Last I heard there was this padded room he was locked away in, something about soft walls so he doesn't hurt himself :rofl:Quote:
Originally Posted by quake6
Printable View
Last I heard there was this padded room he was locked away in, something about soft walls so he doesn't hurt himself :rofl:Quote:
Originally Posted by quake6
chillax dude...Quote:
Originally Posted by raju
illl get a clamp, alright alright... until then, take it easy will ya :P
LOL :DQuote:
Originally Posted by bingo13
hmmmm do you know thomas?
i talked to him a few times via email when i worked for foxconn...
he seemed like a really nice guy, im sure hed share more details with you if youd ask him.
so your not seeing any of the problems at all?
you can bump up vcore on the msi board to 1.4v+ and its fine?
maybe hes testing with an es chip and it pulls more current than retail chips?
or faulty psu?
burned pads.sockets only happens on top boards?
hmmmmmmmmmmmmmm
so maybe it does have to do with droop...
So which boards don't use the foxconn socket? I have no idea where to look for info like that so hoping someone here knows :D raju? :D
I don't have full model specifics because every vendor seems to have some boards in the channel with Foxconn sockets.Quote:
Originally Posted by Dahmer
So far, DFI have dropped Foxconn on 1156 until further notice. But there are some boards out there from early batches using Foxconn.
EVGA are using a mixture of TYCO AMP, LOTES and Foxconn. I was told future batches will focus on using TYCO AMP and LOTES.
ASUS, so far I have only seen Foxconn from them (not heard anything back at all)
Gigabyte, they are using Foxconn too.
MSI have some boards which use LOTES sockets, from budget to GD80.
It's a little hit and miss all round at present.
I should add though that not everyone has issues. It's a case of checking that most pins touch if you have a Foxconn socket. Just a bit dissapointing that there can be a significant amount of unconnected/poor contact pins.
later
Raja
the problem is that foxconn is making tons of them and other socket makers seem to only make low volume...
i think most boards are foxconn and then some are mixed, sometimes foxconn sometimes others... only evga uses lotes only... maybe dfi as well idk..
EDIT: nvm, read rajas post above :D
I have 3 MSI P55 : 2 GD80 with Lotes socket and 1 GD65 with Foxconn socket.
huh? no credit to xs or the people who brought it up in the anandtech article? why?
Just checked all the pictures of the P55 boards on evga site and if I can trust those half of them show foxconn :(
P55 LE, P55 Micro, P55 FTW clearly show foxconn on the piece of plastic in the socket or on the backplate, other ones I'm not sure of...hope they get revisions out soon :D
check the pics on Anand, compare to this post: http://www.xtremesystems.org/forums/...4&postcount=37Quote:
Originally Posted by saaya
check username: raju
check name author at Anand: Rajinder Gill
1+1 ? :clap:
i know that raju=rajinder gill :rolleyes:Quote:
Originally Posted by jmke
i thought he was insipred by this thread so i was surprised he didnt mention sf3d, massman or xs... whats the big deal, geez...
you made it a deal ;)Quote:
whats the big deal, geez...
duh. :D
So what happens to current users of Foxconn-socket boards then? Don't we have a Gigabyte rep on this board? I haven't touched the voltages at all on my board, so hopefully I don't see anything when i check this weekend...but is there any hope for RMA/replacement (as a preventative measure)?Quote:
Originally Posted by Dahmer
Mav451, read all thread :) This happens only at high loads. Something like ~4.5GHz+ and 1.35V+ with a load of linx or similar.
Any reports of dead Asus boards?
It's always nice to get mentioned for whatever you did to help out, be it investigate research or just bringing up the issue. If no one would share findings, there wouldn't be so much to discuss.Quote:
Originally Posted by jmke
That being said, I think Raja's the one who did all the research, so I have no problem with him not mentioning anyone. The article is more than just the 'heads-up', it's sort of an explanation as well.
I was going to ask for you permission to post up some pics. I can possibly add them into the engine and give thanks where needed. I only really pursued this becasue I got caught out twice in rapid succession with the burn (the second time was just plain nasty and on the best CPU I had potentially screwing any future results) and then a close friend on the same weekend. Had no choice to investigate because it delayed the article badly (still not done).Quote:
Originally Posted by massman
Had I used any direct information I would have given direct thanks all round for the input.
I'll be sure to contact next time if I use anything directly.
Yes, two Max III ROGs (one is mine). The second board was worse - only ran 32m at 5G's, power draw was not even that high, but cause of so little pin contact the pins that were making contact got screwed.Quote:
Originally Posted by kiwi
Yikes P55 :slapass: , never had one never will buy one :down:
My thought exactly. I just posted earlier today @ techreaction... P55 is like the handicapped redheaded stepchild of bloomfield...:p:Quote:
Originally Posted by chispy
H'mm just checked my Biostar TPower I55 - Foxconn on mine. :(
Whats the chances of Foxconn getting totally anally sued over this for liable damages? I hope they get screwed seriously. :mad:
Why? This would never happen to anyone unless they were voiding their warranty and pushing excessive voltages and clocks into the processor :shrug:Quote:
Originally Posted by Nelly
eVGA P55 FTW here with LOTES!!!!!!!!!! :D :D :D
Hah nice. Did you check the pin markings anyway, or did u just take the side off your case?Quote:
Originally Posted by KURTZ
just checked the side off mate :)Quote:
Originally Posted by Mav451
Ahh thats alright then I'm not one of those people. ;)Quote:
Originally Posted by TheKarmakazi
New foxconn socket revision corrected that issue... as i heard by Biostar.Quote:
Originally Posted by Nelly
You are referring to the socket protector or the backplate. Not the actual socket which is normal because I think everybody uses their proector.Quote:
Originally Posted by Dahmer
i've the LOTES backplate, it's possible that i've the Foxconn socket? really?Quote:
Originally Posted by Cpt.Planet
I highly doubt you have a foxconn socket. Sometimes they use different vendors for the socket, the socket cap, etc etcQuote:
Originally Posted by KURTZ
So what happens to early boards that have the issue? Thanks for the cash and see you later? And I'm talking about anyone that's affected, Gigabyte/Asus/MSI/eVGA etc.Quote:
Originally Posted by Pt1t
:(
so will all board now use a newer foxconn socket? is this what you mean??Quote:
Originally Posted by Pt1t
Hey , iam going to Buy a MSI P55 GD80 tomorow how can i check if the socket is Lotes or Foxcon , kinda scared about this was hoping to get a I7860 to 4ghz 24/7 , Also would it be beter to go for The ASUS P7P55D DELUXE or does Asus also having this socket issues ?
pushing excessive voltages and clocks into the processor just shows the problem sooner, but the problem is there as it seems that the CPU does not touch all of the pins.......... over time this will become a problem even with normal voltage IMHO.Quote:
Originally Posted by TheKarmakazi
Just flip the mobo upside down. B/c MSI can have either, it'd be best if you were able to see the mobo before purchasing, a la Microcenter/Fry's. If you have neither...you are really entering a lottery then.Quote:
Originally Posted by usoldier
When I flip mine upside down their is no details/info on the backplate, on the front plastic socket protection cover it says Foxconn.
So the Biostar boards have a new revision of the Foxconn socket?Quote:
Originally Posted by Pt1t
I'm guessing when you did your results yours didnt cause any damage to the socket?
cause i asked about it?Quote:
Originally Posted by jmke
hows that making it a big deal?
so if i think an article on madshrimps.be is insipred by a forum thread or somebody elses article, which happens all the time, and i ask you: "why didnt you credit him/them" thats a big deal? how so? if id have said something like "whoa he used this thread/massman and sf3ds findings to create an article without giving them credit, what an sob!" then ok, sure, that would be making it a big deal and out of line... but i know raja for quite some time, so i figured he just forgot it. i used a :stick: smiley in both threads originally but then realized it might make the post look rude and edited the posts to remove them... and i think i forgot it in one of the two replies... but even with that smiley i dont see what the big deal is?
damn... :/ sorry to hear man...Quote:
Originally Posted by raju
and it came out of nowhere?
that really sucks...
you had some damn impressive clocks with that one... :/
probably something like that... with the added line: if you want an EVEN better board with a mega ultra durable socket, hand over some more cash pls... :DQuote:
Originally Posted by Mav451
That was on the ROG. I ran 2 benches that day, Wprime 1024M and 32m (bear in mind the Tyco and LOTES boards passed the bench at higher clocks that what I was running on the ROG). The CPU has some VCC pads that warped from the heat (they did not burn completely). The board fared worse and died when some of the pins melted. At this point I really needed to know what was causing it. A friend of mine rang the next day and he had the smae problem (he only ran 32m), so the heat was on to find out why this kept happening.Quote:
Originally Posted by saaya
I've used the CPU again since and it clocks good for WPrime still so I might have gotten away with it despite some of the VCC pads looking a bit warped. Nearly finished now and I should have it all posted up soon.
later
Raja
Ah ok, I wasn't 100% sure if they'ld have the socket itself too so I pointed out it's the protector having the foxconn label :)Quote:
Originally Posted by Cpt.Planet
And I guess I'll have to get the P55 ftw board :)
cool, cant wait to read that piece :D
so the cpu is still alive?
and it clocks the same?
i have a bad feeling about it though lol, those pads are really thin, and if they are already in bad shape, it makes it more likely for the pins to overheat again...
what do you think, does getting the socket cold enough prevent this from happening?
wouldnt it be alright if the socket and pins would be cooled well enough? then the plastic cant melt, the pins stay in place, plus they get cooled so they keep roughly the same resistance and pressure against the pads... or is it only slowing the chain reaction of higher temps->higher resistance+less pressure against the pad->higher temps?
oh and raja, please enabled pms on xs man... :D
i originally wanted to pm you about not mentioning xs and sf3d and massman but i had to post something cause you have pms disabled :P
Last time I sent a PM on here to someone I got accused of trying to lure away XS staff. So best keep it out in the open, it actually creates less controversy. :) that and I have limited time to answer stuff nowadays.Quote:
Originally Posted by saaya
I'm guessing that would only slow down the chain reaction. Heat is not really the problem, it's just the result of the problem of the pins not having enough contact with the pads.Quote:
Originally Posted by saaya
you asked, people replied, you replied "what's the big deal". :rolleyes:Quote:
Originally Posted by saaya
EVGA P55 F.T.W board with Lotes.
http://img243.imageshack.us/img243/2425/lotes1.jpg
This is what mine looks like, mine has no name on the front plate or backplate either yet some have Foxconn on them. (second pic not my board just for description/info purposes)
http://img389.imageshack.us/img389/9...ri55socket.jpg http://img230.imageshack.us/img230/1...ketcloseup.jpg
Anyway I've just checked my CPU under a light with a magnyfying glass & all the gold plates are firmly inprinted against the socket on my CPU so theirs no problem with mine at all lol. :)
I havent got a digital camera as I would just love to proof this & mine has the Foxconn socket! :shrug:
Quote:
Originally Posted by jmke
...nevermind...Quote:
Originally Posted by jmke
Just checked my ASUS board, definitely Foxconn, but complete pin to pad contact. Purchased on the 9th of Sept., for what its worth, not sure how to determine when it was manufactured.
come back to EU man, you logic is affected;)
you edited your post, never saw this:rolleyes:Quote:
Originally Posted by saaya
you are making a big of this, nobody else is ;)
Ļ[M] always credits source btw! ;)
also noticed for all the damaged sockets the plastic right at the hinge is either unmarked, or marked with a small 1, 2, or 3. my board has a 4 at the same hinge. any thoughts, is this the socket revision number?
The 1 2 3 4 is probably a cav# of the injection mold it came out of. this could be traced down to 1 bad cavity and not all foxcon sockets are bad
Registered User
Join Date: Jul 2007
Posts: 7 Hi I have problems with my 1156 i7 860 at stock speeds with it randomly rebooting and windows explorer crashing and games crashing and multiple blue screens etc. i have an asus p7p55d deluxe with the ressesion made foxconn socket! and yep my cpu has the exact same pin area not touching as the article on anandtech. this is just shoddy workmanship on foxcons part, the motherboard manurfacturers not testing their boards properly and intel leaving an integral part of how their processors operate up to seccond rate manurfacturers! I have rma my board with the instructions of not giving me a repacement board as it will also have a shoddy foxconn socket on it too, i am asking for my money back, and im going with 1366 now!
oh here is my proof i took this picture myself.
http://www.flickr.com/photos/43731058@N08/4018093023/
yeah your right ;)Quote:
Originally Posted by jmke
my bad :rolleyes:
thanks for pointing this out :clap:
thanks :up:
Does the cpu operate at stock speeds Roy or is it a ram problem ? I don't think you will get ya money back like that, also the multitude of pins you are telling are not making contact, I wonder how how on earth the rig can boot ?
More likely poor contact in places than no contact at all. If the contact resistance is altered to some key signal pins, you'd get erratic behaviour.
Poor contact will effect ever1 overclockler or not. Those contacts do more than just deliver juice they also send data. no contact memory might not work. no contact PC! express may be compromized and so on this will effect every1 not just overclockers.Quote:
Originally Posted by raju
This was going to happen sooner than later with the pins in the sockets. to many variables. You know how difficult it is to make those pins perfect every time? doesn't look like it leaves much room for error. Then you have the plastic housing and other plastic parts in the socket that they protrude through. Any1 ever work at a plastics factory? Well I have and it's not exactly the best repeatable process. it could even come down to some moron forgot to turn on the water to the tool, but more likley some (_{_) of a boss wanted the job speed up to make more $$$$ and the parts are warped or flash in the holes pushing the pins in the wrong direction or limiting their movement.
open arms :):(:confused::D;):p::cool: :rolleyes::mad::eek::rofl::ROTF::clap::welcome: :down: :up::shakes::shrug::shocked: :stick::woot::slapass::slap::horse::stop::ocau: :idea: :2cents::kissbutt::smoke::smileysex::hump: :chicken: :flame::weapon::caution::bows::party3::party: :party2:Quote:
Originally Posted by saaya
Hence the term 'erratic behavior' Raju's using.Quote:
Originally Posted by cowpuppy
The socket burnout is caused by VSS/VCC not, or improperly being connected. The pins used for communication (=data) have nothing to do with the burnout. In other words, the poor contact of VSS/VCC pins doesn't cause the system to become unstable because of the poor contact.
wtf? :)
VCC- Vcore
VSS- GND
and poor contact with these pins cause burnout LGA1156. especially poor contact with VCC pins.
Poor contact causes those that are properly connected to overheat because too much current is drawn through them.
I see contact marks on the pads you highlighted. The light marks in the center are from the contacts. the little pin holes are not. They are from the factory that way.Quote:
Originally Posted by 0_o-Roy
The pin holes are from the Foxconn socket. The scuff marks likely when the CPU is first instered into the socket or direct from Intel.
No, I've been waiting on my H50 bracket from Corsair before installing my 860. So I looked at my brand new, never installed i7 and the pin holes were there. I took a pic (not the best, but you can see the marks) I then installed it in my Maximus III Gene w/ Foxconn bracket and now see the light marks in the middle of the pads. Looking at the pins, they are not pointed and they are at an angle so I do not see how they could possibly make little pin holes. I don't see how anyone can think that the CPU would still be working with so many pins not touching and so many pads with 2 pin marks. (unless people think it's from multiple insertions I suppose)Quote:
Originally Posted by raju
Picture of chip before install
Picture of chip after install
edit: And here is a 1336 i7 that someone took a pic of trying to show the pin marks in the center of the pads as being good contact. Well look closer and see that there are atleast 10 pads with no pin marks and multiple pads with 2 marks.
edit: Here's another 1156 i5 chip from hardforum before being installed.
The scuff marked contacts on the bottom of the cpu may be making contact but it would be very poor contact if they are! The cpu should have one (1) mark on every pad, on mine many have one, some have two! I also only had put the cpu into the socket once ever so i cant see how some contacts have two? I dont see why I shouldnt get my money back? its not a Ram issue i tested the ram in my other pc. Also to lathode, a brand new CPU should have zero marks on the contacts, none at all, I dont know whats up with your 860?
The pin marks are NOT from the socket, they are from the factory. The scuff marks are from the socket. Look closely at the socket, the contacts are round, they are not points. Also they are at an angle and bend down when there is pressure on them, they will not put hard indents into the pads. Look at the 1336 CPU I linked to, you will also see multiple pads with 2 pin marks as well as pads with no pin marks. Look at the new i5 I linked to, it also has pin marks. Also look at multiple pads right next to eachother that have 2 pin marks on them, if there were 2 pins marks would a pad adjacent to it have none? Now think about this, if the pin marks were contacts how in the world would a CPU work if there are dozens of pads with no contacts and pads with 2 contacts? Especially considering the PCI-E controller and memory controller are on the CPU. One other thing, I'm betting the marks that are on the CPU from the factory are from testing. So yes I think they should have some type of marks on them and it would make sense that they are not tsted in standard sockets, but custom sockets for testing purposes.Quote:
Originally Posted by 0_o-Roy
my freinds p55 max3 from newegg socket melted down tonight. and he was only on air at 4.1. his system shut down and would not restart. he called me i said just out of curiosity to check his cpu socket and sure enough in the same corner he had the socket melted like these pics shown here and the bottom of the cpu was messed up. he did have the foxconn socket also. he just got his last week also. this is sure making me a bit nervous about buying anything p55 right now. as i said he was only on air. i do not know his settings but he said it was at 4.1 ill try to find out more of what settings he was running tomm
What was he doing when it shut down and how much voltage was he giving it?Quote:
Originally Posted by zfactor
i dont know the voltage settings but he said he was in the middle of a game he didn't mention which one. so id gather nothing that stressful. he said before that he was running prime. so that damage may have been done enough before hand i dont know. if i can get him on here i will. last time we talked he was under 1.4v that was 2 days ago. he is not answering his phone right now so ill find out tomm. all i know is he is PIS-SED off right now.. he has no idea if the chip is fuber now and doesnt know what the rma will be like. back to newegg is what he is hoping but they prob will not take it burned like that
As a quick fix for this problem, what about taking an old plastic card (old school id or whatever, just no raised numbers like a CC) and cutting out a square from the middle. Place it on the processor such that the hole fits just around the heatspreader. It should press the processor into the socket better.
edit: i *think* it would work, but have not tried it, someone else want to give it a shot?
hahaha :D ok, :peace:Quote:
Originally Posted by jmke
lol saaya, you had to quote all the smileysQuote:
Originally Posted by saaya
Yeah the Foxconn socket should leave a slight scuff, but it does not seem you always get contact on all of the pads or it is poor in places (leaving the feintest of marks).Quote:
Originally Posted by Lathode
i've had a lot of poor contact related issues on Socket 1366 boards in pretty much all brands which are usually manifested in dropping one or two channels or not starting at all or getting stuck at early postcodes too
i guess it's the way of the new sockets......
http://img27.imageshack.us/img27/5551/cpumelt.jpg
Again Foxconn socket. OC was around 5G's, but load was Spi 32M (so around 7.5-10amps from EPS 12V)
Is that from a Foxconn socket or LOTES? Have you tried subzero at ~5GHz with a LOTES socket yet?
Foxconn once againQuote:
Originally Posted by Lathode
Zero burnouts on LOTES so far. Even at higher OC's and more VCC (current draw), I've had no problems on the TYCO AMP or LOTES sockets.
I should add that it does not take much current when the contact is bad - 7.5 ~ 10 amps is very low VCC draw (factor VTT out of that) and the subzero die temps cooling the internal power plane. The pins break the thermal inertia and still manage to melt despite the sub-zero die temps.
Quote:
Originally Posted by zfactor
Can you get some pics from him? Interested in the burn pattern.
Word is that Foxconn have changed the socket top plate in a revision to correct the problem. Not sure that it is trouble free now as someone in mainland China had a problem on the revised socket too. What I do know is that DFI, EVGA and MSI have moved production over to other socket vendors.
When were the sockets revised and are they identifiable?Quote:
Originally Posted by raju
thats the part Lathode that is all under wraps. I've had two vendors come forwards and show me what Foxconn sent them. Foxconn claim the issue was identified in May and that Intel were due to revise their design to rectify. Apprently these were supposed to be sample sockets with the issue and all issues were supposed to be in house in some vendor labs. Since then there have been issues reported on what is supposed to be the revised Foxconn socket (or samples that have gone into the wild on retail boards), so I'm not sure what to believe at this point. What some of the vendors have seens themselves is not centric to an earlier model or revision of socket.Quote:
Originally Posted by Lathode
completely moved to LOTES and no other sockets?Quote:
Originally Posted by raju
wow
Yeah, LOTES only for DFI and MSI, while EVGA are using TYCO AMP or LOTES only for P55. DFI actually removed the Foxconn socket from P55 board stock and reworked to LOTES (obviosuly not production/stock scale like some of the bigger vendors which makes this possible).Quote:
Originally Posted by dinos22
Any news form Asus on this? I know that they used to use the LOTES on their high end boards before the P55Quote:
Originally Posted by raju
What is the actual fault really? is it the socket itself or is it the design of the metal plate pressing onto the socket? Can't we just replace the metal retention plate with a LOTES one?
Has anyone tried using pliers to increase the bend a few more degrees so it gives more pressure to the socket?
http://img88.imageshack.us/img88/934...ri55socket.jpg
i will try to get some before he sends it back wherever its going to :down: i sent him the pics from this thread and he said it looks basically just the same top left corner he said not as bad as the worst ones here but it was for sure more than a pin or two. he said this morning he guesses maybe 6-8 pins at leastQuote:
Originally Posted by raju
My MIIIF have the problem, i see my i5 750 and non all pads have the right contact, what a SHAME! :cry
The top plate on the LOTES is slightly different looking, mainly the tabs that put pressure on the CPU have different indents. So if Foxconn is changing the top plate like Raju's source says then I wonder if people can request a new top plate and replace their own.
LOTES
http://home.comcast.net/~lathode/misc/lotes.jpg
Foxconn
http://home.comcast.net/~lathode/misc/foxconn.jpg
Hi@all
Your thread and the article @anandtech makes users worlwide crazy :ROTF: Young OC junkies here in Europe means P55 is :flame: .
It s a nice socket with only little problems :up: .
@ raju
- Can u tell us your proffesion ?
- Why not u show burnout socket with videocam ? ....youtube
- Before and after OC
- Clear new boards with LOTES vs Foxconn socket with new CPU
Only that makes 100% clearance for xtreme overclockers.
If not, 100000 technical people means you and the other "burnouter" are
only crazy panic entertainer vs Foxconn . U know the technical conditions ,that socket 1156 has 75Vcc pins less vs socket 1366 ? ..... its better manufactures limit VCore.
See u in David Letterman Show ? :rofl:
Looks like the altavista translation failed there.
Hi
Does it back plate and socket top plate removeable ? it had 3 screw and I was wonder if I have foxconn top and back plate then remove it and order new LOTES top and back plate kit to install it ? will it works ? easy swap it out ?.
http://i34.tinypic.com/2i8i3d5.jpg
http://i35.tinypic.com/15qpxlt.jpg
dg
Oh come on. Many here now Rajinder and I donīt think thats neccesary.Quote:
Originally Posted by BikerKS
Also heīs not the only person reporting this. Or now itīs not only OC that needs to be video-taped to probe it?
Yup was told the same by DFI they're going Lotes for P55 - I know the DFI P55 i have coming is Lotes based :)Quote:
Originally Posted by raju
Quote:
Originally Posted by BikerKS
Count the VCC pins here;
(Page 86 ~ 88)
http://download.intel.com/design/pro...hts/322164.pdf
(Page 44~47)
http://download.intel.com/design/pro...hts/320834.pdf
It' around 180 for 1156 to over 250 for 1366. So yes around a 75 pin gap.
Why do you need to see it on video? Do you work for David Letterman? (I assume it's cable/boom duties) :rofl:
Seems like we have a Foxconn employee among us :D
It looks easy enough to change the ILM/Backplate ourselves, just a case of sourcing this: http://cn.lotes.cc/product_info.php?id=3289
http://img16.imageshack.us/img16/8784/1156ilmbp.jpg
Alzheimer's best friend ? :ROTF: I say it before.Quote:
Originally Posted by raju
No video- no reliably informed- about your proper working conditions ;)
Any sh*t is on youtube, why not so "meanfuly things" like yours :D ?
Its an easy way to use. When we will report @Anandtech burn socket like yours, we can use 2V Vcore with boxed cooler and send the pics to stupid online warriors too. :ROTF:
We mean u lift LOTES and others not fair in heaven .
That would imply I am talking BS. You might as well ask for video evidence to everything and from everyone. Don't like it? Don't read it, simple.Quote:
Originally Posted by BikerKS
Something strange going on for sure. Now we are talking lies unless we provide video evidence :rolleyes:Quote:
Originally Posted by massman
So even Foxconn uses Lotes sockets.Quote:
Originally Posted by BikerKS
I think Babel Fish Translation is not working well today :ROTF:Quote:
Originally Posted by BikerKS