Putting my money where my mouth is

[SpyderTracks]

So on my new rig, I opted for the H150i on my 7800X3D, it's overpowered, but I always go overkill on cooling anyway, and also wanted the 3 fans for aesthetics.

BUT, I'm having an issue with temps, could be for a few reasons, the new coolers have VERY little pre applied paste in a dot format, I'm sure it is enough, it just doesn't look enough and made me wary before fitting it, but that's the same for everyone with these, and I'm sure there would be widespread complaints if it wasn't enough, so it's likely not related

It could simply be I haven't applied enough torque on the heatsink screws so it's not making good enough contact, that's more likely.

The below is on idle just with Plex server running, but no one connected.

The pump and fans are operating fine, so it's not that, and the coolant temp is fine, so I don't think it's a flow issue. It does suggest the CPU is just building up heat as unable to transfer it to the heatsink.

So I'm guessing it's either the heatsink not suitably screwed down, or the paste. Either way though, I thought I'd take the opportunity to try out the Thermal Grizzly Kryosheet

 

[TonyCarter]

Same CPU and AIO on mine, and the coolant has hit 41ºc in the warmer weather we've had this week (CPU sitting somewhere below 70ºc), and only after playing games (Ghost of Tsushima and Forza Horizon 5) non-stop for 4 or 5 hours.

I can cool it down a bit further with more extreme fan profiles on the AIO, but the fans becomes noticeably noisier over 1500rpm, so prefer to keep them quieter and live with the 41ºc max coolant temp I've seen. With my profile, my amber temp warning starts at 40ºc, and red is at 45ºc, so I think I'm in a relatively safe zone.

With nothing running it sits a lot lower than that though...29-31ºc normally.

Here's mine after 15m of an Aida Extreme stress test, which was still running at the time...

...and here it is 10 minutes after I stopped the stress test...

 

[SpyderTracks]

Same CPU and AIO on mine, and the coolant has hit 41ºc in the warmer weather we've had this week (CPU sitting somewhere below 70ºc), and only after playing games (Ghost of Tsushima and Forza Horizon 5) non-stop for 4 or 5 hours.

I can cool it down a bit further with more extreme fan profiles on the AIO, but the fans becomes noticeably noisier over 1500rpm, so prefer to keep them quieter and live with the 41ºc max coolant temp I've seen. With my profile, my amber temp warning starts at 40ºc, and red is at 45ºc, so I think I'm in a relatively safe zone.

With nothing running it sits a lot lower than that though...29-31ºc normally.

Yeah, the temps I’m seeing currently is what I’d expect under mid load rather than idle so I’d like to get the resting temps down.

 

[SpyderTracks]

Same CPU and AIO on mine, and the coolant has hit 41ºc in the warmer weather we've had this week (CPU sitting somewhere below 70ºc), and only after playing games (Ghost of Tsushima and Forza Horizon 5) non-stop for 4 or 5 hours.

I can cool it down a bit further with more extreme fan profiles on the AIO, but the fans becomes noticeably noisier over 1500rpm, so prefer to keep them quieter and live with the 41ºc max coolant temp I've seen. With my profile, my amber temp warning starts at 40ºc, and red is at 45ºc, so I think I'm in a relatively safe zone.

With nothing running it sits a lot lower than that though...29-31ºc normally.

What idle temps do you see on the cpu, is that the 29-31c you mentioned or is that the liquid temp?

 

[TonyCarter]

What idle temps do you see on the cpu, is that the 29-31c you mentioned or is that the liquid temp?

As per the newly-added screenshots, the 29-31c is the coolant temp.

The CPU hit 71c during the 15m stress test (might have gone higher had I left it longer of course) and settled back to c.47c quite quickly - and the coolant temp reduced to its more normal range after 10 minutes.

My fan speeds are all (except for the fixed speed ones) based off the coolant temp.

 

[SpyderTracks]

As per the newly-added screenshots, the 29-31c is the coolant temp.

The CPU hit 71c during the 15m stress test (might have gone higher had I left it longer of course) and settled back to c.47c quite quickly - and the coolant temp reduced to its more normal range after 10 minutes.

My fan speeds are all (except for the fixed speed ones) based off the coolant temp.

Thanks, gives me a ballpark. I knew the 7800X3D idles a lot higher than previous generations, but think my high 50s can definitely come down a bit.

 

[TonyCarter]

I've still got my Kryosheet that I was going to put on my old AM4 build, but tweaking the fan profiles further solved some of the higher temps I was seeing (nothing serious, but higher that I wanted).

It's a 38x38mm AM4 version, so I guess I can just trim it to 33x33mm if I ever feel the need to fit it!

 

[Martinr36]

Be interesting to see what results you get

 

[SpyderTracks]

Ok, so this was the state of the pre applied paste coverage on the heatsink

And then the transfer on the IHS

Just doesn't look sufficient to me, there's a definite blank hole in the upper middle, and I know you're talking micro level contact, but with a normal paste application there would be a thicker deposit on the IHS, you certainly wouldn't be able to read through to the print.

I do wonder if perhaps my unit had a faulty paste application.

But anyway, cleaned it up with Isopropyl 99% and laid the Kryosheet with some plastic tweasers from my iFixit kit which is the perfect tool for something this fragile

And wow, I do think there was something wrong previously, unquestionably, I didn't even show the system under load as even on about 40% load it was hitting thermal throttling and in the hot weather just serving a few plex clients on very little load it was maxing everything out trying to keep the system in reasonable temps. One night was quite concerning and actually shut it down.

The idle temps are down a good 5 degrees which is a bonus and much more where I'd hope to see it.

But the stress loads are HUGELY down, much much better now, this is only after about 5 minutes, but it gives an idea of where it' heading.

So now it's really only operating within a 15 - 20 degree delta which is quite incredible really!

This case was never a great one for cooling, it's fine, but it was essentially an early prototype model of the eventual O11D, it's more about looks than performance, they put all that engineering work into eventually releasing the O11D, so temps are never going to be incredible.

But actually those stress temps are better than I could have hoped for!!!

I'm seriously impressed with this.

 

[Martinr36]

Ok, so this was the state of the pre applied paste coverage on the heatsink

View attachment 41753

And then the transfer on the IHS

View attachment 41747

Just doesn't look sufficient to me, there's a definite blank hole in the upper middle, and I know you're talking micro level contact, but with a normal paste application there would be a thicker deposit on the IHS, you certainly wouldn't be able to read through to the print.

I do wonder if perhaps my unit had a faulty paste application.

But anyway, cleaned it up with Isopropyl 99% and laid the Kryosheet with some plastic tweasers from my iFixit kit which is the perfect tool for something this fragile

View attachment 41752

And wow, I do think there was something wrong previously, unquestionably, I didn't even show the system under load as even on about 40% load it was hitting thermal throttling and in the hot weather just serving a few plex clients on very little load it was maxing everything out trying to keep the system in reasonable temps. One night was quite concerning and actually shut it down.

The idle temps are down a good 5 degrees which is a bonus and much more where I'd hope to see it.

View attachment 41750

But the stress loads are HUGELY down, much much better now, this is only after about 5 minutes, but it gives an idea of where it' heading.

View attachment 41751

So now it's really only operating within a 15 - 20 degree delta which is quite incredible really!

This case was never a great one for cooling, it's fine, but it was essentially an early prototype model of the eventual O11D, it's more about looks than performance, they put all that engineering work into eventually releasing the O11D, so temps are never going to be incredible.

But actually those stress temps are better than I could have hoped for!!!

I'm seriously impressed with this.

Yeah it sealed to work well on the laptop as well, keep thinking of doing the Berserker

 

[SpyderTracks]

Yeah it sealed to work well on the laptop as well, keep thinking of doing the Berserker

Better than I expected, there's not really a need for it on a desktop, but I'm still surprised just how effective it is, it's still worth it because it really is closing in on Liquid Metal performance wise. Really impressive

 

[polycrac]

I know you only have a sample size of 1 and fixed it with an option PCS don't currently offer, but does this new pattern of paste application give you any pause in terms of recommending the Corsair coolers? Most on the forum have defaulted to saying 'the pre-applied paste is good and likely avoids the risk of errors in putting on the Arctic paste, so save a few quid there'. Is that advice a bit less valid for some of the coolers or am I jumping to conclusions?

 

[SpyderTracks]

I know you only have a sample size of 1 and fixed it with an option PCS don't currently offer, but does this new pattern of paste application give you any pause in terms of recommending the Corsair coolers? Most on the forum have defaulted to saying 'the pre-applied paste is good and likely avoids the risk of errors in putting on the Arctic paste, so save a few quid there'. Is that advice a bit less valid for some of the coolers or am I jumping to conclusions?

No, I think mine was an outlier, Corsair coolers are HUGELY popular and the paste application is done like this now as standard on all units.

I think this one was just a dud in that respect, no idea why, it's possible it was down to me, perhaps I didn't apply enough pressure on the mounting screws which in turn due to higher heat build up led to quicker degredation of the paste.

If it was a concern, we'd be seeing loads of reports on Corsair forums as well as here.

It's more than likely a me issue.

 

[Scott]

I got the sheet myself the other day. I'm going to go a mini-step further with it. I've got it for my laptop but I de-lidded it within about half an hour of owning it

I got a far larger sheet than I needed with the hope that I can cover both the CPU & GPU as well as directly onto the die.

That coating does look less than I've ever experienced. The pattern is right though, it spreads really well and I've actually had a cooler off then back on without a replacement without any ill effects (not recommended, I tested it thoroughly after).

My guess in your situation is it probably wasn't tightened down. As much as it doesn't look great, the paste is actually filling in blemishes.... rather than meant to coat the die sink itself. Yours is showing that it must be very flat and quite good overall. You would only really tell if there was a lack of paste with a before application pic

 

[SpyderTracks]

but I de-lidded it within about half an hour of owning it

Ha ha, awesome!

I got a far larger sheet than I needed with the hope that I can cover both the CPU & GPU as well as directly onto the die.

Just remember it's conductive!

My guess in your situation is it probably wasn't tightened down. As much as it doesn't look great, the paste is actually filling in blemishes.... rather than meant to coat the die sink itself. Yours is showing that it must be very flat and quite good overall. You would only really tell if there was a lack of paste with a before application pic

I think you're right, I think it just wasn't enough torque put on the mounting screws.

 

[Scott]

Good shout on the conductive, it is graphene after all

I'll be extra careful with my cut sizes

 

[SpyderTracks]

I didn't appreciate that apparently pure water is actually not conductive, it's the impurities within most water that makes it conductive. But pure H2O on it's own actually only contains neutral Ions and is non-conductive

This was probably a science lesson 101, but I think I purged that part of my memory some time ago

Is Water Really A Good Conductor Of Electricity?

For electricity to travel through a liquid, a movement of charge must take place through the liquid. In tap water, rainwater and seawater, there are countless impurities, such as salt (Na+), calcium (Ca 2+)

www.scienceabc.com

 

[Scott]

I didn't appreciate that apparently pure water is actually not conductive, it's the impurities within most water that makes it conductive. But pure H2O on it's own actually only contains neutral Ions and is non-conductive

This was probably a science lesson 101, but I think I purged that part of my memory some time ago

Is Water Really A Good Conductor Of Electricity?

For electricity to travel through a liquid, a movement of charge must take place through the liquid. In tap water, rainwater and seawater, there are countless impurities, such as salt (Na+), calcium (Ca 2+)

www.scienceabc.com

Yeah, pure water was what was used for the original fish tank builds back in the 90s

I think they use an oil now, helps with longevity but ironically it isn't as good at cooling. I think it just lasts longer. IIRC the biggest issue with pure water is it doesn't stay pure for long when a system is running. Makes sense as you will always get impurities coming out of any mechanical or electrical part.

Guess they were built for show and tell and then dried off.

 

[SpyderTracks]

Yeah, pure water was what was used for the original fish tank builds back in the 90s

I think they use an oil now, helps with longevity but ironically it isn't as good at cooling. I think it just lasts longer. IIRC the biggest issue with pure water is it doesn't stay pure for long when a system is running. Makes sense as you will always get impurities coming out of any mechanical or electrical part.

Guess they were built for show and tell and then dried off.

Oh wow, I had no idea they used water back then

Luke from LTT at the recent trade show showed a new iteration of a submerged cooling PC at Computex this year, Skip to 4:04, it's pretty cool

 

[Paul1964]

.....IIRC the biggest issue with pure water is it doesn't stay pure for long when a system is running. Makes sense as you will always get impurities coming out of any mechanical or electrical part....

Yep, was going to mention this.

If you get a water spill, with any basically clean water, on your system it should be OK if powered off immediately (or ASAP) dried off and left for a couple of days you should be good to go.

Linus from LTT has had such a thing happen in his basement server rack because of some oddly selective galvanic corrosion incident.