Undecided on PWM or 3-pin fans. Which will spin slower?

[tracerit]

I have a Fractal Design Node 304 and would like to get new fans. This case has a fan speed controller for 3-pin fans that will allow 5V/7V/12V settings. When PWM fans are at their lowest RPM setting is it less than a 3-pin fan at 5V?

I'll be looking at the Noctau Redux fans and they have PWM and 3-pin versions.

 

[Soulkeeper]

Every fan will be different, but generally there is a minimum speed that a fan will spin at (both pwm and 3 pin) which is usually in the 600-800 range. Over coming the inertia on start and the runtime resistance prevents super low rpm.
That being said, after using pwm fans I can't see myself ever going back to 3 pin fans.

 

[tracerit]

how would PWM work spread out among three 92mm fans and one140mm fan?

I plan to have two 92mm fans intake, one 92mm CPU cooler, and one 140mm exhaust. I'd like to use a PWM splitter attached to the CPU_FAN header. How would it modulate the 140mm fan, does PWM work on percentages?

 

[Soulkeeper]

Ok, I see your case now. Your situation is a bit different than mine with those 92mm fans.

http://www.frozencpu.com/products/1...WM_Fan_B9_PWM.html?tl=g36c365s940&id=XsgHjYsZ

Those claim to spin down to 300
Yes, you can use a PWM splitter. I soldered one together for my two 120mm pwm fans.
They work on a duty cycle, so kinda like percentages. 0-255 usually, allowing you to control the speed with software based on temps.

This is what my 140mm pwm viper fan does:

My dual 120mm pwm sp120 intakes with a y splitter:

Here is another fan I used to run, but swapped it out because it wouldn't spin slower than ~1100

Notice how the fans in the second pic stop spinning around 600rpm, the result would be the same with non-pwm fans if you try to spin them super slow too. Depending on the particular fan, some won't stop (ie: 0 would be the same as 50 or whatever). I guess with the smaller 92mm fans you can maintain a slower rpm however. Also if you run a pwm fan on a 3 pin header they typically spin at 100% speed and act like a normal fan.
In software I set my fans to spin at the slowest speed where they won't stop when temps are below a certain level. As my system heats up the speeds increase proportionally from there.

For your situation i'd recommend tying the two intake fans together, then giving the cpu and exhaust fans their own header (assuming you have 3 on your mb).

 

[monkeydelmagico]

^ great post soulkeeper

 

[BonzaiDuck]

I seem to be "promoting" this device lately, since it's cheap, reliable, and it "works for me:"

http://www.frozencpu.com/products/2...-_SATA_Power_8W-PWM-SPL-ST.html?tl=g47c17s424

This is especially nice for limited PWM options on a motherboard, assuming that the CPU_FAN port is a PWM port. The fans are controlled by the CPU_FAN PWM signal, but they are powered directly from your PSU.

Soulkeeper does more to address the minimum RPM issue, but my sig-rig has two sets of fans: two PWM fans (120mm and 140mm), and two 200mm 3-pin units. At the minimum level defined by my "fan-curves" in the mobo software, the 120mm spins at 1,500 (top-end = 4,250) and the 140mm spins at 660 (top-end = 1,600). The large 3-pin fans are running as low as 340 RPM.

 

[Soulkeeper]

I seem to be "promoting" this device lately, since it's cheap, reliable, and it "works for me:"

http://www.frozencpu.com/products/2...-_SATA_Power_8W-PWM-SPL-ST.html?tl=g47c17s424

This is especially nice for limited PWM options on a motherboard, assuming that the CPU_FAN port is a PWM port. The fans are controlled by the CPU_FAN PWM signal, but they are powered directly from your PSU.

Soulkeeper does more to address the minimum RPM issue, but my sig-rig has two sets of fans: two PWM fans (120mm and 140mm), and two 200mm 3-pin units. At the minimum level defined by my "fan-curves" in the mobo software, the 120mm spins at 1,500 (top-end = 4,250) and the 140mm spins at 660 (top-end = 1,600). The large 3-pin fans are running as low as 340 RPM.

That looks nice, originally I had soldered together something similar only powered off molex. Oddly enough my asus mb failed to send a pwm signal if it didn't detect a load on the fan header so I had to redo it to draw the power for both fans from the header.

 

[tracerit]

ah, nice post Soulkeeper. I'll hook the two front fans to a fan controller, the CPU fan to the CPU_header and the exhaust to the CASE_header. May hook the CPU and exhaust fan together though, but we'll see.

I ended up getting this splitter a few days ago: http://smile.amazon.com/gp/product/B009D3KQ54/ref=oh_details_o02_s00_i00?ie=UTF8&psc=1

Seems like it does the same thing as the Swiftech one but is only four cables. Hopefully it works just as good, never heard of the brand before.

 

[BonzaiDuck]

That looks nice, originally I had soldered together something similar only powered off molex. Oddly enough my asus mb failed to send a pwm signal if it didn't detect a load on the fan header so I had to redo it to draw the power for both fans from the header.

I wouldn't know how the Swiftech manages to do what it does -- after all -- it's "just a 'splitter' device." But it definitely works off my CPU_FAN PWM header on my ASUS P8Z68-V Pro board. The only drawback: If the PWM ports on the Swiftech are number 1, 2, 3,. . . 8 -- the motherboard only monitors the actual RPM off the fan connected to #1. Of course, you could move the tach wire for another such PWM fan to a motherboard header, but the motherboard likely doesn't have seven more such headers handy.

When I think of the few hundred clams I'd spent on fan-controllers and other schlock over the last few years, or how I abjured using PWM fans for so long in favor of the 3-pin offerings, this little $10 splitter is really "all my dreams come true" -- for thermal control of fan speed and fan noise.

 

[Carson Dyle]

All else being equal (fans of the same diameter and same maximum speed), I always thought non-PWM fans were known to be capable of spinning slower. Maybe that's no longer true.

In my Fractal Design Define Mini, with an ASUS motherboard, I have the original two Fractal Design PWM fans in front and a Scythe PWM fan in the rear, all controlled by 4-pin headers on the ASUS motherboard. Works incredibly well. All of them are approximately 1300 RPM fans. Below about 850 RPM I don't hear any of them, so it's not necessary for me to spin them at their absolute minimum speed.

 

[BonzaiDuck]

All else being equal (fans of the same diameter and same maximum speed), I always thought non-PWM fans were known to be capable of spinning slower. Maybe that's no longer true.

In my Fractal Design Define Mini, with an ASUS motherboard, I have the original two Fractal Design PWM fans in front and a Scythe PWM fan in the rear, all controlled by 4-pin headers on the ASUS motherboard. Works incredibly well. All of them are approximately 1300 RPM fans. Below about 850 RPM I don't hear any of them, so it's not necessary for me to spin them at their absolute minimum speed.

Maybe I said so . . . I have the standard CM 200mm LED fan in front-panel, an NZXT 200mm in side-panel. Those are 3-pin units. Then I have an Akasa Viper 140 PWM in the middle of my NH-D14 pipe-and-fin towers, and a Gentle Typhoon AP-30 PWM at rear exhaust behind the D14.

Currently at idle, the 200mm fans are at 343 and 848 RPM respectively. [The NZXT tops out at 1,300]. the Akasa (idle) is at 680, and the AP-30 at 1577 RPM.

At CPU temps above 65C, they all go to 90% of top-end: 600, 1,200, 1,500 and 3,850 RPM respectively. The first three fans are barely noticeable even then, while the AP-30 can be heard, although the acoustic mods around the AP-30 make the noise level quite acceptable -- mostly muffled white-noise.

I honestly can't hear anything. IMHO, the concerns about how slow you can get one type of fan to spin as opposed to another isn't any earth-shattering or fundamental issue. Because of my "fan-curve" customizations, the reduced airflow at idle has raised idle temperatures by a few degrees -- to ~ 35C. But if you can't hear anything, a few RPM here or there wouldn't matter . .. IMHO . . And neither would idle temps averaging 35C.

Now . . . it's true that my ducting with the AP-30 still shows average core MAXIMUM load temperatures at ~73C (room ambient 77/78F). That's with a 4.7Ghz over-clock. Another member with an i7-2700K says his CPU peaks at 50C @ 4.8Ghz, and he's using a Corsair H110 AiO cooler.

But if the issue has been equally about noise, my system is more quiet than any H110 would show. And that's with an AP-30 running 300 RPM below the full-bore 4,250.

 

[TakBaseTech]

If you ever want the lowest noise possible at all times, consider ditching a PWN setup. Yes, your fan speeds will be fixed, and temps will rise when more performance is used, but so long as temp rise is within acceptable limits to you, this won't matter. You also won't have to hear the noise that comes with the fans increasing speed (akin to mechanical drives when they have to go from idle to seek). This kind of noise is more noticeable and annoying to me than constant noise, which I can more easy block out.

Better fan controllers out there like the latest ones from Scythe will automatically start the fans at full voltage, then immediately lower the speed to virtually whatever you want. This will allow you to run most fans well below their required startup voltage, resulting in a quieter system. My preference is to use one of these controllers and keep my fans and very low, fixed speed.

 

[Carson Dyle]

I've often wondered why there aren't more good fan controller boards with software control and temperature sensing. Why are people still using clunky manual controllers that require drive bays and/or still have manual knobs on them? We should have PCI or PCIe boards with 4-6 fan headers and complete software control - with both PWM and voltage based control, either manual or temperature based. I should be able to click an app from which I can instantly switch all of the computer's fan profiles from something like "performance" to "normal" to "quiet" and anything else I can think up.

 

[TakBaseTech]

Well, computer cases are starting to eliminate 5.25" bays now (finally), so that is one trend that bodes well for your non-bay controllers.

The knob thing has been solved for a while now, though. My Scythe controller has retractable knobs that can be pushed flush with the surface, allowing me to use it with a closed case door. And plenty of bay controllers have digital touch screens these days. Scythe recently released their Kazemaster Flat II, which is is flat AND will do 12V startup with automatic reduction to lower voltages.

Motherboard manufacturers like Asus and their 5-way optimization have really made strides with mobo fan control software/firmware, too, although I haven't checked lately to see if they can do the 12V startup thing.

 

[Carson Dyle]

Well, computer cases are starting to eliminate 5.25" bays now (finally), so that is one trend that bodes well for your non-bay controllers.

Perhaps, but not by itself. Some mount in a slot opening and have knobs on the back of the computer. Not as whiz-bang, of course, because they have no cool display to satisfy the teenage gamer.

The knob thing has been solved for a while now, though. My Scythe controller has retractable knobs that an be pushed flat with the surface, allowing me to use it with a closed case door.

I wasn't talking about the knobs getting in the way, but for manually controlling something in a computer that could (and should) be controlled through software, they seem archaic.

And plenty of bay controllers have digital touch screens these days. Scythe recently released their Kazemaster Flat II, which is is flat AND will do 12V startup with automatic reduction to lower voltages.

It's still manual.

 

[Carson Dyle]

Motherboard manufacturers like Asus and their 5-way optimization have really made strides with mobo fan control software/firmware, too, although I haven't checked lately to see if they can do the 12V startup thing.

Exactly. Why doesn't someone offer that in an add-in board? The software certainly can't be complicated.

 

[easp]

Exactly. Why doesn't someone offer that in an add-in board? The software certainly can't be complicated.

Because the market is small (that subset of enthusiasts who aren't served by whats offered on the motherboard) and shrinking (because motherboards are incorporating more and more of the functionality?

I think there would still be a market if you kept your costs low enough. Don't make it a PCIe board though, because that limits your market further. Make it USB2, with the option of plugging it into an internal header. Pull power from a molex. Make it small enough to stick wherever it will fit, a cheap microcontroller and hardware for 4 fans. If they want more, they can buy another one. Firmware would probably be pretty simple. Make it opensource and maybe someone will make it even better. The software to config it could be pretty simple too, and ideally, you could get an existing fan control app to support it.

Actually, the firmware probably already exists. I'd be surprised if there aren't at least two different Arduino fan control sketches.

And then there is this... http://geoffg.net/fancontroller.html