[SOLVED] Fan speed configuration

Loadhaul

Prominent
Aug 17, 2019
21
0
510
Hi all,

Currently running a 9 fan setup (6 intake and 3 exhaust) on a Corsair 500D SE case with a front mounted 360mm radiator with 6 fans on it. For the past half a year or so I have been assuming the best scenario to run the fans at with dust being a factor is to run my 6 intake fans at a slightly higher speed than the top and rear mounted (3) exhaust fans which run slightly slowly than the intake. My thinking comes from basic physics where a vacuum is created forcing the air out the cracks of the case as the air pressure flows from a point of high vacuum to low vacuum . I'd like to hear a few preferences with this setup, particularly considering the reduction of dust throughout the case as where I live it can get very dusty. I'm wondering if I should be running my exhaust fans faster than the intake.

Thanks in advance.
 
Solution
So you have a good tool to centralize your fan management. I note that your dust prevention plan is working. As a small side note, you might give thought to how the Commander Pro's systems are deciding on fan speeds while using your custom curves. As I said, the common strategy is to base the speeds of CPU cooling fans (which happen to be all your intake fans) on the CPU's internal temperarture, whereas the case vent fans' control is based on a mobo temp sensor. It is not clear whether that is what you have set up. As I said, though, I realize that placing the two fan groups' speeds under control based on different temp sensors MIGHT cause them not to track each other for air balance purposes, but I actually expect that would not be a...

Paperdoc

Polypheme
Ambassador
Your basic concept is quite correct. I personally prefer to have a slightly positive pressure inside the case so that air flow at cracks is inside-to-outside, specifically to prevent uncontrolled entry of dusty air. Achieving that, though, is difficult because specs never can tell you reality. Even on specs only, six intakes should pull in a lot more air than three exhausts, so my first thought is that setting those fans with a deliberate speed difference may not be needed. However, I also recognize two factors there : you have not specified which fans, so fan count alone does not tell us air flow capacity; and, intake fans on a radiator deliver LESS than their max air flow because of the resistance provided by flow through the rad fins.

You have not commented so I'll add a minor reminder: check the intake area dust filters from time to time and ensure they are kept relatively clean.

I also had this thought - if you are setting your own speeds of the two fan groups, that suggests you are setting fixed fan speeds, and not using the mobo's automatic fan speed control systems. Generally I prefer to have those systems do their job. They manipulate their respective fans' speeds according to the heat generated by your changing workload, which is tracked by temperature sensors in your system, and there are at least TWO such systems. One is guided by the sensor built into the CPU chip, and that is normally used for the CPU_FAN and CPU_OPT headers. The second is based on a sensor in the mobo and used for SYS_FAN or CHA_FAN headers. In your situation, OP, there is a slight "glitch" for the way I would do things. That is that the INTAKE fans all are on the CPU cooling system radiator, so I assume they are being guided by the CPU_FAN header and the internal CPU temp sensor. On the other hand, the EXHAUST fans all likely are connected to SYS_FAN or CHA_FAN headers and guided by the different sensor on the mobo. So, trying to co-ordinate the changing actions of those two fan groups might be tricky theoretically. But there is a helpful factor. CPU heat generation (and intenal temperature) usually correlates well with heat generated elsewhere in your case (and hence the mobo temp sensor reading), so to a reasonable extent you could expect the two automatic fan fan control systems to track each other well.

But all of that is background and predictions. We need facts to really get it tuned properly. So here's a tool to give you a rough but reasonable measure of actual air flow. You need a visible "tracer". For that I use a small smoke source - either a burning cigarette or a smouldering incense stick. Get your system running under normal load. Slowly move the smoke source near any case cracks at a few locations around the outside and watch which diredtion the smoke flows, and how fast. Try to do this under seveeral different workload scenarios if you can. Ideally you want the smoke ALWAYS to flow away from the cracks, and at a modest rate - no need for a really fast stream of air blowing out, since that would indicate more intake air and internal pressure than you really need.

Once you have your info from a smoke tracer study you can decide whether the air flow balance under all operating conditions is acceptable, or whether you need to tweak it. Doing it this way is more effort and takes time, but is based on reality rather than on speculation from specifications.
 

Loadhaul

Prominent
Aug 17, 2019
21
0
510
Your basic concept is quite correct. I personally prefer to have a slightly positive pressure inside the case so that air flow at cracks is inside-to-outside, specifically to prevent uncontrolled entry of dusty air. Achieving that, though, is difficult because specs never can tell you reality. Even on specs only, six intakes should pull in a lot more air than three exhausts, so my first thought is that setting those fans with a deliberate speed difference may not be needed. However, I also recognize two factors there : you have not specified which fans, so fan count alone does not tell us air flow capacity; and, intake fans on a radiator deliver LESS than their max air flow because of the resistance provided by flow through the rad fins.

You have not commented so I'll add a minor reminder: check the intake area dust filters from time to time and ensure they are kept relatively clean.

I also had this thought - if you are setting your own speeds of the two fan groups, that suggests you are setting fixed fan speeds, and not using the mobo's automatic fan speed control systems. Generally I prefer to have those systems do their job. They manipulate their respective fans' speeds according to the heat generated by your changing workload, which is tracked by temperature sensors in your system, and there are at least TWO such systems. One is guided by the sensor built into the CPU chip, and that is normally used for the CPU_FAN and CPU_OPT headers. The second is based on a sensor in the mobo and used for SYS_FAN or CHA_FAN headers. In your situation, OP, there is a slight "glitch" for the way I would do things. That is that the INTAKE fans all are on the CPU cooling system radiator, so I assume they are being guided by the CPU_FAN header and the internal CPU temp sensor. On the other hand, the EXHAUST fans all likely are connected to SYS_FAN or CHA_FAN headers and guided by the different sensor on the mobo. So, trying to co-ordinate the changing actions of those two fan groups might be tricky theoretically. But there is a helpful factor. CPU heat generation (and intenal temperature) usually correlates well with heat generated elsewhere in your case (and hence the mobo temp sensor reading), so to a reasonable extent you could expect the two automatic fan fan control systems to track each other well.

But all of that is background and predictions. We need facts to really get it tuned properly. So here's a tool to give you a rough but reasonable measure of actual air flow. You need a visible "tracer". For that I use a small smoke source - either a burning cigarette or a smouldering incense stick. Get your system running under normal load. Slowly move the smoke source near any case cracks at a few locations around the outside and watch which diredtion the smoke flows, and how fast. Try to do this under seveeral different workload scenarios if you can. Ideally you want the smoke ALWAYS to flow away from the cracks, and at a modest rate - no need for a really fast stream of air blowing out, since that would indicate more intake air and internal pressure than you really need.

Once you have your info from a smoke tracer study you can decide whether the air flow balance under all operating conditions is acceptable, or whether you need to tweak it. Doing it this way is more effort and takes time, but is based on reality rather than on speculation from specifications.
Thanks very much for the detailed reply, I appreciate it.

I should've mentioned before that I'm using a PWM fan hub (Corsair Commander Pro) which I have set a dynamic fan curve for different temperatures, as you said it would've been very tricky to wire these up specifically using the mobo's connections and have full PWM control on each individual fan. The current setup consists of 6 intake 120mm Corsair fans, while the two top mounted exhaust fans are 140mm and the single rear mounted exhaust is a 120mm as well. Currently I'm using the Corsair Commander to adjust my fan curves with the six intakes running approx. 100RPM-200RPM faster than the exhaust fans respectively. It should be mentioned as well that the radiator is between these 6 intake fans (3 on each side of the rad). I actually clean the dust filters out quite regularly, I actually only need to clean out the intake filter. The good thing is that there hasn't been a build up inside the case itself, such as on the GPU.
 

Paperdoc

Polypheme
Ambassador
So you have a good tool to centralize your fan management. I note that your dust prevention plan is working. As a small side note, you might give thought to how the Commander Pro's systems are deciding on fan speeds while using your custom curves. As I said, the common strategy is to base the speeds of CPU cooling fans (which happen to be all your intake fans) on the CPU's internal temperarture, whereas the case vent fans' control is based on a mobo temp sensor. It is not clear whether that is what you have set up. As I said, though, I realize that placing the two fan groups' speeds under control based on different temp sensors MIGHT cause them not to track each other for air balance purposes, but I actually expect that would not be a problem.

My core suggestion, though, is not to make your decisions based on fan speeds. Using actual fan air flow is better, but you have no way to measure them. The next-best idea I can suggest is the smoke tracer study which can give you a good indication for the real balance of air flow forces in and out from the leakage sites due to the air pressure in the case.
 
  • Like
Reactions: Loadhaul
Solution