Agreed. I wonder if OEMs are have problems right now. But, as for us, It is too early to tell lga1156 will have a "dark image".
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
P55 overclockers...check your sockets
- Thread starter Atech
- Start date
Agreed. I wonder if OEMs are have problems right now. But, as for us, It is too early to tell lga1156 will have a "dark image".
Sabrewulf165
2[H]4U
- Joined
- Jun 23, 2004
- Messages
- 2,974
As I have an Asrock motherboard, I feel fairly certain my socket is probably Foxconn, but I can't say I've noticed any problems so far. Of course I am running at stock speed, but all this brouhaha is making me wonder whether or not I should check the socket/CPU...
New Egg won't refund the brand new unopened MSI GD-65 and I5 750 I just purchased. Only send back another of the same model.
That is some monkey shit right there. Not sure what other options I have other than burning the hell out of it and sending it back 400 times.
That is some monkey shit right there. Not sure what other options I have other than burning the hell out of it and sending it back 400 times.
mongoosenco
Weaksauce
- Joined
- Jan 11, 2008
- Messages
- 91
I think maybe people are making this into a much bigger deal than it really is. This happened with some pretty extreme overclocking. I have my MSI GD80 clocked to 3.7Ghz with no issues. I think when I start hearing people with normal overclocks or at stock frying their stuff then I will start worrying. Now if you are in that group of extreme overclockers then yeah I would be a little pissed if that happened to my board.
calebb
[H]F Junkie
- Joined
- Mar 26, 2000
- Messages
- 8,758
This temperature could easily affect the pressure exerted by the socket pins. Weaker/less flexible pins/springs combined with rapidly changing temperatures (load vs no-load) could explain this.
The article seems to imply that this only occurs in [H]ard|core overclocking - and they attribute it to the smaller number of VCC pads (for LGA1156) + potentially defective sockets combined with the higher current required at higher frequencies.
Another theory would be that it's their method of evaporative cooling that is cooling the foxconn connector pins outside the required environmental conditions.
Trepidati0n
[H]F Junkie
- Joined
- Oct 26, 2004
- Messages
- 9,269
This smacks of RROD. It seems minor right now...people can try and rationalize it. But if the results are fairly consistent including NON-OC conditions which people are posting now, it really makes you wonder.
mongoosenco
Weaksauce
- Joined
- Jan 11, 2008
- Messages
- 91
Can you point to a source where NON Overclockers are having the same problems?
Last edited:
Doubtful anyone will be able to. The problem is probably every non overclocker is going to call this a DOA board when a required pin doesnt make contact. Nobody keeps records as to why something is DOA.
I checked my socket and everything to confirm the issue. I reassembled and am running it again. It annoys me, but as I said before the system is stable and I am happy with it. I am going to wait until there are more official responses from the mobo companies and from Intel.
If this actually ends up being a real problem, Intel will put out a press release and start pointing fingers. Otherwise people would start to blame Intel instead of FoxConn.
Is this problem concerning, Yes. Should people stop running their systems or take other extreme steps, Nope. We just need to be aware this is out there and keep our ears to the ground as more information and the official responses are made available.
I mean who knows how much redundency Intel puts into their chips, we really don't know how many firm points of contact are really needed.
My last comment is has anyone checked their 1366 sockets? For all we know this amount of contact could be normal and we just never took the time to look at it. I know this is the first time I have really looked at the CPU so closely...
It's not all 1156 socket that are affected, only the FoxConn's...I will pop the lid of my 1336 when I come home in a week or so, but I doubt the high-end X58 is hit by this.
I have an i7-950 sitting at home. I am going to check it out when I get back home from work.
I guess the point I am trying to make is that 1156 has less points of contact from the start and the sockets are rated at a lower wattage. In these overclock they are pushing them to the same voltages and amperage as the 1366 CPUs. This could effect FoxConn connectors across the board.
For example, lets say both the 1366 and 1156 FoxConn connectors only had a 75% contact rate for the voltage connections and you overclocked them and upped the voltage to the same amount on both. If you started with a possible 18 VCC connections on 1366 and 14 VCC connections on the 1156, which one is more likely to have problems? I don't know the actual numbers, those are just made up. That is what I am getting at.
Actually we know exactly how much "redunency" intel has put into the pins as we have pin outs avaliable here, starting on page 36. You'll note the first 4 pins (as well as lots more) are not redunt and they all must have contact. As far as the non-functioning pins, you can find those here, page 18. Note all of those are around the edge and none of them are in the center.
Here on page 32, you'll find the load requirements. They require 356N to generate electrical contact, which divided out means you'll need .3N per pin. From here, page 15 we have the design of the contact point on the CPU so a little math should give you the minimum area of their requried indent.
hungryJonJon
Limp Gawd
- Joined
- Aug 31, 2007
- Messages
- 202
My MSI P55-GD65 is a Foxconn. awesome.
Got it from newegg on launch of i7 860 or the day after.
Got it from newegg on launch of i7 860 or the day after.
Anandtech mentioned near the bottom of the arcticle that they had a cpu that was showing darkened contacts that had not be under exteme methods. Granted, they also said it had undergone "thousands" of hours of testing, presumably before they discovered the Extreme issue. Still not happy about shoddy sockets, CPU's are complex enough without giving them more potential points of failure.
Well, I can definitely say that this really is just a 1156 problem.
Here is a pic of my 1156 I posted earlier
Here is a pic of my 950 that was from in a FoxConn 1366 connector.
Notice a difference?
Here is a pic of my 1156 I posted earlier
Here is a pic of my 950 that was from in a FoxConn 1366 connector.
Notice a difference?
Well, I can definitely say that this really is just a 1156 problem.
Here is a pic of my 1156 I posted earlier
Here is a pic of my 950 that was from in a FoxConn 1366 connector.
Notice a difference?
I'm checking mine tonight after I get done with other crap. After seeing that, we'll see how mine turns out.
Whisperfang
Gawd
- Joined
- Nov 12, 2007
- Messages
- 860
Well, I can definitely say that this really is just a 1156 problem.
Here is a pic of my 1156 I posted earlier
Here is a pic of my 950 that was from in a FoxConn 1366 connector.
Notice a difference?
Damn!
D
Deleted member 204526
Guest
Thanks for posting that. That's a pretty stark comparison there, and should help folks know what to look for. I still haven't pulled my CPU because my home computer doubles as my work computer, and I'm too busy just now; I did drop the clock back down to stock in the meantime, though.Well, I can definitely say that this really is just a 1156 problem.
Here is a pic of my 1156 I posted earlier
Here is a pic of my 950 that was from in a FoxConn 1366 connector.
Notice a difference?
mastrdrver
Limp Gawd
- Joined
- Dec 8, 2008
- Messages
- 412
Has anyone put forth the idea that the reason that the 1156 chips are having problems and nothing has been seen from the 1366 chips is that the latching mechanism design has changed with the 1156s? Maybe Foxconn didn't do as much homework as the others?
LOL @ everyone, especially Anand, well hopefully. I just looked at my NEVER installed i7 860 (waiting on H50 bracket) and it looks the same as Anands. There are pads with 2 indents (both near the edge of the pad) and pads with no indents. This was done during testing from Intel as far as I can tell. I was thinking earlier about this, doesn't Intel test every processor before it's shipped? If all pads are supposed to show contact points, wouldn't they show them before even being installed? And if two contacts were hitting the same pad on so many of them and no contacts on a lot of them as well wouldn't the CPU most likely not work period? Perhaps there is light contact which shows no indents and that creates a problem for subzero cooling. As far as I'm concerned there's no issue. Then again I could be wrong.. but the evidence is pointing to no issue for me.
edit: I apologize if I came off as an ass in my post.
edit: I apologize if I came off as an ass in my post.
Last edited:
Post a picture of your never installed 860.
Well, I took mine out and found quite a few pins not making contact. The only overclocking for now is turbo mode up to 3.4GHz. Since I freed up some funds from selling off some audio equipment so I bought an EVGA P55 FTW (with the lotes socket) at Micro Center that I really wanted when I built my i7 system. Not wanting to take a chance on frying a CPU and mobo. I'll just wait and see what Asus (along with other brands) plan on doing.
This is why the problem only really shows up during extreme conditions. For any signaling, the contact can be really light, it doesn't really matter. However having a better connection is a good thing for powering the CPU.
Even though I posted pics and bitched about this problem, after the photo I reassembled my system and am using it again. What we need is for Intel to clarify this. The sockets aren't really "bad" the just are making a sub-standard connection. This is the [H], we don't want just average when it comes to our parts, we want the best.
I may be wrong on this, but I don't thing any current 1156 owners like myself have stopped using our systems. I think this whole things serves as an FYI to everyone. As this thing develops, the shit will hit the fan, or all of this could mean nothing. I don't think holding off on a 1156 purchase or turning down an overclock for a little while until this gets worked out is all that unreasonable.
Not the best pic, and not same angle as the other one posted in this thread but you get the point.
Quite a few missing on the top....
Well quite a few people were getting worked up about it (including me). I was more talking about the fact there is multiple contacts on one pad and no contact marks on others. People were jumping to the conclusion that 2 pins on the socket were hitting one pad and being misaligned. Sure there may be a problem with not enough pressure for heavy overclocks or who knows. Just saying it's not as big of an issue as some are thinking.
I got really worked up at first too. But I calmed down after I though about it. With the 1156 socket those contacts are interfacing with the GPU and Ram. If there were so many double contacts and pins with no contact, that stuff wouldn't work at all.
The other thing that got me thinking was an article linked from another thread showing people blowing up their MOSFETs when they overclocked their 1156 systems.
I have a feeling that when everything works itself out, it will be determined that 1156 systems were not build to do massive overclocks while x58 systems are more robust. I mean the 1156 system is the "mainstream" socket and really built for normal conditions, while , while the workstation/server grade 1366 socket is more robust. I don't know, I'm just letting my mind wonder....
My line of thinking is that there is no free lunch. Costs were cut, and the differences are definitely felt. Now I feel silly for not spending the extra $40 on a x58 mobo, $40 on cpu, and $30 on 6GB kit. $110 for peace of mind isn't much when you consider I'll be using this for 1.5-2 years. If you work for a living, $110 over 1-2 years is NOTHING.
I wonder if this is why my new i7-860 and Gigabyte GA-P55-UD4P won't post. I get the short continuous beeps as far as I can tell and according the the manual that is a power supply issue. I have tested the board on two different power supplies, a new one and one that definitely works, with the same results. My eyes suck so I can't see the contact points on the chip well enough to see if I am getting doubles. I guess I will have to see if I can snag someone's camera to see if I can get a good shot to view.
Differences are definitely felt for who? How many fried CPU's/sockets are out there? So far I've only seen Anand's which was cooled to -105c at over 5 GHz.
I have doubles on a chip never installed.
mongoosenco
Weaksauce
- Joined
- Jan 11, 2008
- Messages
- 91
Well I just took a look at my MSI-GD80 and it seems I was lucky enough to get a LOTES. I may have to check it out this weekend to see the pin marks just for the hell of it though.
mastrdrver
Limp Gawd
- Joined
- Dec 8, 2008
- Messages
- 412
Huh??
http://www.xtremesystems.org/forums/showthread.php?t=234723
For those that don't think you can reach the required parameters under air/water
Last edited:
Edit: rereading Anand's article.
Alright so first time I read through quickly (before leaving for work), my mistake. This time I read a bit more and they said they believe there will only be an issue for extreme overclocking, and basically anything up to 4 GHz or so should be fine. Personally i'm fine with that. It's just the people talking about OEM's possibly having issues or 2 pins hitting the same pad that worried me. I'm fine with 3.6-4.0. I'll probably run it stock for a while though since I probably won't notice a difference in the games/apps I use.
Alright so first time I read through quickly (before leaving for work), my mistake. This time I read a bit more and they said they believe there will only be an issue for extreme overclocking, and basically anything up to 4 GHz or so should be fine. Personally i'm fine with that. It's just the people talking about OEM's possibly having issues or 2 pins hitting the same pad that worried me. I'm fine with 3.6-4.0. I'll probably run it stock for a while though since I probably won't notice a difference in the games/apps I use.
Last edited:
I just installed the CPU to see if it would make contact marks on the pads. Sure enough there are noticeable contact marks in the center of the pads. It was the best pic I took, can't see the whole CPU but all the pads that you can see have contact marks. This was on an Asus Maximus III GENE w/ Foxconn socket. Maybe I got lucky, or maybe I can't see the bad ones.
Picture of chip before install
Picture of chip after install
Picture of chip before install
Picture of chip after install
The XS thread sort of suggests the problem might be with the 1156 pinout itself, not with any one socket vendor. Nice to see a before and after pic though.
An XS member was saying how Biostar may have an "updated Foxconn socket" - but that guy never really substantiated it with anything. When did you buy the Maximus? I'm asking since it isn't part of the more advertised "P7P55D lineup" and may actually substantiate that guy's claim about an updated socket.
I'm wondering if those numbers on the socket backplate indicate anything. Of course I could be grasping at straws.
I'm wondering if those numbers on the socket backplate indicate anything. Of course I could be grasping at straws.