brand new 140mm Noctua PWM fan "dying"

tracerit

Senior member
Nov 20, 2007
457
1
81
I bought a 140mm Noctua NF-A15 as my CPU cooler fan and I noticed that sometimes it'll just "die". A reboot would bring it back to life. I then noticed later on that I could also go into Asrock's eXtreme Tuner app and if I change the fan target speed from Level 1 to Level 2 or even keep it at Level 1 and just hit Apply, it'll revive the fan again.

Is an issue like this with my fan or my motherboard? I haven't deteremined when it happens but noticed on two occasions it happens just minutes after I start playing a game and it has to ramp up fan speed and once overnight when i wasn't doing anything at all.

In the mean time i'm going to put back the 120mm fan I had on the cooler and see if it's the fan header plug that's the issue.
 

dma0991

Platinum Member
Mar 17, 2011
2,723
1
0
Every fan has a minimum start up voltage. If the fan header isn't providing that minimum, it will not spin continuously. Try plugging it directly to a molex connector with the LNA or ULNA adapters.
 

BonzaiDuck

Lifer
Jun 30, 2004
16,053
1,681
126
Every fan has a minimum start up voltage. If the fan header isn't providing that minimum, it will not spin continuously. Try plugging it directly to a molex connector with the LNA or ULNA adapters.

Good point.

I am curious as to tracerit's motherboard model -- or -- It would be good to know the number, type [PWM or 3-pin] and motherboard fan amperage spec -- either cumulative total or per port.

I'm not going to look up that Noctua fan specs, so I'll only say that you can control a PWM fan from the CPU_FAN port -- or any other thermally-controllable PWM port -- if you're inclined to spend $10 to get a Swiftech 8W-PWM-SPL-ST, which powers up to eight fans directly from the PSU. But, abjuring a reading of the Noctua spec, I think they only Noctua fans I've ever encountered were 3-pin fans. It may be possible for some motherboards to control a 3-pin fan in the traditional way from a PWM port. The reason I say that: my own board does it. I ran a Panaflo 120x38 "medium" 2,300 RPM/103 CFM 0.40+A 3-pin fan on a PWM CHA_FAN1 port, and the board plus software controlled it perfectly. But that plug is a PWM plug, so any PWM fan would be just as controllable, since control through pulse-width-modulation is the more flexible fan control technology.

If there is a problem with spinning up a Noctua fan, which even Noctua's techs admit is limp ("our customer-base inclines toward a- [*-n anal, nit-picky] standard of noiselessness . . . "), then I'd wonder if the total fan-amperage draw from the motherboard has been exceeded. [*Brackets my own inserted refinement.] On the other hand, settings in BIOS might interfere with the proprietary software, but I've never seen such thing in my ASUS boards, and ASUS and ASRock have common corporate roots. It still might be possible to enter the wrong selections in the BIOS: my ASUS board offers "Turbo," "Standard," and "User" profiles. If you create a profile in the proprietary software, it is saved as a file to be loaded again at startup. But if the BIOS is configured as "turbo" or "standard," it won't read that file, or execute control of the fan by the profile you had saved.
 
Last edited:

tracerit

Senior member
Nov 20, 2007
457
1
81
it's a PWM fan with 4 pins.

I'd prefer not to connect it to the PSU or a temp controller since it wouldn't be able to scan automatically with the CPU temp.

the motherboard is an asrock z77 extreme4. how would i be able to find or monitor the fan amperage for the CPU_Fan header?

i'm testing to see if it's the fan or CPU_Fan header by connecting the 140mm Noctua PWM to another 4-pin Chassis header and connecting a 120mm PWM fan to the CPU_Fan header to see what happens. WIll report if anything comes up
 

BonzaiDuck

Lifer
Jun 30, 2004
16,053
1,681
126
it's a PWM fan with 4 pins.

I'd prefer not to connect it to the PSU or a temp controller since it wouldn't be able to scan automatically with the CPU temp.

the motherboard is an asrock z77 extreme4. how would i be able to find or monitor the fan amperage for the CPU_Fan header?

i'm testing to see if it's the fan or CPU_Fan header by connecting the 140mm Noctua PWM to another 4-pin Chassis header and connecting a 120mm PWM fan to the CPU_Fan header to see what happens. WIll report if anything comes up

But that's just it! Unless ASRock uses some arcane and radically different approach, it WILL scan, scale -- or control automatically with the CPU temp!

Even so, I think you might have a wise approach there. Perhaps you already had an idea as to whether it was a problem of overloaded fan ports, and have dismissed that possibility.

So consider this. If the problem is the power provided to the fan, the splitter would resolve that problem as long as the PWM signal wire and onboard features for it continued to work. If your warranty period isn't expiring, you might observe the reliability of the CPU_FAN PWM signal, but you would have to buy the splitter. This is my opinion: if you can rely on the signal, it's a waste of time to RMA the motherboard; and the splitter would give you expanded fan control for up to eight PWM fans or pumps of different specification -- amperage, size, speed -- because PWM control follows a general duty-cycle "percentage" that is transmitted with the PWM signal. The only drawback: you could only monitor the fan or pump on the first of the eight PWM pinouts. You could, however, run the single monitoring tach wire for some few of those fans to available unused motherboard fan-ports, and I believe it wouldn't matter whether they were 3-pin or PWM 4-pin. But you'd have to find some old fan-wire 3-pin plugs, use a small knife to tease out the yellow wire from the 4-port plug and insert it securely into the appropriate 3-pin plug. Then, connect it to the available port and voila!
 

BonzaiDuck

Lifer
Jun 30, 2004
16,053
1,681
126
Also . . . if you have a multi-meter -- run a continuity test on the two fan power wires -- +12 and ground. If you're using a fan "patch" cable, which usually comes all dressed up in fancy sheathing but can be as unreliable as s***, remove it and do the same.

I bought two of those damn things for a few dollars each, replaced a bad one with a "new" one, and found that I now had two bad ones and an unknown. Live and learn. All the patch cables I've teased and soldered together, in every machine I ever built -- still work. Those other *!@#$*&&!! units have always tended to suck like a Victorian bathtub . . .
 

tracerit

Senior member
Nov 20, 2007
457
1
81
oh wow, that would be great then. i think i'm gonna get the swiftech splitter. would it work with 3pin fans as well? jsut asking because my case fans are three pins. if yes, these three pin fans would spin at 100% though right?
 

BonzaiDuck

Lifer
Jun 30, 2004
16,053
1,681
126
oh wow, that would be great then. i think i'm gonna get the swiftech splitter. would it work with 3pin fans as well? jsut asking because my case fans are three pins. if yes, these three pin fans would spin at 100% though right?

I'm pretty sure the Swiftech and other [[here's a FrozenCPU search page http://www.frozencpu.com/search.htm...oAfh889&searchspec=PWM+splitter&go.x=0&go.y=0
]] varieties would power the three-pin fans, but only at 100% duty-cycle.

Instead, if your ASRock board is similar to the ASUS technology, 4-pin PWM plugs can serve just as well to control 3-pin fans with voltage variation.

This latter assertion is a certainty -- if the ASRock boards . . ASUS . . . etc.

The first point about powering the 3-pinners to 100% was something I'm sure I read somewhere recently. Maybe I'd read a review of these splitters, I can't be sure. But if you look around at the different resellers, their descriptions and their customer reviews, you could verify it.

Everything I've done in the last couple months has focused on these fan-control issues. I purchased two Swiftech splitters and am currently using one of them. Even so, I only have four fans in my case, excluding the dual VGA fans on my GTX 780 graphics adapter. Two are three-pin, because I haven't found any 180 to 200mm fans in the PWM flavor (but grateful if someone has seen any and wants to share).

And the entire objective has been complete fan-control all around, without purchasing a front-panel controller, "smart" controller with USB, bundles of analog and digital sensors and special software -- less likely to read the CPU temperature of the motherboard than not.

So when you design your computer-building project, you have flexibility in addition to picking the case, the size in ATX, mATX or ITX, and the array of fans. The accessory controller and (hopefully) software would then only be a complication for show -- especially if it also came with a front-panel display.

If the motherboard software allows for "fan-curve profiles" and does its job with industrial-strength reliability, then -- so long SpeedFan (but thanks, Alfredo), so long to software that requires the now-defunct MBM motherboard manager or any $80-or-more front-panel de-lights.