Seagate HDD clicking noise and HWinfo

FrozenPhoton

New Member
Hello, I have Hwinfo constantly running in the background on my PC as it feeds the data for needed for my modern gadgets rainmeter skin which I love having in view at all times. Recently, I've noticed that two of my Spinning HDDs are making a clicking noise that sounds like the drive head parking every ~3 seconds, like some kind of digital heartbeat inside my PC. This was driving me insane as I couldn't find the source of it for WEEKS of googling until I came across this LTT forum post that helped me ID the issue as originating from HWinfo (they cite AIDA64 in the thread, but both must be doing the same thing under the hood).

So a bit more detail, I have 4 spinning HDDs in my system:
1x WD 3TB (WDC WD30EZRX-22D8PB0)
1x Seagate 1.5TB (ST1500DL003-9VT1)
2x Seagate 1TB (ST31000528AS)

By removing the SATA power cables one by one, I pinned down the issue to only be occuring with only my 2x 1TB seagate drives from the list above, not the other two. Shutting down HWinfo stopped the noise. Going further into the settings, I realized that I could still get HWinfo to run and not poll the drives constantly by going into Settings -> Safety -> IDE/(S)ATA drive scan and chaing the radio button to "Disable Drive Scan".

While this stops the noise, I no longer have access to the temperature sensor within the drives for my Modern Gadgets plugin. Below are screencaps with HWinfo set to Safe Mode (left) and Drive Scan Disabled (right). Low-Level IO is no different from Safe Mode with respect to the temp sensor and clicking noise.

Drive Scan disabled.PNGSafe Mode.PNG

Alternatively, I found via trial an error that disabling S.M.A.R.T on the drive (leaving the Drive scan in safe mode) stops the noise from occuring, but still prevents access to the HDD temp sensor on ALL my drives.

I'd like to still have access to this sensor, but at a much lower polling rate (Say, every few minutes rather than every 2-3 seconds). Could this feature be user defined like the screencap from AIDA64 in the LTT thread I linked above? Or perhaps is there a way to obtain information from the drives in a passive way, as is the case for my other 1.5TB and 3TB drives that don't seem to park the read head constantly when pulling data from the sensor. Or could it be defined on a drive-by-drive basis? I know this is partially out of your hands and a question for Seagate with respect to what is different about these drives, but looking up this issue in past forum threads its come up before with Seagate drives (I think):

Example 1
Example 2

I've tried disabling the CSMI SAS support box (as I think these two are SAS drives) but that didn't do anything.

Apologies if my question has been asked in the past, but I'm just trying to get to a state where I can monitor my drive temps (at least for the drives that are NOT having this issue) and not deal with this insane clicking noise

Thanks for reading this far!
 
Unfortunately this issue is caused by polling the SMART data which includes the temperature value. For some unknown reason the drive firmware performs operations resulting in such effects, while it should just provide actual status data and not affect drive operation. So I believe the problem is in the drive firmware.
To reduce disk polling frequency you can adjust the "Disk S.M.A.R.T. every n cycles" setting in sensor settings of HWiNFO.
 
Thanks for your response!

tbh I didn't know there was a "sensor settings" menu in the program :oops:. Now that I found that I've got it checking smart status much less often and that should be fine for me.

I can try and get ahold of seagate support to see if there is a firmware update for those drives - but adjusting the timing is a good enough fix.

Thanks again!
 
Unfortunately this issue is caused by polling the SMART data which includes the temperature value. For some unknown reason the drive firmware performs operations resulting in such effects, while it should just provide actual status data and not affect drive operation. So I believe the problem is in the drive firmware.
To reduce disk polling frequency you can adjust the "Disk S.M.A.R.T. every n cycles" setting in sensor settings of HWiNFO.

Same use case as OP. Found this after days of troubleshooting disk noise and failure of spinning disks to sleep when otherwise idle.

I don't believe this is a firmware bug, as I have 3 drives by 2 different manufacturers, and they all exhibit the same issue when HWiNFO is running (which is all the time, by default).

In my opinion, SMART checking should be disabled by default. Running SMART monitoring every second or two is potentially crushing HDD lifetimes in two ways: First by preventing sleep, and second because it resets the drive heads. This exponentially increases wear and tear on the moving parts. Maybe default to once a week, or once a day, or display an option at install-time, but "off by default" is really the safest option to avoid harm to users' devices.

In any case, if monitoring can't be done without interfering with normal/expected disk operation, then it's really important to make that clear to users I think.

Thanks for your hard work on HWiNFO -- it's a great utility and I appreciate it very much, and by no means do I mean to impugn the program as a whole, but I think this is a fairly serious issue that many (most?) users are unlikely to identify or diagnose on their own.
 
Just wanted to add, you can visualize the heads resetting on these graphs with & without HWiNFO running, with SMART polling at 1000ms intervals. (This implies that performance is impacted as well, but the more important issue is that the heads are resetting IMO.)
 

Attachments

  • HDTachLongWithoutHWiNFO.PNG
    HDTachLongWithoutHWiNFO.PNG
    34.3 KB · Views: 9
  • HDTachLongWithHWiNFO.PNG
    HDTachLongWithHWiNFO.PNG
    38.4 KB · Views: 8
Can you please check if other tools (i.e. CrystalDiskInfo) are affected as well?
 
CrystalDiskInfo won't even go below a 1 minute "auto-refresh" interval (as they call it), but yes, it prevents the drives from spinning down if the auto-refresh interval is shorter than the idle timer, and wakes (spins up) the drives if the interval is longer than the idle timer, or if the info is refreshed manually while the drives are "sleeping".
 

Attachments

  • CrystalDiskInfo.PNG
    CrystalDiskInfo.PNG
    61.8 KB · Views: 5
Back
Top