P55 overclockers...check your sockets

Swapped heatsinks yesterday so I took some pics of the chip and socket. I don't have a macro lens unfortunately and had to use the flash so it was incredibly hard to get decent pics. I have an MSI P55-GD65. Looks like contact everywhere.
http://coldgravity.com/misc/860/cpu01.jpg
http://coldgravity.com/misc/860/cpu02.jpg
http://coldgravity.com/misc/860/1156.jpg

I did notice, like some others have mentioned earlier, that pictures don't always pick up the indentations.
From a slightly different angle and no flash IIRC:
http://coldgravity.com/misc/860/badpic.jpg
 
lathode said:
Looks like really good contact. Have you been overclocking?
Click to expand...
Only been able to get 3.5GHz stable (prime95 and LinX all memory 6+ hrs each) but I think it's due to my stock PCH 1.08V and PCH 1.05V voltages. 71/69/68/69 temps with Ultra-120A and 1156 bracket. I need to mess with it a little more, not sure what the PCH amounts should be so I haven't moved them. haha
 
I don't go ballistic when I OC but kinda happy I have a 1366 for the time being.

I hope they sort this out for everyone else, but I doubt it will be fixed for existing users vs just for the new users as this would not happen when not running an OC.

Kind of a tough luck scenario.
 
hungryJonJon said:
Only been able to get 3.5GHz stable (prime95 and LinX all memory 6+ hrs each) but I think it's due to my stock PCH 1.08V and PCH 1.05V voltages. 71/69/68/69 temps with Ultra-120A and 1156 bracket. I need to mess with it a little more, not sure what the PCH amounts should be so I haven't moved them. haha
Click to expand...
PCH voltage is for the south bridge, you shouldn't have to touch it.
 
Wow another thing I somehow went forever without knowing about, I was considering a i5 upgrade but now I'm not sure.
 
dario03 said:
Wow another thing I somehow went forever without knowing about, I was considering a i5 upgrade but now I'm not sure.
Click to expand...

until new revisions come out that don't use foxconn sockets you need to pay a premium for a manufacturer that already doesn't use them. And at the prices they ask I would just buy socket 1366.
 
I've read at least one report of someone who got the cheapest EVGA board from Newegg and it had the Lotes socket. $139 after rebate. I don't know if you can count on that, but I'll know my luck in about 2 days...
 
Sojuuk said:
until new revisions come out that don't use foxconn sockets you need to pay a premium for a manufacturer that already doesn't use them. And at the prices they ask I would just buy socket 1366.
Click to expand...

Yeah I'm not to far from a Microcenter so a i7 and Asrock mobo are only about $90 more than a i5 and mobo I was looking at. Though I wasn't upgrading right now anyways and I need to stay on the cheap since I'm broke.
 
Ugh. I wish I had read this all earlier. My parts for my new computer are en route and schedule to be here on Monday. Now I have to decide if I want to keep them all or send them back for a refund.
 
Drinyth said:
Ugh. I wish I had read this all earlier. My parts for my new computer are en route and schedule to be here on Monday. Now I have to decide if I want to keep them all or send them back for a refund.
Click to expand...

If they are from Newegg, you should contact them via email or phone before UPS attempts delivery, since motherboards can only be replaced if defective with the same model. Newegg used to be quite lax and easy about this, but lately they are quite strict.

Newegg will then instruct you to refuse delivery if you are home, or call for an RMA number if the delivery is made. Your best option is not to open the shipping carton at all, but your call and/or email prior to delivery will put you on record with them so that they will accept the motherboard back for a refund.

If you know that someone will be home, and that you won't miss catching UPS, you can simply refuse delivery without calling Newegg in advance.
 
aamsel said:
If they are from Newegg, you should contact them via email or phone before UPS attempts delivery, since motherboards can only be replaced if defective with the same model. Newegg used to be quite lax and easy about this, but lately they are quite strict.

Newegg will then instruct you to refuse delivery if you are home, or call for an RMA number if the delivery is made. Your best option is not to open the shipping carton at all, but your call and/or email prior to delivery will put you on record with them so that they will accept the motherboard back for a refund.

If you know that someone will be home, and that you won't miss catching UPS, you can simply refuse delivery without calling Newegg in advance.
Click to expand...

Thanks for the heads up! I checked more into it and it looks like some motherboards from NewEgg have a "limited replacement only" policy while others have the "standard return policy". The Gigabyte board that I ordered coincidentally has the standard policy, so hopefully I'm covered.

Even still... I'm not sure what boards out there are guaranteed to have the LOTES type connector. I seem to be getting conflicting reports from all manufacturers. I suppose I could just return my parts and then wait to get the board later when things straighten themselves out. Or when the USB3 boards hit the market here soon...
 
Drinyth said:
Thanks for the heads up! I checked more into it and it looks like some motherboards from NewEgg have a "limited replacement only" policy while others have the "standard return policy". The Gigabyte board that I ordered coincidentally has the standard policy, so hopefully I'm covered.

Even still... I'm not sure what boards out there are guaranteed to have the LOTES type connector. I seem to be getting conflicting reports from all manufacturers. I suppose I could just return my parts and then wait to get the board later when things straighten themselves out. Or when the USB3 boards hit the market here soon...
Click to expand...

i believe USB 3.0 has been delayed by intel until next year.
 
michael19 said:
It will be on GIGABYTE boards coming out before the end of the year.
Click to expand...

It will be on some Gigabyte and some Asus boards via an NEC USB3.0 controller. I think we'll have to wait for the next round of chipsets for native USB3.
 
According to a PC Perspective review, ASUS will also be releasing an add-on card with
the same NEC USB3.0 and SATA 6.0 on the same card for a retail price of $29.99.
Not a lot of money to add this onto a motherboard that doesn't have it. The jury is still
out on SATA 6.0 benefits but USB 3.0 looks like it will be good.
 
aamsel said:
According to a PC Perspective review, ASUS will also be releasing an add-on card with
the same NEC USB3.0 and SATA 6.0 on the same card for a retail price of $29.99.
Not a lot of money to add this onto a motherboard that doesn't have it. The jury is still
out on SATA 6.0 benefits but USB 3.0 looks like it will be good.
Click to expand...

Yup. I've mentioned this in the other USB 3 theads, just buy a board with a PCI-E x4 slot. You'll be able to easily add the new standards to your board.
 
For anyone who was interested, the EVGA P55 LE board I just got from Newegg (the cheapest one, which was $139.99 last week and is 149.99 at the moment) came with the Lotes socket. Time to install!
 
Apparently first instance of burn on a LOTES socket, on eVGA P55 FTW
http://outofspecs.gr/forum/showthread.php?t=3094

See post #12.

*Guess I should add a note:
Foxconn burns so far have been found in the top left corner of the socket
while the LOTES burn from this incident is on the top right and mid-right.

Different places of burn meaning different things?
 
Last edited:
Mav451 said:
Apparently first instance of burn on a LOTES socket, on eVGA P55 FTW
http://outofspecs.gr/forum/showthread.php?t=3094

See post #12.
Click to expand...


OK people, be careful about drawing too many conclusions. The problems with the foxconn connectors was able to be duplicated, a single failure on a LOTES socket doesn't mean anything.

At work we use all HP 7700DC desktops. All socket 775. I had one of those melt down the socket at stock speeds not too unlike the FoxConn boards. I'm sure others have had the same issue as me; but it was caused by a bad individual socket, not a systemic problem.

A couple/few people having problems with a LOTES socket doesn't automatically point to a huge problem. If we end up having repeated meltdowns during testing by someone like [H] or anandtech then feel free to get upset. But for now it doesn't mean shit.
 
IMHO, LGA1156 overclocking is playing with fire (literally and figuratively). The platform draws as much power as LGA1366 when you hit the 4.0+ GHz and 1.3+ Vcore range but only has 66% or 75% the number of power delivery pins.
 
Hmmm. That leads me to wonder... if you're OCing and not increasing beyond the stock voltages, are you at any higher risk of running at OC speeds vs stock?

I can get all the way up to 3.6 GHz on my i5-750 without having to increase voltages at all (in fact, I can hit 3.6 GHz at only 1.125v which is under stock). Would I be any more prone to meltdown at 3.6 GHz than at 2.66 GHz given that the both operate at the same voltage?
 
I had a customer (i work at CompUSA) come in with an 1156 build and his CPU socket was completely fused with the CPU on one side. I actually still posted but was very unstable.
 
YES! Wooohooo, not just Foxconn. :) I guess it just means, be careful how much you raise the voltage.
 
Drinyth said:
Hmmm. That leads me to wonder... if you're OCing and not increasing beyond the stock voltages, are you at any higher risk of running at OC speeds vs stock?

I can get all the way up to 3.6 GHz on my i5-750 without having to increase voltages at all (in fact, I can hit 3.6 GHz at only 1.125v which is under stock). Would I be any more prone to meltdown at 3.6 GHz than at 2.66 GHz given that the both operate at the same voltage?
Click to expand...

+1.

I am interested in this answer...
I am running 3.5GHz at below stock voltage on an 860.
 
Update - After countless hours of running P95 or the IntelBurn Test @ 4ghz (1.368 vcore), I'm pretty happy to report that I have no burnt areas within my socket / cpu. This makes the 3rd time that I have removed my mobo / cpu cooler / cpu to check for any abnormal wear / defects.

So, for me, I'm very comfortable leaving it "as is". I doubt that I will tear it down for a 4th time anytime soon. My temps: great.. my performance: wonderful.. my piece of mind: priceless. ..hehe.
 
I'm not trying to completely resurrect an old thread but I was wondering if anyone has found out anything else? Is FoxConn still making faulty sockets? If I just ordered a GigaByte P55-UD3R motherboard from newegg am I screwed for extreme overclocks? Does checking the pin indents suffice??
 
neybis said:
I'm not trying to completely resurrect an old thread but I was wondering if anyone has found out anything else? Is FoxConn still making faulty sockets? If I just ordered a GigaByte P55-UD3R motherboard from newegg am I screwed for extreme overclocks? Does checking the pin indents suffice??
Click to expand...

I think you should be fine.. If you are phase changing, Dry Icing or Liquid Nitrogen cooling, you probably should be a little hesitant. Otherwise, I think you will be fine if you are 4ghz or less.. I too have a Gigabyte P55 and I have pulled my processor 3 times to make sure. I am at 4ghz and I have had ZERO issues.
 
So just to put this thread to rest, I just talked with Gigabyte and they said they have only ever used one revision of the foxconn socket so if the P55's were having problems and the P55A's were not then the problem is in the latching mechanism. Has anyone had any burns on the P55A??
 
Drinyth said:
Hmmm. That leads me to wonder... if you're OCing and not increasing beyond the stock voltages, are you at any higher risk of running at OC speeds vs stock?

I can get all the way up to 3.6 GHz on my i5-750 without having to increase voltages at all (in fact, I can hit 3.6 GHz at only 1.125v which is under stock). Would I be any more prone to meltdown at 3.6 GHz than at 2.66 GHz given that the both operate at the same voltage?
Click to expand...

yes, you will be drawing more current... voltage is irrelevant really
 
Mav451 said:
Apparently first instance of burn on a LOTES socket, on eVGA P55 FTW
http://outofspecs.gr/forum/showthread.php?t=3094

See post #12.

*Guess I should add a note:
Foxconn burns so far have been found in the top left corner of the socket
while the LOTES burn from this incident is on the top right and mid-right.

Different places of burn meaning different things?
Click to expand...

The whole point of the original article was based on faulty/inconsistent pin/pad connections, which are only present in Foxconn sockets. So if there are now instances of burning on other sockets maybe it's just a limitation of the overclock. It is actually possible to fry things with overclocking :p Doesn't mean the socket is defective. That one instance pretty much means zip.
 
argh. i'm just now finally getting around to building my new rig. i've got a gigabyte p55-ud4p. it has a foxconn socket. i'm putting a xeon x3440 into it. i dropped in the cpu and closed the latching mechanism, then opened 'er back up and looked at the pads on the cpu. what i'm seeing pretty much matches up with the pics in anand's article.

the thing is, i'm not planning on any overclocking in the foreseeable future. so should i be worried about this issue?

this build has already had countless delays due to wrong parts being delivered and life just being too busy. i would hate to delay yet again to rma my mobo and then pick out a different one.
 
Jason said:
argh. i'm just now finally getting around to building my new rig. i've got a gigabyte p55-ud4p. it has a foxconn socket. i'm putting a xeon x3440 into it. i dropped in the cpu and closed the latching mechanism, then opened 'er back up and looked at the pads on the cpu. what i'm seeing pretty much matches up with the pics in anand's article.

the thing is, i'm not planning on any overclocking in the foreseeable future. so should i be worried about this issue?

this build has already had countless delays due to wrong parts being delivered and life just being too busy. i would hate to delay yet again to rma my mobo and then pick out a different one.
Click to expand...

the foxxcon is in spec imo the issue with high vcore and VERY low temp on the socket that take the socket out of spec then you have less pin contact higher amps on less pins
and then you get a melted socket
so if your not overclocking or only overclocking air mild water your fine
 
My guess is that at -100*C the thermal expansion/contraction is much larger than designed for.

the sockets are designed to go from room temp(20ish) to 90*C (which it won't see without overclocking, and any overclocker will be watching their temps and keep them well within that range) a more typical range would be 20 to 60*C (so about 40) whereas with the systems anand describes you're going through a range of 120*C so about 3 times more.

plus those EOC boxes are run intermittently so the machine will be thermally cycled alot more than say my box which runs 24/7
 
mike82 said:
My guess is that at -100*C the thermal expansion/contraction is much larger than designed for.

the sockets are designed to go from room temp(20ish) to 90*C (which it won't see without overclocking, and any overclocker will be watching their temps and keep them well within that range) a more typical range would be 20 to 60*C (so about 40) whereas with the systems anand describes you're going through a range of 120*C so about 3 times more.

plus those EOC boxes are run intermittently so the machine will be thermally cycled alot more than say my box which runs 24/7
Click to expand...

heating isnt an issue as metal expands and would increase contact the issue is the high Vcore and amp draw at < 0c temps were the metal contracts
 
You must log in or register to reply here.
Back
Top