I use the fan headers on the board, but I also use a Cosair iCUE Commander PRO for most of my fans in the same machine. There is a github to support linux based fan speed modifications
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Fan controlling in both zones (CPU and HD) can be switched on and off independently. But the FANs in the zones defined by IPMI and it cannot be changed AFAIK. When you control CPU zone all FAN1-FAN4 fans will change their rotation speed (and same goes for HD zone with FANA-FANB).
Last edited:
the fan control script needs to run via cron and at boot to set the fan speeds. The motherboard will override the fan speeds on a reboot
I'm just sitting in a Windows install now to copy some files off NTFS disks (unraid is being annoying because they have a windows software mirror on them) and the fan speeds keep resetting. It's ****ing irritating. Per the thermal sensors, nothing is overheating to the point that it should be triggering a reset.
Anyone have any tips? Again, the thresholds are set.
edit; so apparently I didn't read the OP article that closely, lol, but if you have anything other than "full" selected, the system will always still try to manage the fans. Yikes. So I set it to full, and applied the commands, and boom everything's great.
Last edited:
Interestingly I got threshold problems in low temperature situations (with an open window winter time) in the past, meaning the the fan reached the lower threshold value.
Some fellows are using Optimal mode and let the system find the proper fan rotation speed, like in this case.
I prefer Full mode and I control the fan rotation speeds with my scripts. In both ways the precise setup of the thresholds are compulsory.
Thanks for creating this thread, it's very useful!
I'm about to relocate my X11SCM-F/CPU/cooler into a different chassis, and learning about the two fan zones has been useful. I've got a Noctua cooler on the CPU and it has a 92mm fan with a max speed of 2500PM. The chassis (SC-825) has three case fans which are currently the standard Nidec 6300RPM units, although I may replace them.
This motherboard only has two "HD zone" fan connectors, and four "CPU zone" fan connectors, but given the descriptions in this thread and the massive differences between my fans I don't think I want to put them all into the CPU zone. Would I be better off getting a PWM fan splitter so that I can connect all three case fans to the "HD zone"?
I'm about to relocate my X11SCM-F/CPU/cooler into a different chassis, and learning about the two fan zones has been useful. I've got a Noctua cooler on the CPU and it has a 92mm fan with a max speed of 2500PM. The chassis (SC-825) has three case fans which are currently the standard Nidec 6300RPM units, although I may replace them.
This motherboard only has two "HD zone" fan connectors, and four "CPU zone" fan connectors, but given the descriptions in this thread and the massive differences between my fans I don't think I want to put them all into the CPU zone. Would I be better off getting a PWM fan splitter so that I can connect all three case fans to the "HD zone"?
Ofc, you can use splitters for your fans. In the definition of the Super Micro IPMI modes:
On these boards there are 4 levels of speed control:
- Standard: BMC control of both fan zones, with CPU zone based on CPU temp (target speed 50%) and Peripheral zone based on PCH temp (target speed 50%)
- Optimal: BMC control of the CPU zone (target speed 30%), with Peripheral zone fixed at low speed (fixed ~30%)
- Full: all fans running at 100%
- Heavy IO: BMC control of CPU zone (target speed 50%), Peripheral zone fixed at 75%
Right, but I don't want the case fans' speed to be based on the CPU temperature, and they will be if they are plugged into the numbered ("CPU zone") connectors. I'll order a PWM fan splitter 
You may read the very first post in this thread for better understanding.
In the user manual (on page 13) there is also a reference on this:
FAN1 ~ FAN4
FANA, FANB CPU/System Fan Headers
Yea I have since noticed.
But anyway, is there any way to set the FAN speeds on a server running ESXi? There's no ipmitool.
But anyway, is there any way to set the FAN speeds on a server running ESXi? There's no ipmitool.
You may use ipmitool remotely since IPMI/BMC is an independent subsystem. You can configure things from a different Windows or Linux PC.
You need the IP address for the BMC interface and ADMIN credentials.
Oh, ok.
But I am slightly confused with the first post/article. The author says he has a script apply the settings every time during boot.
But I am slightly confused with the first post/article. The author says he has a script apply the settings every time during boot.
Those scripts were created for local execution indeed, but you can hack ESXi with ipmitool remote access if you really want to.
You may also google any existing solution for ESXi:
Setting up IPMI fan control mode, or fan thresholds are persistent configurations. If you configured once they will stay while the BMC is not reset.
If you want to have a dynamic fan speed control based on temperature then use can use a script (it must be executed locally, and be running continuously on your machine) or use Optimal mode and let BMC configure the fan speed (I just shared the link in post #126).
Can anyone explain the thresholds to me? I understand the lower ones are to ramp up the fans if the speed drops below the values, BUT:
1) what is the difference between the three low ones
2) what's the purpose of all the upper ones?
1) what is the difference between the three low ones
2) what's the purpose of all the upper ones?
1. Reaching of the lower thresholds will generate different assertions at BMC web interface "Server Health > Health Event Log". These thresholds should be lower than the minimal physical rotation speed of your fan. In case of assertion the BMC will try to change fan speed. Read more about this here.
2. I think it is here for theoretical purpose (maybe for special fans), I did not see any practical use
2. I think it is here for theoretical purpose (maybe for special fans), I did not see any practical use
Great wealth of information here. Thanks for everyone pitching in on this topic that started 6 years ago. Helped me resolve my issues.
FYI, on my X13SAE-F the "standard" fan speed seems to have the SYS headers tied to CPU temp. The SYS headers are behaving just like CPU_FAN1 in that they scale with CPU temp in a 1:1 ratio. As in the past, Optimal and Heavy I/O put the SYS headers at a fixed speed.
Not sure if this is a change for all X13 mainboards or not.
I was able to alter the critical thresholds using IPMI Util (I am on Windows, no Linux system handy on the network). This was how I set a new lower threshold:
ipmiutil sensor -N <IPMI IP> -U <IPMI username> -P <password> -n 41 -l 280 -h 2240
Where sensor #41 is the fan of interest.
The command for listing the sensors (posted elsewhere) is:
ipmiutil sensor -t -N <IPMI IP> -U <IPMI username> -P <password> -c
This netted actual low thresholds of "lo-crit" = 140 RPM and "lo-unrec" of 0 RPM. This BMC rev seems to not use the "lo-noncrit" setting or "hi-noncrit" setting and the above command does not change. They only display when no fan is connected to the header.
After moving a fan I had to reset the BMC to get it to recognize the header was now ABSENT and not just at 0 RPM.
Not sure if this is a change for all X13 mainboards or not.
I was able to alter the critical thresholds using IPMI Util (I am on Windows, no Linux system handy on the network). This was how I set a new lower threshold:
ipmiutil sensor -N <IPMI IP> -U <IPMI username> -P <password> -n 41 -l 280 -h 2240
Where sensor #41 is the fan of interest.
The command for listing the sensors (posted elsewhere) is:
ipmiutil sensor -t -N <IPMI IP> -U <IPMI username> -P <password> -c
This netted actual low thresholds of "lo-crit" = 140 RPM and "lo-unrec" of 0 RPM. This BMC rev seems to not use the "lo-noncrit" setting or "hi-noncrit" setting and the above command does not change. They only display when no fan is connected to the header.
After moving a fan I had to reset the BMC to get it to recognize the header was now ABSENT and not just at 0 RPM.