Noctua fan does not spin after reboot

[Braden Fontaine]

I have two NF-F12 iPPC 3000 PWM fans on my radiator, connected to a splitter and then connected to my CPU fan header on my motherboard. One of the two fans does not kick in after a reboot and I do not know why. I have tried different splitters and eliminated that as the cause, I hooked the two noctua fans up to a fan controller that has a half speed and full speed preset. At half speed after a reboot, the fans will not spin up, but once switched to full speed and then half speed they will spin. Is this how these fans are supposed to work? With just one fan connected to the CPU header it will spin normally, but not with two.
Motherboard is the Gigabyte 990FX-UD3

 

[pebbleberries]

It's to do with power distribution. running 3 fans on one header can take a lot of power. Try running it on a different header or off the power supply.

When a computer starts up there is very little load on the CPU so it s not sending much power out of the CPU header so when you have 3 fans on the header, there is not enough power to spin them all.

Fans will only spin when there is a certain amount of power going to them so that's why they will not spin and only 2 do.

 

[JackNaylorPE]

No. power is not your problem ... no way / no how. Your symptoms are simply not indicative of a power issue. You're drawing 0.6 sustained maximum amps (fan is rated for 0.3 max and 2 x 0.3 = 0.6) and the header is rated for a minimum sustained amperage of at least 1 amp. Given the above, have to veto the above answer as a solution.

In my personal rig, I have 16 fans on 3 headers using properly wired fan PCBs attached to 3 different MoBo headers..
6 on 420 rad
4 on 280 rad
6 case fans

a) You don't have 3 fans only 2
b) The 0.6 amp is no issue for your MoBo Header and certainly no issue for any fan controller.

MoBo Headers are typically rated at 1 amp, 2 amps on some high end boards. So even if you did have 3 fans ... 3 x 0.3 amps = 0.9 amps ... you are good.

Ya don't see 3000 rpm fans much these days as that kind of fan was typically associated with high fpi radiators of yesteryear. At 3000 rpm, your minimum speed will be around 600 - 750 rpm which is quite high.

Splitters are a poor solution and, in my experience with same , are not a wise choice for situations like yours. In all likelihood, the splitter isn't wired correcty with regard to the return signal ... just 1 reading should be sent back to the MoBo header. I also find the wire clutter unattractive.

The other issue which much be considered is the power / current draw. moBo headers are limited to 1 amp or 12 watts ... but you are OK there with just 2 fans and 0.6 amps. I recommend leaving a little leeway there proportional to fan speed for inrush and sustained current .. at 3k rpm, 50% is more than comfy if you wanna be "super-safe" accounting for a low budget MoBo, and 0.6 x 1.5 = 0.9 amps .... so again OK. I have run up to 8 fans (1.12 amps) on a single header w/o issue and I don't doubt you could run 3 ... but rather be on the safe side. In any case, if wired properly, they would either all run or the header would blow.

http://noctua.at/en/nf-f12-industrialppc-3000-pwm/specification

Max. Input Power = 3,6 W
Max. Input Current = 0,3 A

I would suggest a Fan Hub. if it's just the 2 fans, would suggest a 2-4 connector PWM hub from ModMyToys

http://www.frozencpu.com/products/23519/ele-1299/4-Pin_PWM_Power_Distribution_PCB_4x_Way_Block_MMT-PCB-4P-44P.html?tl=c121s424b214
http://www.performance-pcs.com/modmytoys-4-pin-pwm-power-distribution-pcb-4-way-block.html

Though this has the potential to connect 4 fans, with those power suckers, not knowing the MoBo or other components, I'd limit it to just 2

 

[Paperdoc]

I agree with JackNaylor PE above - his suggestions are a thorough solution. But I'll suggest four other ideas.

1. Why are your radiator fans connected to the CPU_FAN header? That depends on exactly what liquid cooler system you have, of course. Many (but not all) I have seen control their own fans when you plug them into sockets on the pump unit. Maybe yours does it differently?

2. What you describe for the one fan is typical of a fan that is being fed a too-low voltage at start-up. This MAY be a problem only with the start-up power conditions, but it may also be a symptom of a fan with bearings wearing out that requires an unusually high start-up voltage. So consider checking the fan's condition.

3. Anyway, to consider the start-up, here is what NORMALLY would happen. The fans you have are PWM. At start-up the mobo header normally would send the fans a set of signals to start up at full power and speed. After a couple of seconds when systems are stable, it gets a reading of the CPU temperature and re-specifies to the fans what speed they should slow down to. This means they always should be able to start with full power. Now, it is interesting that, when you tested the fans with a different controller system (using Voltage Control Mode, I bet, not PWM Mode), neither could start at half voltage, and both could start at full voltage. So MAYBE in your normal set-up the voltage being supplied is low. Why? Well, first off, let's check for one error many make. Did you install in your fan supply connections one of those "Low Noise Adapters" that come with fans, to make them quieter? If you did, REMOVE IT NOW! Those things drop the voltage to the fans so they run quieter and slower, and do much less cooling. BUT they should never be used in an automatic fan speed control system - that system will do the slow-and-quiet job for you MUCH better anyway, PLUS, the Adapters can drop the fan voltages so low they fail to start up, as you are experiencing. In fact, that effect is much worse when you use a splitter, and Noctua specifically says NOT to use the Adapter with a splitter system.

4. Last suggestion is this. The mobo CPU_FAN header can be configured in BIOS (manual p. 53 under "CPU Smart Fan Mode") to "Auto", Voltage" or "PWM". Make sure it is set to PWM; if you change this, make sure to SAVE and EXIT.

 

[JackNaylorPE]

1. CLCs have multiple methods of connection. If the water block connected to the rad is on the CPU, then the appropriate starting point would be the CPU Header. That is often connected to a PCB mounted on the CLC which in turn sends separate signals to the pump and fan(s). The EK Predator is a good example.

If your unit is not so equipped, ya can USE CPU_1 for the pump (don't forget power connector) and CPU_2 for the fans.

4. By default it is set to auto ... which automatically detects what type of fan it is connected to. This setting should work just fine. And don't forget to install the fan control utility that is on the MoBo CD as BIOS options for speed control are somewhat limited.

It's also important to note that power saving modes (whether basic Intel / BIOS settings or software utility) generally don't engage until Windows is loaded.

 

[Braden Fontaine]

This solved my problem

https://www.amazon.com/gp/product/B00VNW556I/ref=oh_aui_detailpage_o03_s00?ie=UTF8&psc=1

 

[JackNaylorPE]

Glad it worked. I think you will also find that hub's are a lot "neater" than splitters, reducing clutter

The ModMyToys ones linked above do the same job for about 1/3 to 1/2 the price ... the Swiftech is one I have also used ... but, again, if you don't have 8 fans, less expensive options available.

http://www.swiftech.com/8-WayPWMsplitter-sata.aspx
http://www.newegg.com/Product/Product.aspx?Item=9SIAAXT4V60523&cm_re=swifteck_pwm-_-2MV-0035-00001-_-Product

Since the best answer was not the one that worked for you, I undid the best answer selection so as future forum users won't zero in on inadequate power as the likely solution.