[SOLVED] Can i connect AIO Pump and fan on cpu_fan?

AlxR25

Prominent
Apr 18, 2022
28
0
530
So lately I've been thinking of getting an AIO cooler. Specifically the "Cooler Master MASTERLIQUID ML240L RGB V2" and i noticed that my motherboard does not have the CPU_OPT header and I've been searching abt it for almost 2 hours. I've been thinking of trying to connect both the fans and the pump into cpu_fan since my pc does not include CPU_OPT and i've used all my SYS_FAN headers. Will it work?
 
Solution
Regarding pump speed in such an arrangement, there is no problem. I did not give details of the "quirk", but this may help. An older 3-pin fan has its speed controlled only by changing the Voltage supplied to it on Pin #2. In the new PWM system, Pin #2 always supplies 12 VDC, and Pin #4 carries the new PWM signal. A 4-pin fan has an extra component, a chip that uses that PWM signal to modify the flow of current from the fixed 12 VDC supply line through the windings to achieve speed control. So when connected to a 4-pin header sending out signals in the new system, a 3-pin fan receives from Pin #2 a constant 12 VDC. It does not get the PWM signal from pin #4, and it does not have a chip to use that, anyway. So the 3-pin fan can NOT alter...

kanewolf

Titan
Moderator
So lately I've been thinking of getting an AIO cooler. Specifically the "Cooler Master MASTERLIQUID ML240L RGB V2" and i noticed that my motherboard does not have the CPU_OPT header and I've been searching abt it for almost 2 hours. I've been thinking of trying to connect both the fans and the pump into cpu_fan since my pc does not include CPU_OPT and i've used all my SYS_FAN headers. Will it work?
You would be better off running the pump directly from the power supply with an appropriate adapter. The pump should run at full speed all the time. Put the fans on the CPU_FAN and let the CPU temp change the speed of the fans.
 

Paperdoc

Polypheme
Ambassador
I disagree for one reason. Plus I'll point to another POSSIBLE arrangement.

Any mobo fan header has an important secondary function - it monitors the speed signal set back to it from the fan on Pin #3 for NO signal, indicating fan failure. On most mobos the CPU_FAN header pays extra attention to that protection task and will pop up an immediate warning on screen if it detects failure of the CPU cooler "fan". Many will go further than that - after a very short time it may shut down your system completely to prevent possible CPU overheating and damage without even waiting for the CPU internal temp sensor to show a high temp. Some will not allow you to boot up if it gets no CPU fan speed signal within a VERY short time of booting.

On an AIO system, failure of the rad FANS (even all of them at once) can be tolerated because the result is a slow rise of loop temperature and CPU internal temperature, which the normal Windows temperature monitor will catch to cause throttling of CPU speed and possible shut-down if the rise continues. But failure of the PUMP completely stops all heat removal and cannot be allowed to continue for significant time. So the PUMP is the unit that must be monitored for failure, and the one header that does that is CPU_FAN.

You CAN connect both the pump and the rad fans all to the CPU_FAN header. That header can two jobs, of course: control the rad fan speeds according to temperature inside the CPU chip, as well as monitor the pump for failure.You do this using a 3-output fan SPLITTER like this

https://www.amazon.com/Cable-Matters-2-Pack-Way-Splitter/dp/B07PXLHNZ6/ref=sr_1_3?crid=18XXJ7COTZ0QH&keywords=fan+splitter&qid=1650312134&sprefix=fan+splitter,aps,74&sr=8-3

That's a 2-pack. Note that a Splitter has one "arm" that plugs into a mobo header, and three (in this case) output "arms" for fans, but NO other connections. A HUB is different and you do not need it. It has a third "arm" that must plug into a SATA power output from the PSU for fan power. On the Splitter note that only ONE of its male outputs has all four pins in it, and the others are missing Pin #3. That is because the host fan header can only deal with a speed signal from ONE "fan" and must not receive others. So you MUST plug the PUMP connector into the one output arm with all four pins so that the PUMP speed signal is the only one fed to the CPU_FAN header for monitoring. The two rad fans plug into the other outputs and their speed WILL be controlled by the CPU_FAN header; however, their speeds will never be "seen" anywhere. This does NOT affect ability to control their speeds.

Regarding power, the PUMP uses a quirk of the design of 4-pin fans. In this system any older 3-pin fans plugged into a header actually using the new PWM Mode of control will always run full speed. So the PUMP is wired like those older fans. When connected to the CPU_FAN header it will always run full speed, as almost all AIO system designs want, even though the PWM FANS for the rad, connected to the same header by a Splitter, WILL have their speed controlled. To ensure that, you really need to go into BIOS Setup for the CPU_FAN header and ensure that it is set to PWM Mode, not to DC Mode.

If you do it this way, the CPU cooling will be controlled automatically by the CPU_FAN header since it alters the speed of the rad fans according to the CPU's internal temp sensor. Meanwhile the Pump is running full speed all the time as designed, AND its speed signal is monitored for possible failure.

There is another alternative for SOME mobos - you have not told us which you have. On some mobos for some or all of its SYS_FAN or CHA_FAN headers you have an option of which temperature sensor it uses as its guide. Normally you have no choice about the CPU_FAN header - it always uses the temp sensor inside the CPU chip. And normally all the SYS_FAN headers use the sensor on the mobo. BUT some allow you to choose to use the CPU's internal temp sensor for a SYS_FAN or CHA_FAN header. Essentially this converts that one header into a CPU_OPT header, and it can then be used for the FANS cooling the CPU, including the fans on the RAD of an AIO system. IF you do this, there is one other adjustment you need to check. Each fan header has a default "fan curve" set for it. MANY mobos allow you a choice for each header to change that to a customized version of what speed the fan should run for what measured temperature. I fully expect that the default curves for the CPU_FAN and SYS_FAN headers are different. So go into the CPU_FAN header and copy down that curve's settings. Now go to the one header you have changed for use with the rad fans and customize its fan curve to match the default CPU_FAN curve. Then you can connect your rad fans to this customized SYS_FAN header, and not use them on the CPU_FAN header. This way you WILL "see" the speed of one of the rad fans at that header, and their speeds will be controlled according to the CPU's internal temp sensor.
 
Last edited:

AlxR25

Prominent
Apr 18, 2022
28
0
530
I disagree for one reason. Plus I'll point to another POSSIBLE arrangement.

Any mobo fan header has an important secondary function - it monitors the speed signal set back to it from the fan on Pin #3 for NO signal, indicating fan failure. On most mobos the CPU_FAN header pays extra attention to that protection task and will pop up an immediate warning on screen if it detects failure of the CPU cooler "fan". Many will go further than that - after a very short time it may shut down your system completely to prevent possible CPU overheating and damage without even waiting for the CPU internal temp sensor to show a high temp. Some will not allow you to boot up if it gets no CPU fan speed signal within a VERY short time of booting.

On an AIO system, failure of the rad FANS (even all of them at once) can be tolerated because the result is a slow rise of loop temperature and CPU internal temperature, which the normal Windows temperature monitor will catch to cause throttling of CPU speed and possible shut-down if the rise continues. But failure of the PUMP completely stops all heat removal and cannot be allowed to continue for significant time. So the PUMP is the unit that must be monitored for failure, and the one header that does that is CPU_FAN.

You CAN connect both the pump and the rad fans all to the CPU_FAN header. That header can two jobs, of course: control the rad fan speeds according to temperature inside the CPU chip, as well as monitor the pump for failure.You do this using a 3-output fan SPLITTER like this

https://www.amazon.com/Cable-Matters-2-Pack-Way-Splitter/dp/B07PXLHNZ6/ref=sr_1_3?crid=18XXJ7COTZ0QH&keywords=fan+splitter&qid=1650312134&sprefix=fan+splitter,aps,74&sr=8-3

That's a 2-pack. Note that a Splitter has one "arm" that plugs into a mobo header, and three (in this case) output "arms" for fans, but NO other connections. A HUB is different and you do not need it. It has a third "arm" that must plug into a SATA power output from the PSU for fan power. On the Splitter note that only ONE of its male outputs has all four pins in it, and the others are missing Pin #3. That is because the host fan header can only deal with a speed signal from ONE "fan" and must not receive others. So you MUST plug the PUMP connector into the one output arm with all four pins so that the PUMP speed signal is the only one fed to the CPU_FAN header for monitoring. The two rad fans plug into the other outputs and their speed WILL be controlled by the CPU_FAN header; however, their speeds will never be "seen" anywhere. This does NOT affect ability to control their speeds.

Regarding power, the PUMP uses a quirk of the design of 4-pin fans. In this system any older 3-pin fans plugged into a header actually using the new PWM Mode of control will always run full speed. So the PUMP is wired like those older fans. When connected to the CPU_FAN header it will always run full speed, as almost all AIO system designs want, even though the PWM FANS for the rad, connected to the same header by a Splitter, WILL have their speed controlled. To ensure that, you really need to go into BIOS Setup for the CPU_FAN header and ensure that it is set to PWM Mode, not to DC Mode.

If you do it this way, the CPU cooling will be controlled automatically by the CPU_FAN header since it alters the speed of the rad fans according to the CPU's internal temp sensor. Meanwhile the Pump is running full speed all the time as designed, AND its speed signal is monitored for possible failure.

There is another alternative for SOME mobos - you have not told us which you have. On some mobos for some or all of its SYS_FAN or CHA_FAN headers you have an option of which temperature sensor it uses as its guide. Normally you have no choice about the CPU_FAN header - it always uses the temp sensor inside the CPU chip. And normally all the SYS_FAN headers use the sensor on the mobo. BUT some allow you to choose to use the CPU's internal temp sensor for a SYS_FAN or CHA_FAN header. Essentially this converts that one header into a CPU_OPT header, and it can then be used for the FANS cooling the CPU, including the fans on the RAD of an AIO system. IF you do this, there is one other adjustment you need to check. Each fan header has a default "fan curve" set for it. MANY mobos allow you a choice for each header to change that to a customized version of what speed the fan should run for what measured temperature. I fully expect that the default curves for the CPU_FAN and SYS_FAN headers are different. So go into the CPU_FAN header and copy down that curve's settings. Now go to the one header you have changed for use with the rad fans and customize its fan curve to match the default CPU_FAN curve. Then you can connect your rad fans to this customized SYS_FAN header, and not use them on the CPU_FAN header. This way you WILL "see" the speed of one of the rad fans at that header, and their speeds will be controlled according to the CPU's internal temp sensor.

thanks for your advice, but if i use a splitter wouldn't it control both the fans and the pump so when my temps get lower the fan speed decreases but it'll as well decrease the pump speed and it'll not be able to run at 100% all the time as i've read that it must do. Also my motherboard is a gigabyte z490m rev 1.0 Micro ATX if that helps you explain to me what i can do.
 

Paperdoc

Polypheme
Ambassador
Regarding pump speed in such an arrangement, there is no problem. I did not give details of the "quirk", but this may help. An older 3-pin fan has its speed controlled only by changing the Voltage supplied to it on Pin #2. In the new PWM system, Pin #2 always supplies 12 VDC, and Pin #4 carries the new PWM signal. A 4-pin fan has an extra component, a chip that uses that PWM signal to modify the flow of current from the fixed 12 VDC supply line through the windings to achieve speed control. So when connected to a 4-pin header sending out signals in the new system, a 3-pin fan receives from Pin #2 a constant 12 VDC. It does not get the PWM signal from pin #4, and it does not have a chip to use that, anyway. So the 3-pin fan can NOT alter its speed - it must run full speed all the time. Since common AIO pumps are wired just like 3-pin older fans, you can connect them using a Splitter to a 4-pin header using the PWM control signal system and it still runs full speed, irrespective of what the PWM signal on Pin #4 calls for. But 4-pin fans connected to that same set of signals DO have their speeds controlled by the PWM signal they use.

Note that it is important that the header DOES use the new PWM Mode of control. On some mobos the configuration options for a CPU_FAN header include an "auto" setting that tests what type of fan is plugged in there and sets its control system according to the result. I believe what that really does is start out by using the new 4-pin PWM mode and trying to reduce its fan's speed. A 4-pin fan will slow down, but a 3-pin fan will not. So if the fan does not slow down, it changes to using the older Voltage Control Mode. (Note that a true 4-pin PWM style fan WILL have its speed controlled in this system, even if it is not technically ideal.) So when you connect a PUMP and some 4-pin fans to a CPU_FAN header using a Splitter AND you make sure that the PUMP is the one item that sends its speed back to the header (for failure monitoring), letting the header make its own test and configuration setting will result in the header's deciding it is dealing with a 3-pin fan, and changing to the older Voltage Control Mode. That is what you do NOT want. So it is important NOT to leave the header set to auto-adjusting its Mode. Set it specifically to PWM Mode.
 
  • Like
Reactions: Excelso and AlxR25
Solution