Please report all spam threads, posts and suspicious members. We receive spam notifications and will take immediate action!
Results 1 to 9 of 9

Thread: Z87 Extreme4 PWM Fan control




  1. #1
    Join Date
    Dec 2013
    Location
    USA
    Posts
    21

    Default Z87 Extreme4 PWM Fan control

    I asked in my previous "RAID" thread about this, but as it was a minor concern, I didn't harp on it and it was overlooked by others.

    The UEFI of my board has several settings for fan control (Custom, Quiet, Standard, Performance, and Full Speed) for each motherboard fan header, but none of these settings seems to offer the standard PWM variable speed control that auto-adjusts fan speeds based on CPU/Motherboard temperature readings. Not even the Standard option will do that. Instead, the various settings just run a fan at a set RPM, or rather a percentage of a fan's maximum RPM. (Running at a percentage means the board is using a PWM-based level of control, though...) It's as if the only means to achieve variable fan speeds is to create custom fan profiles, which makes the fact that the fan supports PWM control somewhat irrelevant. (It also serves to make the PWM connectors on the motherboard irrelevant in the first place.)

    Like I said, it seems that aside from creating a custom profile for each fan on each header, there's no variable control at all. The manual sure doesn't explain it, and neither does the accompanying Software Setup Guide. All the guide mentions is the ability to set 5 levels of speed control based on thermal readings. Plugging a PWM fan into a PWM header has always done that automatically on every other board I've owned, including the ones where you had to run a configuration utility within the BIOS so the system had the fan's min & max RPM. I'm just puzzled by the fact there's no automated control or configuration for PWM fans at all...

  2. #2
    parsec's Avatar
    parsec is offline Senior Member
    Join Date
    Jun 2012
    Location
    Third stone from the sun
    Posts
    4,824

    Default Re: Z87 Extreme4 PWM Fan control

    So you are saying when using the BIOS setting in H/W Monitor for the CPU fan, and choosing any of the settings you listed, the PWM fan on your CPU cooler does not change its speed at all as the CPU temperature changes? What PWM fan(s) are you using?

    I have a Z87 Extreme 6 board, and I'm pretty sure our fan speed control option are similar if not identical. I've always used the Custom option in the UEFI/BIOS, and my PWM CPU fan does change its speed as the CPU temperature changes. I configured a fan speed percentage for a given temperature, as the Custom option provides, and it works for me. The Critical Temperature setting also in that screen, must be set to a relatively high CPU temperature, or the CPU fan will run at full speed at whatever temperature that option is set to.

    Also, in the A-Tuning software, in Fantastic Tuning, a PWM CPU fan can be calibrated, which must be done IMO. It only needs to be done once, and A-Tuning does not need to be run every time Windows runs, IMO.

  3. #3
    Join Date
    Dec 2013
    Location
    USA
    Posts
    21

    Default Re: Z87 Extreme4 PWM Fan control

    Well, I may have been mistaken, but only under one condition -- using the "Silent" mode. It seems that if I select silent mode, the currently installed NoiseBlocker eLoop B-12 P PWM fan (800-2000rpm) runs at ~775rpm at idle. Under load, it will speed up to about 980rpm once the CPU hits about 60-degrees Celsius while running a Prime95 small FFT torture test. So, it would seem that PWM is working, and that my Cooler Master Hyper 212 EVO works far better on my 4670K than I expected it would.

    While using the Standard, Performance, or Full-speed settings, the fan never increased in RPM under the same Prime95 stress tests. Why? It seems obvious (now) that the fan was already running at an RPM above what was being called for through PWM control for the 60-65 Celsius temperatures it was reaching. (I experienced the same phenomena using the stock Cooler Master PWM fan while the fan control was set to Standard, Performance, and Full speed prior to replacing it.)

    So, I guess if I used the stock HS/Fan, I never would have noticed an issue. But because the Hyper 212 EVO works so much more effectively, there was never any need for the fan to speed up.

    I guess we can consider this one solved... Maybe I'll try a 400-1500rpm NoiseBlocker eLoop B-12 PS fan instead. So long as the CPU temps don't exceed 70C, I doubt the current P-model will ever exceed 1000rpm.

  4. #4
    parsec's Avatar
    parsec is offline Senior Member
    Join Date
    Jun 2012
    Location
    Third stone from the sun
    Posts
    4,824

    Default Re: Z87 Extreme4 PWM Fan control

    As long as you're using stock clocks, or not enabling AVX instructions in a stress testing program, Haswell processor temps are manageable.

    I don't use the built in settings, since they don't adjust CPU fan speed enough to keep CPU temps where I want them to be. Also, all fans are different in their response to input voltage. Several ASR board users have threads complaining that their CPU fan speeds were to high, and could not reduce it enough with any setting available.

    Your NoiseBlocker eLoop B-12 has a rated starting voltage of 4.5V, which is in the low range for PC fans. Did you run the calibration process in A-Tuning?

    As I said earlier, the Custom option provides the most fan speed control out of them all.

  5. #5
    Join Date
    Dec 2013
    Location
    USA
    Posts
    21

    Default Re: Z87 Extreme4 PWM Fan control

    The newest version of Prime95 uses AVX instructions, so it does cause some heat. I considered testing with AIDA64 to truly punish the CPU, but AIDA64 uses a completely unrealistic load -- I'll never encounter a real-world situation where I'd be doing so many things that every supported instruction set would be thrown at my CPU simultaneously. AIDA64 seems to be the CPU equivalent to FurMark for GPUs.

    My model B12-P eLoop fan has a variable RPM rating of 800-2000. Again, in silent mode, it sits in its minimal ~800RPM range until the CPU hits 60C, at which time it slowly begins to creep toward 1000RPM. The cooler works so well at such a low fan speed that unless I really push this 4670K with a 4.5GHz or higher OC, the fan will probably never need to go over 1000RPM. If it should start to hit 70C or so, then I can manually adjust the Silent profile so it begins to ramp-up its speed a little earlier. I have played with A-Tuning's calibration, and I've also messed around with some customization. As such, I've pretty much got that potential modification covered. But for now, there's no need to make any adjustment.

    Oh... just to clarify. All NB eLoop-series fans have a start-up voltage of roughly 3.0V, which is quite low, but seems to be quite common for high-end, high-performance fans.

    Thanks for the reply.

  6. #6
    parsec's Avatar
    parsec is offline Senior Member
    Join Date
    Jun 2012
    Location
    Third stone from the sun
    Posts
    4,824

    Default Re: Z87 Extreme4 PWM Fan control

    Quote Originally Posted by RazberyBandit View Post
    The newest version of Prime95 uses AVX instructions, so it does cause some heat. I considered testing with AIDA64 to truly punish the CPU, but AIDA64 uses a completely unrealistic load -- I'll never encounter a real-world situation where I'd be doing so many things that every supported instruction set would be thrown at my CPU simultaneously. AIDA64 seems to be the CPU equivalent to FurMark for GPUs.

    My model B12-P eLoop fan has a variable RPM rating of 800-2000. Again, in silent mode, it sits in its minimal ~800RPM range until the CPU hits 60C, at which time it slowly begins to creep toward 1000RPM. The cooler works so well at such a low fan speed that unless I really push this 4670K with a 4.5GHz or higher OC, the fan will probably never need to go over 1000RPM. If it should start to hit 70C or so, then I can manually adjust the Silent profile so it begins to ramp-up its speed a little earlier. I have played with A-Tuning's calibration, and I've also messed around with some customization. As such, I've pretty much got that potential modification covered. But for now, there's no need to make any adjustment.

    Oh... just to clarify. All NB eLoop-series fans have a start-up voltage of roughly 3.0V, which is quite low, but seems to be quite common for high-end, high-performance fans.

    Thanks for the reply.
    I was quoting specs for the eLoop fans, and given your comment is true (not saying it isn't, 3V is low) that's the part that does not make sense IMO.

    Given that sensitivity, the fan should speed up more than it does. But those ASR fan settings must really be slow to change. Or did you set the "Critical Temperature" beyond 80C? A guy posted that his Nexus PWM fan with similar specs to yours (800 - 2000 RPM) was always running to fast for his taste.

    Curious how some people say Prime95 is more difficult than AIDA64, and vice versa. Might be the new CPU ucode that was updated for Haswell CPUs in a recent ASR BIOS update, since my 4670K would hit 100C in AIDA64 with AVX on, and throttle. Mine stayed much cooler recently using AIDA 64. Or is your CPU delidded?

  7. #7
    Join Date
    Dec 2013
    Location
    USA
    Posts
    21

    Default Re: Z87 Extreme4 PWM Fan control

    Spec sheets be damned, this fan starts at 3V. :D

    I pulled my old Sunbeam Rheobus out along with my voltmeter to test it. The fan began twitching at 2.7-2.8V and was going strong by 3.1V. (Well... if 800RPM can be considered strong.) Someone on Youtube tested all 6 NB eLoop models. All models except the B12-1 were spinning by 3.0-3.5V according to his test station's reading. (B12-1 started at 5.5V, but it twitched a few times before that.) Here's a link to youtube tester's videos: CoolingTechnique - YouTube

    And no, my CPU hasn't been de-lidded. And yes, the Crit Temp in the fan curve is set to 80C. The Silent fan setting simply has a very slow-to-rise curve. It has a flat 30% duty cycle (~800RPM on eLoop) until the CPU temperature reaches 60C, at which point it begins a sharp climb to 60% duty cycle (~1400RPM on the eLoop) at 75C. It then jumps to 100% fan speed at 80C. Personally, I'd prefer something slightly more aggressive. Perhaps 30% until 45C with a climb to 60% at ~65C, 75% at 70C, then the normal jump to 100% at 80C. So no, the fan shouldn't be spinning up more aggressively... It's just following the predefined curve.

    The Standard profile starts at a flat 50% duty cycle, while the Performance profile starts at a flat 70%, and neither will ever drop below those speeds. I feel that's wrong. If PWM control and temperature readings would allow it, then they should allow the fan to spin down. Starting at 30% with a jump to 50% or 70% once the CPU temp reaches 40C would be acceptable. With Haswell's power-saving capabilities, they run very cool at idle, meaning there's little need for the fan to spin at 50-70% constantly.

    Using the Standard profile with my CPU in its stock configuration, it never topped 62C while subjected to Prime95's Small FFT load. The Standard fan profile spins the NB eLoop at ~1250RPM, and I believe the stock CoolerMaster fan spun at about 1400RPM using the Standard profile. I'm using the Silent profile w/ the eLoop installed, now. With the CPU in the stock config, it hits a max of 66C under the same Prime95 load, with the fan topping out at ~975RPM. Interestingly enough, the CPU clocks at 3.6GHz during that test instead of the base 3.4GHz, so it's actually boosting while under that load.

  8. #8
    parsec's Avatar
    parsec is offline Senior Member
    Join Date
    Jun 2012
    Location
    Third stone from the sun
    Posts
    4,824

    Default Re: Z87 Extreme4 PWM Fan control

    It's good to have a low starting voltage fan, that indicates good design and build, plus it's not power hungry. I would say 800 RPM is not slow, many quiet case fans max out at 1000 - 1200 RPM. I've run several different PWM fans at 800 RPM or less, which is fine for a CPU at idle or low load.

    The new CM JetFlo fans (which are nice, I'm using two now) are rated at 0.40A, which is high for a PWM fan. A pair will draw enough current to fry a Corsair H100's built in fan speed controller (which is poor, not PWM) according to posts on the Corsair forum.

    The CM fan you mentioned is either a Blade Master, or the newer version, right? I've used them, they are good PWM fans.

    Those ASR profiles are not for me, I like Custom since that lets you up the fan speed at any temperature you like.

    I know what you mean about the processor running at 3.6GHz at otherwise stock settings. The ASR BIOS causes that AFAIK, several settings affect that. I wasn't sure if you were quoting temps with stock multipliers.

    If your board's latest BIOS has the new negative offsets for the CPU voltage (and others), give that a try with Adaptive voltage. You can really get the voltage down and retain stability.

  9. #9
    Join Date
    Dec 2013
    Location
    USA
    Posts
    21

    Default Re: Z87 Extreme4 PWM Fan control

    Yes, the stock 212 EVO fan is the new version of the Blade Master. It has a smoke-colored clear plastic instead of standard ABS plastic. It's not a bad fan at all. It offers great airflow and has good pressure. But, it's a bit too noisy in its upper RPM range. I'm using it in my other system, which never gets hot.

    JetFlo fans look fantastic, and seem to be very solidly built. But I have no need for fans that approach 100 CFM, especially those that approach 40 dB when they do so... Not with so many fans available in the 70-90 CFM range with 19-24 dB noise level ratings.

    Speaking of such fans... In my Fractal Define R4, I'm using two BitFenix Spectre Pro 140mm fans as intake, and one of the stock Fractal 140mm fans for exhaust. My case is filtered, so I elected to use a positive air pressure configuration to ensure all air comes in through those filters instead of through any available crack or vent. Instead, the extra air pressure pushes air out through the minor cracks. I hate cases with dirty interiors... Having worked as a PC repair tech previously, I've seen my fair share of nasty cases, and no case of mine will never get anywhere close.

    I did test its overclocking ability. It was stable at 4.5GHz @ 1.2V, which is respectable. But I have no need to overclock it at this stage. It chews through everything thrown at it in its stock configuration, so under-volting might be worthwhile.

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •