Deep Talk About Bent CPU Pins

[JBRONCFAN]

Might recognize my handle since I've been running around here with a RAM allocation problem that was solved by discovering my CPU had two bent pins. Noticed something troubling. Seems to be an awful lot of people with bent CPU pins. I mean, been running around pc forums now for about 15 years and I can't remember ever seeing so many bent CPU pin threads. Made me do some extra thinking and......

I believe that this "bent CPU pin" epidemic may be attributed to the heavy use of aftermarket coolers. CPUs only needed a small heat sink and small fan to handle heat. Last 5 years or so have seen the advent of some pretty powerful CPUs which of course means a lot more heat. Whole new market of CPU coolers is created. Only problem is no one has seen the long term effect of these large coolers strapped to our MOBOs and wrenched onto our CPUs. CPU coolers' screws are spring loaded which are meant to keep a constant pressure. Add to that YEARS of 100+ dF temps and bending pins does not sound so crazy.

After having my CPU out and in my hands I learned how fragile those pins are. Looks like they are made of copper which is a very soft metal.

I had to build a new rig cause after a hard drive seemingly died that system wouldn't power back up. Now I wish I'd had pulled the CPU to inspect it.

I understand since I needed to read posts about bent CPU pins I found a ton of threads about bent CPU pins but still, it still seems like there is WAY too many people out there with bent CPU pin problems.

 

[huvik]

i dont think your logic works. its not the CPU pins holding the cpu onto the board. so no matter how heavy your cooler is, its not putting more stress on the pins. on the motherboard, yes. not the pins tho.

I think the answer is just people not being careful during install.

 

[mcnumpty23]

i would put it down to more people than before trying to build their own pcs

 

[TJ Hooker]

Are you sure you aren't just noticing more threads about bent pins because you're searching for the solution to a problem caused by bent pins?

 

[k1114]

Bent pins are a rare occurrence and usually an issue with improper handling of cpus or mobos, nothing to do with heatsinks. Maybe see one thread once a month compared to the thousands of other common issues so it's far from an epidemic. I don't think tdp and aftermarket coolers are the issue if you go back 10+ years starting with the ghz war and see the tdp on cpus has always stayed 80-90ish for higher end ones. Compare the threads you found with the cpus' tdp and what cooler they have. If you are lucky, you may see if they are new to building.

 

[JBRONCFAN]

"Are you sure you aren't just noticing more threads about bent pins because you're searching for the solution to a problem caused by bent pins?"

The reason why I posted this - "I understand since I needed to read posts about bent CPU pins I found a ton of threads about bent CPU pins but still, it still seems like there is WAY too many people out there with bent CPU pin problems."

Ah, who knows. Maybe the pins were bent before I installed the CPU......?? It took a jewelers loop for me to discover the two bent pins I found. Maybe I should just get in the habit of doing that, inspecting the CPU.

 

[Rogue Leader]

Pins are just so tightly spaced that its an easy thing to do if you aren't VERY precise when you're installing your CPU. For people who have never done it before I can see how this can happen. Especially if you try to install it the wrong way first. Then you may slightly bend a pin and not even notice it. That said, once the processor is secured, the cooler or whatever else you do to it will have no effect on it as the pins are well secured within the socket.

 

[synphul]

It could be a mixture of newer cpu's having more pins (whether the pins are on the processor like amd or the socket for intel). That along with people who aren't the most careful with things. Not everyone is mechanically inclined and some folks who have a hard time getting a cd/dvd centered in a drive tray decide they can slap their own system together.

I won't say everything is user error but after seeing many of the posts all over the internet that I have, not much surprises me anymore. You ask anyone and the first words out of their mouth is 'of course I was careful, it couldn't have been me'.

Then look at all the multiple stories of people with thermal paste oozing out all over the motherboard, spilling into the socket, thermal paste all over the pins UNDER the cpu. Or people who scratch their pcb's with tools they either dropped or a screwdriver that slips off of a screw, gouging their boards. Breaking things off (and of course capacitors and things mysteriously 'fall off' on their own). People who talk about cracking the substrate of their new cpu after putting it in a vise and chiseling at it with a hammer.

Evidence would suggest there's a lot of user error going on here. There are way too many problems period and I'd venture to say while some may be manufacturing issues, consider all the thousands of successful builds as well. It's just as likely some folks shouldn't even be allowed to peel the security sticker off their cpu box.

 

[Ecky]

Intel's sockets don't allow the CPU to be pulled out when removing the heatsink. There is a bracket that holds the chip in place. I can't imagine how a cooler would bend the pins of a CPU that's locked into a socket.

Additionally, every generation since Nehalem has gotten cooler - CPU power usage is on the decline. Intel's Skylake i7 CPUs have a rated TDP of only 65w. As such, the stock cooler included has also gotten smaller over the years as power consumption has declined.

Perhaps you're thinking of AMD's recent CPUs? They haven't released a new 8-core since 2012, and so have resorted to adding massive voltage so as to ramp up the clocks in order to (still fail to) compete with Intel's more recent CPUs in raw performance?

 

[k1114]

The heatsink is exactly the same size for many years now although they've had different copper or aluminium cores. I don't remember exactly when but I know that is still the same size as nehalem i7 and I know it was used on core 2 duo as well. They had bigger versions on a couple series including core 2 quad and nehalem for 1366. I wouldn't say intel got cooler since they switched to tim instead of soldering the ihs on some cpus. Also tdp is not declining every gen and actually been on the rise since ib. Nh 95w, sb 95w, ib 77w, hw 88w, the i7 6700k is 91w as is the i5 6600k. Still tdp remains in the same general 80-90ish area.

But he does have amd which is pga and not lga like intel and that's easier to bend pins on. We still get threads with bent pins in the mobo socket for intel. It probably isn't all user error. I've seen some cheap mobos with poor quality over the years too. They should be tested and a bent pin should show up in good qc. Even with thinner pcbs, the backplate and mounting bracket gets most of the weight and the cpu is secured in the socket. A lot of pins would be bending at the same time if it was a pressure from a heatsink.

 

[JBRONCFAN]

Well if I am talking about bent pins ON the CPU then clearly I am referring to AMD. Do a search yourself and see the many different threads started on this. Could be just more people are building their own rigs than in the past. It just blew me away the amount of threads that are out there.

All that extra electricity being sent to the newer CPUs can also contribute to this problem as was mentioned previously.

 

[Rogue Leader]

Well if I am talking about bent pins ON the CPU then clearly I am referring to AMD. Do a search yourself and see the many different threads started on this. Could be just more people are building their own rigs than in the past. It just blew me away the amount of threads that are out there.

All that extra electricity being sent to the newer CPUs can also contribute to this problem as was mentioned previously.

Extra electricity has nothing to do with bending CPU pins. Nothing at all and that is not what was said. Now advancements in cpus giving them more pins than ever of course make it easier to bend one, as they get thinner and tighter. However it all comes down to the installation, putting the CPU in place properly and exactly. 1 wrong move can bend it, and sometimes just enough to cause a problem but not enough to notice it. Its extremely delicate.

The only thing that needs to change is with these delicate pin setups more emphasis should be put in guides as well as instruction manuals how important it is to get it right the first time, and that if you do make a mistake you should be thoroughly examining the processor before trying to install it again.

 

[Ecky]

The heatsink is exactly the same size for many years now although they've had different copper or aluminium cores. I don't remember exactly when but I know that is still the same size as nehalem i7 and I know it was used on core 2 duo as well. They had bigger versions on a couple series including core 2 quad and nehalem for 1366. I wouldn't say intel got cooler since they switched to tim instead of soldering the ihs on some cpus. Also tdp is not declining every gen and actually been on the rise since ib. Nh 95w, sb 95w, ib 77w, hw 88w, the i7 6700k is 91w as is the i5 6600k. Still tdp remains in the same general 80-90ish area.

But he does have amd which is pga and not lga like intel and that's easier to bend pins on. We still get threads with bent pins in the mobo socket for intel. It probably isn't all user error. I've seen some cheap mobos with poor quality over the years too. They should be tested and a bent pin should show up in good qc. Even with thinner pcbs, the backplate and mounting bracket gets most of the weight and the cpu is secured in the socket. A lot of pins would be bending at the same time if it was a pressure from a heatsink.

Depends on if we're talking about K series or non-K.

i5 750 - 95w
i5 2500 - 95w
i5 3570 - 77w
i5 4670 - 84w
i5 5775c - 65w
i5 6500 - 65w

So, not a linear drop, but there's a trend. I'm getting off topic though.

 

[JBRONCFAN]

You don't think heat generated by all that extra electricity plus the constant pressure of a large cooler onto very soft....very thin copper pins would make a difference????

 

[Rogue Leader]

You don't think heat generated by all that extra electricity plus the constant pressure of a large cooler onto very soft....very thin copper pins would make a difference????

There are 938-942 pins in an AM3+ cpu, its tightly set into its socket with negligible clearance, and the cooler is braced by a mount attached to the socket, and compressed into place. As well copper's melting point is 1,085 deg C, meaning the 70 deg C your CPU generates barely even registers to it. There is no chance in the world that those CPU pins are bearing 1 bit of the load from that cooler.

 

[JBRONCFAN]

DUDE.....melting point???? No one is talking "melting point". Christ I can bend those pins with minimal effort. I just think a little bit of heat and constantly pushing on the CPU thru the cooler is enough to facilitate a bend in pin/pins.

 

[Rogue Leader]

DUDE.....melting point???? No one is talking "melting point". Christ I can bend those pins with minimal effort. I just think a little bit of heat and constantly pushing on the CPU thru the cooler is enough to facilitate a bend in pin/pins.

And I'm telling you the cooler is braced to the socket not the processor, its compressing the processor, and even if it wasn't the processor is tightly held in the socket completely flat against the deck of the socket. And my point about the melting point is that the amount of heat which is minimal in relation has no effect on the pins.

There is not a chance in the world it is bending the pins in the slightest bit, and especially when the processor is completely installed and secured in place where even if the pins somehow bent after the fact their contact with the base of the socket has already been made.

 

[JBRONCFAN]

Don't know if you got a chance to review my thread about this in the Memory section of this forum. Before I discovered the bent pins I decided to lightly unscrew the screws holding the 212 cooler. After doing this the performance of my pc got better. 3 times I went into my pc and lightly unscrewed each screw and each time the pc performance got better. In fact it was in doing this that made me realize something was up with my CPU.

 

[Rogue Leader]

Don't know if you got a chance to review my thread about this in the Memory section of this forum. Before I discovered the bent pins I decided to lightly unscrew the screws holding the 212 cooler. After doing this the performance of my pc got better. 3 times I went into my pc and lightly unscrewed each screw and each time the pc performance got better. In fact it was in doing this that made me realize something was up with my CPU.

I was one of the people regularly replying to you there.

And that was a case of overtightening the cooler. In theory that can bend the pins (except that it would bend ALL of them not just 2), but there are other things this can do as well and what you probably did was crush the motherboard by tightening it too much, it could have actually bent the board, pulled stuff apart, warped the base of the heatsink, or damaged the metal die top of the processor. Again this isn't the fault of larger coolers, this was specifically caused by the installation (which can happen with a stock cooler too). One other thing to keep in mind if you did bend the pins doing this, they would stay bent and not retract back, so even backing it off wouldn't have an effect which leads me to believe other damage may have been caused, luckily now that the pressure has been released full contact is being made on the socket of your board and its not permanent (hopefully).

Cooler should be just snugged down. And again this is not caused by the weight of the cooler, but instead it being pressed mechanically into place more than it ever should be.

 

[turkey3_scratch]

Well if I am talking about bent pins ON the CPU then clearly I am referring to AMD. Do a search yourself and see the many different threads started on this. Could be just more people are building their own rigs than in the past. It just blew me away the amount of threads that are out there.

All that extra electricity being sent to the newer CPUs can also contribute to this problem as was mentioned previously.

"Electricity" does not exist. It is unclear what you are saying here. There is no such thing as electricity because there is no agreement on what electricity is. Are you talking about electromagnetic energy? Are you talking about electric charge? Are you talking about voltage?

There are so many branches grouped under "electricity" and so many inconsistencies on what electricity is that all in all it has no absolute definition but is a subjective term. I'm assuming you are talking about charge, which moves from voltage. Moving charge is a current. How is charge moving through a pin going to bend the pin?

 

[JBRONCFAN]

I guess I meant the voltage going to the CPU. Sorry I thought it was obvious.

Hey Rogue...oh yeah, I remember and appreciate your assistance in that thread.

If you remember I said I <mod edit> up the first CPU (in my defense I was celebrating the birth of my new rig with a few cocktails) by rough housing my 212 cooler. When the second CPU came in, I babied that and the screws for the cooler. Taking my time and only turning each screw 3 times going in the classic "X" pattern and still bent two screws.

The 212 cooler I am using is about 5 years old and the springs supporting each screw LOOK five years old. Springs are made using very thing metal and my hypothesis is the springs are not doing their job exactly like they used to. Their job is to maintain a proper and consistent pressure and if they fail or are failing then the pressure is expressed onto the CPU (or at least that's what it seems like).

Cooler Master doesn't have any data on the long term usage of these oversized coolers. I should have gotten a brand new cooler but for now, I do NOT have the screws for the cooler tightened all the way. They are at least half way, maybe a tad more than half way tightened.

<Let's watch the language in these forums>

 

[TJ Hooker]

"After doing this the performance of my pc got better. 3 times I went into my pc and lightly unscrewed each screw and each time the pc performance got better."

In what way did it get better? Did your temperatures improve? Or do you mean you ran a benchmark each time, and simply loosening the screws caused your score to increase?

 

[Rogue Leader]

I guess I meant the voltage going to the CPU. Sorry I thought it was obvious.

Hey Rogue...oh yeah, I remember and appreciate your assistance in that thread.

If you remember I said I <mod edit> up the first CPU (in my defense I was celebrating the birth of my new rig with a few cocktails) by rough housing my 212 cooler. When the second CPU came in, I babied that and the screws for the cooler. Taking my time and only turning each screw 3 times going in the classic "X" pattern and still bent two screws.

The 212 cooler I am using is about 5 years old and the springs supporting each screw LOOK five years old. Springs are made using very thing metal and my hypothesis is the springs are not doing their job exactly like they used to. Their job is to maintain a proper and consistent pressure and if they fail or are failing then the pressure is expressed onto the CPU (or at least that's what it seems like).

Cooler Master doesn't have any data on the long term usage of these oversized coolers. I should have gotten a brand new cooler but for now, I do NOT have the screws for the cooler tightened all the way. They are at least half way, maybe a tad more than half way tightened.

But even if the screws were loose or putting more pressure to one side than another, the most pressure it would exert would be the tightness of the screws and the weight of it. Thats not going to bend a pin that is securely in the socket.

I'm sorry but there is no way this theory holds water. An improperly installed cooler can cause many things, but other than tightening it a lot AND twisting really hard it at the same time (a nearly impossible scenario), bending the pins is not one of those things.

 

[Barty1884]

If you remember I said I <mod edit> up the first CPU (in my defense I was celebrating the birth of my new rig with a few cocktails) by rough housing my 212 cooler. When the second CPU came in, I babied that and the screws for the cooler. Taking my time and only turning each screw 3 times going in the classic "X" pattern and still bent two screws.

From what you're saying there, you babied the CPU cooler......not the CPU, which is where the problem lies.
Over-tightening a cooler cannot bend pins..........if the CPU is seated correctly.
Sounds to me the CPU wasn't seated correctly and then, by tightening the cooler, it magnified the bent pin issue.

It's been said already, that over-tightening a CPU cooler can cause issues, but related to the socket/motherboard opposed to the CPU itself. Once the CPU is seated correctly, the pins do not have any pressure applied to them.

The electricity argument is mere speculation. Rogue was pretty clear there, 1,085'C melting point so the 70'C temps of the CPU would give it no concerns. If you're talking charge (as Turkey mentioned), the simple passing of a charge will have a minimal (if any) impact in temperatures, nowhere close to impacting the structural integrity of the copper pins.

The pins are thin, and you can physically bend pins, absolutely. But that does not lead to a the conclusion that electricity could do the same thing

 

[JBRONCFAN]

From my post: "When the second CPU came in, I babied that and the screws for the cooler."

I babied THAT (the CPU) AND the screws for the cooler. Clearly I inferred I babied both.