could i solder my IHS to my CPU cooler?

[Nogghan]

so, this is just an idea and i probably will not be doing this but i am curious how well would it work if i were to solder my IHS to the coldplate of my cooler? just to replace thermal paste was my only idea, still using LM on the die

but would i bust a heatpipe when i soldered it to the coldplate? would it even fit back in the socket? or would i just not be able to use the socket retention mechanism?

 

https://buy.solder.com/3-Solder-Alloys-Ribbon-Kit/P1035_1014/ id use something like this if i were to do it but obviously not $400 worth

[knightslugger]

you would literally be marrying your cooler (whatever that might be) to that CPU, 'til death do the (not?) part. ...for what, another 5*C?

 

that's not to mention the insane amount of problems you'd encounter by (necessarily) removing the CPU bracket from the socket. i have no idea how you'd get the pressure required to adequately regain the pressure needed.

 

might as well just say the MB, CPU and cooler are now a single unremovable "Thing". try reselling THAT...

[Nogghan]

Just now, knightslugger said:

you would literally be marrying your cooler (whatever that might be) to that CPU, 'til death do the (not?) part. ...for what, another 5*C?

 

that's not to mention the insane amount of problems you'd encounter by (necessarily) removing the CPU bracket from the socket. i have no idea how you'd get the pressure required to adequately regain the pressure needed.

 

might as well just say the MB, CPU and cooler are now a single irremovable "Thing". try reselling THAT...

well no id be marrying the cooler to the IHS until i decide to heat it up again and remove it

[knightslugger]

5 minutes ago, Nogghan said:

well no id be marrying the cooler to the IHS until i decide to heat it up again and remove it

are you REALLY going to go through THAT MUCH trouble just to shed a few more degrees?

 

...assuming you can figure out a reliable way to get the CPU to clamp the IHS down to the socket...

[Nogghan]

Just now, knightslugger said:

are you REALLY going to go through THAT MUCH trouble just to shed a few more degrees?

 

...assuming you can figure out a reliable way to get the CPU to clamp down to the socket...

i already said i wasnt doing this but, arent you aware that coolers apply a good amount of force when you screw them down, especially because it goes around the socket the retention bracket only keeps the CPU from being ripped out when you remove a cooler, it also helps keep it in place but AMD cpus dont have one and dont seem to need one, again because it mounts to the motherboard and still clamps down just the same

 

[Jurrunio]

why not just go on-die water cooling, that's as effective as it can be

[Nogghan]

5 hours ago, Jurrunio said:

why not just go on-die water cooling, that's as effective as it can be

well because then id have to buy a loop and set it up, right now im just using a massive AF heatsink that i can run passively 

[Neo-revo]

-nvm-

 

of course you could do  this, if you could mark out exactly where to line up the cpu on the base,  and pray that you dont bend the pins/ w/e trying to get this in place.

 

also assuming you  do this away from your mobo, cause most soldering kits would mess up everything else around the area if you had lots of wasted heat spread.  not to mention you may over heat a tube like your concerned if you heat the base up to much.

 

there are better ways to make this connection more permanent, like the ceramic epoxy that ArcticSilver makes...  i know you think you'd bea ble to take off a soldered CPU,  but TBH i think epoxying it in would be  a bajillion times easier to line up the cpu/base/pins and then you could in theory remove it... although you'd have to remove the retention arm unit or hope you can actuate it with the cpu/heatsink monster in the way.

[the dutch guy]

If you have the tools to do it then please do it. I would love to see the difference.

 

But before you do: lap the IHS flat first. Those things are more crooked then a 7th term politician.

[knightslugger]

3 hours ago, the dutch guy said:

If you have the tools to do it then please do it. I would love to see the difference.

 

But before you do: lap the IHS flat first. Those things are more crooked then a 7th term politician.

lapping doesn't level the IHS.

[suchamoneypit]

I for one, want to see this man go through with this.

[SolarNova]

One would have to ensure the IHS doesnt warp from the heat used to solder the 2 together.

 

other than that i dont see a problem.

 

Installing it onto the CPU would be a little more tricky , you will need to make sure you can still close the retention arm, the soldered cooler may get in the way.

 

In the end though, i dont see a reason to do this unelss your already running a custom loop.

[the dutch guy]

14 hours ago, knightslugger said:

lapping doesn't level the IHS.

getting the top flat is the whole point of lapping it.

[knightslugger]

23 minutes ago, the dutch guy said:

getting the top flat is the whole point of lapping it.

The point of lapping is to remove imperfections in the surface to make it uniform. It is not a squaring process (straighten, eg to make un-crooked), which you are not going to accomplish with your fingers and some sand paper.

[knightslugger]

On 10/12/2018 at 8:52 PM, Nogghan said:

i already said i wasnt doing this but, arent you aware that coolers apply a good amount of force when you screw them down, especially because it goes around the socket the retention bracket only keeps the CPU from being ripped out when you remove a cooler, it also helps keep it in place but AMD cpus dont have one and dont seem to need one, again because it mounts to the motherboard and still clamps down just the same

 

AMD CPUs are PGA, and use Zero-Insertion Force sockets. Coolers often get stuck to the IHS on AMD cpus and will be pulled straight out of the socket, even when the socket clamp is locked down.

 

INTEL CPUs are LGA, and use mechanically secured sockets. Contact between the pads on the CPU and pins in the socket is assured through spring pressure between the socket pins and the socket clamp. When folks delid an intel CPU and go water-on-die cooling, the amount of cooler pressure required to reliably get contact between the pins and pads is SIGNIFICANTLY higher than normal cooler pressure (which is really just to get even spread of TIM across the IHS and to get as thin a layer as possible). Because the cooler mounting was never designed to hold a CPU down into the socket, board warpage is often encountered resulting in many system I/O errors, usually from RAM traces warping.

[Salacious B Crum]

3 minutes ago, knightslugger said:

AMD CPUs are PGA, and use Zero-Insertion Force sockets. Coolers often get stuck to the IHS on AMD cpus and will be pulled straight out of the socket, even when the socket clamp is locked down.

Yep this happened to me very lucky pins did not bend or rip out of the cpu

[the dutch guy]

1 hour ago, knightslugger said:

The point of lapping is to remove imperfections in the surface to make it uniform. It is not a squaring process (straighten, eg to make un-crooked), which you are not going to accomplish with your fingers and some sand paper.

the point is to make as flat as possible, this is not a new process but has been done as long as intel and amd came with IHS chips.

there are plenty of articles describing the process and its advantages.

here is a really old one but still relevant, especially with solderd IHS being the new standard.

http://www.silentpcreview.com/article1366-page1.html

[WoodenMarker]

25 minutes ago, the dutch guy said:

the point is to make as flat as possible, this is not a new process but has been done as long as intel and amd came with IHS chips.

there are plenty of articles describing the process and its advantages.

here is a really old one but still relevant, especially with solderd IHS being the new standard.

http://www.silentpcreview.com/article1366-page1.html

The goal of lapping is to create a specific surface roughness for better contact between the cpu and cooler and not necessarily flatness. https://metalcutting.com/what-is-lapping/ 

A flatter heatsink will prefer a flatter cpu whereas the same might not apply to one that is more convex. 

[knightslugger]

34 minutes ago, the dutch guy said:

the point is to make as flat as possible

i'm curious, what exactly do think "crooked" means?

 

Because it does not mean what you think it means.

[Tristerin]

31 minutes ago, knightslugger said:

i'm curious, what exactly do think "crooked" means?

 

Because it does not mean what you think it means.

PRINCESS BRIDE

 

*zips up so his age isn't showing*

[Neo-revo]

1 hour ago, WoodenMarker said:

The goal of lapping is to create a specific surface roughness for better contact between the cpu and cooler and not necessarily flatness.

couldn't you in theory, lap it with something so fine you get a mirror finish

[Nogghan]

5 minutes ago, Neo-revo said:

couldn't you in theory, lap it with something so fine you get a mirror finish

I have some 7000 grit sandpaper that I use on heatsinks sometimes, it makes them mirrior Esq in 20 -30 mins 

[Neo-revo]

like brushed stainless? but to teh poitn you still get most of a reflection.  

 

have you noticed better dissapation with that grit run?

[Nogghan]

Just now, Neo-revo said:

like brushed stainless? but to teh poitn you still get most of a reflection.  

 

have you noticed better dissapation with that grit run?

well I can show you I have one laying around right now 

 

I dont notice much difference, I just like the way it looks, and it makes thermal pads stick better 

[Neo-revo]

nice,  good to know if i want to buff something to a nice finish