Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Solved Fans on Asus Z97-A
#1
Hi

Just got this board and I've found that HwInfo is only showing CPU fan and Chassis 4, when I've got a fan attached to Chassis 3 as well, which Asus' AI Suite and the BIOS do show. The readings that HwInfo is showing seem to be a bit off as well.

[Image: Fans.jpg]

EDIT: There's also some discrepancy between CoreTemp and HwInfo for the Current and Max core temps.

[Image: Core_Temp_Hw_Info.jpg]


Attached Files
.zip   HWiNFO64.zip (Size: 72.29 KB / Downloads: 2)
Reply
#2
According to the dump, the Chassis 3 should be shown. Is it maybe hidden, or on a different position?
You might try to "Reset Preferences" to see if that maybe fixes it.
Reply
#3
Ah yes, it was right down the bottom under the NIC for some reason.

Any thoughts on the discrepancies in Core temps and fan speeds compared to the other programs?
Reply
#4
I'm afraid, I don't know why the fan speeds differ. HWiNFO should be reporting the correct fan speed measured by the sensor. I'm not sure about ASUS though. Any idea how to prove who's right? Wink
It's possible there are small temperature differences between various tools. The internal (Core/DTS) temperatures can change very quick (within a fraction of second) and it depends when the particular tool reads the temperature. For example default HWiNFO scan interval is 2 seconds, so it's theoretically possible that the higher core temperature was only a quick peak, which wasn't captured. You might try to reduce the "Scan Interval" in HWiNFO.
Reply
#5
(07-12-2014, 09:17 PM)Martin Wrote: I'm afraid, I don't know why the fan speeds differ. HWiNFO should be reporting the correct fan speed measured by the sensor. I'm not sure about ASUS though. Any idea how to prove who's right? Wink
It's possible there are small temperature differences between various tools. The internal (Core/DTS) temperatures can change very quick (within a fraction of second) and it depends when the particular tool reads the temperature. For example default HWiNFO scan interval is 2 seconds, so it's theoretically possible that the higher core temperature was only a quick peak, which wasn't captured. You might try to reduce the "Scan Interval" in HWiNFO.

I'd imagine the ASUS tool would be able to accurately read the sensors on the ASUS board but I wouldn't put any money on it Wink I just checked again actually and HwInfo and AI Suite are reporting pretty much the exact same fan speeds at the moment. I did tweak the fan profiles in the BIOS a bit, so perhaps that helped some how Confused

I did try setting the 'Scan Interval' down to 500ms, which seems to be more frequent than CoreTemp but still HwInfo reports current and max. core temperatures 6-8c lower than CoreTemp, as you can see in these two screenshots, which is quite significant (ignore any Min. temp discrepancies, they'll mainly be because I started the tools at different times):

[Image: New_Fans_Tweak1_Aida64_44_38_ambient_28c.png]

[Image: New_Fans_Aida64_44_1_225v_ambient_28c.png]

At idle however, things swap round and HwInfo reports higher readings than CoreTemp!

[Image: New_Fans_Tweak1_Idle_afte_Aida64_ambient_28c.png]

I've also noticed the Vcore0-3 readings seems to be a bit nonsensical, often reporting 0.000 V, whilst the VID readings stay steady at 1.22x V:

[Image: Hw_Info_Idle_Vcore0_3.jpg]

HwInfo has recently started getting stuck for a while at 'Detecting Sensors: EC #0', so does it matter if I disable EC Support? I recall it was popping up a warning about something when launching before (which it isn't doing now) but I always told it not to disable whatever it was warning me about and I wasn't seeing this delay until the last day or so.


Attached Files
.zip   HWiNFO64.zip (Size: 37.38 KB / Downloads: 0)
Reply
#6
Hmm, that's quite interesting. A few degrees difference on Core temperatures would be OK. Can you maybe check if CoreTemp has some kind of offset set, or to compare with another tool ?
As for the EC sensor, yes, try to disable it and check the results then. You'll loose a few values from sensors, but the system response should be better when not reading values from the EC.
Reply
#7
OK, tested with HwInfo, HwMonitor and CoreTemp all running. Strangely, now all three seem to show the same, more or less.

This is idling:
[Image: Hw_Info_Hw_Monitor_Core_Temp_Idle.png]

Whilst the tools all seem to take their samples at different times, so one might show 58c whilst the other shows 71c, like this:

[Image: Hw_Info_Hw_Monitor_Core_Temp_Aida64.png]

at the next sample these might swap over:
[Image: Hw_Info_Hw_Monitor_Core_Temp_Aida64_b.png]

and they all show about the same max temps, which is most important. This is at stock clocks/voltage, so I'll have to retest when overclocked as it was before.

I note the Vcore readings are accurate when the CPU is loaded, it's only at idle that they show 0.000v, even though VID shows 0.719v. At idle, HwMonitor shows Vcore: 1.760v, VIN3 0.840v and VIN4: 0.080v, which HwInfo shows as VCCIN, VCCSA and Vcore, so maybe some things are mislabelled?

EDIT: By the way, running all three tools isn't advisable, as it repeatedly made my PC unusable when I closed and restarted CoreTemp whilst the other two were running and I had to hard reboot each time!
Reply
#8
I believe HWiNFO is correct with those labels.
Reply
#9
(07-22-2014, 04:59 PM)Martin Wrote: I believe HWiNFO is correct with those labels.

VCCIN matches what the BIOS shows for CPU Input Voltage and VCCSA matches CPU System Agent Voltage, so those seem fine.

I'm not sure if CPU Core Voltage is considered the Vcore or VID but when I set it to 1.225v when overclocking it shows as Vcore 1.232v in the BIOS (on Auto it shows Vcore 0.9v as I recall) and in HwInfo it appears to be reflected by VID, as under load that goes to what I've set for CPU Core Voltage, whilst Vcore is often a bit lower. In the last screenshot, Core #3 VID is 1.176v whilst Vcore 3 is 1.104v (that was when I wasn't overclocking though).
Reply
#10
Vcore measuring on Haswell is tricky, since the voltage regulator is integrated inside the CPU (FIVR). Some mainboard vendors claim they can measure per-core Vcore values, but these values seem unrealistic especially when the core is in a low-power state. Though I'm attempting to show sensor values as they are reported by particular mainboard vendor's tools.
Reply
#11
(07-22-2014, 05:34 PM)Martin Wrote: Vcore measuring on Haswell is tricky, since the voltage regulator is integrated inside the CPU (FIVR). Some mainboard vendors claim they can measure per-core Vcore values, but these values seem unrealistic especially when the core is in a low-power state. Though I'm attempting to show sensor values as they are reported by particular mainboard vendor's tools.

OK, I guess it's not really important as I can see the VID anyway Smile
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)