Optimus Foundation AMD WB Mounting. Even enough?

V

VirtualMirage

Limp Gawd
Joined
Nov 29, 2011
Messages
470
This may be a silly question, I am just a little OCD.

So last night I have mounted my Optimus Foundation AMD water block to my board and followed their instructions, crisscross and finger tighten as much as possible. I did it in small amounts, which felt like it took forever and may have forgotten the order I was working in but always remained in a crisscross pattern and finger tightened until I could no longer tighten them any more. Paranoid that due to me having a lapse in the crisscross order, I was concerned that maybe it was tightened unevenly. So I decided to try and measure the height of each thumbscrew to see if they are even or not. It isn't an exact science, but hopefully is close enough.

My measurements came down to roughly 31mm from the top of the motherboard PCB to the top of the thumbscrews. When I say roughly, it appears to be less than .5mm +/- of 31mm.

  • Top left looked to be about spot on 31mm or just a hair over (maybe .1-.2mm)
  • Top right was just over 31mm (maybe .3-.4mm)
  • Bottom left was just over 31mm (maybe by .25-.33mm)
  • Bottom right looked to be about spot on 31mm
Again, the measurement wasn't a complete science and even itself could be susceptible to be slightly off to account for the slight measurement differences at each post. I took a stiff piece of paper, stood it up against each post and then used a pencil to mark where the top of the post sits in relation to the paper. I then labeled the marks so I know which post they represent. I then took a ruler to measure the length but my ruler only measures as small as millimeters. But seeing that the measurements were pretty close to each other, it seems like the measurement method was consistent.

So my question to those with experience: Is it good enough to leave it alone or should I try and adjust it further to get them closer to the same measurement?

Being that they are NOT spring tensioned, it's a bit nerve-racking tightening it down as much as possible via just the finger tighten method even though that is what Optimus instructs and ensures is fine. They feel very confident that if you are just using your fingers to tighten and no tools that there is no concern for overtightening (and their posts have no way to used tools on them anyways). I've always been used to spring tensioners and thread stops preventing you from going too tight.
 
Last edited:
When I tighten blocks down I tighten 2 at a time, opposite corners, and crisscross that way. If it makes you feel better, adjust a bit. The springs are designed to not allow damage as long as you hand tighten. Maybe if you are Dwayne Johnson you may need to worry. Otherwise it should be just fine.
 
Smoked Brisket said:
When I tighten blocks down I tighten 2 at a time, opposite corners, and crisscross that way. If it makes you feel better, adjust a bit. The springs are designed to not allow damage as long as you hand tighten. Maybe if you are Dwayne Johnson you may need to worry. Otherwise it should be just fine.
Click to expand...
Haha....I am nowhere near like Dwayne Johnson, no worries there.

Sorry, I just realized there is a typo in my original post. They are NOT spring tensioned is what it should have said, but that Optimus ensures that with their design that it is fine so long as you only do the finger tighten method. Hence why it is nerve-racking.

I did think about the two at a time method, but my installation method wouldn't permit it. Because of my setup, the motherboard was already in the case and the rear mounting bracket on the motherboard will not stay in place if laid down (I tried initially to tape or secure it to the back temporarily but it wouldn't hold from the bracket weight). The screw posts Optimus has you install into the motherboard's rear mounting bracket are thin enough to slide right through the holes that the rear mount rests in. So instead, I had one hand firmly pressing on the mounting block as evenly as possible as I secured the thumbscrews in a crisscross pattern, at least until it reached the initial snug. Thinking about it now, I guess I could have done the two at a time method afterwards since it was now in place, but then the moment passed and I just continued to press forward with the method I started with.
 
Smoked Brisket said:
Do not get me wrong, I think it is great that you want to get it perfect. But you have nothing to worry about, trust the hardware, fire it up, it will be just fine.
Click to expand...
You're probably right. It's just my OCD kicking in and dealing with a mounting method I have not used before. I just needed someone to tell me it'll be fine.

I've done numerous builds in the past and have a nearly 20 year career in the IT world. So working on PCs isn't an issue. It is mainly when doing something different it gives me pause to make sure it is done right the first time. While I have done water cooling in the past (AIO kits), this will be my first custom water cooling loop and, to add to the complexity, I am working in a fairly small space with 2 240mm x 54mm radiators and a EKWB Quantum Kinetic TBE 200 D5 pump/reservoir combo (using a Fractal Design Node 804 mATX case). It'll be impressive when complete, but it will be a pain to take it apart for a silly mistake. So I am taking my time and making sure all my i's are dotted and t's are crossed.
 
If you just want someone to tell you that it's fine and it'll work, then alright.

It's fine and it'll work.

But if you want to know the TRUTH...

Then let me explain what needs to happen for your heatsink to work properly.

All air and gaps between the heatsink and the CPU must be squeezed out and all pits and bumps must be equalized by the paste.

The best way to do this is the pea method, which creates a "shockwave" type motion as it's tightened, pushing out all air from the center, and perhaps even overflowing a bit as the paste itself is completely crushed flat against the heatsink and CPU.

This will ensure that heat transfer is optimal. It does not need to be incredibly tight, but the tighter you can get it, the better.

Applying the pressure slightly unevenly is not going to completely undermine your attempt to install a heatsink, but it is not optimal.
Essentially, you get only one absolutely perfect chance to install the heatsink, and if there is even a tiny mistake, it will not be as perfect as it would be if you had done it perfectly.

OCD has its purpose. Perfection pays off.

But we are only mere mortal humans, so there is a limit to possible perfection.
If you have done your best to ensure optimal installation settings and minimal mistakes, then it should be fine just as it is.

But do not be afraid to remove the heatsink, clean off the thermal paste, and then completely redo it from scratch. OCD has its place, and science and engineering is that place. Don't doubt yourself. Measure twice cut once. Be the ultimate. Even if perfection is relative, always do your best and don't settle for any less.

I believe in you if you believe in you.
 
Aegir said:
If you just want someone to tell you that it's fine and it'll work, then alright.

It's fine and it'll work.

But if you want to know the TRUTH...

Then let me explain what needs to happen for your heatsink to work properly.

All air and gaps between the heatsink and the CPU must be squeezed out and all pits and bumps must be equalized by the paste.

The best way to do this is the pea method, which creates a "shockwave" type motion as it's tightened, pushing out all air from the center, and perhaps even overflowing a bit as the paste itself is completely crushed flat against the heatsink and CPU.

This will ensure that heat transfer is optimal. It does not need to be incredibly tight, but the tighter you can get it, the better.

Applying the pressure slightly unevenly is not going to completely undermine your attempt to install a heatsink, but it is not optimal.
Essentially, you get only one absolutely perfect chance to install the heatsink, and if there is even a tiny mistake, it will not be as perfect as it would be if you had done it perfectly.

OCD has its purpose. Perfection pays off.

But we are only mere mortal humans, so there is a limit to possible perfection.
If you have done your best to ensure optimal installation settings and minimal mistakes, then it should be fine just as it is.

But do not be afraid to remove the heatsink, clean off the thermal paste, and then completely redo it from scratch. OCD has its place, and science and engineering is that place. Don't doubt yourself. Measure twice cut once. Be the ultimate. Even if perfection is relative, always do your best and don't settle for any less.

I believe in you if you believe in you.
Click to expand...
For the thermal paste, I ended up using the method that both Optimus and Thermal Grizzly suggested, spreading it as thin and even as possible across the whole heatsink. I have used a similar method in the past, but on Intel processors, with decent results. I understand that AMD heatspreaders tend to be more convex while Intel's are more concave, so maybe this will make them less susceptible to retaining air pockets when installing the heatsink?

For application, I used the supplied spreader/dispenser that came with the Kryonaut I applied. I applied an even thin coat across the whole heat spreader as thin as possible but still evenly coating the heat spreader. I know the application is too thin if the thermal paste would get picked back up by the spreader. The applicator is pretty nice, the tip/spreader is very soft and flexible (which I wasn't expecting) and allowed for a very easy spread without creating lumps or noticeable air pockets. This thermal paste was certainly one of the easiest to work with versus my past experiences, Gelid Extreme and Arctic Silver 5 being used before. In the past I used to pretreat the heatsink by rubbing in thermal paste on to it then wiping it off and then apply thermal paste to the top of the CPU heat spreader with the method above. But that was also during a time when heatsinks tend to have "rougher" surface". This Optimus water block was like a mirror and their claim is an extremely flat surface with only 0.000025mm of roughness, I didn't dare touch it.
 
VirtualMirage said:
For the thermal paste, I ended up using the method that both Optimus and Thermal Grizzly suggested, spreading it as thin and even as possible across the whole heatsink. I have used a similar method in the past, but on Intel processors, with decent results. I understand that AMD heatspreaders tend to be more convex while Intel's are more concave, so maybe this will make them less susceptible to retaining air pockets when installing the heatsink?

For application, I used the supplied spreader/dispenser that came with the Kryonaut I applied. I applied an even thin coat across the whole heat spreader as thin as possible but still evenly coating the heat spreader. I know the application is too thin if the thermal paste would get picked back up by the spreader. The applicator is pretty nice, the tip/spreader is very soft and flexible (which I wasn't expecting) and allowed for a very easy spread without creating lumps or noticeable air pockets. This thermal paste was certainly one of the easiest to work with versus my past experiences, Gelid Extreme and Arctic Silver 5 being used before. In the past I used to pretreat the heatsink by rubbing in thermal paste on to it then wiping it off and then apply thermal paste to the top of the CPU heat spreader with the method above. But that was also during a time when heatsinks tend to have "rougher" surface". This Optimus water block was like a mirror and their claim is an extremely flat surface with only 0.000025mm of roughness, I didn't dare touch it.
Click to expand...

I don't mind using some 3000 grit sandpaper on CPUs, but a "mirror" on the heatsink? Nice. That sounds wonderful to me.

I still stand by the pea method, but only due to watching some videos of people pressing transparent glass against CPUs, to show the real-world spread of thermal paste.

The spread method seems troublesome to me, but it's a lot harder to mess it up. The pea method, if done wrong, can end up squirting it in just one direction leaving huge amounts of the CPU untouched, which can be preeeetttyyy bad. The best method is the method that you're more comfortable with, especially when you understand the science of it, as well as having lots of practice.
 
Aegir said:
I don't mind using some 3000 grit sandpaper on CPUs, but a "mirror" on the heatsink? Nice. That sounds wonderful to me.

I still stand by the pea method, but only due to watching some videos of people pressing transparent glass against CPUs, to show the real-world spread of thermal paste.

The spread method seems troublesome to me, but it's a lot harder to mess it up. The pea method, if done wrong, can end up squirting it in just one direction leaving huge amounts of the CPU untouched, which can be preeeetttyyy bad. The best method is the method that you're more comfortable with, especially when you understand the science of it, as well as having lots of practice.
Click to expand...
Here's an oldy, but goody:
https://www.pugetsystems.com/labs/articles/Thermal-Paste-Application-Techniques-170/

And here is a more recent one:
https://www.gamersnexus.net/guides/3346-thermal-paste-application-benchmark-too-much-thermal-paste

The thin spread method is more work but seems to be more consistent in both coverage and temps. Same goes for the x method. But the conclusion to both articles across all application methods used is that so long as you have even coverage and applied enough thermal paste (e.g., don't use too little), they all will work pretty well and the temp difference between the different methods is fairly negligible. The pea/rice/line methods tended to not offer full heat spreader coverage when not enough is used, but luckily still covered the areas of the heat spreader that needed it the most.
 
VirtualMirage said:
Here's an oldy, but goody:
https://www.pugetsystems.com/labs/articles/Thermal-Paste-Application-Techniques-170/

And here is a more recent one:
https://www.gamersnexus.net/guides/3346-thermal-paste-application-benchmark-too-much-thermal-paste

The thin spread method is more work but seems to be more consistent in both coverage and temps. Same goes for the x method. But the conclusion to both articles across all application methods used is that so long as you have even coverage and applied enough thermal paste (e.g., don't use too little), they all will work pretty well and the temp difference between the different methods is fairly negligible. The pea/rice/line methods tended to not offer full heat spreader coverage when not enough is used, but luckily still covered the areas of the heat spreader that needed it the most.
Click to expand...

I saw some tests done, and apparently the only thing that would make it go badly is not enough.

Too much is just enough, imo. It's not like thermal paste is that expensive. I guess when I say "pea," I mean a really BIG pea. =p
 
You must log in or register to reply here.
Back
Top