How do I tell if my AIO is working properly?

[Krosis]

Hi all,

 

I recently switched to a NZXT X53 after using air cooling all my life (mostly due to pet hair issues).

I am usually idling between 35-43C on my Ryzen 5 3600 (with PBO). On a Prime 95 small FFTs test for maximum heat, my max CPU temp is 78C.

This would indicate that my AIO is doing it's job properly. However, my issue is that the NZXT CAM software is telling me that the liquid temp in the AIO doesn't go higher than 35C during this stress test.

Also, with a Deepcool Gammax GT air cool, I used to idle around 30C.

How is that an AIO is worse than a $45 air cooler?

 

I have 2 Deepcool 120 intake fans in the front, one exhaust in the back. NZXT X53 240mm rad exhausting up top.

 

I am aware my current setup has basically become negative pressure, as opposed to positive pressure with my air cooler. Could this be enough reason?

[.Apex.]

10 minutes ago, Krosis said:

Hi all,

 

I recently switched to a NZXT X53 after using air cooling all my life (mostly due to pet hair issues).

I am usually idling between 35-43C on my Ryzen 5 3600 (with PBO). On a Prime 95 small FFTs test for maximum heat, my max CPU temp is 78C.

This would indicate that my AIO is doing it's job properly. However, my issue is that the NZXT CAM software is telling me that the liquid temp in the AIO doesn't go higher than 35C during this stress test.

Also, with a Deepcool Gammax GT air cool, I used to idle around 30C.

How is that an AIO is worse than a $45 air cooler?

 

I have 2 Deepcool 120 intake fans in the front, one exhaust in the back. NZXT X53 240mm rad exhausting up top.

 

I am aware my current setup has basically become negative pressure, as opposed to positive pressure with my air cooler. Could this be enough reason?

Assessing idling temps are difficult without measuring the power usage in each scenario, I'd look more into the maximum temperature as that's more consistent to measure and having a loop temperature of 35C sounds about right, it shouldn't be as hot as your CPU as then it would be worthless as there's no temperature difference.

 

I can't tell you if the AIO is working optimally or properly but it's definitely working, if not then you'd thermal throttle instantly.

 

Comparing a 240mm AIO with a single-tower aircooler should be similar in performance, I think..

[Krosis]

2 minutes ago, .Apex. said:

Assessing idling temps are difficult without measuring the power usage in each scenario, I'd look more into the maximum temperature as that's more consistent to measure and having a loop temperature of 35C sounds about right, it shouldn't be as hot as your CPU as then it would be worthless as there's no temperature difference.

 

I can't tell you if the AIO is working optimally or properly but it's definitely working, if not then you'd thermal throttle instantly.

 

Comparing a 240mm AIO with a single-tower aircooler should be similar in performance, I think..

I agree with your comment. Unfortunately, I don't remember if I ever ran a Prime 95 stress test on my air cooler, otherwise I would be able to tell the max temp. I do think it will go over 80 though, which in itself would show that the AIO works properly.

 

I think I am just being a little bit paranoid because I feel like the liquid temp should at least show an increase in temp to note that heat is getting transferred out effectively. But then again, as my fan curve is set up, the fan pump speed and all my system fans would ramp up with CPU temp increase to cool the liquid, which would make sense why it doesn't go up too much.

[.Apex.]

13 minutes ago, Krosis said:

I agree with your comment. Unfortunately, I don't remember if I ever ran a Prime 95 stress test on my air cooler, otherwise I would be able to tell the max temp. I do think it will go over 80 though, which in itself would show that the AIO works properly.

 

I think I am just being a little bit paranoid because I feel like the liquid temp should at least show an increase in temp to note that heat is getting transferred out effectively. But then again, as my fan curve is set up, the fan pump speed and all my system fans would ramp up with CPU temp increase to cool the liquid, which would make sense why it doesn't go up too much.

That loop temp is about average for liquid coolers, and you're right in some sense that the temperature difference between the loop and the CPU would indicate the heat transfer efficiency, but I would expect that from a Zen 2 CPU as the die is much smaller and therefor the heat is concentrated so it doesn't dissipate the heat as well, you might improve that if you used better thermal paste but I doubt it'd be affected by much, this cooling performance is inline with other liquid cooled systems I've seen using a Zen 2 CPU.

 

Though I should mention that Prime95 Small FFTs is extremely unrealistic and you would never reach those temperatures in real world applications, I suggest you use Cinebench instead, and since Small FFTs pushes your CPU to the limits then it might be that once again because the die is smaller the CPU IHS cannot dissipate the heat as it's saturated.

[.Apex.]

1 minute ago, .Apex. said:

Though I should mention that Prime95 Small FFTs is extremely unrealistic and you would never reach those temperatures in real world applications, I suggest you use Cinebench instead, and since Small FFTs pushes your CPU to the limits then it might be that once again because the die is smaller the CPU IHS cannot dissipate the heat as it's saturated.

In my system with an 8700K I cannot for the life of me manage to tame the temperatures that Prime95 Small FFT generates, and it's mostly because my CPU has thermal paste between the die and the IHS, so I stopped caring about reaching that type of standard, the fact that your CPU is under reasonable temperatures at that load is way more than good enough.

[Krosis]

1 hour ago, .Apex. said:

That loop temp is about average for liquid coolers, and you're right in some sense that the temperature difference between the loop and the CPU would indicate the heat transfer efficiency, but I would expect that from a Zen 2 CPU as the die is much smaller and therefor the heat is concentrated so it doesn't dissipate the heat as well, you might improve that if you used better thermal paste but I doubt it'd be affected by much, this cooling performance is inline with other liquid cooled systems I've seen using a Zen 2 CPU.

 

Though I should mention that Prime95 Small FFTs is extremely unrealistic and you would never reach those temperatures in real world applications, I suggest you use Cinebench instead, and since Small FFTs pushes your CPU to the limits then it might be that once again because the die is smaller the CPU IHS cannot dissipate the heat as it's saturated.

I completely agree. I was just running Prime 95 to get the unrealistic max to test my thermal headroom. Cinebench is max 65, which is obviously quite good.

 

I am already using Arctic MX-4, so the only upgrade would be to go liquid metal.

 

I think I am just going to stop worrying about this, and let my AIO do its job.

 

Thanks for your input!