Question Fans randomly ramp up to max rpm ?
- Thread starter kaishi00
- Start date
it's not responding to temps. i'm using some lian li SLv2 fans that are connected to motherboard headers. didn't use any hubs.
Using 7700x with asrock x670e steel legend.
AIO pump is on CHA 1/WP
AIO Fans (3 chained) is on CPU 1
Rear fan is on CHASIS 2
Bottom fans (3) CHASIS 3
Front fans (3) CHASIS 4
Not exactly overloading headers, should be well within spec.
35below0
Commendable
- Jan 3, 2024
- 1,148
- 512
- 1,590
I think you definetly need a fan hub. You are likely stressing the headers too much. 3 fan fans on one header can even damage the motherboard itself. The CPU_FAN header is more robust than system fans but it's not recommended to use more than two fans, or even one.
With a fan hub, fans draw power from the PSU and the motherbord is safe.
However, this is a separate issue and my suggestion for you. for a healthier system. I'm not at all sure if it's connected to your RPM problem.
What you can do to test this, is connect no more than one fan to each header and see if they still go to max RPM or not.
Malware scan turned up nothing?
With a fan hub, fans draw power from the PSU and the motherbord is safe.
However, this is a separate issue and my suggestion for you. for a healthier system. I'm not at all sure if it's connected to your RPM problem.
What you can do to test this, is connect no more than one fan to each header and see if they still go to max RPM or not.
Malware scan turned up nothing?
Your first post contains two intriguing items that say the issue is NOT too many fans on one header. First, when they go to full speed they stay there and do not slow down again. But MORE importantly, you say all of them are set to a fixed Quiet speed option. So NONE of the headers involved should be able to change that and send out signals for the fans to go full speed. Unless, of course the BIOS has some odd protection system that overrides any Quiet setting if the CPU gets too hot. I've never heard of that option, but even that ought NOT to affect ALL the fan headers at once.
In fact, affecting ALL headers at once would be really weird for almost any cause. Are you SURE the all do this, or only the fans on ONE header?
A possible cause for ONE header might be if the PWM signal from Pin #4 of that header somehow does NOT get through - maybe from a dirty or loose connection. In the PWM system for 4-pin fans if the fan gets NO PWM signal it WILL run full speed all the time.
In fact, affecting ALL headers at once would be really weird for almost any cause. Are you SURE the all do this, or only the fans on ONE header?
A possible cause for ONE header might be if the PWM signal from Pin #4 of that header somehow does NOT get through - maybe from a dirty or loose connection. In the PWM system for 4-pin fans if the fan gets NO PWM signal it WILL run full speed all the time.
Just for clarification, the silent profiles are a fan curve in BIOS, so the fans do respond to temperature.
I know ALL the fans went max speed because I monitored it through AIDA64, which is also why I know they aren't responding to temperature because the temps are normal.
Another quirk is after a while after restarting (the amount of time it takes to do this is random), my RPM sensor on CPU fan 1 will just up and disappear on AIDA64 (and other reporting programs). The fan (and rgb) continues to function despite sensor not being there.
Check some of these settings on the headers your fans use.
CHA_FAN1 / WP Switch set to PUMP, and Temp Source set to Monitor CPU. This ensures the AIO pump connected there will always run full speed as designed.
CPU_FAN1 you may have no options to adjust, other than setting to Silent Mode for a pre-set low-speed fan curve as you have done.
CHA_FAN2 (and 3 and 4 - set all of these the same):
All should have the Switch set to Fan, not Pump
Control Mode set to PWM assuming your fans are 4-pin type
Setting to Silent or Customize or whichever Profile you want.
Temp Source to Monitor MB, not to CPU. THIS may be why ALL your fans are doing the same thing IF you have these case fan headers controlled by the temp sensor inside the CPU.
The fact the speed reading on CPU_FAN1 disappears randomly will NOT affect the mobo's ability to control that fan. But it MIGHT be a hint about the quick change of fan speed. For any fan header that signal is monitored to detect FAILURE by getting NO fan speed. Normally the automatic reaction of the header to a failure is to change the speed setting to full speed to ensure the fan really does start up, then watch for a good speed signal. IF the speed signal is restored then the fan speed signal is returned to whatever it was before. However, if that signal is set so low that it actually caused the fan to stall, the whole process will repeat continuously. On the other hand, if the speed signal does NOT come back under a full-speed signal condition, the header will pop up a screen warning of fan failure and, in the case of the CPU_FAN headers, take more drastic action to shut down your whole system to protect the CPU from damage by overheating due to NO cooling. You do NOT report such warnings or any shut-down, so we should assume that any actual fan stall IS recovered by the automatic re-start process.
That leaves us with the question: Does at least ONE of the three fans daisy-chained on the CPU_FAN1 header actually stall and trigger this process, repeating that sequence? YOU need to watch the actual fan action, NOT just look at the screen display of fan speed, to check this. IF the one fan of that trio is somehow stalling, it is possible that the Quiet pre-set fan curve you are using is set too go too slow at low temperatures and causing the fan stall. For that you would need to alter the fan speed setting at the lowest temperatures.
Another thing to check in that group. You have three fans connected in a daisy-chain system to the CPU_FAN1 header. Only ONE of those (quite likely the first in the chain) can report its speed back to the header. So two items might cause a bad speed signal there. ONE is that the connection from the first fan back to the header is faulty or dirty. So try this with your system shut down completely and unplugged from the wall. At each end of that cable from the header to the first fan, disconnect the cable, then re-connect it. Do this several times. This action may "scrub" any dirt or oxidation from the contacts in the connectors. Re-connect everything, plug your machine back into the wall, and try how it works.
Another possibility is that the cables are OK, but the one fan whose speed signal is sent back to the header (probably first in the daisy chain) is failing to send its signal. Try re-arranging the cables between the three fans so that first fan is last in the connection sequence and a new fan of those three is now the only one sending back its speed signal. See if that makes a difference.
CHA_FAN1 / WP Switch set to PUMP, and Temp Source set to Monitor CPU. This ensures the AIO pump connected there will always run full speed as designed.
CPU_FAN1 you may have no options to adjust, other than setting to Silent Mode for a pre-set low-speed fan curve as you have done.
CHA_FAN2 (and 3 and 4 - set all of these the same):
All should have the Switch set to Fan, not Pump
Control Mode set to PWM assuming your fans are 4-pin type
Setting to Silent or Customize or whichever Profile you want.
Temp Source to Monitor MB, not to CPU. THIS may be why ALL your fans are doing the same thing IF you have these case fan headers controlled by the temp sensor inside the CPU.
The fact the speed reading on CPU_FAN1 disappears randomly will NOT affect the mobo's ability to control that fan. But it MIGHT be a hint about the quick change of fan speed. For any fan header that signal is monitored to detect FAILURE by getting NO fan speed. Normally the automatic reaction of the header to a failure is to change the speed setting to full speed to ensure the fan really does start up, then watch for a good speed signal. IF the speed signal is restored then the fan speed signal is returned to whatever it was before. However, if that signal is set so low that it actually caused the fan to stall, the whole process will repeat continuously. On the other hand, if the speed signal does NOT come back under a full-speed signal condition, the header will pop up a screen warning of fan failure and, in the case of the CPU_FAN headers, take more drastic action to shut down your whole system to protect the CPU from damage by overheating due to NO cooling. You do NOT report such warnings or any shut-down, so we should assume that any actual fan stall IS recovered by the automatic re-start process.
That leaves us with the question: Does at least ONE of the three fans daisy-chained on the CPU_FAN1 header actually stall and trigger this process, repeating that sequence? YOU need to watch the actual fan action, NOT just look at the screen display of fan speed, to check this. IF the one fan of that trio is somehow stalling, it is possible that the Quiet pre-set fan curve you are using is set too go too slow at low temperatures and causing the fan stall. For that you would need to alter the fan speed setting at the lowest temperatures.
Another thing to check in that group. You have three fans connected in a daisy-chain system to the CPU_FAN1 header. Only ONE of those (quite likely the first in the chain) can report its speed back to the header. So two items might cause a bad speed signal there. ONE is that the connection from the first fan back to the header is faulty or dirty. So try this with your system shut down completely and unplugged from the wall. At each end of that cable from the header to the first fan, disconnect the cable, then re-connect it. Do this several times. This action may "scrub" any dirt or oxidation from the contacts in the connectors. Re-connect everything, plug your machine back into the wall, and try how it works.
Another possibility is that the cables are OK, but the one fan whose speed signal is sent back to the header (probably first in the daisy chain) is failing to send its signal. Try re-arranging the cables between the three fans so that first fan is last in the connection sequence and a new fan of those three is now the only one sending back its speed signal. See if that makes a difference.
It's funny you mentioned the system might just shut down, because that also does happen, I just thought it was a separate issue. I will try redoing the fans connected to the CPU header and see if that solves anything. Or set minimum speed of the fans higher. The system is really new, just built it back in December and hasn't had issues until recently. I do have an extra set of fans so I could change the whole set out if needed.
Before you get too involved in many changes, you could try this simple test. For all the fan headers you are using, set the PROFILE to Standard Mode rather than your Silent Mode choice. The default cooling fan curve VERY likely will NOT try to set the fan speeds too low. If this setting completely eliminates the problems., you know the signals and connections etc. are good and the system WILL work if the signals sent out by the headers are reasonable. If that is what happens, you will know you just need to adjust the minimum speed settings.
So I set all the fans to standard vice silent and the problem persists. It's also always just the CPU header fan going missing, this time it took around 3 hours before it did.
edit: computer random shut down just now, about 4 hours after cpu fan sensor went poof. Wasn't even doing anything, just browsing web sites. Fans weren't going up or down speeds or anything.
Last edited:
So the problem is not with ALL fans, but only with the set of three rad fans connected (daisy chain) to the CPU_FAN1 header. Try this test to see if the problem is with that header. You have NOT used the CPU_FAN2 / WP header. Configure that header just like the CPU_FAN1 header, except you need to ensure three settings. Make sure it is set to FAN, not PUMP. Ensure the MODE is PWM Mode. And choose Monitor CPU for the Temperature sensor. Move the cable feeding those three fans to this header and see whether the problem is solved. If it runs with no problems, then the CPU_FAN1 header was the source of the problem. But if the problems continue with the second header, then the problem is in the set of 3 fans and/or the cables used for them.
I took a few minutes to just straight up change out the 3 fans while the computer was shut down. If that doesn't change anything I'll move it over to CPU 2 header.
edit: it lasted a while this time, but left it overnight and at some point the RPM sensor disappeared. Then a few hours later computer shut down. Switched to CPU 2 header.
Last edited:
This is probably the final report from me on this subject.
So after switching to CPU 2 header, it APPEARED fine, at least after a few hours. Until the computer randomly shut down again. I said screw it and installed a new motherboard and reinstalled windows. Now everything seems fine, at least for now.
I'm not completely sure if the issue was for sure motherboard related or software, but motherboard seems likely. No other hardware was changed.
ReveurGAM
Prominent
- Sep 28, 2022
- 389
- 21
- 695
Just as a note, the problem might have been software for the mobo, or another program that can control fans. They override BIOS settings. I had a really big headache with Asus Armoury Crate not only doing this but refusing to relinquish control and subverting registry edits I made that stopped the problem. Even after using Windows to unmistakable AAC, it left behind the program that controlled the fans. It was so like a Trojan that I went on the Asus forum and complained. Perhaps coincidentally, they released an update that seems to have mitigated the problems, but I've been too busy to verify that.
So the system has been running stable for 2 days now and no crashes and the fan sensors are still there. I also didn't have any mobo software installed previously (still don't, not a fan). Pretty sure the motherboard somehow just decided to crap itself somehow.
ReveurGAM
Prominent
- Sep 28, 2022
- 389
- 21
- 695
TRENDING THREADS
-
-
Question I need help updating my motherboard's BIOS ?
- Started by Everesttheraccoon
- Replies: 52
-
Question New build, but when I try to turn it on this happens (see video), and then it restarts and does it again ?- Started by Tachyie
- Replies: 41
-
Facebook
Twitter
Instagram