P55 overclockers...check your sockets

Eagle156 said:
those are all shit mainstream boards that aren't really meant for overclocking. it's not that surprising that they failed.
Click to expand...

Those were the "value" boards that ASrock, ECS and MSI each picked for THG to review and they sure as heck knew they were going to be overclocked. :p

The only thing we are not certain of is if they also used Foxconn sockets or not. No one can say for certain that the poor pin contact did not also contribute to mobo failure - if these issues are related.

UPDATED: i asked at Toms and got the answer that the Foxconn socket issue is probably unrelated to their issues with the mobo failure. And that ASRock did indeed fix their mobo with a BIOS update
It's fixed, I tested it, you'll see an update in around a week. And you need to know that Asus and Gigabyte, the two boards that didn't have any issues, used Foxconn sockets.
Click to expand...
 
Last edited:
apoppin said:
Those were the "value" boards that ASrock, ECS and MSI each picked for THG to review and they sure as heck knew they were going to be overclocked. :p

The only thing we are not certain of is if they also used Foxconn sockets or not. No one can say for certain that the poor pin contact did not also contribute to mobo failure - if these issues are related.

UPDATED: i asked at Toms and got the answer that the Foxconn socket issue is probably unrelated to their issues with the mobo failure. And that ASRock did indeed fix their mobo with a BIOS update

It's fixed, I tested it, you'll see an update in around a week. And you need to know that Asus and Gigabyte, the two boards that didn't have any issues, used Foxconn sockets.
Click to expand...

That's VERY good news for me.. I have pulled my processor and inspected the socket twice so far.. It looked rock solid to me. I have the Gigabyte P55 UD4P w/ an i7 860 @ 4ghz. I have had ZERO issues and I'm glad to see some feedback from a reviewing website. With that said, I will probably pull my processor a few more times (one time at the end of this month and one time during the month of November), just to make sure everything is still in good working order.
 
Here's an update from Raju over at XS:
Stance from companies using Foxconn sockets is that there's no real problem. The tech dept's of some companies are putting it down to operator error/extreme OC (some are telling users they've never even had an issue..lol).

That being said, it is possible ASUS may move their P55 line up over to another socket vendor (they are looking at qualifying other sockets). The Foxconn sockets did pass Intel qualification back in June. So it's really a case of - are there sockets out there from pre June qualification that should not be, or is the overhead of the Foxconn socket limited by deisgn in comparison to those from LOTES or TYCO AMP?

I had a big conversation with one of DFI's tech guru's about this and he's as stumped about the cause as anyone (they've had a socket burn issue on their end too). Every reported case has happened under overclocked state so far. Until we get this happening at stock it's going to be difficult get answers from those in the know. The only way to test the sockets themselves would be to use tools that only Intel have access to (a transient load tester deisgned around the socket to measure the voltage drop via the socket).

I've got one Gigabyte board here using the Foxconn socket and it has not burned out under extreme OC, so it's still random I'm afraid.

regards
Raja
Click to expand...
 
So I got paranoid about this too, took apart my build, and the pins all look to be making good contact. That brings about the question, what was wrong with these motherboards? I heard it mentioned before in another thread, but I'm also wondering if these burnt sockets are due to operator error. VTT max is specified at 1.21V, which is much lower than on Socket 1336 boards, and I wonder if they just ended up frying the sockets using 1.35V+.
 
officermartinez said:
That's VERY good news for me.. I have pulled my processor and inspected the socket twice so far.. It looked rock solid to me. I have the Gigabyte P55 UD4P w/ an i7 860 @ 4ghz. I have had ZERO issues and I'm glad to see some feedback from a reviewing website. With that said, I will probably pull my processor a few more times (one time at the end of this month and one time during the month of November), just to make sure everything is still in good working order.
Click to expand...

Frankly, if it looked good now I would tend to leave it alone.
Opening up the connector repeatedly will only weaken the spring
effect. Less use of the socket is better.

If you keep opening it up and closing it, you are headed to socket issues.

You have a multi-year warranty on the CPU and motherboard, so
I would use it until it dies, which (based on what you saw) probably won't
happen. I think that you are fine.
 
Looks like the only way to settle this is to hammer your CPU into the socket. Not like anyones gonna upgrade their CPU anytime soon. :rolleyes:
 
Trepidati0n said:
Well...seems there is a nice case of stock burnout. Because just running prime burnt out his socket.
http://forums.hardwarezone.com.sg/showthread.php?t=2540820
Click to expand...
But what were the conditions prior to the burn? If people are improperly overclocking and not following Intel's provided voltage specifications, there's not much else to blame. I'm not saying there aren't mating issues, as the pictures posted clearly show there's an issue on some motherboards. That said, I don't know if you can blame every socket failure on a manufacturing defect.
 
Mr. K6 said:
But what were the conditions prior to the burn? If people are improperly overclocking and not following Intel's provided voltage specifications, there's not much else to blame. I'm not saying there aren't mating issues, as the pictures posted clearly show there's an issue on some motherboards. That said, I don't know if you can blame every socket failure on a manufacturing defect.
Click to expand...

Make you a deal. My i5 system is about to go live. I'll use the stock heatsink and run it at stock. If it burns out running just prime on all 4 cores will you buy me an i7 system? Really...are you that confident?
 
Oh great, I just invested in a P55 and this shit happens? I guess I better not overclock my system then. I have a nice GA-P55-UD6 board and it cost me like $319, better not ruin it.

This is not good news at all.
 
Trepidati0n said:
Make you a deal. My i5 system is about to go live. I'll use the stock heatsink and run it at stock. If it burns out running just prime on all 4 cores will you buy me an i7 system? Really...are you that confident?
Click to expand...
Actually, somebody does need to do this step.
 
Trepidati0n said:
Make you a deal. My i5 system is about to go live. I'll use the stock heatsink and run it at stock. If it burns out running just prime on all 4 cores will you buy me an i7 system? Really...are you that confident?
Click to expand...
Do you really think they'd let a major fuck up like you're suggesting go unnoticed? There are QA folks that check this stuff. I'm not saying it's impossible, just highly unlikely. As I said, I've been running my i5 since launch, and a few days ago I pulled off the heatsink and inspected the CPU. All of the pins were making good contact and there were no signs of burning or any other kind of damage.

Have there been any reported instances of failure due to this issue while running stock settings? The thread you linked to doesn't make any mention of running at stock speeds. The Anandtech article linked to reports from extreme overclocking attempts.

EDIT: Read through some of that thread, it was mentioned that there were reports of stock failures in the (I'm guessing) associated user comments of the Anandtech article, so I'm reading that now. Now I have a Foxconn socket on my P55M-UD4 and I don't seem to have any issues, so go figure. IIRC, I read somewhere that there was a mention that this affects a specific socket model that was used on many budget focused boards (ECS, etc.).

EDIT 2: In 22 pages of comments, IIRC, one user reported returning his system due to "instability" :)rolleyes:) and there were a few reported failures of air-cooled overclocking by Raja himself (one at 4.1GHz and a suicide run at 4.7GHz). I didn't see any evidence of stock failures (one user linked back to the same thread you linked to, which, again, I didn't see any stock failures, maybe I'm wrong). None of these are convincing arguments for widespread problem. One thing I noticed when I looked at my chip is that the pins don't all contact the center of the pad, some of the contact points are off to the side, and it might be easy to miss at a quick glance. Anyway, I still don't see any reported failures at stock, no reported failures while staying under manufacturer recommended limitsunder precautions while overclocking. There is physical evidence, such as in the photos posted at Anandtech (I don't see any contacts on those pads), but there is no data regarding what specific boards/lines are affected, nor the conditions under which failures occurred. Overall, there is an incredible lack of data that makes it impossible to make a conclusion at this time. There's way too much FUD being spread without any concrete evidence to back it up (reminds me of the microstutter scare, when neophytes heard the phrase and reported microstuttering in pretty much every card that was on the market). I guess we'll all just have to wait and see, but as far as I can tell, I'm unaffected.
 
Last edited:
I'm only interested in trying to get a moderate OC out of my 860, I'd be perfectly fine with 3.5 Ghz at hopefully decently under 1.4v. What kind of voltage are you i7 OCers having to use?
 
bitgod said:
I'm only interested in trying to get a moderate OC out of my 860, I'd be perfectly fine with 3.5 Ghz at hopefully decently under 1.4v. What kind of voltage are you i7 OCers having to use?
Click to expand...

Stock voltage at 3.5Ghz on a i7 920.
 
my i5 and MSI P55-CD53 earlier this week (though not setup - waiting on PSU).

i was reading this thread before my parts arrived and i was almost certain that my low end MSI board's socket would be foxconn manufactured. but when it came and i checked it, to my surprise its a LOTES socket. looks like i dodged a bullet.
 
Atech said:
Stock voltage at 3.5Ghz on a i7 920.
Click to expand...

Thats awesome.. I've achieved decent results with my i7 860..

4ghz-Everest.jpg
 
Hi!

In fact, I'm not worried about my P7P55D Deluxe, it does not intend to OC higher than 3.6GHz, simply because don't need all this clock.
Neither will use LN2 or others, to see if the material used in the socket FOXCONN resists temps very negative, etc. Or put 1,64 vcore, like a guy in XS thread,

What scares me so far, based on the facts reported, mainly in the XS, is the poor settlement of the CPU. That is, the contact between the pins of the socket and CPU. What causes small "sparks" and consequently the short-circuit and melt the socket, pins and CPU. This is very serious!

The reports noted that until now, speak over the 8XX series CPUs, perhaps because these are the most sought after by so inclined OC, really extremes.

With my combination of P7P55D Deluxe + i5 750, reports have not noticed yet ... but I'm watching.

I contacted Asus Brazil reporting the case, including informing this links (from XS, Anandtech etc) and asked for explanations when a possible RECALL, from Asus.

They responded but, just sent a "standard e-mail response" to all who ask about warranty of any Asus product that is not manufactured in the Brazil, How to uses warranty, when it's is possible etc. No objective response on this case specifically.

Now, I contacting FOXCONN Brazil and China Headquarter, to ask then about this issue.

This is mine "version" of FOXCONN socket:





and sorry my poor english, I'm an illiterate, almost.
.
.
.
.
.
.
.
 
So, I took some MACRO HIRES pics from socket, just to observe yours details.












Thanks!
 
For those still following this thread.. I contacted Gigabyte USA to see if this was an issue with their P55 boards and this is the reply I received a few minutes ago:


Untitled.jpg
 
Sojuuk said:
It's a form letter. Can't be trusted.
Click to expand...

Exactly! Haha. I was like.. what tha.. It's a d*mn generic response letter. I'm just sitting here wondering how I'm gonna handle it. I have plenty of reference material for the P55 socket (in general), however, specifically I can't find any socket failures on the Gigabyte P55-UD4P (my specific motherboard). Maybe they know something that I don't. Ughhhh! Hehehe.
 
Mav451 said:
Hah I suppose so. stasio @ XS just posted this today:
http://www.xtremesystems.org/forums/showpost.php?p=4077751&postcount=174

Now I don't know how to interpret this, but if there's already an updated P55A and it includes a LOTES socket (replacing the faulty Foxconn) that's a pretty big admission that the Foxconn was crap.
Click to expand...



I wonder what Foxconn is actually going to do about the problems with their sockets? DFI has already switched and now Gigabyte. I am sure Asus and MSI are sure to follow in the near future. I guess we'll see :)
 
What worries me is that with a new SKU/product out there, Gigabyte has no reason to support users with the older boards? Bad/burnt socket? What are we supposed to do, RMA? With a new SKU, that means there's no reason for Gigabyte to update the UD3R line with LOTES socket, much less do anything to help users who have the old board (and now phased out line).

The trump card is if someone buys a UD3R now, and goes "oh wow, Gigabyte did a ninja edit and now has LOTES sockets on them." That however is...being naively hopeful and almost idealist.
 
Mav451 said:
What worries me is that with a new SKU/product out there, Gigabyte has no reason to support users with the older boards? Bad/burnt socket? What are we supposed to do, RMA? With a new SKU, that means there's no reason for Gigabyte to update the UD3R line with LOTES socket, much less do anything to help users who have the old board (and now phased out line).

The trump card is if someone buys a UD3R now, and goes "oh wow, Gigabyte did a ninja edit and now has LOTES sockets on them." That however is...being naively hopeful and almost idealist.
Click to expand...

Maybe check with revision on board ?

phatbx133
 
Looks like the 1156 is now the Vista of motherboards, because no matter how big (or not) the issue is, people are going to remember there is/was a problem.
 
its worse than vista at least with that you could roll back to xp with vista not buy yet more new hardware
 
Meh, I'm not returning any hardware until it stops working as advertised, then I'll take advantage of my warranty and get a new one. No point in worrying myself to death about it or ripping my computer apart and trying to get a new motherboard.
 
FoxhoundOp said:
Meh, I'm not returning any hardware until it stops working as advertised, then I'll take advantage of my warranty and get a new one. No point in worrying myself to death about it or ripping my computer apart and trying to get a new motherboard.
Click to expand...

Ya you are right to a certain point unless a) the mboard manufacture rejects your claim because you are OCing or b) it frys your CPU in the process of burning out.

But ya I have my system running now.. not a lot of point in worrying about it till it dies.
 
Eh, in my case, I always like to have an excuse to upgrade. "I didn't want to put together a new system, but my old one died". :)
 
You must log in or register to reply here.
Back
Top