Question Radiator fan push or pull

[Atiriug]

Hi, so I have some questions about fan intake and exhaust on the radiator.

Basically I can't put the arctic freeze iii 240mm ( 2 fans) on top bc it's not thick enought and occupies the rams space.
Having it in front I left the fans on exaust position thinking it would be better to send hot air out then in. But after reading it seems I should make then intake so fresher air comes in through the radiator.
On my setup I have two top and one rear fan as intake.
First question:
So I should make this all go the opposite way right?

The radiator should be intake and the tops and rear exaust.
My board does support more fans.

Second question:
Should I remove one of the top fans, maybe the front top and leave only 2 outtake to not unbalance the intake and outtake?

Third question
Also should the radiator fans stay inside and intake or should i move them to the mesh side so they will blow air through the radiator instead of pulling air through the radiator or this doesn't make any influence?

My setup:

AMD Ryzen 5 7600X 4.7 GHz 6-Core Processor

ARCTIC Liquid Freezer III 240 A-RGB 48.82 CFM Liquid CPU Cooler

Asus PRIME B650M-A WIFI II Micro ATX AM5 Motherboard

TEAMGROUP T-Force Delta RGB 32 GB (2 x 16 GB) DDR5-6000 CL30 Memory

Crucial P3 Plus 1 TB M.2-2280 PCIe 4.0 X4 NVME Solid State Drive

KIOXIA EXCERIA G2 1 TB M.2-2280 PCIe 3.0 X4 NVME Solid State Drive

Gigabyte GAMING OC Radeon RX 7800 XT 16 GB Video Card

MSI MAG A750GL PCIE5 750 W 80+ Gold Certified Fully Modular ATX Power Supply
Phanteks Eclipse P400A Digital ATX Mid Tower Case

 

[Aeacus]

Third question
Also should the radiator fans stay inside and intake or should i move them to the mesh side so they will blow air through the radiator instead of pulling air through the radiator or this doesn't make any influence?

Best to run AIO fans in pull.

Reason why here (at 3:45):

View: https://www.youtube.com/watch?v=UyC3lZ5WFMk#t=3m45s


(Works better with top mounting but also beneficial for front mounting.)

Second question:
Should I remove one of the top fans, maybe the front top and leave only 2 outtake to not unbalance the intake and outtake?

Changing fan RPM you can achieve neutral pressure. E.g run intake fans faster than exhaust, for neutral pressure. But on cooling aspect, negative pressure is the best (more exhaust than intake).

First question:
So I should make this all go the opposite way right?

Airflow rule of thumb: front & bottom - intake; top & rear - exhaust.

Since you can't put AIO as top exhaust, you are left with it as front intake. It won't make much internal temp difference.
So, you can configure your fans in pull and call it a day.

 

[Paperdoc]

By question no. MUCH of this is "common way", some meet particular needs.

1. Yes, reverse them all. Intake at front, exhausts top and rear.
2. Yes, remove front fan of the top three, for two reasons. One, as you say, it to rebalance air flow capacities. The second is that, at the case top front, you get an aiflow "short circuit" in which intake air from the top front fan is sucked right out of the case by the top fan right above it, and those fans do little good for your case. Removing the frontmost fan of the top three eliminates this so air from the rad top fan will flow though the upper part of the case. By the way, I disagree with Aeacus above on one item: I believe small positive pressure in the case is best for dust prevention. Pressure itself has no impact on cooling efficiency.
3. Yes, mount the rad fans on the outside of the rad so they blow air into the rad. This may make little difference, so it is not essential. One way it IS important is if those rad fans have LIGHTS in their frames. Then where the lights are positioned (and how you get to see them) has a big impact on how they are mounted on the rad.

On the last item I note the video Aeacus linked and will comment. Very good point made about access to clean the fins of the rad. Note that the video does NOT tell us where the dusty rad was positioned , but it may have been in a top mount with NO dust filter on the fan intake side. Normally I would not call that a dust problem, but we don't know other details. For FRONT mounted fans on the outside of the rad as INtakes, there MUST be dust filters on the outer (intake) side of the fans, and they MUST be cleaned from time to time. This normally reduces dust intake and accumulation a LOT, but I recognize that SOME dust will accumulate on the rad fins (intake side) slowly. As that happens the user may NOT notice, even, because it is hidden behind the fans. Alternatively if the fans are on the INSIDE of the rad as PULL fans, the intake side of the rad is easy to see, and still is protected by the dust filters in front of it. But when dust does accumulate there, it will be "see-able" when the dust filters are accessed for cleaning and CAN be accessed easily to clean it off. And I agree that MOST fans supplied for use on rads are equally efficient no matter which way their intake side is mounted (free air intake, or against fins).

 

[Atiriug]

I thought positive pressure inside the case was better to avoid dust coming in from other holes.

About the top front fan it would go as exaust, removed or try to put it in the front bottom as a 3rd intake.

Its more like mid top then front top.
Here is the image:

 

[Paperdoc]

Not sure there is room for a third front intake fan in that case. IF you can, that would be better than leaving it as a third exhaust, and probably better than having 2 intake (rad) and 2 exhaust. Note that, because the rad fans must pull against the rad fins, their intake air flow is LESS that their max specs, so the 2 intake (rad) and 2 exhaust configuration means there is LESS intake than exhaust and hence a small negative pressure in the case. I see the case manual says you CAN mount a 360 mm rad in the front IF you "remove the mid-plate cover", whatever that is. That MAY be the perforated plate across the case bottom above the PSU mounting bay. It appears removing that would allow air from a bottom front fan to flow in and up into the case.

 

[Aeacus]

By the way, I disagree with Aeacus above on one item: I believe small positive pressure in the case is best for dust prevention. Pressure itself has no impact on cooling efficiency.

Negative pressure essentially creates a vacuum effect inside the PC case, whereby as soon as any hot air is generated, it is expelled fast. And fresh air is sucked in from all the small gaps PC case has. It can lead to higher dust intake but in the same time - provides better cooling performance (best of the three).

Positive pressure can lead to hotspots inside the PC case, where air only circulates around, while the excess air is then pushed out of all the openings case has (including exhaust fans). And when all intake fans are filtered, once can manage dust intake and keep it low. Downside is worse cooling performance.

A nice video about different airflow pressures:

View: https://www.youtube.com/watch?v=sh6F2eccMec


There is no golden: no dust, best cooling solution.
Either: best cooling with some dust (negative pressure);
decent cooling with little dust (positive pressure);
or the one between the two (neutral pressure).

 

[Atiriug]

Not sure there is room for a third front intake fan in that case. IF you can, that would be better than leaving it as a third exhaust, and probably better than having 2 intake (rad) and 2 exhaust. Note that, because the rad fans must pull against the rad fins, their intake air flow is LESS that their max specs, so the 2 intake (rad) and 2 exhaust configuration means there is LESS intake than exhaust and hence a small negative pressure in the case.

Seems I have to go with pull since I noticed I don't have access to all the radiator holes to put the fans on the outside.

But the good side is that I can move the top front fan to front bottom.

Sharing image to show:

 

[Aeacus]

Its more like mid top then front top.
Here is the image:

I see two airflow conflicts with your rear&top intake, front exhaust;
1. In horizontal mount of a GPU, GPU fans take air from the bottom and push it up, through GPU heatsink and essentially through backplate holes upwards, to AIO pump. With this the top fans actually are working against GPU fans, since they push the air back down, against GPU.
2. The front bottom fan sucks the air in front of the GPU fans, leaving GPU fans less air to use, to cool the GPU. It's essentially starving GPU of air. Not as extreme as i've seen elsewhere, but still, will affect GPU temps (which would be higher than normal).

But if you reverse the overall airflow configuration, things improve vastly for GPU. Because then;
1. Top fans as exhaust, actually help GPU fans to pull hot air away from the GPU.
2. Front bottom intake fan also helps GPU fans, by providing fresh air directly to the front of GPU fans.

Another detail i noticed; your PSU shroud has a hatch near the AIO rad, that you can remove. This would enable you to mount your AIO rad lower within the PC case.

 

[Paperdoc]

A further thought with a configuration of: 2 Intakes on the front rad, 2 exhausts top rear and rear, plus one intake (free) bottom front. How to connect / control.

You will connect the Arctic Freezer AIO system to the CPU_FAN header and that gives control of rad fans to that header. You then have three CHA_FAN headers. I suggest you use a simple Splitter to connect both the rear and top exhaust fans to one header, then the bottom front intake fan to a second by itself. This will allow you to control that front intake separately from the others. You MIGHT decide at some point to give that one a custom "fan curve" to set up the balance of air flows.

Re Aeacus' post above, I agree but will add this. Removing that plate ALSO means you can move the rad and fans to the highest position in the front and mount that third fan as intake at the bottom.

 

[Atiriug]

A further thought with a configuration of: 2 Intakes on the front rad, 2 exhausts top rear and rear, plus one intake (free) bottom front. How to connect / control.

You will connect the Arctic Freezer AIO system to the CPU_FAN header and that gives control of rad fans to that header. You then have three CHA_FAN headers. I suggest you use a simple Splitter to connect both the rear and top exhaust fans to one header, then the bottom front intake fan to a second by itself. This will allow you to control that front intake separately from the others. You MIGHT decide at some point to give that one a custom "fan curve" to set up the balance of air flows.

Re Aeacus' post above, I agree but will add this. Removing that plate ALSO means you can move the rad and fans to the highest position in the front and mount that third fan as intake at the bottom.

I actually have an extra fan, didn't know I could use a splitter to add it to the same +5 volt pin.
Thanks for the advice. Might do this 😉👌🏻

 

[Paperdoc]

Woops! What "+5V pin"?

A simple FAN has either 3 or 4 holes in the connector on its cable for the MOTOR. Either way, the power to the fan is at 12 VDC max. The 3- or 4-hole difference is all in whether it is an older Voltage Control Mode fan or a newer PWM Mode fan.

IF the fan also has LIGHTS in its frame then it will have a second cable with a wider connector on its end. This may have 4 holes (simpler plain RGB system supplied with 12 VDC power) or 3 holes (looks like it had 4 but one is plugged, the more complex ARGB light system with power at 5 VDC).

One can get a Splitter for either type of power cable, but they are different of course. A SPLITTER is a simple device that merely connects ALL of its outputs to ONE shared mobo header, so all fan motors (or fan lights) on the Splitter get control from one header and must share the power available from that header. There are limits on how much power one header can supply, but two or three fans on one header using a Splitter is not any real problem. HOWEVER, many fans together requires a different device called a HUB. It is a LOT like a Splitter, but it has a distinguishing added component. It always has a THIRD type of connection that must go to a PSU power output (either SATA or Molex) to get added power and NOT rely on the limits of the header.

So, what type of fans do you have - lighted or not? And what mobo header(s) are you going to use for these fans?

 

[Atiriug]

Woops! What "+5V pin"?

A simple FAN has either 3 or 4 holes in the connector on its cable for the MOTOR. Either way, the power to the fan is at 12 VDC max. The 3- or 4-hole difference is all in whether it is an older Voltage Control Mode fan or a newer PWM Mode fan.

IF the fan also has LIGHTS in its frame then it will have a second cable with a wider connector on its end. This may have 4 holes (simpler plain RGB system supplied with 12 VDC power) or 3 holes (looks like it had 4 but one is plugged, the more complex ARGB light system with power at 5 VDC).

One can get a Splitter for either type of power cable, but they are different of course. A SPLITTER is a simple device that merely connects ALL of its outputs to ONE shared mobo header, so all fan motors (or fan lights) on the Splitter get control from one header and must share the power available from that header. There are limits on how much power one header can supply, but two or three fans on one header using a Splitter is not any real problem. HOWEVER, many fans together requires a different device called a HUB. It is a LOT like a Splitter, but it has a distinguishing added component. It always has a THIRD type of connection that must go to a PSU power output (either SATA or Molex) to get added power and NOT rely on the limits of the header.

So, what type of fans do you have - lighted or not? And what mobo header(s) are you going to use for these fans?

The two top fans and the extra fan i got have 2 holes. The rear as 4 holes.
Does the splitter allow to connect both 2 holes and 4 holes together?

 

[Paperdoc]

There is no "standard" fan motor connector with 2 holes. So connecting those to anything is going to need some adapter or custom work. On those two top fans, what can you tell us form their labels? maker and model would help. They may also show Volts and / or Amps and / or Watts.

 

[Atiriug]

There is no "standard" fan motor connector with 2 holes. So connecting those to anything is going to need some adapter or custom work. On those two top fans, what can you tell us form their labels? maker and model would help. They may also show Volts and / or Amps and / or Watts.

My bad it's 3 holes.
They are original from the phanteks CPU case.

Another question since we are on the topic.

I'm leaving computer idle to see what you temperature it goes to without any activity like gaming.

What should be the ideal temperature at iddle?

I should tune the fans rpm to keep it always at a certain Celsius right?

Not on for 10min and already at 42.7 Celsius degrees.

 

[Aeacus]

There is no "standard" fan motor connector with 2 holes.

Actually, there is. 2-pin DC connector. Same as 3-pin but without RPM feedback. Most current day Delta Industrial fans (server grade) are 2-pin DC fans. Heck, if you buy Delta fan, oftentimes you even won't get fan connector with the fan. Just two wires (power and ground) ending bare.

What should be the ideal temperature at iddle?

There is no ideal idle temp. Since idle temp vastly differs by ambient temp, cooling solution, fan RPM, fan size and fan amount. (Dust too.)

Hence why, there are no idle temp charts in component reviews.
E.g your CPU review temp chart: https://www.techpowerup.com/review/amd-ryzen-5-7600x/25.html

But this much i can tell, that you will never get lower temps than ambient temps. Since PCs are cooled by ambient air.
Also, idle temps doesn't matter because even the smallest cooler can keep CPU/GPU near ambient temps. What matters, is under load temps, since there you can see the effectiveness of your cooling solution (usually balance between noise and temperature target).

E.g in my Skylake build (full specs with pics in my sig), idle temps currently are:
* 27.2C - ambient
* 30C - CPU (i5-6600K)
* 42.7C - GPU (GTX 1660 Ti)
* 41C - MoBo chipset (Z170)
* 49C - OS drive (970 Evo Plus 2TB)
* 35C - OS backup drive (960 Evo 500GB)
* 33C - data drive (870 Evo 2TB)
* 36C - data backup drive (MX500 1TB)

But these values of mine will tell you 0, since you do not have the same ambient, components and airflow setup as i have.

 

[Atiriug]

Actually, there is. 2-pin DC connector. Same as 3-pin but without RPM feedback. Most current day Delta Industrial fans (server grade) are 2-pin DC fans. Heck, if you buy Delta fan, oftentimes you even won't get fan connector with the fan. Just two wires (power and ground) ending bare.


There is no ideal idle temp. Since idle temp vastly differs by ambient temp, cooling solution, fan RPM, fan size and fan amount. (Dust too.)

Hence why, there are no idle temp charts in component reviews.
E.g your CPU review temp chart: https://www.techpowerup.com/review/amd-ryzen-5-7600x/25.html

But this much i can tell, that you will never get lower temps than ambient temps. Since PCs are cooled by ambient air.
Also, idle temps doesn't matter because even the smallest cooler can keep CPU/GPU near ambient temps. What matters, is under load temps, since there you can see the effectiveness of your cooling solution (usually balance between noise and temperature target).

E.g in my Skylake build (full specs with pics in my sig), idle temps currently are:
* 27.2C - ambient
* 30C - CPU (i5-6600K)
* 42.7C - GPU (GTX 1660 Ti)
* 41C - MoBo chipset (Z170)
* 49C - OS drive (970 Evo Plus 2TB)
* 35C - OS backup drive (960 Evo 500GB)
* 33C - data drive (870 Evo 2TB)
* 36C - data backup drive (MX500 1TB)

But these values of mine will tell you 0, since you do not have the same ambient, components and airflow setup as i have.

The fact the ambient temperature affects the case temperature is what worries me and makes me try to make all this Colling system more efficient, since we just entered spring and it's still cold overall (around 16° today).

I also noticed that Adrenaline edition gives me CPU package temperature and the board application gives me both, but had the cpu temperature as Standart. I had no idea there were two cpu temperatures and reading about it right now 😅.

Tomorrow I'll receive my splitter and change all the fans direction and add the other one under the radiator.

I appreciate all the help and knowledge guys 👍

 

[Aeacus]

I also noticed that Adrenaline edition gives me CPU package temperature and the board application gives me both, but had the cpu temperature as Standart. I had no idea there were two cpu temperatures and reading about it right now 😅.

Better use HWinfo64 for telemetry. It is the most extensive and gives you every data your PC usually monitors,
link: https://www.hwinfo.com/download/

For temps, there are actually plethora of more sources, than just "two" CPU temps. Package temp, individual core temp, temp to Tjmax etc.
Also, HWinfo64 has logging feature, where you can enable logging, play a game (or bench the system) and afterwards, you can see how system behaved during logging (temps, frequency, voltages etc.). Moreover, HWinfo64 shows: current, min and max values. Very handy to have good overview of your system.

 

[Paperdoc]

OP, you asked, "I should tune the fans rpm to keep it always at a certain Celsius right?"

Answer: NO!

Why? Well, the system that alters fan speeds is really a TEMPERATURE control loop. Its focus is on keeping the temperature at a sensor at or near a target, and it manipulates the speed of the relevant cooling system to achieve that. The basic model of control used is called a "Feed-Forward" system. That means that it uses the current measured temperature to predict what fan speed SHOULD manage to keep the temp from going higher. If that works perfectly, the temp will NOT change without any change in actual workload, so the fan speed will stay unchanged. If the speed it sets allows the temp to rise again, then the speed will be increased. Repeat if necessary until the rate of heat removal (by fan speed) equals the rate of heat generation (from workload). If workload changes, this will alter the heat generation and thus the temp measured, and the system will adjust fan speed accordingly. NOTE that, to get the fan speed to increase and cool more, the measured TEMP must increase first. The impact of ambient temperature (room air) is small. The control strategy does not take that into account, so cooler air will result in slightly better heat removal and thus a tiny lowering of CPU temp, but that will trigger a tiny reduction in fan speed.

The detail of this strategy is in the "fan curve" set up for each fan header. It is a graph of what fan speed signal should be requested for what measured temperature. IF you think of this as a way to make the temp stay at a particular setting, it is not quite that. It really is a way to choose a temperature target that changes, but tries to keep it from going too high by merely stabilizing the temp at something acceptable. It does NOT try to keep ONE target temp for all conditions. You CAN alter the fan curve to have it run the fan faster at any specified temp and it will do that, thus making the CPU run cooler. If you do that for all temps in the fan curve range you can have the CPU run cooler than the mobo maker's default settings over the whole range, but that will NOT make it run at a fixed temp for all workloads. Certainly you could make the CPU run cooler at idle workloads, but for no benefit - the CPU is perfectly capable of running at those low temps with NO impact on its lifetime. At the other end - at the max temp and fan speed end - you may or may not be able to lower the CPU temp. That depends on whether your CPU cooling system (fans or whatever) running at max speed actually can remove heat faster than your CPU can generate it at max workload. Often that is not the case, so at max fan speed your CPU actually might run a little hotter than "ideal" max temp. In the case of a cooling system not adequate for the CPU, the CPU temp at max workload may actually overheat enough to trigger a protection system that forces the CPU to slow down its processing rate and thus generate less heat to match the cooling system capacity to remove it. Most people do not have this problem, but it can occur.

Back to your query, OP, about the two top fans that have THREE-hole connectors. This means they are the older-style 3-pin fan. Control of speed on that type is done only by changing the VOLTAGE supplied to the fan from Pins 1 and 2 of the header. Because there are now so many fans on the market of a new design called 4-pin or PWM, makers always install 4-pin headers on the mobo. But then they ALSO add a feature in the options for configuring each header to set the type of control signal sent out on the pins. These are the signal MODE settings and include choices of Voltage (or DC), PWM, and possibly Automatic. As the user, YOU should adjust this item for each fan header you use, according to the fan type connected there. For your 3-pin fans, set this to Voltage or DC. If you leave it on PWM, those fans will run full speed all the time with no control. For the Arctic Freezer AIO scooler system for your CPU, the CPU_FAN header for that should be set to PWM, not Voltage and not Auto.

 

[Atiriug]

Hello once again.

Update
I changed all the fans directions.
Didn't manage to add the 3rd front fan yet since I'm missing to long radiator screws since I need to us the last two to put the upper 3rd fan screws.

I also united the rear fan and top back fan to same chan fan. So this way I can control the top front fan and in the future the 3rd front fan independently.

What did I noticed? After leaving the computer on Abit like I testes yesterday I get cou package temperature of 39-40°Celcius and yesterday it was 41-42° Celsius.
Also the day yesterday the house was Abit colder.

Conclusion:
Better to get fresh air coming in with pull configuration on radiator rather then pushing/exhausting.

At the moment I got negative pressure since I haven't turned the fans yet.

Will update once I manage to get the 3rd fan mounted.

Thanks for all the help guys 🫡

 

[Atiriug]

OP, you asked, "I should tune the fans rpm to keep it always at a certain Celsius right?"

Answer: NO!

Why? Well, the system that alters fan speeds is really a TEMPERATURE control loop. Its focus is on keeping the temperature at a sensor at or near a target, and it manipulates the speed of the relevant cooling system to achieve that. The basic model of control used is called a "Feed-Forward" system. That means that it uses the current measured temperature to predict what fan speed SHOULD manage to keep the temp from going higher. If that works perfectly, the temp will NOT change without any change in actual workload, so the fan speed will stay unchanged. If the speed it sets allows the temp to rise again, then the speed will be increased. Repeat if necessary until the rate of heat removal (by fan speed) equals the rate of heat generation (from workload). If workload changes, this will alter the heat generation and thus the temp measured, and the system will adjust fan speed accordingly. NOTE that, to get the fan speed to increase and cool more, the measured TEMP must increase first. The impact of ambient temperature (room air) is small. The control strategy does not take that into account, so cooler air will result in slightly better heat removal and thus a tiny lowering of CPU temp, but that will trigger a tiny reduction in fan speed.

The detail of this strategy is in the "fan curve" set up for each fan header. It is a graph of what fan speed signal should be requested for what measured temperature. IF you think of this as a way to make the temp stay at a particular setting, it is not quite that. It really is a way to choose a temperature target that changes, but tries to keep it from going too high by merely stabilizing the temp at something acceptable. It does NOT try to keep ONE target temp for all conditions. You CAN alter the fan curve to have it run the fan faster at any specified temp and it will do that, thus making the CPU run cooler. If you do that for all temps in the fan curve range you can have the CPU run cooler than the mobo maker's default settings over the whole range, but that will NOT make it run at a fixed temp for all workloads. Certainly you could make the CPU run cooler at idle workloads, but for no benefit - the CPU is perfectly capable of running at those low temps with NO impact on its lifetime. At the other end - at the max temp and fan speed end - you may or may not be able to lower the CPU temp. That depends on whether your CPU cooling system (fans or whatever) running at max speed actually can remove heat faster than your CPU can generate it at max workload. Often that is not the case, so at max fan speed your CPU actually might run a little hotter than "ideal" max temp. In the case of a cooling system not adequate for the CPU, the CPU temp at max workload may actually overheat enough to trigger a protection system that forces the CPU to slow down its processing rate and thus generate less heat to match the cooling system capacity to remove it. Most people do not have this problem, but it can occur.

Back to your query, OP, about the two top fans that have THREE-hole connectors. This means they are the older-style 3-pin fan. Control of speed on that type is done only by changing the VOLTAGE supplied to the fan from Pins 1 and 2 of the header. Because there are now so many fans on the market of a new design called 4-pin or PWM, makers always install 4-pin headers on the mobo. But then they ALSO add a feature in the options for configuring each header to set the type of control signal sent out on the pins. These are the signal MODE settings and include choices of Voltage (or DC), PWM, and possibly Automatic. As the user, YOU should adjust this item for each fan header you use, according to the fan type connected there. For your 3-pin fans, set this to Voltage or DC. If you leave it on PWM, those fans will run full speed all the time with no control. For the Arctic Freezer AIO scooler system for your CPU, the CPU_FAN header for that should be set to PWM, not Voltage and not Auto.

I understand, so overall the cooling system efficiency is what actually matters.
I played with the fan curves and actually putting them on max for CPU temperatures of 80°C looking for an strategy to try and compensate the temperature on the CPU.

Also you made a good point. Worrying about the temperature at idle doesn't really matter since it's in the safezone.

The splitter actually made the 3 hole pin fan go full speed. So I opted per leaving them each in dif chan fan headers until I mount the 3rd forward fan.

About the CPU_Fan Header:
My board doesn't have pump header so I only plugged one of the pins on the CPU fan header. So this header will control AIO fans and pump instead of having two separated headers.

The fact Im using one header for both should I still set it PWM?

 

[Paperdoc]

I omitted a note about using Splitters when a fan is 3-pin. If you MIX 3- and 4-pin fans on one Splitter and header, you have to choose which MODE setting to make on the header. Setting it to PWM is ideal for the 4-pin fan, but makes the 3-pin fan run full speed all the time. Setting header Mode to Voltage (or DC) will give you full control of speed of the 3-pin fan. But the design of a 4-pin fan actually means it ALSO can have its speed controlled by that Voltage Mode of signals. So BOTH fans will have their speed controlled that way.

The alternative in your case, since you (will) have three case fans and three CHA_FAN headers, is to connect each fan to its own header. It appears that is what you have done. Initially configure all of those headers identically, except setting the Mode for the type of fan. Set all to default fan curves. Later you can go to the front bottom fan you add and adjust that fan curve to suit you.

IF you have changed somehow and will be using FOUR case vent fans, then arrange to have both 3-pin fans on one Splitter and header, OR both 4-pin fans on one Splitter and header. Then the header confgurations can be set for each fan type.

That AIO system comes with two different cable sets to plug into the pump unit and make mobo connections. You choose to use only one of them. One has only a connection to the CPU_FAN header, and that is the easier way to do this. The other cable has THREE arms that must go to separate headers, and allows you as user to customize cooling of three components - pump, rad fans and VRM fan (the small fan on top of the pump) - via those separate headers. I think you are using the single-arm cable to the CPU_FAN header. But in either case, whichever headers you use for this system, all the ones used for AIO cooler connections should be set to PWM - that is what this system requires. As an aside for your info, in this system no matter which way you do this, the speed reported to a mobo header is only the speed of one of the rad fans. It does not have any way to report the PUMP speed to the mobo, nor to report the speed of the little VRM fan.