P55-UD6 socket burn

[cyclo]

You do understand the CPU sits on top of the bottom plastic of the socket, it is impossible to make it sit lower, unless you use a hammer, so a heat sinks weight is not the problem.

If the socket was not pressing the CPU tight enough and there is an unequal pressure brought to bear by the HSF on the CPU which is heavy on the bottom/south side of the CPU, light on the top side (I am thinking of the big tower type HSFs with vertical heat pipes), then it is possible that there might be "less contact" on the top part of the CPU. "Less contact" in this case might mean a small gap might even be forming towards the top part of the CPU/Pins compared to the bottom part which under pressure.

The scenario I am describing is the result of a combination of factors including, the tighter spacings (and tolerances) of the pins on the new CPUs, the "loose fitting" Foxcon sockets, and top heavy HSFs that overclockers are using to cool the CPU. Note that most of these arc burns appear on the pins on the upper end of the CPU.

Again I am just throwing out my 2 cents here based on observations of pictures I have seen on this thread and in the Anandtech article.

[So_BAD]

So, I took some MACRO HIRES pics from socket, just to observe yours details.

Front and back plate:



All views:








.
.
.
.
.
.

[dctokyo]

If the socket was not pressing the CPU tight enough and there is an unequal pressure brought to bear by the HSF on the CPU which is heavy on the bottom/south side of the CPU, light on the top side (I am thinking of the big tower type HSFs with vertical heat pipes), then it is possible that there might be "less contact" on the top part of the CPU. "Less contact" in this case might mean a small gap might even be forming towards the top part of the CPU/Pins compared to the bottom part which under pressure.

The scenario I am describing is the result of a combination of factors including, the tighter spacings (and tolerances) of the pins on the new CPUs, the "loose fitting" Foxcon sockets, and top heavy HSFs that overclockers are using to cool the CPU. Note that most of these arc burns appear on the pins on the upper end of the CPU.

Again I am just throwing out my 2 cents here based on observations of pictures I have seen on this thread and in the Anandtech article.

Very interesting, I understand what your saying, just wish we had more details on the setup`s

[Nelly]

So apart from the original author of this thread & Raja, has anyone else suffered socket burn on their setup? If so has your motherboard manufacturer accepted RMA & replaced your board? same goes for your Intel CPU?

When you get hardware for free it doesn't really effect financially does it?

[Raja@ASUS]

There are a couple of consumer burnouts linked over the last few pages of this thread (you must've missed them).

[cowpuppy]

Just some more thoughts here as I know the plastic injection molding process that is used to make the plastic housing is a complete crap. These are out sourced to the lowest bidder who just don't care and will speed them up causing warp and other issues. It could also be a tool & die issue where one side just wasn't made the same. A lot going on here.

If its a clamping pressure issue as they are saying. I think once we mount our heat sink it would eliminate this issue. Hell half of us tighten down the sink till it warps the motherboard. I'm sure that's more pressure than any clamp would exert.

so if that's my theory than the only alternative would be something is limiting the downward movement of the cpu in the socket. The only thing that might cause a issue of downward movement might be a broke detail or maybe the socket isn't so level on the flat surface that the bottom of the cpu sits causing unevenness.

Someone needs to set their cpu in the socket apply pressure and see if it sits level no high sides or rocks back and forth someone may have to rip the black piece from the board to check it properly.

[stasio]

New upcoming models(USB 3.0,SATA 3,SLI or CrossFireX) P55A (EX58A?) coming with LOTES socket.

http://www.xtremesystems.org/forums/...&postcount=174

[dctokyo]

Also could this be a hardware design problem?

The LGA 1156 Socket has only 66% of the current capacity of the LGA 1366 Socket.......because it only has 175 VCC pins instead of over 250 VCC pins on the LGA 1366........... therefore each of the VCC pins on the LGA 1156 carries more of load then the LGA 1366 socket.

[Raja@ASUS]

Also could this be a hardware design problem?

The LGA 1156 Socket has only 66% of the current capacity of the LGA 1366 Socket.......because it only has 175 VCC pins instead of over 250 VCC pins on the LGA 1366........... therefore each of the VCC pins on the LGA 1156 carries more of load then the LGA 1356 socket.

In terms of absolute maximum current handling and assuming similar pin/pad strutcure, yes 1366 has more current capacity than socket 1156.

Trouble is that some of these burnouts have taken place under the load of a single thread of super pi 32M. Bear in mind that at 5GHz and the CPU IHS around -100, power draw is no more than 50-60w to VCC.

later
Raja

[Mav451]

New upcoming models(USB 3.0,SATA 3,SLI or CrossFireX) P55A (EX58A?) coming with LOTES socket.

http://www.xtremesystems.org/forums/...&postcount=174

Lol @ me for being an early adopter. Updated USB3.0/SATA spec, and to add insult to injury, LOTES socket (and no more Foxconn). I thought I could avoid the revision game by avoiding Asus, but I just re-entered it with my first Gigabyte purchase.

[safaxs]

Hello all,

I have ordered just last weak a new Intel i5 and Asus P7P55D. And then I read article about burned processors and motherboards on 1156 Foxconn chipsets.

What should I do? I don't have in plan to overclock processor but still as people say it can happen once. And no warranty will be there.

Do you think it will work since I won't overclock a processor?

Thank you.
Safaxs

[Dahaka]

Yeah, (swear bypass removed - STEvil) onn rulz!!

That pictures are from a customer's new ride, what a shame.

[stasio]

That pictures are from a customer's new ride, what a shame.

It's shame,but for FOXCONN.

[safaxs]

Hello everyone,

I just got my new Asus P7P55D and i5 so please tell me what do you think:



The red area does not have a contact in my opinion.

[STEvil]

There's nothing wrong with either of you guys' CPU's.

[RPGWiZaRD]

Well I uninstalled the CPU today on my Gigabyte UD5 to check and for me all pins made contact. Damn it sux checking through all those small 1156 pins ("I see pins!" O_o).

But then I've been having another issue tho, random reboots without any bsod despite memtest, and linx are fine and OCCT PSU stress test and also tried with my old TT 750W PSU without luck but then I started suspecting my Samsung F3 hdd incompability as that's a very new device for me and swapped from the intel ICH to gigabyte GSATA ports and at least for 3~4hrs straight it's been able to play games without a reboot.

I did suspect the Foxconn socket issue at first but all pins made contact and don't find any burnt pins etc either.

[Homeless2]

New upcoming models(USB 3.0,SATA 3,SLI or CrossFireX) P55A (EX58A?) coming with LOTES socket.

http://www.xtremesystems.org/forums/...&postcount=174

Any chance they could update the original p55 ud3r with the lotes socket? While the P55A version has nice features, I don't want to spend extra on something that hardware doesn't exist for yet

[RPGWiZaRD]

Well critical "kernel-power" issue random reboot issue still persists making me think it has to be the mobo (already ruled out PSU by testing with my old TT 750W PSU and memtest doesn't report any probs when running it over the night so it can't be ram, it could possibly be CPU that's taken a toll by the 1.4v vcore but I'm 70% sure it's a mobo issue) with all the foxconn socket failures around I wouldn't be suprised if this is the issue for me too but then again at least the pins looked all ok for me and made contact etc.

Would I have known the trouble this would follow I had definitely went i7-920 route instead.

So which mobos aren't using the foxconn socket except for EVGA?

EDIT: BTW, I wonder if some1 with a Gigabyte P55 can mention what the CPU-Z reads for vcore contra what's set as BIOS value, I think I may have a similiar problem as http://www.tomshardware.com/reviews/...d,2436-15.html For me there's quite a difference between what's set in BIOS and what cpu-z reads and becomes bigger the higher I go, at say 1.44 - 1.45v bios set, it reads 1.408v in CPU-Z at idle but only 1.36v during load at "standard" LCC setting? That can't be right? The random reboots when playing happens also at stock settings now but I think it may happen more rarely.

But then I don't know why I'm able to run for example LinX or OCCT without getting any reboots, only when playing games.


Any1?

EDIT: OMG this drives me crazy with all the restarts for no reason, prolly gonna dump p55 and buy a proper i7 setup, ie X58.

[Drwho?]

Just make sure you are pushing with the right pressure on the top of the socket.
1) too much pressure is not good
2) asymetrical pressure is not good.
3) not enough pressue is not good.

For those who remove the bottom metallic plate, make sure the board does not bend.

If you pay attention to all of this, there is no problem at all, with Thermal dummy CPUs, this socket can work up to 2V without issues. (Just don't play stupid on the voltage ... )

For those who like it very technical: http://download.intel.com/design/pro...nex/322167.pdf (Page 13, and 32)

Francois

[Leeghoofd]

Gigabyte engineers told us the early socket models had a bad mix of plastic composites, that couldn't resist to the heat... it has been fixed... it was a bad batch , problem has been recitifed... move on... there's far too much speculation and theories floating around...

[saaya]

so the problem was the plastic? 0_o i find that hard to believe...
if plastic inside a -100C socket melts, then its not a problem of the plastic melting too quickly, its def a problem of the pins getting too hot...

[Leeghoofd]

Talk to Hicookie then, this is what we officially got at the Giga event as explenation for this... you don't have to believe me... there was a bad batch of sockets, which tended to deform slightly under heat/cold and massive Vcore use : this leading to bad contacts : even more heat generated. End result : finally socket burn... and damaged CPU's

How many in here complaining about bad contacts have ended up with burned sockets ?

This is a snowball effect in action here. The same with VTT in few years back...

One starts a respectable theory and then suddenly everyone looks at their cpu's sees "bad" contacts and want an RMA... lol ...

[Mav451]

I'd be more re-assured if there was an updated revision, something, anything that would guarantee that these problems do not arise again. Saying that it was a bad batch and that "it was rectified" isn't giving me an inch of confidence in Gigabyte. I bought my board at Newegg (late Sept) and considering they're one of the biggest retailers, who knows how many stock in their warehouses were "bad batch" and which ones were "rectified".

[Migi06]

Well critical "kernel-power" issue random reboot issue still persists making me think it has to be the mobo (already ruled out PSU by testing with my old TT 750W PSU and memtest doesn't report any probs when running it over the night so it can't be ram, it could possibly be CPU that's taken a toll by the 1.4v vcore but I'm 70% sure it's a mobo issue) with all the foxconn socket failures around I wouldn't be suprised if this is the issue for me too but then again at least the pins looked all ok for me and made contact etc.

But then I don't know why I'm able to run for example LinX or OCCT without getting any reboots, only when playing games.

More likely Windows 7 mysterious reboot problems..
Could be drivers like chipset, audio, GPU etc..
Powersuply problems like bad contact, rail issues..
Incompability software..

EDIT: OMG this drives me crazy with all the restarts for no reason, prolly gonna dump p55 and buy a proper i7 setup, ie X58.

Me too but dont wanna buy "antic" x58 stuff X58 need new mobos and cpu´s like 32nm quads or this i7 930 (Got i5 750 & UD3R combo.. awesome price/performance ratio.. With this system it's good to wait something new and better )

[massman]

so the problem was the plastic? 0_o i find that hard to believe...
if plastic inside a -100C socket melts, then its not a problem of the plastic melting too quickly, its def a problem of the pins getting too hot...

Doesn't that make sense?

Pins are attached to socket plastic; pins get hot and heat up plastic. Plastic can't hold heat and melts.