[motherboard] ga-h55-usb3_v2.0 bent cpu pin?
Pretty sure that's the case. Socket 1156 with memory controller onboard the cpu. The symptom is one of the dual ddr3 channels does not work. When either of its slots is occupied with a dimm, alone or in combination with another dimm in the other channel, the boot process starts but loops endlessly after 6 or 8 seconds in. System works when one or two dimms are in the other channel, only. I've updated the bios to latest and been through the menu to ensure adequate voltage, standard settings.
Hope the pics upload with enough resolution to see. If not, I can put them on Dropbox. In the second column from the right, about even with the top of the central pin void, is the out-of-place looking offender.
I'm about ready to yard on that puppy with a tweezers or pen knife but am soliciting guidance. My friend was not too happy with how I cut some traces on the Biostar motherboard he bought and built my first PC with when I attempted to open the dimm lock with a kitchen utensil, earning my nickname, "Icepick".
It shouldn't be too difficult to straighten it up, especially if you work under a magnifier.
|reply to berserken |
Definitely looks bent. Try the edge of a credit card... that has been suggested before. I've been fortunate enough to never have to do this.
Those who would give up Essential Liberty to purchase a little Temporary Safety, deserve neither Liberty nor Safety.
|reply to berserken |
Mechanical pencils work great for straightening out bent pins.
Credit card gave me the sense that these pins are like bent-over stamped shapes of a thick metal foil: flexible, springy, delicate. However, it's thicker than the spacing between the pins. Mechanical pencil was a good fit for holding in my right hand, rested on the pcie-x16 connector, magnifying glass in my left hand, to push on the bent pin from outside right toward the center as, fortunately, it appeared only a bit torqued in a clockwise direction.
^Before and after:
Feeling lucky, I enabled XMP and an Extreme memory performance pre-set and it booted right up. From not working @ ddr3-1333 to dual-channel xmp @ ddr3-1600, so far, so good.
Thanks for the tips!
|reply to berserken |
I know the feeling! I had 2 or 3 bent pins on an 1155 Asus Z68 MB I had to straighten out.
This was after Newegg returned my RMA as "Damaged" and I literally had no other choice but to attempt fixing the problem myself or scrapping the board.
My fault for not checking the condition of the socket BEFORE installing the CPU and cooler and documenting it.
I believe the pins were "bent" somehow somewhere before I got the board? (This was NOT my first LGA MB!) No way to prove prior damage, and Asus was NOT going to take my word for it.
At any rate, I have built a Z77 system since, and you can be sure I checked and double checked the socket for ANY abnormalities.
I even went as far as purchasing from a brick and mortar CompUSA for a bit higher cost, since they guaranteed me I could physically return the MB if there was anything wrong with the board within 30 days.
The socket and build went fine, but I did note however that Intel (Asus) actually changed their CPU insertion procedure.
On the Z68 1155 MB they had you first remove the protective cover from the socket, then lift the socket arm up and out of the way and insert the CPU, and then bring down the arm to lock and secure the CPU.
On my Z77 1155 MB, they had you lift the arm WITH THE COVER PLATE STILL ATTACHED, then install the CPU. After the chip was located in the socket properly, bring down the arm with the cover plate attached, presumably helping in the securing?
And finally remove the protective cover from the arm AFTER the chip in the socket had been seated and secured!
I believe Asus (Intel) was somehow responding to the ever increasing numbers of "Bent Pins" and never admitted there might be a problem in their procedure? Either removing the cover plate prematurely,
or having the cover plate assist in keeping the chip from moving while securing? These are my speculations, but it`s funny that this cover plate removal procedure seems to have been widely accepted now.
The only thing I might mention to you from those nice (difficult to make) pictures you submitted, is that it appears there is quite a bit of dust/debris on the board, or at least near the socket.
I would take care to blow away any dust/debris BEFORE pulling the cooler and CPU to ensure no dust/debris has any chance to enter the exposed socket!
Again, great feeling rescuing a MB/system back from the "dead"!
This was a freecycle score, with the owner saying FedEx knocked the HSF fan off in transit, gave him a settlement with which he bought a different system. He could boot but Windows "did not recognize the cpu". I thought it sounded like there was some life left in it. With an i5 661 included, 4G G.SKILL Ripjaws he didn't need due to the upgrade generously thrown in, it should make for a decent gift to family or friends, maybe a project bench rig.
Mountain View, CA
|reply to berserken |
Glad to hear you fixed it -- had I seen the thread earlier I would given my full agreement with others' that the socket pin was bent.
But I figured I'd go one step further and tell you exactly what pin was bent / what its purpose is/was. The CPU model you have is an Intel Core i5 661 (thank you for the memtest86 shot).
From the photos, it looks like it's the 12th pin down from the top, in the 2nd column of pins from the right.
If that's correct, then according to the Intel Core i5-600, i3-500 Desktop Processor Series and Intel Pentium Processor G6950 Series Datasheet, Volume 1 that would be pin AH2, which has the pin label SA_DQ. Per the same specification, SA_DQ is one of the many DDR3 RAM data lines:
-- Data Bus: Channel A data signal interface to the SDRAM data bus.
So yes, this pin being bent would cause exactly what you experienced. :-)
It took me a while to work out the orientation of the socket compared to what's in the specification PDF, but thankfully the photo you took matched up with the PDF perfectly ("upper-right" area of socket).
Why looking up pin role is useful/important (and sometimes fun) -- story time! Many years ago I sent a Core 2 Quad CPU to a friend of mine in Sweden. While he was building his system, friends of his came over, and in the process bumped his workbench. The CPU flew up and off the table and landed on the floor -- with two pins bent. One pin he was able to fix by bending it back into position, but the other broke off. He wasn't sure whether or not to use the CPU (my inclination is always to RMA such things, but in this case the CPU was used and had suffered the equivalent of "customer error", so Intel wasn't going to RMA it). We read Intel's technical PDFs and it turns out the pin was one of many GND pins. He tried it and the CPU worked reliably (and still works reliably to this day). He got lucky -- real lucky.
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.
Very informative back-story, thanks!