraser. We're only talking one lone pad on mine, no blemishes elsewhere (how likely is that?). I know you're trying to be thorough - but believe me I already checked for everything.No burned socket pins on mine; just the single CPU pad and everything around it is perfect. if the socket were at fault btw, it would certainly have been nailed by my other CPU which I gave more voltage and clocked higher.

3 guys with insulation problems within a few days and all of us at the same part of the CPU? Clamping pressure was what I normally use for my other CPU too, no problems there..

Quote:

---

Originally Posted by SF3D

In this UD6 case, insulation and condensation is not the problem. I believe, that these pins will get too hot under heavy conditions and they will just burn.

Well.. it can't be the whole truth, cause I did not even push my board yet. 1.57V was the max voltage in tests :shrug:

---

hmmmm maybe current is the prob and not volts?

Quote:

---

Originally Posted by raju

It was selaed - the whole socket with eraser. We're only talking one lone pad on mine, no blemishes elsewhere (how likely is that?). I know you're trying to be thorough - but believe me I already checked for everything.No burned socket pins on mine; just the single CPU pad and everything around it is perfect. if the socket were at fault btw, it would certainly have been nailed by my other CPU which I gave more voltage and clocked higher.

3 guys with insulation problems within a few days and all of us at the same part of the CPU? Clamping pressure was what I normally use for my other CPU too, no problems there..

---

well condensation probs are likely to happen at the same spot...
hmmm then what do you think the problem is?

bad socket quality i guess? i wouldnt be surprised cause foxconn had probs with the 1366 sockets as well, the early batches were sensitive to get the pins bent easily...

but how can a bad socket cause this? how can all the sockets burn at the same space? maybe thats where the cpus draw the most current through the socket? or thats where the pressure is the highest so the pins make the best contact, other pins make bad contact and even more current flows through those pins, they heat up and bam?

I would say bad connection or to much draw for those thin pins.....Burn means its getting hot.

Think about the burned traces on the classies running 4x 4890.....to small a trace for to much of a draw...

A similar test could be duplicated by trying to start your car with 10 guage wire....it would just burn right up and in very short time.

chew

if that's the case you would think then that 1.7vcore with 1.66VTT would just vaporise the pins right :shrug:

all my testing has been done on this UD6 so far

Quote:

---

Originally Posted by saaya

hmmmm maybe current is the prob and not volts?


well condensation probs are likely to happen at the same spot...
hmmm then what do you think the problem is?

bad socket quality i guess? i wouldnt be surprised cause foxconn had probs with the 1366 sockets as well, the early batches were sensitive to get the pins bent easily...

but how can a bad socket cause this? how can all the sockets burn at the same space? maybe thats where the cpus draw the most current through the socket? or thats where the pressure is the highest so the pins make the best contact, other pins make bad contact and even more current flows through those pins, they heat up and bam?

---

It could be a in CPU grid distribution issue. If the socket were at fault the failures would not be 'random' in that X amount of current would cause a pin meltdown (the amount of current for critical meltdown is static, not a 'random factor'). If you use 0.1 VID more on one CPU than another and you've used a clamp meter to measure line draw - CPU B draws more than CPU A - why does CPU A suffer form a burned out pad if the socket has BGA soldering issues or weak pins?

That top side of socket is easy to insulate on the board I have here, in fact it is on the GB board too. For condensation to isolate a lone pad in my case if air is getting to the CPU in sub zero conditions is a weird supposition.


http://img185.imageshack.us/img185/3945/burnedpad.jpg

the part of that pad that is missing was kinda stuck to the socket pin (weakly fused). It came off when I pulled it, the socket pin is fine and has no damage or signs of distress. Like I said, very isolated for condesation and no pin damage, points to a CPU side issue rather than the board IMO.

so you think the pin grid in the cpu package doesnt distribute the current load well enough, so current draw can vary and spike in a single pin/pad contact and then burn out?
makes sense... could also be the cpu socket not distributing the load among the pins well enough... that would be odd for an ultra durable board with thicker power and ground layers tho :D

Quote:

---

Originally Posted by dinos22

chew

if that's the case you would think then that 1.7vcore with 1.66VTT would just vaporise the pins right :shrug:

all my testing has been done on this UD6 so far

---

Hmm well resistance can also cause heat......maybe the Problem lies not in the socket but the traces in the cpu itself.....The socket is just where the problem manifests itslef.

Quote:

---

Originally Posted by saaya

....the socket pins are bad hipro? what exactly is wrong with them? bad shape or cheap metal?

---

I believe that they were constructed in WRONG shape for these currents... ;)

between the processor and motherboard pins short-circuit if the water does not.
Single stage When the humidity can sometimes be, I think this is the case.

Quote:

---

Originally Posted by massman

I had the burn a couple of days ago when I used my single stage cooling. Re-ran the setup yesterday evening under LN2 and it didn't got worse. As long as it works, I'll keep using the platform ;-).

Btw, I'm not saying this is a general UD6 error. Maybe it's a end-user :banana::banana::banana::banana:up, maybe I did something that caused this. I just posted this because apparently SF3D has had the same issue.

---

hehehe you little killer :ROTF::up:

I had to take some "medicine" for my nervers yesterday ;)

http://www.aijaa.com/img/b/00644/4914204.jpg

OC Drink:

5cl Some genuine Vodka
Intel "Adrenaline on a chip" energy drink
Ice
LN2 for cover

Then I decided to look the socket one more time today:

http://www.aijaa.com/img/b/00603/4914247.jpg

It did look quite nasty and the pins were melted down in to the plastic.

I took sharp needle and started to fix it.

After 30 minutes it was looking like this:

http://www.aijaa.com/img/b/00384/4914273.jpg

Board and the CPU are both booting and working like this tragedy never even happened :up:

Burned UD6 board count = 0 :D

A couple of questions that crossed my mind.....

Were both chips '750s?

Have they been tried in a different board?



Are the 7 and 8 series on identical substrates or are the 8 series reinforced because of the HT?



It could be a dodgy batch of packaging...

Did anything jump out on first inspection of the 1156 socket when you received the boards?

Quote:

---

Originally Posted by K404

A couple of questions that crossed my mind.....

Were both chips '750s?

Have they been tried in a different board?

---

Mine was, I don't SF3D sees anything but the high-end chips :D

I tried this chip in the GD80 before, no issues. Nothing special about the socket.

rma the cpu and see what intel says happened :D

well does the ud6 use a special socket version other boards dont use?
i doubt gigabytes ud6 is the only board using the same version of this socket, so im not convinced its the socket...
im not sure it has anything to do with the board either... couldnt it just be that, since this happened at the edge of the cpus, they had a fingerprint on them, possible with thermal grease so there was a thin layer of grease on it, preventing good contact and then resulting in higher resistance, more heat, even higher resistance, and then the plastic melted? the area this happened is exactly the part where i at least always hold the cpus before placing them into the socket... and its really easy getting a finger print or a little thermal paste on the back of the cpu and onto the pads... even if you wipe it off theres still a thin layer of fat or thermal grease there...

and if the contact area of the pins and pads is small, the chance of having a bad contact is even higher... so maybe its indeed a problem of the pins and socket, like hipro suggests... in combination with not 100% clean pads?

why sf3d can i see only half immages ? :/

Quote:

---

Originally Posted by eva2000

rma the cpu and see what intel says happened :D

---

I can't, because funnily enough I got this cpu from MSI. :p:

Quote:

---

Originally Posted by saaya

well does the ud6 use a special socket version other boards dont use?
i doubt gigabytes ud6 is the only board using the same version of this socket, so im not convinced its the socket...
im not sure it has anything to do with the board either... couldnt it just be that, since this happened at the edge of the cpus, they had a fingerprint on them, possible with thermal grease so there was a thin layer of grease on it, preventing good contact and then resulting in higher resistance, more heat, even higher resistance, and then the plastic melted? the area this happened is exactly the part where i at least always hold the cpus before placing them into the socket... and its really easy getting a finger print or a little thermal paste on the back of the cpu and onto the pads... even if you wipe it off theres still a thin layer of fat or thermal grease there...

and if the contact area of the pins and pads is small, the chance of having a bad contact is even higher... so maybe its indeed a problem of the pins and socket, like hipro suggests... in combination with not 100% clean pads?

---

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

i've never liked LGA sockets.