This build addresses the issue of the Fractal Design Node 202's temperatures while on the vertical stand. It compares temperatures of Node 202 builds with no additional fan; 1 additional fan in the GPU fan in both push and pull; and 1 additional fan in the CPU compartment in both push and pull. It shows how a single fan in the GPU compartment, working to pull hot air out of the case diminishes temperatures of around 8°C; further it demonstrates that an additional fan in the CPU area can be counterproductive. Also, a look is given at the Noctua NH-L9i heatsink, it's compatibility with other standard 92mm fans, and it’s behavior within the Fractal Design’s Node 202.
Fractal Design's Node 202 is a Mini-ITX case with two distinguished compartments: a GPU area, and a CPU area. These are divided by a metal "wall". This "wall" has rather small holes that allow the passage of power cables from the PSU, hosted in the CPU area, to the GPU area; and for the passage of SATA and power cables for the extra hard drive and extra fans respectively. As imaginable, these cable obstruct these holes almost completely, and close this "wall" off completely. As a consequence, air in the GPU compartment cannot easily transition to the CPU compartment or vice-versa. The GPU and CPU areas are considered in this case two separate systems.
In a standard tower case, good temperatures can be simply obtained with a push-pull fan system: a fan pulls fresh air into the case, a fan pulls the air warmed by the CPU and GPU out of the case.
In Fractal Design's Node 202, this thermal solution cannot be adopted. The Node 202 comes without fans. In a vertical setup, the warm air simply flows out via the holes on top of the CPU department, and, supposedly, the GPU fan should autonomously push the air out.
Using this solution proves to be very inefficient. The CPU, with the Intel heatsink, constantly reached temperatures of 78/79°C when the dust filters were not present. With these on, the temperature would reach 81/82°C.
The GPU, with no additional fans in the case, would reach 76°C.
Although not worrying temperatures, an no throttling ever occurred, the noise produced by the fans to cope with the heat was at unacceptable levels, both while gaming and while watching high definitions movies.
Having seen that Fractal Design's Node 202 does not excel in it's pure form, without additional fans that is, it is time to see how the situation can be improved.
The first thing to do is examine the Node 202, where one will quickly notice plenty of "holes" around the case. Air can thus move in and out naturally, but the absence of fans does not help the circulation.
The theory here is that the Node 202 should do fine in pulling air from outside if a fan would be creating a negative pressure within the case. In other words, fans should be pushing air out, rather than pulling air in. Although the most natural theory is that the more cool air one pushes inside the case, the cooler the recorded temperatures will be.
As mentioned, the CPU and GPU areas are separate and independent. Therefore the test have to be carried separately. We'll start with the GPU.
Breath in, breath out
With a single case fan, an Arctic F12 PWM PST, placed on the right side of the case (not directly over the GPU then) The recorded temperatures for push and pull air scenarios were the following:
The GPU is 8°C cooler when the case fan is pulling air out of the case.
No / On (Push) 500rpm / 69°C Yes / On (Push) 500rpm / 73°C -- No / On (Pull) 500rpm /65°C Yes / On (Pull) 500rpm / 68°C -- No / On (Pull) 1000rpm / 63°C Yes / On (Pull) 1000rpm / 66°C
Two things are constant: the dust filters consistently rise the temperatures of 3°C whenever present; and the temperature are up to 6°C lower when the fan is in pull mode compared to push.
63°C/66°C on the GPU while under stress if acceptable, and a certainly an improvement over 69°C/73°C.
1000rpm is the limit for keeping the Arctic F12 silent. That is the limit for this experiment: Temperatures diminished and the system remained quiet.
The CPU area is the most complex problem of the case. This area is warmed up by the CPU, but also from the PSU hosted in the same compartment. The only two fans present in a standard scenario are the CPU heatsink’s fan, and the PSU's intake fan. The PSU's hot air naturally blows outside of the PSU, mounted on top of the Node 202, and does not disperse into the case. Yet, its heat also radiates inside the CPU compartment.
On the vertical stand, heat from the CPU and PSU can only flow through the top and the side meshed side of the case. There is no official slot for a fan other than the CPU’s.
CPU compartment’s unorthodox approach
With the standard Intel i5 heatsink, the CPU would easily reach the 81°C in standard tests. Although not a dangerous temperature, the issue was the noise which would reach unacceptable levels while testing and when when the system was under stress.
With no specific slots for an additional fan, the only element that can be changed to improve the situation is the CPU cooler.
When searching for a heatsink to replace the original Intel i5 cooler, the candidates came down to two: the Cryorig C7 and the Noctua NH-L9i. The Noctua prevailed in the end due to the similar price and the fact that standard 92mm fans could be used to replace the fan should it fail and need to be replaced, also in the case the famous “Noctua-brown” does not please the viewer. With that in mind, a Arctic F9 PWM was ordered along with the Noctua CPU cooler.
Quick facts: The Noctua NH-L9i + 2.5cm Arctic F9 PWM are compatible; as expected. The Noctua NH-L9i + NF-A9x14 PWM + 2.5cm Arctic F9 PWM fan will fit together in the Node 202, stacked on top of each other, perfectly.
Since PCpartpicker does not allow tables to be embedded, here is a quick summary: NH-L9i + NF-A9x14 PWM average temperature after 15mins: 64.9°C - Noctua NH-L9i + Arctic F9 PWM average temperature after 15mins: 66.5°C.
In a non overclocked scenario, the Arctic F9 PWM mounted on top of the Noctua NH-L9i performed similarly as the Noctua fan. Both scenarios were in any case better than the i5 stock cooler, and considerably quieter. Noise levels between the two CPU fans were also comparable, with the noisiest fan being, in both cases, the added Arctic F12 in the GPU compartment.
As mentioned in the quick facts, the two CPU 92mm fans will fit in the Node 202 at the same time, one on top of the other. A quick test was run by offsetting the fans, with the improvised Noctua "case fan" set in pull and in later in push mode. The CPU temperatures did not change while the Noctua was pulling, but in push the temperature did rise by 3°C maximum temperature).
What if you pull it?
As a final test, the both the Arctic and the Noctua fans were set to pull on the Noctua NH-L9i heatsink. The theory here was that this would force the cool air through the bottom of heatsink, and then be pushed outside the Node 202, keeping the temperatures low.
In both cases this was not true: a higher maximum temperature, and a higher average were achieved in both experiments. Although also these performed better than the Intel heatsink, the fans pushing into the heatsink were the clear winners.
The Noctua NF-A9x14 PWM in pull over the Noctua NH-L9i should also be considered an inconclusive result; the fan cannot be screwed tightly to the heatsink with either the screw types provided by Noctua. The results noted here are therefore to be taken as an indication only. Yet, compared to how the Arctic F9 in pull mode behaved, though, the results are similar indicating that in the Node 202 the CPU temperatures are better managed in push mode as mentioned above.
As a final test, the Noctua NH-L9i without any fan reached 90+°C, and throttling, within 1.57 mins of testing.
All the scenarios were run with Aida64 stress test for the CPU, and 3DMark Fire Strike Extreme Stress Test for the GPU.
From the results above the argument can be made that Fractal Design’s Node 202, on the vertical stand, behaves better with the following setup:
To be tested Further tests that could be carried out:
12cm fan on the exterior of the case, on the CPU side, pull and push.
2 fans in the GPU compartment: one pull, one push; both pull; both push.
CPU fan: pull instead of push.
Do RGB fans help temperatures? And if so which colour is best?
Everything in horizontal, rather than vertical.