Question Concerned about temps.

[Deleted member 2720853]

Here I am. AGAIN. Worrying about temps instead of enjoying my build. Only built this a year ago and I'm still paranoid.

==

I have a Ryzen 9 3900X, stock operation with no PBO. I recently bought a NH-U12A to replace my NH-D15S (based on Noctua's NSPR rating). The D15S was too big for me to maneuver my hands inside my case, and I have to say I am very satisfied with how the U12A cooler looks and sounds. Very silent at 1350 rpm on both fans. Can actually move my hands around in the case now.

However, I'm a bit concerned about my temps. I get around 82-84C in Prime95 Small FFT with AVX disabled. I get around 85-87C in AIDA64. And in Cinebench, 76-78C. All of these after 5 minutes. Gaming temps are around 65C in GTA 5, 63C in PUBG.

I applied NT-H2 on the CPU using the 5 dots method described in their video. 4 small dots in the corners and a bigger dot in the middle. Tightened the screws until they wouldn't turn anymore, did 4 turns on each screw then went to the other screw, and repeat.

I have a be quiet! Pure Base 500DX case with three Silent Wings 3 140mm PWM fans running at 900 RPM, 2x front intake 1x rear exhaust. No top fans, top is closed.

Should I be concerned or is this normal behavior? Thanks.

 

[Phaaze88]

I believe this is the difference in fans being used + your fan curves.

I get around 82-84C in Prime95 Small FFT with AVX disabled. I get around 85-87C in AIDA64. And in Cinebench, 76-78C. All of these after 5 minutes. Gaming temps are around 65C in GTA 5, 63C in PUBG.

What did the D15S get, and what was/were the fan speeds?


I read through Noctua's NSPR guide, and noticed they don't mention the fan curves they used in their testing, so I can only assume when they did do their testing, it was only at 100%.
Not doing it that way adds many variables and is really complicated.

 

[Deleted member 2720853]

My fan curves are: intakes and exhausts to 50% max speed (~900 RPM), CPU fans to 53% max speed (~1250 RPM). Case fans are set to CPU temp, and only ramp up after 60C on the CPU, otherwise 30% idle.

The D15S had some concerning temps for what it is, about the same but slightly worse by a few degrees. But that's because when I took off the D15S, two corners of the CPU weren't even covered. So this time I did the 5 dot method and hopefully I got it right. But seeing almost the same temps is making me uneasy. The D15S had about 950 max RPM on the fan. Same configuration on the other fans.

I could feel serious airflow in the case with my current configuration and the side panel off, so I doubt my case airflow isn't up to par. I want it to be silent, so I'm willing to sacrifice a few degrees, but this seems a bit off.

 

[Deleted member 2720853]

Well, going by this, it seems the 3900X is just naturally a hot boy:

I have my fan curves set a little bit more relaxed for added silence. So far this cooler is performing better than the D15S which apparently had a bad thermal paste application, and because of that I did the 5 dot method this time around. So I doubt my application or mounting is the problem. I learned all my lessons from not installing my D15S properly the first time around.

My case is a cramped midtower, with a direct front to back airflow, components are really close to the front 140mm fans. Nothing in the way of the fans, aside from the dust filter and front mesh.

It would honestly be easier to worry less about temps on Intel at this point, since Ryzen is so unpredictable. Are these temps good or bad? Who knows.

 

[Phaaze88]

The 3900X isn't exactly a power hog, so you could never really 'soak' the D15S' large heatsink with it.
The U12A should be more effective... at least at idle, but no one really cares about idle, right?

There's also the thing about 140mm fans moving more air than 120mm at a given speed, but fan curves below 100% makes things difficult to determine.
That the 2 aren't much different kind of lines up with their NSPR...

 

[Deleted member 2720853]

The 3900X isn't exactly a power hog, so you could never really 'soak' the D15S' large heatsink with it.
The U12A should be more effective... at least at idle, but no one really cares about idle, right?

There's also the thing about 140mm fans moving more air than 120mm at a given speed, but fan curves below 100% makes things difficult to determine.
That the 2 aren't much different kind of lines up with their NSPR...

I'm just wondering if I should worry or not. So far I'm seeing similar and sometimes improved temps in the games I play. I got this cooler to be able to easily get to my PCIe lock and top NVMe slot, and have a less cramped case, and not to worry about temps, but here I am worrying about temps.

The 1250-1300 RPM speeds I have on my NF-A12x25 fans should be about the same as on my NF-A15's 950 RPM.

 

[Phaaze88]

I'm just wondering if I should worry or not.

I don't believe you should.

I figure the reason for the top panel being closed is noise, right? Leaving it open would allow the air coolers to draw air from there.

The 1250-1300 RPM speeds I have on my NF-A12x25 fans should be about the same as on my NF-A15's 950 RPM.

You know you shouldn't do that, because the curves below 100% aren't linear...

 

[Deleted member 2720853]

You know you shouldn't do that, because the curves below 100% aren't linear...

I didn't base my curves over that don't worry, I just don't like my fans spinning above a certain speed where noise is intrusive.

I don't believe you should.

I figure the reason for the top panel being closed is noise, right? Leaving it open would allow the air coolers to draw air from there.

Exactly, and also because I notice an increased amount of dust getting in the case with the ventilated top as opposed to the closed one. But in such a cramped midtower like this (compared to other midtowers) I don't think streamlined front to back airflow is an issue.

 

[Deleted member 2720853]

Uh, I switched the tops back to the ventilated one and am noticing warm air when I move my hand over my case now. Gonna keep it like this and see if anything improves. Trapping hot air isn't a great idea.

 

[Phaaze88]

If I look at your fans running at 100%, what I could say with certainty is this:
"The Silent Wings 3 fans aren't keeping up with the pace of the NF-A12x25 fans. The A12s are recirculating some of the air they've already pushed through the heatsink back through it."
[If the chassis fans can't at least keep pace with the fans on the heatsink, the cooler chokes in a sense.]
I'd do away with the SW3 fan in the back entirely, because it's just in the way; blowback, backflow... that sort of thing.

This is something I picked up on when testing a Thermalright Silver Arrow IB-E Extreme Rev. B(2x 2500rpm fans) in the H500P Mesh, which had at the time, a 360mm front intake using NF-F12s, and a rear NF-A14.
When the F12s and A14 were changed for IPPC 3000s, things improved - mind the jet though.

The A12x25s have a lot of static pressure and don't lose much of it depending on how much you crank the speed down, I'd imagine. The airflow is probably still pretty strong even at 1300rpm.
I don't think I can say the same for the Silent Wings 3, which has a max of 1000rpm, and the static pressure isn't very high either.


Since you don't run any of the fans at 100%, none of the above I posted applies... or does it?

 

[Deleted member 2720853]

There's the problem with removing the rear case fan: I'm gonna despise looking at my PC, which sits right next to me on my desk.

 

[Phaaze88]

OH.
Sorry!

 

[Deleted member 2720853]

I see what you're saying, these fans are insanely strong based on what my palm told me after putting my hand next to them, and could probably do as an exhaust fan by themselves. But I can't have a fan grill just sitting there like that in my PC.

During loads, the fans are still pushing a lot of air out of the case. I'm probably worrying over nothing here and changing the top may have been the only thing I needed to do to get some of that hot air pocket out.