Custom loop worse than AIO

[harry_x_ren]

To experienced water cooling folks:

I have a EKWB custom loop kit, EK-KIT X240, and also an Arctic Freezer 240 All-In-One liquid cooler. I compared them with the same CPU (i7 7700K, no-overclocking) using the same thermal compound. I was doing CPU only cooling.

Testing with Prime 95, small FFT option, the EK-kit reached 81 C while the Arctic Freezer AIO stayed at 72 C.

I am very puzzled by this result. The only advantage that the Arctic Freezer AIO has is the number of fans: 4 vs EK's 2 fans. But I hear that the number of fans do not make much difference. On the other hand, the EK kit has a thicker radiator (60 mm vs 38 mm of Arctic Freezer, both are 240mm radiators), and a stronger pump, so it should outperform the AIO.

I would like to find out why the temperature difference is so large, or if there is anything wrong with my EK kit, before I further invest in custom water cooling.

 

[rubix_1011]

Push/pull is almost always better than push or pull only, so this information you heard was incorrect.

Radiator thickness does help, but it also depends on what fans are used. Thick radiators will often need the benefit of higher static pressure fans or fans in push/pull. Also, this could also depend on how well thermal paste is applied and how well the CPU block is seated. Plus, some thermal compounds also need thermal loading/cooling cycles before they are fully cured or spread.

 

[_Sensa]

the thicker radiator is good and bad. If you have the air power, the thicker rad should yield a lower operating tempature. But you do need ample airflow. if your case does not have great air input and output points (think of chassis's like the NZXT H440 where inbound airflow is very limited). In this case, you're fans may not be up to par and with the larger rad, you will need a little more airflow across the fins versus a smaller radiator. The EK on it's own is a great kit so I believe you made the right choice equipment-wise. To me it feels like you're missing some additional airflow across the rad to be as optimal as with your AIO kit.

hope this helps...

 

[harry_x_ren]

Thank you guys for the comments. I am only testing them on an open bench, but airflow could still be the main factor that makes the EK kit underperform, if the pump and the cpu block are functioning normally. I am using the two Vardar F3-120 fans that come with the kit. Maybe they are not strong enough to cool the radiator. What do you think would be the best way to improve performance? Change to stronger fans? Add another radiator? I would like to overclock to 5Ghz one day, but the current temperature is too high for that.

On the other hand, the Arctic Freezer 240 is an excellent AIO, very likely the best in its class. I may have to accept that this AIO is much better optimized than the EK kit in its default setting.

 

[rubix_1011]

Are you running the fans on the EK radiator in push or in pull?

Pull is usually all around better for most fan speeds on most radiators, if only by a small margin.

 

[_Sensa]

Thank you guys for the comments. I am only testing them on an open bench, but airflow could still be the main factor that makes the EK kit underperform, if the pump and the cpu block are functioning normally. I am using the two Vardar F3-120 fans that come with the kit. Maybe they are not strong enough to cool the radiator. What do you think would be the best way to improve performance? Change to stronger fans? Add another radiator? I would like to overclock to 5Ghz one day, but the current temperature is too high for that.

On the other hand, the Arctic Freezer 240 is an excellent AIO, very likely the best in its class. I may have to accept that this AIO is much better optimized than the EK kit in its default setting.

You may be right around the efficiency differences between EK and the Arctic Freeze. Honestly the only AIO I ever used before making my custom loop was the Corsair H100i at the time. I'll say I was not really impressed especially after I started to dabble in overclocking. Push/Pull is going to offer more positive airflow across the radiator but I'm not sure in the case of the AIO if that cost is really impactful. if we are doing a custom loop, then push/pull is probably the best-practice. Another thing I just thought is for both systems, take note of the pump speed being shown between AIO builds. I have seen (corsair in my case) that the pump was running a little slow. In the fan header, I made sure it was cranked up all the way and that helped a lot. This to me is the problem I have with AIO is troubleshooting. Custom loop we check for pump speed, clear radiator throughput, tight bends on the tubing, stuff like that. I'm not sure if this helpful be it may spark some other troubleshooting questions to ask.

 

[harry_x_ren]

The fans of the EK kit are in the push position now.

At one point, I even removed the pump's PWM connector to the chasis fan plug on the mother board, so the pump was powered by the PSU only and it ran at full speed. No difference in temperature.

I am thinking to check the flow rate and see if the pump is working properly. Any idea how to do that?