[SOLVED] Using 2 fan controllers in a build (1 for intake fan group, 1 for exhaust fan group)?

[Wrathburn]

Going to be building a PC for the first time in 10 years, so I'm very out of the loop.

I've laid out my case to have 10 fans (6 intake at the bottom and front, 4 exhaust - 3 top + 1 at the back, all of them being the exact same fan, the Noctua NF-S12A PWM)

I have never worked with things like fan curves and setting your own fan speeds, so this is all new to me.

Does the following fan connection scenario make sense:

• Plug CPU fan into its dedicated CPU Fan header on the motherboard
• I want to control the speeds of the 6 intake fans grouping and 4 exhaust fans grouping separately
• Using the Deepcool FH-10, it can only sense the fan speed of Fan 1 that's plugged into it, so, all fans plugged into this controller will have the same fan speed percentage applied to all of them (i.e. if I set Fan 1 to 60%, all other fans connected to this controller will also run at 60%),
• Connect all 6 intake fans to one controller, and all 4 exhaust intakes to another controller - allowing me to running each grouping at different speeds to achieve my desired levels of thermal management and noise levels.

Does that make sense? Is there a cleaner solution?

 

[Paperdoc]

Your concept is right, but could be simpler. It depends on what hardware you buy.

First, ALL mobos have headers on them for fan power and control, in two main groups. There is always a CPU_FAN header, and there may be similar ones associated with the CPU, like CPU_OPT, AIO_PUMP, etc. Then there is a second group (of 1 or more) of case ventilation fan headers called CHA_FAN or SYS_FAN. There is a temparture sensor built into the CPU chip itself, and that is used to guide the CPU_FAN header actions. There is a separate sensor built into the mobo at a spot the maker considers best for monitoring the mobo temperatures, and that is normally used to guide the CHA_FAN or SYS_FAN headers. (Some mobos allow you to choose to use the internal CPU chip sensor for this, but that is not usually done.) The mobo may also have a few extra temp sensors on particular mobo compenents (e.g, the Voltage Regulators or the Northbridge chip) in case you want to position a fan particularly to concentrate on that item. If you do that, you can choose that special sensor to guide one SYS_FAN header where you plug in that dedicated fan. All of the headers normally have several options for how they control their fan speeds, and the normal default setting is Automatic or Standard, or some such name. This mode really is an automatic TEMPERATURE control system. That is, it manipulates the speed of the fan on that header to keep the TEMPERTURE of that sensor in the proper range. It does not aim for a particular fan speed. It changes the fan speed to whatever it takes to keep the sensor on the temperature target. The other options usually offered are Turbo (full sped, max cooling), Quiet (low speed, noise and cooling) or Custom (you ge to specify the settings of fan speed versus measurerd temp, rather than using the pre-progammed values). For most users, these facilities already included in your mobo are all you need.

A third-party Fan Controller is a module you add to your case to use instead of the mobo facilities. Normally they give you total MANUAL control of the fans the control - there is NO link to the mobo temp sensors. There are a few that include their own temperaure sensors that you must find out how to mount somewhere, and then you must decide what the speed-versus-temp control curve should be for that. But most do not have that much automation, and YOU are the controller who must monitor your system somehow and decide when to change fan speeds.

Most mobo fan headers can supply up to 1.0 A current in total to all fans connected to the the header. You can connect more than one fan to one header by using either a SPLITTER or a HUB. These are differeent, BUT the makers and sellers mix up those terms badly. Based on actual design and performance, here's the difference. A SPLITTER simply connects all its fans in parallel to the header liines, so all the fans get their power solely from the header and the 1.0 A limit applies to the group. A Splitter has one input arm that plugs onto a header, and two or more outputs where you plug in fans, and that's all. This is the only type of device that can be used with 3-pin older fans. A HUB has an extra "arm" that must plug into a PSU output to get the power for all its fans, and it does NOT use power from the mobo, so the 1.0 A limit does not apply. It does distribute to all its fans the PWM signal from the header to control their speeds all the same. Thus a HUB can only be used in a 4-pin fan system (yours are this type) AND with a header that is using the new PWM Mode method of controlling fan speed - most can do that. So for you, OP, with many case fans, ONE hub might do for all of them off one SYS_FAN header IF you wanted them all to be at the same speeds. BUT for your plan, you want two groups whose speeds may differ, so using two HUBS and two mobo headers can do that. BUT to make them operate at different speeds (I don't quite know why?) you would have to change at least one of the headers to using a custom "fan curve" with speed settings different from the other.

If you need more info, post back with questions.

 

[Wrathburn]

Thank you so much for the thorough explanation. This clears up almost everything I had on my mind.

Few more questions/clarifications:

1) I watched a video on Youtube that explained that it's generally accepted that the most ideal setup is either neutral or slightly positive air pressure inside the case. With the fan setup I have proposed, based on the CFM ratings, I will have a net positive +126.54 CFM pressure inside my case. Figured that if I can control both groups separately, I can speed up/slow down the fan groupings to achieve a net CFM rating for my case closer to 0. Maybe I'm looking at this all wrong, so any insight would be much appreciated.​

2) You mentioned that the SYS_FAN header have a 1.0A limit. I was looking over the mobo manual (GIGABYTE Aorus Master X570) and I can't seem to find where it states this limit. Is this simply an industry standard that everyone "just knows"? Or should this be listed somewhere in the mobo manual?​

 

[Paperdoc]

1. I prefer the slightly positive pressure approach, too. Why? Well, if you have positive pressure in the case, air leakage at cracks and mounting holes will be from inside to out, preventing influx of unfiltered external air. NOTE that this only works IF you ensure that your air intakes at the fans do have dust filters installed, AND you take the time to clean them from time to time. Why slightly? Only because much higher intake vs exhaust gives you no further improvement in dust prevention, but it does mean you are using up fan resources and power for them with little benefit. However, there still is SOME benefit in over-doing the intakes, because that will mean slightly higher total air flow though the case. Achieving the "right" balance of intakes vs exhausts starts with caalculations such as you've done already, but that is not the complete answer. That is because there are other factors that you cannot predict, you can only measure when you're done. Among these, the biggest is the various impediments to air flow in the system. A significant one is those dust filters on the intakes. You need them, but they do reduce the air flow from those intake fans by an unknown amount. And of course, when they get partly dirty, they impede air flow more. Other items like interior components in the air flow path are unpredictable. The best you can do is to plan for slight excess of intake over exhaust capacity by fan specs. Then when your system is assembled and running, you can do a simple smoke tracer test that is not quantitative, but a reasonable guide. With the system running you need a small source of smoke, like a cigarette or a smoldering incense stick. Move that around the outside of your case, especially near small gaps at mounting slots, and observe the smoke flow pattern. If it blows out at a moderate pace, you have good outwards flow. If it really blows away fast, you have more than you need. If it is sucked in, you have inflow from negative pressure. If you are not satisfied with the result, you have some info on what change you want to make.

2. You are right, there does not appear to be any spec in the manual. But yes, this is almost an "industry standard" because apparently all mobos use virtually the same chip for their fan header outputs. A common convention is that the 1.0 A limit applies unless you are told otherwise. SOME mobos include one SYS_FAN header with a higher output limit of 3.0 A intended for use with high-capacity pumps for liquid cooling systems. I have even seen a few mobos that have a lower limit of 0.5 A on their fan headers, but that's rare.

I understand why you might want to have separate controls for the two fan groups. As I said, the only real advantage to that is reducing slightly an over-supply of intake air. To do that you would have to arrange different fan performance "curves" of fan speed versus temperature measured. For this, though, you really have only ONE temperature sensor on the mobo. So if you arrange the two groups on separate mobo SYS_FAN headers that use that one sensor and the pre-programmed "Normal" curve, both headers will send out to their fans identical signals. In the manual on p.53 where these items are discussed, it notes you can customize the curve for each header using a tool they provide. If you do that, just be sure that at ALL possible measured temperatures, the intakes DO provide more air flow than the exhausts.

 

[Wrathburn]

Can't thank you enough for the superb explanations. You've cleared everything up for me.

 

[Paperdoc]

You're welcome, and thanks for Best Solution.