file it under 'diseases of the cpu' in 'My Hardware Almanac'. Holy Schnitzels, Batman! I did not know until today that warped cpus was a thing!
See
$4.35 fix for Intel thermal problems from gamer's nexus on youtube. You can get a frame that's meant to address this the gist of it is that the $4.35 version from Ali Express is ok for everything but the most extreme overclocking scenarios but apparently that physical mounting pressure can affect memory performance.
What I don't know is if you can use this to de-warp an already warped cpu if you see above flick at 15:00 mins, 'Mike shows how to install thermaright frame', you can see that these days installing a cpu is becoming much of a skill and I don't know how widespread of a problem it has become but clearly something's been done to deal with it.
Note the process of how he painstakingly torques the screws 1/4 of a turn a time to ensure that the frame descends level and is flush to the socket.
What it also doesn't show is the entire process of removing the old mobo cpu retaining assembly and what cooler brackets are compatible with it, I assume that it doesn't affect the position of the cooler mount mechanism and the cpu is at the same height so any lga 1700 cooler mechanism will work with it the same as it does the original mobo bracket.
So the $300 dollar question is could you use this to slowly torque down your CPU and flatten it without the pcb breaking? Well when I say slowly I mean really slower even slower than Mike does I mean hypothetically just slowly go round with that 90 degree per screw in opposite corners until you feel any resistance and then leave it for 10 mins and then tighten it a bit more with one more turn, let it rest 10 mins, turn it again... slow to give the cpu pcb a chance to adjust rather than suddenly bending it back.
What could go wrong? Well maybe being uneven as it is, the uneven pressure will put the socket pins at risk if the cpu slides around while it's being screwed down. So then the $300 dollar question becomes a $500 dollar question if it winds up borking the socket as well.
So the best thing to do would probably be to TRY and straighten the CPU in a mobo that has already seriously bent or broken pins since the aim is really to just try and flatten it without breaking the pcb.
It probably warped because of uneven mounting pressure AND operational heat, yes?
Since you're screwing it back in cold it's not going to be as flexible as when it's warm.
So my hypothesis would be, it is quite risky to attempt to flatten the cpu.
You Don't Have a mobo which is already borked to test the process with. Could take you all day and it will likely end in failure.
As for bending back pins I think it's a matter of technique. The shaky hands question, well I saw something a while back about just making sure that you have a support to lean on so maybe a couple of books stacked on one side to rest on or some kind of holder or workbench with an elevated edge, since in 3d space most everyone's hands shake and oft as not the experts with 'nerves of steel' do actually employ some kind of support to keep their hand steady. Think I saw it when I was watching a video about suturing. Well don't know if that's enough to bust any myths I do recall hearing it from some med school instructors that's all I'm saying about shaky hands.
Presumably if you can manage to apply a gentle force in the opposite direction to the force that bent the pin(s) they should go back? I'm not sure there are bent pins in this socket though it could just be the light.
Ding on the mobo trace well actually looks like it's the edge of a ripple like maybe just the coating that was applied to the mobo didn't go on right, if you look to the left of the red arrow kindly provided by Kerberos, where the traces curve a bit, could just be the light because of the angle, but it does also look like a ding on the trace.
It may or may not be breaking the circuit - if it is simply because the coating then probably the trace is ok underneath. If it was a ding because something contacted it well I'd consider it a defect. How that could've possibly happened, slipped screwdriver maybe - but then that ripple-riffle pattern makes it look like the coating randomly didn't cover it entirely.
Such a dilemma about whether the cpu could be flatted. Maybe you could try placing the cpu on a flat surface under the thermalright contact frame and placing some weights around the edges the issue here is that putting weight top down on the cpu die won't do anything it has to be on the narrow edges of the frame and then you need to start with light weights again and slowly add weight to give the pcb a chance to adjust while it's cold rather than warm so I don't know if it can be done since you basically need precision.
It might be possible if you had the tools - an already broken mobo socket that you can play with, with no risk, or some kind of square wooden frame you could make to go around the edges of a contact frame so you could weigh it down slowly with something on a flat surface.
If the CPU isn't working because it's warped and not making good contact well just putting in the socket and attaching a cooler is basically taking a similar risk to trying to screw it flat with a contact frame anyway so could it be worse, well maybe could screw up mobo pins if they aren't already.
Trying it an an already broken mobo however, can only make the cpu worse if the pcb cracks while trying to flatten it. The warping must've occurred due to uneven pressure while it was operationally warm .
But you can't warm it back up with the power off. I don't know of anything that would apply a steady heat of 50c while you tried to bend it back I mean it could be possible to make something to do that if you had a bunch of warped cpus and a process that worked like maybe a peltier if you could regulate it to produce only 50c to keep the pcb warm to bending back.. get the idea? So attach a peltier to a metal plate to warm up a flat surface to 50c, put the cpu on it to warm it up to 50c, or 60 or even 70c (who knows?) and then use a contact frame, and a frame to go over the top of the edges of the contact frame and some weights to slowly bend it back - and some sort of vice to ensure it does not move - such a tool does not exist.
If you had the tools it might be possible to de warp it or you could take a risk and try and flatten it in the socket with a contact frame. How it could be warmed up? Well maybe placed on a warm surface but the question is how to keep a surface warm long enough to warm up the components enough, and not so much as to exceed their operational limits.
So it's whether the naysayers... could be right. Basically have to invent a reliable method to de-warp it if it's even possible. People must've thought about it already and developed the contact frame yes?
But this was a preventative measure which you would install before the cpu warped, on day 1 rather than a remedial process.
So first you would've needed to know about the cpu potentially warping issue and then bought the contact frame and installed it on day 1 I don't think it's really meant to remediate it after it happened so you are looking at a question that hasn't been addressed.
I just searched the term 'can you bend back a warped cpu pcb' on youtube and all the returned hits were 'how to fix bent pins' - the search engine retranslated my query into how to fix bent pins so presumably
nobody has produced a documentary on that.
That must mean nobody has invented a tool to de-warp cpus - maybe some think it's a problem, maybe it isn't. The problem is basically simply not knowing whether the materials the cpu are made of can tolerate being flexed back.
You, home user, most often buy one CPU - you don't work with them all day. You work with one cpu for years. So that question is what do manufacturers or in this case what the Manufacturer of the CPU which is of course Intel says and how they the issue of warped cpus- what do they blame? Can you rma the cpu because of this or do they say, well you have a naff mobo mounting mechanism, or an uneven cooler ?
because if they blame the mobo manufacturers, or the cooler manufacturers, or user error installing the cpu, or unevenly mounted cooler ...
haven't seen how prevalent of an issue this might be... haven't heard of any official policies - haven't seen it before today... it's a new one on me...
Only seen that some very experienced folks like gamer's nexus know about it for months.
Hey I just noticed there is no dot over the first i in official - who cares about things like this until it affects them?
--
Facebook
Twitter
Instagram