Associate
Dodgy LGA 1156 connectors on some boards!
- Thread starter OrphanBoy
- Start date
for the simple fact that I have owned all of these at some point and i have never had any problems....you get what u pay for...i will always buy the best...1.the fact is i will because i can 2.They have never let me down in the past...simple
saying that my bfg card had problems when i bought it but i sent it back to bfg and i had a brand new one within 3 days...awesome service...my next GPU(s) will be BFG also
Last edited:
Associate
for the simple fact that I have owned all of these at some point and i have never had any problems....you get what u pay for...i will always buy the best...1.the fact is i will because i can 2.They have never let me down in the past...simple
saying that my bfg card had problems when i bought it but i sent it back to bfg and i had a brand new one within 3 days...awesome service...my next GPU(s) will be BFG also
So because YOU have never had any issues with them, EVERYONE should buy them and only them ? Well, Ive had issues with ASUS boards assigning the wrong voltages to pretty much all components. Gigabyte board that straight up didnt work (see my link). Never had any issues with EVGA but I have read about 1 or 2.
There are loads of people who've experienced problems with all of them.
Stating your biased and relatively uninformed opinion as fact aint too smart.
Soldato
you will have problems with anyway, if you have had probkems iwth say gigabyte and asus in the past but not EVGA OR DFI than you would tend to go with those brands and have a push to change that
Soldato
what i dont get about all this is the double connections that are showing up on the cpu , i mean surely every contact point should make contact with one pin and one pin only, how is it that cpus are still working when it looks like two pins are contating one pad on the cpu
Associate
People are reporting the same issues with UD5 boards too, (just found this out whilst waiting for delivery of my UD5
Thought i'd let you know, I've just got my i5, and there are contact points on nearly all pads, and there are a fair few with double contact points. I have NOT put this in a motherboard yet, this is how it came from intel. My motherboard hasn't even arrived yet 
Edit - Any suggestions on how to tell once i mount this in my asus board if this is making a good connection, given the fact most of the pads have contact points already?
Edit - Any suggestions on how to tell once i mount this in my asus board if this is making a good connection, given the fact most of the pads have contact points already?
Last edited:
Thought i'd let you know, I've just got my i5, and there are contact points on nearly all pads, and there are a fair few with double contact points. I have NOT put this in a motherboard yet, this is how it came from intel. My motherboard hasn't even arrived yet
After more research i've found the following information. The pin like dots found on the cpu will be on all cpus to some degree because they are made by what ever machine tests them in the factory. You can find this out by looking at any new cpu unopened and confirm that you will find pin holes.
The motherboard socket should not make these marks on your cpu but might make a light surface scuff where contact is made. All pins from your motherboard need to be making contact with the cpu or it won't function correctly.
So if you put your chip in your board, it posts and is stable at sensible voltage then you have no problem.
As for what happens when you put lots of volts through the system in regards to what is causing problems on a technical level, then you'll need to ask someone with a more electronics based background.
Nice one spazzfish. To me, it seems odd that this problem has only just come about, I mean, turbo just overclocks the processor. I would have thought this kind of problem would have come about when the processors were first released, or even in previews. PS you have the exact same stuff i was looking at getting 
Associate
Thats odd, when I got my 920 there were no marks on the bottom (i even took a photo, a habbit i got into for later reference) but after taking it out of my UD5 there were all sorts of dodgy doubled up and missing contact points.
What i find odd is that if these contact marks are made by intel during testing, and to be honest, what else could they be, then why are there double contacts and some pads without any contacts.
I guess we just don't know enough about what intel does before the chips leave the factory.
Edit - Maybe lower the voltage so that it draws less power? Any thoughts guys?
I guess we just don't know enough about what intel does before the chips leave the factory.
Edit - Maybe lower the voltage so that it draws less power? Any thoughts guys?
Last edited:
Associate
Some kind of monitoring equipment on the doubled up pins ?
Im still looking for that photo of my chip before I put it in the socket. Im certain the marks werent there before I mounted the chip
Associate
There is a lot of bull going on in this thread and this is one of them I am afraid. I am p&ssed off with the reaction to anantech. He pointed out that at extremely high currents some sockets burnt. But if you read the article he even points out that this could be due to careless processor installation.
There is so far, no proof that this is due to a fault in the socket. Its all just heresay and forum rumour. Intel choose the Foxconn socket becuase it was available in spec first. Do you really think that Intel dont check these things? The use multi million dollar FAB plants and they check there production with extremely expensive equipment. More than just a fews noobs looking at pin pricks on a CPU.
Its just not true that all pins need to make contact. Infact the chip and socket is designed with redundancy. Go read the intel spec for the 1156.
http://download.intel.com/design/processor/datashts/322164.pdf
You will see that there are 175 pins for VCC alone. It does not matter a dot if some of the pins are not connected as the load will be spread around the other pins. Thats by design.
Everytime you insert a cpu it will align slightly differently with the socket. The worst thing you could do is repeatedly insert the CPU remove and inpect the contact plates. Each time you do it you are increasing the electrical resistance between pin and contact. Also you are exposing the chip and socket to contamination.
Even a slight finger print on the CPU contacts will increase the resistance by say 1ohm. Which is well out of spec for the CPU. Thats why the base plate of the CPU and the socket is protected by plastic screening plates.
Have you ever been in a clean room? I have. There is a reason for the space suits and goggles etc. Its to stop contamination.
Please stop spreading nonsense about CPU inspection and pin ***** counting.
There have been no major returns of boards or CPUs with burnt out sockets or pins. All we have seen is a few boards from poland using LN at extreme currents. Big deal.
Last edited:
Nobody's disagreeing with you. I've looked at that spec and I won't pretend that I understand it because I don't.
The only thing that interests me though is if as your saying there is scope for lack of pins to make contact with the chip (which to me seems unusual because the way sockets are designed they should touch the pins as a matter of course anyway) Then in effect contamination of the chip is doing the same thing?
For example one chip makes all the connections but is contaminated effecting x pins.
While the other chip isn't contaminated but has x pins not making good contact. Both chips in this example have same quantity of effected pins so then there is no problem due to both falling into failure tolerance or am I missing something?
problem is though what is most concerning re: anantech article.
looking at cpu removed from foxconn socket their are pads with seemingly no contact, yet when put into lotes/tyco socket, there seems to be contact on all pads.
I think that this is very concerningm from what i read elsewhere it is worrying when considering new build, as even if you dont overclock, you dont want to worry about possible damge over x period.
They say DFI have now stopped using the foxconn sockets, why is this? There is a lot of suggestions perhaps the actual clamp/plate is faulty etc... who knows.
But from looking at this, it has completely stopped my initial plan to upgrade since my copy of win7 come through the letterbox
looking at cpu removed from foxconn socket their are pads with seemingly no contact, yet when put into lotes/tyco socket, there seems to be contact on all pads.
I think that this is very concerningm from what i read elsewhere it is worrying when considering new build, as even if you dont overclock, you dont want to worry about possible damge over x period.
They say DFI have now stopped using the foxconn sockets, why is this? There is a lot of suggestions perhaps the actual clamp/plate is faulty etc... who knows.
But from looking at this, it has completely stopped my initial plan to upgrade since my copy of win7 come through the letterbox
Associate
Take the following example:
Lets say the cpu in its socket has 95% of VCC CPU pads with "perfect" contact with the socket. If 3% have an off contact, eg low pressure or slightly off pin then these 3% will probably have 0.1 Ohm more resistance than the perfect contacts. Then lets imagine 2% have no contact and are not conducting. The non conducting pins we can ignore.
If you look at Ohms Law the heating effect is proportional to the resitance.
So let us say a resistance of 0.1ohm will generate a heating of say X.
A contamination will either block the pin, eg fluff and thus there is no conduction and heating effect.
But a conductive resitance eg, sweaty finger print or LN residual may produce 1 ohm of resitance.
1 Ohm resistance will generate a heating of 10X.
So a sweating finger print will result in 10 times as much heat as an off connection.
Not a good idea especially if localised into 5 square millimeters etc.
BTW one interesting point was the 1366 had 250 VCC contacts. It needed more becuase is TDP and current draw was higher.
So we are talking about VCC and power pins. But if you look at the spec there are also other pins, eg control, data pins. These are what the CPU uses to "talk". If these dont contact then we have a no go situation, eg no data comms through the pins. But if these are not in perfect contact then it doesnt matter as the current through these pins is much lower and thus the heating effect is also much lower.
CPUs are designed to be removed from there packets, inserted once into sockets and left alone.
Dont remove them, reseat them etc. If you do then you will be breaking down the conductive material on the CPU. In this case I estimate a 0.2 -> 0.5 ohm increase in resistance and thus a 2->5 times heating effect across those pins.
So please guys listen to me. Dont reseat your CPUS. The sockets clamps have an extremly high pressure. You are more likely to increase thermal burnout from reseat when overclocking than a once off seating.
Associate
- Joined
- 26 May 2008
- Posts
- 1,774
Clean rooms shrooms, sorry but how clean you keep em is not the discussion.
It's pad contact. now you say there is redundancy, but I don't think saying there is redundancy is quite the same as random pad contact or double contact in the same pad which is effectively a short in some instances.
Many here have probably done BSEL mods on chips, and I am sure you can't just do any old pin in an attempt at guessing a rough overvolt.
Intel may use multi dollar fab plants, it does not control the fabrication of Foxcon sockets though, and those fab plants do not manufacture them. Having two pins contact some pads and some pads with no contact is in no way controlled on these sockets, nor is it acceptable, nor can anyone state with certainty it is not detrimental long term. People don't have to pay for something that works "sort of" and your own post contradicts it's points.
Next it will be OK to wire the house electrics up willy nilly and guess the specs.
Associate
The best situation is 100% contacting pins. No one will argue with that. But as I said, a few off pins are within spec, even when overclocking it wont make any difference. If you are doing extreme overclocking and putting 250 watts through a dodgy pin like in the Anandtec article (5ghz plus with LN) then it might!
For "normal" over clocking eg 3.8-4ghz on an i5/i7 I dont see a problem. These clocks are even within the MOBO manufacturers "well within" specs, eg OC Genie ETC. BTW if you re read the Anandtec article he says that no problems have been spotted using air cooling. Now air/water cooling will only allow the overclocks I have mentioned.
But I say it again. The worst, most stupid thing you can do is to insert your CPU, remove it, count the conducting pins and reinsert. Thats going to add extra resistance and heating effect.
BTW. One thing to NOTE. And this is BIG style, is that you have no idea if your socket is actually a Foxconn. There are three pieces to a socket.
1) The backplane
2) The Socket
3) The Socket front plane
Each mobo can have a combo of the above.
So just becuase you have Foxconn written on the back or front or on the plastic cover, doesnt mean that the socket is Foxconn.
Associate
Its by design. Its a little bit more complex then wiring up your mains sockets. The CPU is "DESIGNED" so that it doesnt need all off its 175 VCC pins in perfect connection to the socket. It has built in redundancy. If you chop through you mains cable at home then your lights will go out, because there is no redundancy. If you dont have some VCC CPU pins with perfect contact then the lights will be on. Its by design.
BTW, BSEL mods
I am sure that an "expert" like yourself could keep a CPU clean from contamination etc, whilst BSELing his 775 3.6ghz p4. But my advice is to the worried many who obviously dont understand the tech as much as you and the other "experts" on this board do
Last edited: