Socket Pin Burnout Returns to Haunt LGA1155?

RavenXXX2 · 17 Jan 2011 at 09:34

I received two motherboards from Gigabyte recently for testing, one P67A-UD4 and one P67A-UD7. Upon arrival, I did a thorough visual inspection, and found the UD4 had obvious signs of “socket burn”….a phenomenon discovered last year with the LAG1156 socket area. After notifying my contact with Gigabyte about the issue 8 days ago, I’ve yet to get any word back. Hopefully this issue is not widespread.

With the last generation, it appeared that socket burn was most likely to occur under extreme conditions. But with Sandy Bridge, “extreme conditions” and 24/7 use are very close together. Will this problem be a major issue for a large number of users? Time will tell.

http://www.techreaction.net/2011/01/14/blog-socket-burn-is-back-with-sandy-bridge/

reygjabla · 17 Jan 2011 at 09:55

Wow!

That's gonna haunt the early adopters. :rolleyes:

d_brennen · 17 Jan 2011 at 10:05

I'll stick with my toasty Q6600 ta. At least I KNOW everything's burning

3t3P · 17 Jan 2011 at 10:58

haha brennen

Muel · 17 Jan 2011 at 11:00

Think I'll definitely be sticking with AM3 now. I can push this chip to the ragged edge and beyond and it isn't dead yet!

CAT-THE-FIFTH · 17 Jan 2011 at 11:05

I wonder if OcUK will also cover motherboards damaged due to excessive voltage too??

Despite the disdain for the maximum safe voltage set by Intel it does look like it would be better to keep to lower voltages for most users. Benchmarkers probably don't really care as they will factor in hardware going pop as part of their hobby.

It will be interesting to see if this is more widespread problem or not TBH.

I suspect that for most users the sockets should be fine.

joeyjojo · 17 Jan 2011 at 11:17

3t3P said:
haha brennen

Brennen is burn in German *cue X-Files music*

d_brennen · 17 Jan 2011 at 11:22

joeyjojo said:
Brennen is burn in German *cue X-Files music*

The boy's name Brennen \b-rennen, br(en)-nen\ is a variant of Brendan (Irish, Gaelic, Celtic) and Brennan (Irish, Gaelic), and the meaning of Brennen is "prince; teardrop".

Prince Tear-drop, thank you!

ThomasTheTankEngine · 17 Jan 2011 at 14:11

This always happens when a new socket or cpu comes out. People ignore manufacturer guidelines, break their components, and thus establish the "safe" baselines for everyone else to go by.

Hagar · 17 Jan 2011 at 14:14

It always seemed a little stupid to me to have pins just bearing on the ends. At least with AMD cpu pins, they are sleeved into the socket, the bearing area is on the side of the pins.

Also if there is a nano gap for any reason, erosion or slightly short pins, it will probably cause an arc creating further erosion and burning.

andy.

HazardO · 17 Jan 2011 at 14:17

I really don't understand how 140W bloomfield processors which overclocked can draw in excess of 200W have never suffered from this, when low voltage 95W CPUs are suffering from this.

Is the manufacturing process for the 2 sockets really that different?
Are there just more pins for power on the LGA1366 boards?
Have any OCUK techs or customers seen anything like this?

dave_beast · 17 Jan 2011 at 14:30

Duplicate thread

http://forums.overclockers.co.uk/showthread.php?t=18230681

EniGmA1981 · 17 Jan 2011 at 14:36

d_brennen said:
Prince Tear-drop, thank you!

Better than Princess Pear Drop isn't it? lol

MooMoo444 · 17 Jan 2011 at 14:55

Read and heard a lot of horror stories about these early Sandy Bridge CPU's, I guess it pays to wait till they 'perfect' them, like B3/G0 Q6600's for example!

HazardO · 17 Jan 2011 at 15:10

MooMoo444 said:
Read and heard a lot of horror stories about these early Sandy Bridge CPU's, I guess it pays to wait till they 'perfect' them, like B3/G0 Q6600's for example!

I agree, why jump from a known stable D0 i7 for the sake of a few extra MHz? The C0 i7s were really only good for upto 3.6-3.8GHz, the D0s are 4-4.2. I wouldn't be surprised if the next Sandy bridge stepping is 5GHz stable.

Purgatory · 17 Jan 2011 at 22:34

Praying LGA2011 does not suffer from this... :mad:

More pins they add the smaller they get it seems and the more chance of this happening. They may need to look for a better material for the pins to stop this happening they just look way too fragile and seem to burn up :eek:

....

Chaotic.Wolf · 18 Jan 2011 at 07:38

From dave_beast "On Bit Tech they're saying it's Foxconn sockets, Lotes (MSI and Asus) are not affected. "

So how accurate is this assessment? I chose the Asus Pro motherboard, I also will not even overclock my CPU past 4-4.3GHZ.

However still this kind of scares me

hexagonsun · 19 Jan 2011 at 19:34

Is this only if you overclock?

dave_beast · 19 Jan 2011 at 19:50

I would guess this is only if you seriously abuse your socket, i.e putting 1.6v+ through it.

Remember, these testers are not using their own paid for motherboards and processors, it's engineering samples and freebies for testers so will max everything for benching etc.

ATIorNvidia · 20 Jan 2011 at 15:14

HazardO said:
I really don't understand how 140W bloomfield processors which overclocked can draw in excess of 200W have never suffered from this, when low voltage 95W CPUs are suffering from this.

Is the manufacturing process for the 2 sockets really that different?
Are there just more pins for power on the LGA1366 boards?
Have any OCUK techs or customers seen anything like this?

1366 pins vs 1155/1156 pins.