H170M-ITX/DL CPU fan control

I have 4 computers deployed with H110M boards.  At least two have exhibited this behavior.  One of the the excessive fan noise went away on its own.  Another one the fan noise was better today.  I updated the UEFI a few days ago and set the computer aside.  When I went in to put the UEFI in 4-pin CPU Fan mode, I did notice the fan speed was at 2000 RPM, but no noise.  For all 4 computers putting the UEFI in 4-pin mode has slowed the fans down to below 2000 RPM when the system is sitting still.  

An update on the fan speed issue with the Intel Skylake stock CPU cooler.

Long story short, I received a Skylake CPU cooler from Intel. I was hoping to get one with the NIDEC fan, which I thought was understood. Turned out, it wasn't...

I received a Skylake CPU cooler made by Foxconn. Disappointed at first, I decided to try it anyway, and see what happens. I connected it to my Z170 Extreme7+ board, to the CPU_FAN1 header, and ran the Fan Test in FAN-Tastic Tuning, in my board's A-Tuning utility program.






So, this Skylake CPU cooler's fan works fine. I did not have the CPU Fan Type Setting set to 4 Pin.

I received a PM from another ASRock board user that was having this problem with the NIDEC Skylake CPU cooler. He borrowed an Intel stock CPU cooler from an earlier CPU generation (not Skylake) and told me the fan on that CPU cooler worked fine. It did not behave at all like the NIDEC Skylake CPU cooler.

At this point, I am 99% certain that it really is the NIDEC Skylake CPU cooler, or at least a large sample of them, that have this problem. I would have liked to have a NIDEC CPU cooler to test myself, but given the information in this post, I am confident that the NIDEC cooler is the problem.

For those of you with the NIDEC CPU cooler, I suggest contacting Intel for a replacement. If you have access to any stock Intel CPU cooler from any processor that fits in the following CPU sockets, they can be used in your Skylake board:

1156, 1155, 1150, 1151.

Now to try to convince Intel they have a problem with the NIDEC Skylake CPU coolers. They told me they were not aware of any problems with them. I found a thread in the Intel Processor forum about this same issue, and had the OP check the type of Skylake CPU cooler he was using. It was a NIDEC.

MTR wrote: Another H170M-ITX/AC user here with the same problem regarding the NIDEC E97379-001. Similar to previous posters the fan stays OFF below 50% PWM signal and once ON remains invariably at 2000RPM. When using the added 4pin setting in the BIOS however  it is ALWAYS ON, even below the set PWM spinup threshold/temp. When ON (in 4pin) it starts at 1000RPM and ramps up to 2000RPM according to the settings. When replacing the cooler with DELTA E97378-001/0.60A, it does spin up in BIOS AUTO mode, without the need for the 4pin setting, but it also is ALWAYS ON and runs at am minimum of 1000RPM and  further similar to the graph posted by parsec above. Apart from the fact that the NIDEC seems faulty, the behavior of both my DELTA and parsec's Foxconn E97378-001 coolers also is not correct. They should be able to stay OFF until a certain PWM/temp signal is reached and subsequently ramp from xRPM to 2000RPM, instead of being ALWAYS ON and starting at 1000RPM. I'll see if I can get another manufacturer's  -79 or -78 cooler to test with, but I am curious to hear ASRock's comments on this and if they have been able to run one of these coolers as they should: OFF until threshold is reached and then ramping up according to the set schedule (starting from a minimum below 1000RPM). My main issue with the -78 is that the the fan is ALWAYS ON, even at low temps.

I imagine you know that operating a CPU cooler with its fan off is not considered a standard usage situation by Intel or any mother board manufacture.

Whether or not that is appropriate is debatable, but it is a well known general standard to prevent the CPU coolers fan from not spinning without specific intervention by the user, IF that is even allowed.

No offense, but your comment about running the CPU cooler's fan as it "should" be, off until a temperature threshold is passed, is not how any mother board manufacture's CPU fan speed control is configured by default. If you can show us documentation that describes this as the standard method of operation for a CPU cooler, I would like to see it.

This has been enforced in various ways by mother board fan speed control for as long as I have been building PCs. I've had boards that would display a message during POST if the CPU cooler's fan was not spinning at a certain RPM, meaning it was considered to slow.

The main motivation for this is to protect a less than knowledgeable user from using a CPU fan speed configuration that causes their processor to over heat, causing the PC to shutoff. This has also been applied to chassis fan headers on most if not all boards.

Another reason is all fans used in PCs are different in their starting voltages, minimum rotation speed, and rotation speed at voltages below 12V (three pin fans) or at different PWM signal levels (four pin PWM fans.)

The CPU cooler itself is another major variable in this situation. Some large CPU coolers can be used passively, with no fan or the fan off or at low RPM. But the Intel inbox CPU cooler does not qualify for use with its fan off.

The Windows utility program ASRock supplies with their boards (AXTU, A-Tuning, F-Stream, etc) will allow configuring the CPU fan header to shut off the CPU coolers fan on some of their boards. I can do that using FAN-Tastic Tuning in A-Tuning on my X99 Extreme6/3.1 board. Again, that does NOT apply to all versions of FAN-Tastic Tuning in the ASRock Windows utility program.

I can also do this with FAN-Tastic tuning in this board's UEFI. That is done with a Custom fan speed profile created by the user. Your board has the FAN-Tastic Tuning custom fan speed profile feature in the UEFI, H/W Monitor screen. If it is identical to my board, I don't know, but you can try to configure the CPU fan profile as you want with that feature.

MTR wrote: parsec wrote: MTR wrote: My main issue with the -78 is that the the fan is ALWAYS ON, even at low temps. Or at least (if being OFF completely is deemed unsafe) spin at a lower RPM setting. I can also do this with FAN-Tastic tuning in this board's UEFI. That is done with a Custom fan speed profile created by the user. Your board has the FAN-Tastic Tuning custom fan speed profile feature in the UEFI, H/W Monitor screen. If it is identical to my board, I don't know, but you can try to configure the CPU fan profile as you want with that feature. I do happen to know a little about the constraints/safety standards regarding CPU fans (see full quote of last sentence), thanks. That's why my post concerns (1) yet another confirmation of the NIDEC/-79 issue and (2) details on the runnings of another version of the -78 made by Delta. I like that the Delta fan behaves better than the NIDEC, but am unpleasantly surprised that it ALWAYS starts at >1000RPM instead of a lower RPM setting, basically disregarding the user setting s in either the BIOS or A-tuning. In the relevant system, the CPU@idle/low desktop usage is less than 10-15 degrees Celsius over ambient & only requires moderate fan speeds. Especially with the copper core & curves fin configuration of the -78 doing a better cooling job than the -79 heatsink. Every fan has a low limit RPM lower than which it will not spin, it just would surprise me that 1000RPM is that minimum on this fan. That would make it by far the highest minimum on any stock Intel fan I have seen lately. I'll check the low RPM on the stock Haswell cooler from another system for reference, but that one is so low as inaudible from 3ft.

My comments were based on your initial, pre-editted statement, not that it really matters.

I am also surprised the minimum speed of the Delta fan is 1000 RPM. You can see in my screen shot of FAN-Tastic tuning above, that the Fan Speed test shows 988 RPM at a 10% (apparently) PWM duty cycle. I've tried multiple PWM fans (Noctua, Cryorig, Scythe) on my Noctua CPU cooler, connected to the CPU_FAN1 header on my ASRock Z170 board, and I can assure you they were all capable of running at well under 1000 RPM when configured with the FAN-Tastic Tuning feature in the UEFI. The Noctua fans will spin at less than 400 RPM, the others down to ~600 RPM. I have three pin fans that stopped spinning at the 10% level during the fan test on the Chassis fan headers.

Do not use the inbox Skylake CPU coolers as an example of how the ASRock fan speed configuration software works. They use poor fans, particularly being PWM speed controlled fans. Every other PWM fan I have used worked exactly as it should.

Given the ~1000 RPM speed at a 10% fan power level, I don't understand how that is disregarding the user setting. It can be looked at that way, but it really is the behavior of the fan itself.

Can you imagine using the Delta fan stock cooler configured to be off up to say 50C, but then running above that temperature immediately at 1000 RPM? Or if it thrashes on and off as the CPU temperature hovers around the on temperature, if set to 40C?

IMO, the Skylake inbox CPU coolers are very cheaply made compared to previous models of the Intel inbox CPU coolers. I recently read someone complaining about the inbox cooler provided with a Haswell processor. I wonder if that was a late production processor, using the newest production inbox CPU coolers.

I would expect earlier model Intel inbox CPU coolers to have a much better range of fan speeds. I have quite a collection of them, unused, going back to the Pentium D and Core 2 Duo processors.

Hi, all. All this thing is very strange. I had a time (few weeks), when my NIDEC fan worked ok (quietly) and fan control was working! (It just started working). Then after one power on it started to be loud again. I have not touched anything. I really get angry so I have replaced fan with old core2 duo fan which I found and FINALLY there is QUIET!