[SOLVED] Arctic Liquid Freezer II 360 temps worse than on Noctua NH-D15, and increasing slowly in Idle

[Pandemiya]

Hi,
I've been using this AIO since yesterday. It's a pretty cool AIO, but there's one thing that concerns me about it.

The AIO cools my Ryzen 3900X connected to my Crosshair VIII Hero WiFi motherboard, in the chassis I have 3 140mm fans connected at the front for good airflow (the fan on the back of the case I had to dismantle because it hooked into the AIO) AIO I have connected on top of the chassis. Freezer works in factory push mode. The thermall paste I use is Noctua NT-H2, which is one of the best on the market. PC case is SilentBase 801 from BeQuiet.

And what happens?

Well, after turning on the computer (first time after installing this AIO) my temps in Idle were - 35.5-36 degrees Celsius, although these temperatures in Idle slowly, steadily increase. After a few minutes the constant temperature in Idle was 37 degrees, then 38 degrees and so it reached 40-41 degrees in Idle. I mean here the lower possible temperature in Idle, I'm skipping the voltage and clock jumps which is normal for Ryzens. Today starting temperature in Idle is always 39 degrees, slowly raising to 41.

I have the cooler fans set to 100% in bios.

Is that normal for this AIO? I thought the temperature in Idle will be constant, but you can see that everything is gently warming up, so it can't reach the initial low temperatures in Idle.

Is this normal behavior for an AIO? Or could be some problem with pump or something?

Before I bought this cooling, I had an NH-D15 that kept better temperatures. I thought this AIO, which is so praised, would cool better, i hoped for 31-32 in Idle. Ambient temp is 28-29

EDIT:
Im also using Windows power Plan from Computerbase and i have negative offset for CPU - 0.1V, without that temps would be even higher

 

[Phaaze88]

Did you max the pump speed(recommended)?

Why'd you even switch? Oh right - some benchmark, which likely had completely different test conditions too...

Don't make a big deal out of idle thermals - it's the load thermals that matter.

 

[Pandemiya]

Did you max the pump speed(recommended)?

Why'd you even switch? Oh right - some benchmark, which likely had completely different test conditions too...

Don't make a big deal out of idle thermals - it's the load thermals that matter.

I switched from NH-D15 because i was curious if 360mm AIO will do better temps. I like to have temps as low as possible But i dont want to make custom loop.

About pump speed - arctic is not sharing any software for setting things in this aio, even there is only one 4-pin cable to connect with motherboard. I just set in bios CPU FAN and AIO PUMP for 100%

 

[Phaaze88]

About pump speed - arctic is not sharing any software for setting things in this aio, even there is only one 4-pin cable to connect with motherboard. I just set in bios CPU FAN and AIO PUMP for 100%

I have the Celsius S36, and it's the same. It doesn't need any stupid software to run it, so that's why I got it. I don't get dynamic control, but that's not a loss to me.
I have it plugged into the AIO_PUMP header and have it set to 100%.

I switched from NH-D15 because i was curious if 360mm AIO will do better temps. I like to have temps as low as possible But i dont want to make custom loop.

You do know there's diminishing returns with that, right? There's also other factors involved here, and I will cover some of them:

1)3900X
Not a particularly power hungry cpu, considering it's a 12 core/24 thread part. Even in thread heavy workloads, it manages to consume less than 200w of power. You're not going to see that with a 12 core Intel equivalent.
Neither cooler is going to have a hard time keeping that cpu cool, but...

2)The 2 coolers have strengths and weaknesses
NH-D15
-air doesn't transfer heat as well as liquid does, so it's likely to incur higher and more frequent temp spikes.
-it's 165mm tall, and made even taller with ram kits greater than 31mm in height = possible chassis compatibility issues
-while a great high TDP cooler, there's limits to what todays air coolers can reasonably handle, to the point that they just can't compete: ultra high TDP cpus that require a 360mm+ liquid cooler or custom loop

Arctic Liquid Freezer 2 360
-it's a 360mm liquid cooler = possible chassis compatibility issues
-the base plate size - the copper fins on the side you can't see
-pump speeds are low
-the radiator presents another obstacle for airflow

3)Silent Base 801
The design of this chassis is going to affect what/which fan and cooler setups work best.

4)The temperature of the air entering the 2 coolers. No, not just the room temperature - the air in the chassis is always warmer.

Anyway, what I THINK is going on here:
Your Silent Base 801 is too restrictive for the Liquid Freezer 2 here. Liquid coolers are just as dependent on good airflow - if not more so, than air coolers.
The Liquid Freezer 2's fans either aren't getting enough fresh air, or they can't get the warm air out fast enough - I'm thinking the latter...
So they're 'choking' a little bit, and the liquid leaving the radiator and returning to the cpu block didn't get cooled off as much.

I mean, I'm looking at images of that top panel... and wow, how's the exhaust from the LF 2 supposed to get out of there effectively?
It MIGHT work better mounted in the front with the rad fans as pull, but the gpu will run a little warmer though.
The NH-D15 isn't suffering from this issue though: it's drawing air from the front and some from the top and exhausting out the back with less issue.

Add the gpu into the equation - IF it's an axial fan model, because they dump most of their waste heat inside the chassis - and this happens:
-both coolers are absorbing some of the gpu's waste heat, but the LF 2 is having a harder time getting rid of it, on top of the cpu's own heat.

Sorry, I'm a bit on the slow side.

 

[rubix_1011]

Also, your idle temps might vary day to day depending on ambient room temperature.

Yesterday it might have been much cooler than today, which impacts the idle temps.

The Arctic Liquid Freezer does not have a software suite for temp management, pump curves, etc. All BIOS settings, which you should be familiar with if you previously had a Noctua...just means adjusting curves....do what you did before.

There are SO MANY items that go into the performance of both air and liquid cooling that people get so hung up over the actual cooler itself. The best coolers in world can still overheat if there is very poor airflow within a case or if they are setup wrong.

 

[Pandemiya]

I have the Celsius S36, and it's the same. It doesn't need any stupid software to run it, so that's why I got it. I don't get dynamic control, but that's not a loss to me.
I have it plugged into the AIO_PUMP header and have it set to 100%.


You do know there's diminishing returns with that, right? There's also other factors involved here, and I will cover some of them:

1)3900X
Not a particularly power hungry cpu, considering it's a 12 core/24 thread part. Even in thread heavy workloads, it manages to consume less than 200w of power. You're not going to see that with a 12 core Intel equivalent.
Neither cooler is going to have a hard time keeping that cpu cool, but...

2)The 2 coolers have strengths and weaknesses
NH-D15
-air doesn't transfer heat as well as liquid does, so it's likely to incur higher and more frequent temp spikes.
-it's 165mm tall, and made even taller with ram kits greater than 31mm in height = possible chassis compatibility issues
-while a great high TDP cooler, there's limits to what todays air coolers can reasonably handle, to the point that they just can't compete: ultra high TDP cpus that require a 360mm+ liquid cooler or custom loop

Arctic Liquid Freezer 2 360
-it's a 360mm liquid cooler = possible chassis compatibility issues
-the base plate size - the copper fins on the side you can't see
-pump speeds are low
-the radiator presents another obstacle for airflow

3)Silent Base 801
The design of this chassis is going to affect what/which fan and cooler setups work best.

4)The temperature of the air entering the 2 coolers. No, not just the room temperature - the air in the chassis is always warmer.

Anyway, what I THINK is going on here:
Your Silent Base 801 is too restrictive for the Liquid Freezer 2 here. Liquid coolers are just as dependent on good airflow - if not more so, than air coolers.
The Liquid Freezer 2's fans either aren't getting enough fresh air, or they can't get the warm air out fast enough - I'm thinking the latter...
So they're 'choking' a little bit, and the liquid leaving the radiator and returning to the cpu block didn't get cooled off as much.

I mean, I'm looking at images of that top panel... and wow, how's the exhaust from the LF 2 supposed to get out of there effectively?
It MIGHT work better mounted in the front with the rad fans as pull, but the gpu will run a little warmer though.
The NH-D15 isn't suffering from this issue though: it's drawing air from the front and some from the top and exhausting out the back with less issue.

Add the gpu into the equation - IF it's an axial fan model, because they dump most of their waste heat inside the chassis - and this happens:
-both coolers are absorbing some of the gpu's waste heat, but the LF 2 is having a harder time getting rid of it, on top of the cpu's own heat.

Sorry, I'm a bit on the slow side.

You're probably right, so I guess if the NH-D15 cools better, I'll sell LF2 and buy industrial Noctua fans for the NH-D15 and thus have optimal temperatures.

 

[Phaaze88]

buy industrial Noctua fans for the NH-D15...

Don't do that either. You'll gain nothing from doing so, because you'll need to 'industrialize' all your other fans to get any benefit... then you'll have completely destroyed the reason you bought the Silent Base 801 in the first place: low noise.


Are you even having temperature issues to begin with? Does your 3900X run over 80C under load?
If not, there's nothing that needs 'fixing' here. You've got your low noise PC, don't make a mountain out of a molehill.