Positive or Negative Pressure for an SFX Custom Loop?

[rahulr1]

I understand that in an air cooled SFX system, negative pressure is often optimal to encourage intake from GPU fans and to exhaust hot air from the system. 

 

In a liquid cooled setup would the passive heat dissipation from the water blocks mean negative pressure is still better? Or would the fans be more ideal as positive pressure intakes? 

 

For context I'm waiting on my SSUPD Meshlicious to arrive and I'm using a 280mm front mounted rad for an RTX 3080 and 5600X. There are no other fans in this system. 

[Dujith]

As long as there is good airflow it just wont matter. In the end you just want air to flow through the radiator.

What u need to prevent is air getting stuck in your system, ie bad airflow. How you achieve that wont really matter.

[For Science!]

4 minutes ago, rahulr1 said:

I understand that in an air cooled SFX system, negative pressure is often optimal to encourage intake from GPU fans and to exhaust hot air from the system. 

 

In a liquid cooled setup would the passive heat dissipation from the water blocks mean negative pressure is still better? Or would the fans be more ideal as positive pressure intakes? 

 

For context I'm waiting on my SSUPD Meshlicious to arrive and I'm using a 280mm front mounted rad for an RTX 3080 and 5600X. There are no other fans in this system. 

In my opinion, in an SFF PC (and compact non-SFF PCs too), where the heated air goes is the larger consideration.  Take my Ncase M1 system for example, it is essentially a very negative pressure system, both 240 mm radiators are on exhaust, but not because of air pressure per-se but that it is the only way that makes sense. The heated air has to find a way out of the case, usually this is done by a rear exhaust fan in a standard tower, but that path does not exist in SFF PC. If I did a bottom intake on this case, it would literally just re-heat the internals and just warm the underside of the GPU. Similarly the hot air of the top radiator would just blow onto my motherboard components (which isn't necessarily a bad thing in this case) but I have a rear intake fan to cool the board.

 

So in SFF PCs, my opinion is that unless there is a compelling alternative reason, radiator heat should be directly exhausted out of the chassis, otherwise the time and energy spent pump water/energy from the CPU block to close to the outside is wasted if you then just blow the heat back into the case.

 

[Levent]

Depends on how free flowing your case is, mine for example benefits from negative air pressure. I also found out that, smaller the case is more exhaust fans you need. Experiment for a day or two, it really doesnt take long to figure out what is better. I would at least keep one intake fan if you decide go with negative pressure setup.

[rahulr1]

42 minutes ago, For Science! said:

In my opinion, in an SFF PC (and compact non-SFF PCs too), where the heated air goes is the larger consideration.  Take my Ncase M1 system for example, it is essentially a very negative pressure system, both 240 mm radiators are on exhaust, but not because of air pressure per-se but that it is the only way that makes sense. The heated air has to find a way out of the case, usually this is done by a rear exhaust fan in a standard tower, but that path does not exist in SFF PC. If I did a bottom intake on this case, it would literally just re-heat the internals and just warm the underside of the GPU. Similarly the hot air of the top radiator would just blow onto my motherboard components (which isn't necessarily a bad thing in this case) but I have a rear intake fan to cool the board.

 

So in SFF PCs, my opinion is that unless there is a compelling alternative reason, radiator heat should be directly exhausted out of the chassis, otherwise the time and energy spent pump water/energy from the CPU block to close to the outside is wasted if you then just blow the heat back into the case.

 

 

Thanks a lot! So from what I understand, the goal is to exhaust all the hot air out of the system and avoid recycling heated air. If I don't have any fans near the VRM though or decide to use tempered glass side panels, exhausting the heat from a front rad still be better than recycling the air from the radiator fans? For context this would be the internal layout in my intended build, except obviously I'm using a custom loop and an SFX PSU. 

[For Science!]

1 minute ago, rahulr1 said:

Thanks a lot! So from what I understand, the goal is to exhaust all the hot air out of the system and avoid recycling heated air. If I don't have any fans near the VRM though or decide to use tempered glass side panels, exhausting the heat from a front rad still be better than recycling the air from the radiator fans? For context this would be the internal layout in my intended build, except obviously I'm using a custom loop and an SFX PSU. 

I would think so, One option you should consider is getting a spot fan so that you can have some directed airflow over critical components like VRMs. You can get 92 mm or smaller fans, and mount them on a flexible arm so they point directly at hot spots. This way you have a bit more directed cooling than just relying on passive dissipation.

 

Sorry for the weird photo, but you can see my spot fan pointing at the VRM heatsink

[rahulr1]

45 minutes ago, For Science! said:

I would think so, One option you should consider is getting a spot fan so that you can have some directed airflow over critical components like VRMs. You can get 92 mm or smaller fans, and mount them on a flexible arm so they point directly at hot spots. This way you have a bit more directed cooling than just relying on passive dissipation.

 

Sorry for the weird photo, but you can see my spot fan pointing at the VRM heatsink

Ah gotcha I'll keep that in mind. My board already has a small VRM fan but its probably because it's got weak 50A power stages so maybe a small 92mm will help!