Official GIGABYTE Forum
Questions about GIGABYTE products => Motherboards with Intel processors => Topic started by: MindLESS on October 15, 2009, 01:48:51 pm
-
Does anyone know offhand, or without having to take their machine to bits, who makes the CPU socket on the Gigabyte boards?
I belive the main people are Foxconn, Tyco AMP and LOTES.
I read an article about the Foxconn not working well with extreme overclocking due to poor contact with the base of the CPU and the current draw being split over fewer pins and burning up the CPU. I don't have plans on overclocking much, it's just curiosity.
I've got a P55-UD4, in case different sockets manufacturers are used on different tiers of boards.
-
Foxconn according to the socket on my P55 UD3R board. I am fairly certain that all Gigabyte P55 motherboards have their sockets made by Foxconn.
-
Hi MindLESS
total surprise :o for me as Foxconn is recognized as being very good in connectors quality, that's why used by Gigabyte.
could you provide the link to this article?
-
Original article:
http://www.anandtech.com/mb/showdoc.aspx?i=3661
It does state that it is to do with extreme overlocking (seems to be >4.4GHz clock speeds)
Fairly sure the Foxconn would be within Intel's tolerances - this is a bit beyond specs and requires some fairly heavy duty active cooling to acheive (air+water won't be cutting it!)
-
If you look here http://forum.giga-byte.co.uk/index.php/topic,638.0.html (http://forum.giga-byte.co.uk/index.php/topic,638.0.html) Bex has got her x58 i7 920 at 4.5ghz prime stable. My GA-EP45-UD3P with an E8600 can reach 4.53ghz prime95 stable. Both boards use the foxconn cpu sockets (as far as i'm aware).
-
I don't think it's a problem with Foxconn in general. As the topic titlesaid it's regarding the P55 (therefore LGA1156 socket) specifically.
Also, Bex's is an X58 i7-920 therefore it's got an LGA1366 connector. According to the Anandtech article, a good number of the pins are used for providing power to the CPU.
The other machine you mentioned is a P45 motherboard, so an LGA775 socket. Pins may be making better contact on that due to a lower density (speculation on my part - but from memory the rough package sizes have stayed constant between Core2 and i5/i7)
The problem experienced (across a number of board manufacturers) was that the LGA1156 has fewer power pad and a number of these weren't making good/any contact and so the current draw was being split across the remaining ones. This lead to too much power being put through a too few pins/pads and the pins/socket melting.
The problem isn't as bad with the 1366 as it has more power points, so lower load across each point. Lower load means less heat.
Personally I'm completely happy with my board - but I'll bear it in mind when overclocking. I'll probably take out the CPU and check the contact it's had with the socket before doing anything heavy with it.
BTW - What's the bike next to you? Keep trying to figure but it's a bit pixelated to work out.
-
Ohh interesting article! Did not know that!
Oh and its a 09 Kawasaki Ninja 250R ;)
-
Yeah, I was rather interested when reading it.
Makes sense thinking about it - having 125+ watts on one of those little CPU pads seems excessive in hindsight. Splitting it up over the pins makes much more sense. Would never occur to me without having to think about it really.
Nice, I've got an 02 Triumph Daytons 955i CE =D
-
Your socket will look like this:
(http://i154.photobucket.com/albums/s261/Massman-/ud6/l.jpg)
Another source:
http://www.xtremesystems.org/forums/showthread.php?t=234723
-
Here is information from Foxconn:
(http://forum.giga-byte.co.uk/upload/files/2009-10%20LGA1156%20Foxconn%20Quality%20Letter.jpg)
-
My GA-P55-UD4P just arrived and it says:
Foxconn 115XDBP 9701
on the back of the Motherboard on the CPU Socket Mounting Plate.
-
And some of them CPU's can draw 140 watts Too,
O/C does carry some risks to your kit
-
Yup - increase the voltage and you'll be upping your current draw a lot.
Overclocking can always be risky as you're operating items outside of their intended threshold.
Thanks for your response, runn3R - I'll be careful if I decide to overclock my chip now =)
-
seems like LOTES sockets will arrive with new mobos, see these links:
http://forum.giga-byte.co.uk/index.php/topic,734.msg3306.html#new
http://www.clunk.org.uk/forums/news-hardware-tech-news-around-globe/26369-gigabyte-p55a-mobos-sata3-usb3-3x-usb-power.html#post74623
-
It's a shame I already have a stock P55 =/
Oh well - it's good to see that Gigabyte have responded to the reports (even though the original sockets are fit for purpose still)
-
One wonders why the change if the original Foxconn sockets are 'fit'... ???
-
I don't know about the P55A - they could be a 'premium' line of P55 motherboards, instead of a new revision of the current P55 range.
The sockets fail *at extreme conditions*. For people who want to try and get the fastest overclock possible, they'll want the best sockets possible.
You can get a car that is 'fit' for driving - but you won't be able to make it do as much as a car with higher specification parts inside the engine. Even if the engine is the same layout.
To be fair this is being blown up a lot on the 'net - people experiencing this kind of failure were often using LN2 to cool the CPU and the voltage cranked up heavily. This is hardly 'normal operating conditions' and if the socket failed under that kind of use it's your own fault. You choose to push it and take responsibility for the consquences.
-
Not all the failures have been under extreme cooling conditions. Read the threads at xtremesystems again and you will see that there are a number of boards with the same damage that were not run at extreme conditions.
My motherboard has the Foxconn socket. I have it running at stock. But why should I have a dodge socket that can potentially damage my CPU some time in the future, especially if Gigabyte have changed socket supplier. Those overclocked systems simply bring to the fore the problem sooner. The problem is still their regardless if you over-clock or not. At the end of the day we have a 1156 socket that does not make contact with all the pads on the CPU. That means those pins that are making contact have to work harder even under stock conditions. I see it as a concern for the long term.
I personally think that Gigabyte should offer a swap out for those of us who do worry about a sub standard socket. If others are happy with their motherboards then they do not need to swap them out. A full recall might be over the top, but possibly appropriate.
I am very happy with the Gigabyte brand, this is not their mess up. My next board will most certainly be another Gigabyte but lets not do the head in the sand thing here. We have a potential problem that could in time kill or permanently damage both the motherboard and the CPU. Prevention is always better than cure I believe.
-
I think the only difference with P55A is it has USB3.0 and SATA 6gb/s. I'm hoping its just a precaution changing socket clips to apease the masses. (ie: sense of security)
-
I don't know about the P55A - they could be a 'premium' line of P55 motherboards, instead of a new revision of the current P55 range. (...)
I think the only difference with P55A is it has USB3.0 and SATA 6gb/s. I'm hoping its just a precaution changing socket clips to apease the masses. (ie: sense of security)
Hi
Here is some info about P55A mb's, they also have more power in USB (3 times more):
http://forum.giga-byte.co.uk/index.php/topic,763.0.html
-
I did have a look at the specs on the P55A - not something that would suit me so I'm happy with what I have for now.
I'm more than happy to wait for onboard gfx in processors to come out and air before I worry about upgrading again and get USB3.0 etc.