First water cooled pc in two decades

[ShrimpBrime]

I'm going to pop into Starfield and get a screenshot of the cpu temps in game. When on air cooling I used to see an average of 65c.

Do that 30 times

 

[67Elco]

A large improvement for sure...

 

[67Elco]

Since I'm a little OCD about how things look I will take the loop apart and tweak things a bit more to my liking. I'll get some more modern, better performing paste and take care of that at the same time. Being hard lined makes simple maintenance difficult and that should take care of it for a long spell. Many thanks to all for the input!

 

[ShrimpBrime]

Since I'm a little OCD about how things look I will take the loop apart and tweak things a bit more to my liking. I'll get some more modern, better performing paste and take care of that at the same time. Being hard lined makes simple maintenance difficult and that should take care of it for a long spell. Many thanks to all for the input!

Sorry for the late thoughts. But sometimes the don't pop up till after you see something assembled.

You can buy flexible copper piping. Then you would have better ability removing the waterblock. Sorry for the belated idea.

 

[Bobaganoosh]

Just to give some context on the paste, here are some examples. It wasn't 100% controlled, but I took the same CPU, the same paste pattern, and ran the same test with different pastes. I'm going from memory a bit here, so bare with me, but what I saw was the three pastes I gave before performing within 1-2°C of each other. Thermal Grizzly Kryonaut (grey, non-extreme) about 4-5°C worse than those. Corsair XTM-50 performed about 10-12°C worse. I bought the Corsair because I was troubleshooting an issue, did a dozen re-pastes in a weekend and ran to my local Best Buy to get whatever they had so I could keep testing...that's what it was...and I wasn't going to blow up my CPU, but it was garbage paste. This was all on a 12900k running an all-core stress test in the BIOS drawing about 285W. So it isn't going to be a direct relation to your use, but the relative performance would be the important factor.

Other available ones I've heard great things from: Noctua NH-T2 and Arctic MX-5/6. I think most newer pastes are going to perform similarly in your case, but you'll want to make sure you don't use anything that's been sitting around for years and has separated (might have been part of the issue with the Corsair stuff I used, I'm not sure). Years ago, you just bought Arctic Silver 5 and called it a day lol...not anymore.

 

[dgianstefani]

Since I'm a little OCD about how things look I will take the loop apart and tweak things a bit more to my liking. I'll get some more modern, better performing paste and take care of that at the same time. Being hard lined makes simple maintenance difficult and that should take care of it for a long spell. Many thanks to all for the input!

Kingpin KPX, Kryonaut/Conductonaut Extreme, Gelid GC-4, MX-6.

 

[67Elco]

I used this Arctic Alumina 20 years ago when I was active in extreme cooling/overclocking. It just worked for me and my purposes at the time and still does as far as I can tell. The Noctua pastes I have on hand are much newer fare, but I have never had reason to open one and try them. I have both the Noctua NT-H1 and NT-H2...assuming the latter is the latest variant. All I have seen on the Noctua pastes suggest they are very long lasting and likely about on par with most pastes concerning conductivity. Considering just giving the Noctua a go.

Performance comparison NT-H1 vs. NT-H2

Designed in Austria, Noctua's premium cooling components are renowned for their superb quietness, exceptional performance and thoroughgoing quality.

noctua.at

 

[wNotyarD]

I used this Arctic Alumina 20 years ago when I was active in extreme cooling/overclocking. It just worked for me and my purposes at the time and still does as far as I can tell. The Noctua pastes I have on hand are much newer fare, but I have never had reason to open one and try them. I have both the Noctua NT-H1 and NT-H2...assuming the latter is the latest variant. All I have seen on the Noctua pastes suggest they are very long lasting and likely about on par with most pastes concerning conductivity. Considering just giving the Noctua a go.

View attachment 333634

20 years ago? Damn right, that's about when I started learning about some hardware and about the legendary Arctic Silver 5, the former AS-3 and the good for the time Alumina and Céramique. They do still work, granted (especially the AS-5), but there's so much more concentrated heat generated by today CPU's and GPU's that's no wonder those great pastes of yesteryear start faltering.

Give those Noctuas a try. In theory, you should only need the H2.

 

[67Elco]

From what I am reading the H1 might get the nod. Seems the H2 is slightly better, but it is a thinner paste.

 

[wNotyarD]

From what I am reading the H1 might get the nod. Seems the H2 is slightly better, but it is a thinner paste.

Fair. I only looked up some numbers, but haven't read about the paste consistency.

 

[67Elco]

Just considering the long term and not necessarily the absolute best performance.

 

[Bobaganoosh]

...I have both the Noctua NT-H1 and NT-H2...assuming the latter is the latest variant. All I have seen on the Noctua pastes suggest they are very long lasting and likely about on par with most pastes concerning conductivity. Considering just giving the Noctua a go.

Those should be fine, and I'd definitely pick either of those over the alumina.

 

[dgianstefani]

For long term you want this.

Thermal Grizzly KryoSheet Review - Tested on RX 7900 XTX with 475 W

Thermal Grizzly KryoSheet is an interesting alternative to thermal paste. The pre-cut sheet doesn’t dry out over time and you don’t have to try spreading it in a thin layer. In our review we test thermal performance on the AsRock RX 7900 XTX, at up to 475 W.

www.techpowerup.com

 

[67Elco]

I'll wait 5 years for user comments before going that route.

 

[ThrashZone]

Hi,
I'd use mx-6 personally
It's well priced and good performer
NT-H1 is eol to me frankly h2 is okay I had a tough time spreading it and it seemed wet at first and dried way to fast which made it hard to spread evenly so I'd pass on it.
Glid I've always wanted to try but never could remember hehe so do it for me

 

[ThrashZone]

The three fans on the back are intake...reverse blades. The one on the side is conventional exhaust and the three on the radiator are pushing air thru the rad and out the top of the case. The psu is mounted upside down and is exhausting thru the bottom.

Seems most vendors don't list the fpi of their radiators. Found it - The rad I have is 20 fpi and rated for 1200w. Specs lead one to believe it is a pretty decent radiator.

Can't go wider than my current rad (120mm) in this case so it is pointless to consider another radiator.

Hi,
Reversible fans hehe got me again

 

[67Elco]

Working on it...

 

[67Elco]

The original mount must have been solid...seeing no difference in temps thus far. Tubes are nice and straight now. That friggin inlet to the reservoir is still wacky tho...

 

[67Elco]

No change in operation so I guess I can leave it be for awhile. Turns out it was not all that difficult to drain after all.

 

[ThrashZone]

lol yeah draining is the ez part
Working out the air again takes time.

 

[Bobaganoosh]

I just noticed (and love) the scissor jack! That's awesome. We use those on optical benches at work all the time.

No change in operation so I guess I can leave it be for awhile. Turns out it was not all that difficult to drain after all.

Did you try anything different with the mount? Different paste?

Also, it looks like your power went up 2W, score up 30 points, and temps are a little lower across all cores. So not a huge change and could be margin of error, but it does look slightly different comparing to this run (post #137). I'm not sure if all your settings are the same though.

 

[67Elco]

I just noticed (and love) the scissor jack! That's awesome. We use those on optical benches at work all the time.

Did you try anything different with the mount? Different paste?

Also, it looks like your power went up 2W, score up 30 points, and temps are a little lower across all cores. So not a huge change and could be margin of error, but it does look slightly different comparing to this run (post #137). I'm not sure if all your settings are the same though.

Nothing different with the mount other than changing to Noctua NT-H1. I did change the platform thermal limit from 95c to 85c which showed about a 5c drop in temps under load.

 

[ThrashZone]

Hi,
Thinking your mount is a little funky and not making very good contact or equal contact with the chip
Your core temp spread max is pretty different 85c and the coolest is 65.5c what 19.5c difference so not great.



All this water loop stuff got me hooking back up my optimus foundation hehe
Only have a single ek 280pe rad on with gpu and cpu and mx-6
No where near your clocks I'm at stock pretty much and just got rid of the silly intel timers so it only turbos to 4.9 for the most part on CB 2024
But my core temp spread on my 10 core 10900k was only 8c so my contact is pretty darn good.

 

[67Elco]

Hi,
Thinking your mount is a little funky and not making very good contact or equal contact with the chip
Your core temp spread max is pretty different 85c and the coolest is 65.5c what 19.5c difference so not great.

View attachment 333812

All this water loop stuff got me hooking back up my optimus foundation hehe
Only have a single ek 280pe rad on with gpu and cpu and mx-6
No where near your clocks I'm at stock pretty much and just got rid of the silly intel timers so it only turbos to 4.9 for the most part on CB 2024
But my core temp spread on my 10 core 10900k was only 8c so my contact is pretty darn good.

View attachment 333823

You may be right...if that's the case I got it wrong two times in a row as it's reading exactly the same as before. I'll try loosening it up and go back with less pressure to see how that effects it.

 

[ThrashZone]

Hi,
Heatkiller Intel mount is garbage at least the amd mount has studs and cap nuts design now days but you are supposed to tighten until stops.
They still probably don't use enough pressure though
So loosening ? you can try I'd opt to also try a little thicker washers.