Question My fan speeds are very diffrent

[Edvin fryklund]

IThe fan speeds on my fans is very diffrent. One of them is at 638rpm and another is at 4500rpm.

 

[Paperdoc]

4500 RPM is very hard to believe. Very few fans can go close to that, so the number may be wrong.

We need more info. First, how do you know what speeds? If you are using third-party utility tools to see fans, some of them tell you the wrong info until they are calibrated. I recommend you start by looking for fan speeds in BIOS Setup where the fan headers are configured. Those numbers are accurate, but you only get to see them in BIOS Setup and not when you are dong normal work. The second tool that you CAN use during normal work under Windows is a utility provided by your mobo maker on the CS of drivers, etc. that came with it. It will have a system monitring app you can run just like any other app under Windows, and that MAY already be installed on your HDD to use.

Now, also post back here this info.

1. What maker and exact model number is your mobo?
2. How many fans? Split them into two groups - how many on the CPU cooling system, and how many used for case ventilation?
3. For each fan, what maker and model number? And, where is that fan plugged in for power?

With that info we can advise.

 

[MadsModsat]

IThe fan speeds on my fans is very diffrent. One of them is at 638rpm and another is at 4500rpm.

One possible cause of the high RPM reading, could be if you are using an AiO CPU cooler, and the pump tachometer is connected to the header that reads 4500 rpm in your monitoring software.

One possible example of such an AiO, could be the Corsair Hydro H60, which supposedly has a pump RPM around - or a little below - 4500 RPM (although, if I remember correctly - the H60 doesn't report pump RPM, only fan RPM, so this is just meant as an example).

Other AiO pumps could possibly operate at similar RPM.

If the reading is not from an AiO pump, I suspect the reading is inaccurate as suggested above by Paperdoc

 

[Edvin fryklund]

4500 RPM is very hard to believe. Very few fans can go close to that, so the number may be wrong.

We need more info. First, how do you know what speeds? If you are using third-party utility tools to see fans, some of them tell you the wrong info until they are calibrated. I recommend you start by looking for fan speeds in BIOS Setup where the fan headers are configured. Those numbers are accurate, but you only get to see them in BIOS Setup and not when you are dong normal work. The second tool that you CAN use during normal work under Windows is a utility provided by your mobo maker on the CS of drivers, etc. that came with it. It will have a system monitring app you can run just like any other app under Windows, and that MAY already be installed on your HDD to use.

Now, also post back here this info.

1. What maker and exact model number is your mobo?
2. How many fans? Split them into two groups - how many on the CPU cooling system, and how many used for case ventilation?
3. For each fan, what maker and model number? And, where is that fan plugged in for power?

With that info we can advise.

In the bios it says that it is the gpu1 that has the high rpm.

 

[MadsModsat]

If the fan header says CPU_FAN1 and not GPU_FAN1, the high RPM you see would most likely be the CPU cooler pump RPM (if the 3-pin connector for the pump tachometer is connected to the CPU fan header).

 

[Edvin fryklund]

One possible cause of the high RPM reading, could be if you are using an AiO CPU cooler, and the pump tachometer is connected to the header that reads 4500 rpm in your monitoring software.

One possible example of such an AiO, could be the Corsair Hydro H60, which supposedly has a pump RPM around - or a little below - 4500 RPM (although, if I remember correctly - the H60 doesn't report pump RPM, only fan RPM, so this is just meant as an example).

Other AiO pumps could possibly operate at similar RPM.

If the reading is not from an AiO pump, I suspect the reading is inaccurate as suggested above by Paperdoc

I have the Corsair Hydro H100i RGB Platinum

 

[MadsModsat]

I have the Corsair Hydro H100i RGB Platinum

If the fan header says CPU_FAN1 and not GPU_FAN1, the high RPM you see would most likely be the CPU cooler pump RPM (if the 3-pin connector for the pump tachometer is connected to the CPU fan header - which is fairly common).

The pump rotates much faster than a large fan like the 120mm ones on the radiator or a lot of case fans, which is why the RPM can appear unusually high.


EDIT:
If it is actually the GPU_FAN1, the reading is quite high, depending on what graphics card you have. But I can't remember seeing a BIOS being able reading GPU fan RPM other than on laptops

 

[Edvin fryklund]

If the fan header says CPU_FAN1 and not GPU_FAN1, the high RPM you see would most likely be the CPU cooler pump RPM (if the 3-pin connector for the pump tachometer is connected to the CPU fan header).

The pump rotates much faster than a fan, which is why the RPM can appear unusually high

I dont know how it is connected because i didnt build the pc and i dont know much sbout pcs. The pc probably work as it should right?

 

[MadsModsat]

Yes, the PC is fine.

I'm almost certain that the CPU cooler (Corsair Hydro) has the 3-pin pump tachometer connected to CPU_FAN1 header - this is very common, and would explain the high number of RPM reported.

If any fan in your PC is actually spinning at 4500 RPM, it would be producing a lot of noise and be quite loud all the time - you would definately notice something like that

 

[Edvin fryklund]

Thank you for your help!

 

[MadsModsat]

Thank you for your help!

I'll have to correct myself.

When looking up the usual RPM of the pump on a Corsair Hydro H100 RGB, it should only be around 2200 - 2400 rpm.

So 4500 rpm is quite a lot more.

Are you experiencing any problems related to CPU temperature / overheating?

What are the fan and pump RPM reported to be in Corsair iCUE or Corsair Link software (whichever you use for controlling the cooler) ?

 

[Karadjgne]

Sometimes if there's 2x fan speed reporting software, the actual speeds can be doubled. Can also happen if the splitter used to seperate 2x fans is not a good splitter and has 2x tach wires instead of the normal master/slave arrangement.

 

[MadsModsat]

Sometimes if there's 2x fan speed reporting software, the actual speeds can be doubled. Can also happen if the splitter used to seperate 2x fans is not a good splitter and has 2x tach wires instead of the normal master/slave arrangement.

That would certainly explain it, and at the same time I learned something new today

 

[Nigel Spike]

In the first reply you were asked about the make and model of your motherboard. We still haven't been told. This becomes even more interesting when you say your BIOS reports fan speed for gpu1.

In an older post you stated that you have/had a MSI B450-A Pro. The manual for that board mentions nothing about gpu fan settings or speed in BIOS.

 

[Paperdoc]

Let me echo the skeptical comments on reports of a GPU fan speed. There is NO industry-standard way for a mobo to acquire the speed of a fan on a graphics card mounted in a PCIe slot, so most mobos can NOT tell you a GPU fan speed in BIOS Setup. Now, third-party utilities sometimes do attempt to access such info without using mobo BIOS data, and these CAN be wrong if the utility is not configured correctly. But you said you observe this in BIOS, not via some other utility.

This is NOT to say such info does not exist. I simply say there is no standard method of communicating it from video card to mobo. However, the makers of the GRAPHICS cards these days very often include on the card the ability to measure the card fan's speed (if it has one) and the GPU chip internal temperature. They also always include with the card a CD of drivers and utilities that includes a tool for monitoring the cards' status and making custom adjustments to some features, including the fan operation. Those utilities are written specifically for that graphics card and this CAN know how to use the non-standard data communiaction features of that card - something the mobo does not know. So you COULD "see" a video card speed using such a tool, but again, NOT in BIOS Setup

There ARE some exceptions to this. There are a few mobo makers who also make and sell their own graphics cards. Then they offer an unusual feature: IF you have the right mobo AND same-maker graphics card paired up, the mobo BIOS CAN access the graphics card info because it is written for the specific non-standard methods of data exchange in that pair. If you get around to telling us the maker and exact model number of your MOBO, and ALSO of your Graphics Card,we can check whether this is how you got that info.

The other possibility, of course, is that you mis-read the screen and the speed was NOT said to be for the GPU card fan.

The suggestion above that an odd speed like that could result from using a badly-designed Splitter is wrong, I believe. Virtually all fan motors generate a speed signal and send it back to the mobo header on Pin #3. That signal is 5 VDC pulses, two per revolution, and the header simply counts the pulses per second. If you were to connect the speed pulse signals from TWO fans both back to the header, that system would receive two different trains of pulses overlapping, and sure, that would deliver twice as many pulses. BUT what really happens is that the two different trains are NEVER exactly the same pulse rate, so the two trains constantly shift with respect to each other. The reult is that, for any given short time interval for counting the pulses, the number actually appears to change significantly because some of them arrive so close together the counting system cannot distinguish them. The practical result is the the speed your BIOS shows you VARIES substantially, so the speed will appear to be changing constantly over a wide range. You have not reported that highly variable reading, OP - your post suggests that high speed is stable. Is that right, or does it really change a LOT?.

 

[Edvin fryklund]

Let me echo the skeptical comments on reports of a GPU fan speed. There is NO industry-standard way for a mobo to acquire the speed of a fan on a graphics card mounted in a PCIe slot, so most mobos can NOT tell you a GPU fan speed in BIOS Setup. Now, third-party utilities sometimes do attempt to access such info without using mobo BIOS data, and these CAN be wrong if the utility is not configured correctly. But you said you observe this in BIOS, not via some other utility.

This is NOT to say such info does not exist. I simply say there is no standard method of communicating it from video card to mobo. However, the makers of the GRAPHICS cards these days very often include on the card the ability to measure the card fan's speed (if it has one) and the GPU chip internal temperature. They also always include with the card a CD of drivers and utilities that includes a tool for monitoring the cards' status and making custom adjustments to some features, including the fan operation. Those utilities are written specifically for that graphics card and this CAN know how to use the non-standard data communiaction features of that card - something the mobo does not know. So you COULD "see" a video card speed using such a tool, but again, NOT in BIOS Setup

There ARE some exceptions to this. There are a few mobo makers who also make and sell their own graphics cards. Then they offer an unusual feature: IF you have the right mobo AND same-maker graphics card paired up, the mobo BIOS CAN access the graphics card info because it is written for the specific non-standard methods of data exchange in that pair. If you get around to telling us the maker and exact model number of your MOBO, and ALSO of your Graphics Card,we can check whether this is how you got that info.

The other possibility, of course, is that you mis-read the screen and the speed was NOT said to be for the GPU card fan.

The suggestion above that an odd speed like that could result from using a badly-designed Splitter is wrong, I believe. Virtually all fan motors generate a speed signal and send it back to the mobo header on Pin #3. That signal is 5 VDC pulses, two per revolution, and the header simply counts the pulses per second. If you were to connect the speed pulse signals from TWO fans both back to the header, that system would receive two different trains of pulses overlapping, and sure, that would deliver twice as many pulses. BUT what really happens is that the two different trains are NEVER exactly the same pulse rate, so the two trains constantly shift with respect to each other. The reult is that, for any given short time interval for counting the pulses, the number actually appears to change significantly because some of them arrive so close together the counting system cannot distinguish them. The practical result is the the speed your BIOS shows you VARIES substantially, so the speed will appear to be changing constantly over a wide range. You have not reported that highly variable reading, OP - your post suggests that high speed is stable. Is that right, or does it really change a LOT?.

The speed changes alot yes.

 

[Paperdoc]

That is a lot. A single normal fan speed will vary ± 10 to 50 RPM over several seconds. This DOES suggest a faulty Splitter that is feeding TWO or more speed signals back to the mobo header. When Splitters first hit the market many were made that way (faulty design) and caused this type of issue. In some cases people did not buy Splitters, they just did custom wiring and made that mistake.

So now the issue becomes: what IS connected to the CPU_FAN header, and what is connected to that, etc? A simple Splitter looks like this

https://www.newegg.com/p/N82E16812423169?Description=coboc fan splitter&cm_re=coboc_fan_splitter--12-423-169--Product&quicklink=true

That is a 3-pin model, Note that it has ONE input end with a female connector (with 3 holes) that plugs into a mobo header. Then it has two male (with 3 pins each) output connectors to plug your fans into. It is not apparent in the photos, but on ONE of the male outputs, there is no wire coming to Pin #3.

Just for reference, the photos for this item

https://www.newegg.com/black-startech-12-others/p/N82E16812200455?Description=fan splitter&cm_re=fan_splitter--12-200-455--Product

show a WRONG design, with all THREE wires from the female going to BOTH male outputs. The YELLOW wire to Pin #3 should be missing on one of the male outputs.

Here's another example of a correct Splitter, this one a 4-pin design which can work with both 3- and 4-pin fans.

https://www.newegg.com/p/N82E16812162026?Description=fan splitter&cm_re=fan_splitter--12-162-026--Product

Note on this that the way they avoid sending back a second fan's speed signal is that one of the two male outputs is MISSING Pin #3, so it cannot connect to that fan's speed signal source.

So, is there a thing like that plugged directly into your CPU_FAN header? Are there then two (or more?) fans plugged into its output arms?

Now look at the actual CPU chip cooler, and examine the wires coming out of its motor to the connector on the end of the cable. Does that plug into the CPU_FAN header somehow, or is it plugged into a different place?

Let us know what you find and we can advise further.

 

[Edvin fryklund]

That is a lot. A single normal fan speed will vary ± 10 to 50 RPM over several seconds. This DOES suggest a faulty Splitter that is feeding Two or more speed signals back to the mobo header. When Splitters first hit the market many were made that way (faulty design) and caused this type of issue. In some cases people did not buy Splitters, they just did custom wiring and made that mistake.

So now the issue becomes: what IS connected to the CPU_FAN header, and what is connected to that, etc? A simple Splitter looks like this

https://www.newegg.com/p/N82E16812423169?Description=coboc fan splitter&cm_re=coboc_fan_splitter--12-423-169--Product&quicklink=true

That is a 3-pin model, Note that it has ONE input end with a female connector (with 3 holes) that plugs into a mobo header. Then it has two male (with 3 pins each) output connectors to plug your fans into. It is not apparent in the photos, but on ONE of the male outputs, there is no wire coming to Pin #3.

Just for reference, the photos for this item

https://www.newegg.com/black-startech-12-others/p/N82E16812200455?Description=fan splitter&cm_re=fan_splitter--12-200-455--Product

show a WRONG design, with all THREE wires from the female going to BOTH male outputs. The YELLOW wire to Pin #3 should be missing on one of the male outputs.

Here's another example of a correct Splitter, this one a 4-pin design which can work with both 3- and 4-pin fans.

https://www.newegg.com/p/N82E16812162026?Description=fan splitter&cm_re=fan_splitter--12-162-026--Product

Note on this that the way they avoid sending back a second fan's speed signal is that one of the two male outputs is MISSING Pin #3, so it cannot connect to that fan's speed signal source.

So, is there a thing like that plugged directly into your CPU_FAN header? Are ther then two (or more?) fans plugged into its output arms?

Now look at the actual CPU chip cooler, and examine the wires coming out of its motor to the connector on the end of the cable. Does that plug into the CPU_FAN header somehow, or is it plugged into a different place?

Let us know what you find and we can advise further.

What and where are the fan headers?

 

[Paperdoc]

You still have not told us your mobo. The only clue is that someone else says you had an MSI B450-A Pro in another thread somewhere. So I downloaded its manual from here

https://www.msi.com/Motherboard/support/B450-A-PRO#down-manual

If that is NOT your mobo, ignore the rest of this and tell us what IS your mobo, because I'm using that manual.

See p. 25, at the top of the board next to the top of the RAM slots. That is where the CPU_FAN1 header is. It is 4 pins sticking up, with a plastic finger beside three of the pins, and they will be covered by a connector body plugged in there. See also the diagrams on p.34.

 

[Edvin fryklund]

Yes i use the MSI B450-A Pro. There is one wire on the Fan header. What do you mean by output arms?

 

[Paperdoc]

Look back at the photos I linked to of Splitters. Most look like two "arms" of wires (loose or in a sheath) each ending in a male (with pins) connector that you can plug your fan into. Those are the output "arms" they both come from a common single input connector that has holes to fit onto the mobo fan header.

So, that wire (cable of wires inside a sleeve?) that is pluued into the CPU_FAN header. Where does it go? Does it go directly to a single fan motor? Or, does it split into two or more outputs to fans? Or, does it go to a circuit board or a box that several fans are plugged into?

 

[Edvin fryklund]

It goes to the Corsair Hydro H100i RGB Platinum. And there is only one wire.

 

[Paperdoc]

OK, that's a bit more info. The item connected to the CPU_FAN header is the SPEED signal from the PUMP of the AIO system. At this point I am going to assume the your earlier post that said you were seeing a very high speed of 4500 RPM and varying a lot for GPU1 is really for CPU1 and that is the PUMP. Further, there is no Splitter involved so it is not receiving double speed pulses to deal with. The possible cause, then, may be a poor contact where that single wire from the pump connects to the CPU_FAN header. If it is a "dirty" contact the header might receive an irregular signal that includes a lot of "static" that might appear to be extra pulses.

Try this. Reach in to that connector and carefully pull it off the header, Then just plug it back in again. Do this several times, and you may clean off the contacts. Then see if the speed settles down to a stable reading. Next, what is that reading? 4500 RPM os a lot faster than that pump is supposed to go. Although Corsair deos not give the spec for that, many commenters and reviewers says it is in the range 2400 to 2900. What do you see now?

 

[Edvin fryklund]

Its often around 3500 sometimes and it pretty constantly jumps to around 4700 and down to around 2700

 

[Paperdoc]

It's that jumping around desricption that makes me think there may be a glitchy connection that creates "static" in the speed signal, causing it to change so much and so rapidly. That is why I suggested the disconnect / reconnect process.

Just as another check, you could temprarily unplg the wire from the pump from the CPU_FAN header and plug it into another header. Then watch what THAT feader reports for the speed of its new "fan" (pump, really). If it reports just one steady speed, then you will know the problem is not a bad signal being generated by the pump. After you do this, return it to the normal connection on the CPU_FAN header.