Login    Forum    Search    FAQ

Board index » Upgrading and Repairing Forum » Desktop PC Hardware




Post new topic Reply to topic  [ 8 posts ] 
Author Message
 Post subject: Acoustic analysis of HDD to ascertain if it's defective
 Post Posted: Sun Oct 11, 2015 12:17 pm 
Offline

Joined: Wed Mar 26, 2008 12:35 pm
Posts: 17
Location: Germany
Hello everyone, it's me again, the weirdo who writes those long posts. Do I hear people heaving a sigh of exasperation in front of their screens?

Scott, I just purchased two new hard drives because I needed more disk space. Both are HGST drives. One of them was relatively inexpensive, probably because it is not a current model, but that doesn't bother me so I bought it anyways. This inexpensive hard drive is the reason for me writing this post. It is an HGST Ultrastar 7K3000 2 TB, HUA723020ALA641; P/N: 0F12470; date of manufacture: April 2012. The datasheet can be found here.

I provisionally installed the new hard drive in an external enclosure to power it on and see if it was working properly. The enclosure is a Sharkoon Rapid-Case 3.5” eSATA. After switching on the power and letting the hard drive spin up to its operating rpm, I heard a low frequency grinding noise, somewhat similar to the sound heard during read/write activity of the drive. However, the enclosure wasn't even connected to a computer and the activity LED on the front of the enclosure wasn't flashing, which means that there was no drive activity and that the drive was just idling. (For the record: the grinding noise is also present when the enclosure is connected to a computer, either via USB or eSATA, and the drive is idling.) The grinding noise was also present after I installed the hard drive in another external drive enclosure (Raidon GT1660-SB3), so that excludes the enclosure firmware as a potential culprit.

I should probably mention two things: 1. I read the drive's SMART data with CrystalDiskInfo and it shows that the drive is new and wasn't used previously. 2. I ran HGST's Drive Fitness Test for Windows and executed the quick test and the drive passed the test without any issues. There is also the option of running an extensive test, but I haven't done that yet.

I didn't like the sound this drive was making, so I decided to try and record the noise that I suspect shouldn't be there and make it available for download in the form of an audio file.

My recording setup is somewhat cumbersome, but nevertheless I think it is effective and the resulting audio is of sufficient quality. I pressed the sound input side of a Sony ECM-TL3 earphone style microphone against one of the ear pads of a Lisle mechanic's stethoscope (Lisle part # 52500). The Sony ECM-TL3 was plugged into my Sony ICD-UX522 digital voice recorder. The microphone sensitivity was set to "high". I set the recording quality to "LPCM 44.1 kHz/16bit", which creates a WAV file. During recording, the top of the hard drive enclosure was contacted at different points with the stethoscope's metal rod, so that I could later choose the portion of audio where the noise could be heard most clearly.

I realize that this is probably not the best way of recording the sounds a hard drive makes. After doing some research, I believe it would be better to use a homemade contact microphone made with a piezo disk. See this Youtube video to see what I'm talking about. I might make one of these contact mics in the future, but right now I don't have the time or the necessary parts available.

Since it apparently isn't allowed to upload WAV or MP3 files as forum post attachments, I have made the WAV files available under this link: WAV files of sounds made by HGST hard drives.

Scott, when I tried to upload a WAV file as an attachment to my forum post, I received an error message telling me that “The extension wav is not allowed.” I then converted a WAV file to MP3 to see if I could upload that. I couldn't. I received an analogous error message: “The extension mp3 is not allowed.” Why aren't these file types allowed to be uploaded?

I have provided two WAV files of the HGST Ultrastar 7K3000. On one you can hear the drive spinning up, on the other you can hear the drive idling. The files are named accordingly. For comparison purposes I have also provided some other audio files. There are two WAV files of my other newly purchased hard drive, an HGST Ultrastar 7K4000, which is installed in a Raidon GT1660-SB3 external HDD enclosure. There are also two WAV files of one of my old Deskstar E7K1000 drives, also mounted in an external enclosure (Icy Box IB-366StUS2-B), which has a power on time of 87 hours and a power on count of 141. This is the hard drive I use for back-ups and therefore it hasn't been used much, despite its age. Both of these drives do NOT make the grinding noise that is present in the Ultrastar 7K3000.

Some people might think that the audio recordings would be better (i.e. the sound clearer) if the top cover of the hard drive is contacted directly with the metal rod of the mechanic's stethoscope. However, I made several recordings while contacting the drive directly with the stethoscope and the results weren't better; if anything, they were slightly worse than when I contacted the top of the enclosure with the stethoscope. I also didn't want to contact the top of the drive directly, because even though the top cover appears to be made of quite rigid sheet metal, I didn't know if the light pressure that would be applied to the cover with the stethoscope's metal rod could cause damage or at least have some kind of momentary detrimental effect on the drive, even if it would only be a slight change in the drive's acoustic signature.

The HGST Ultrastar 7K3000 (HUA723020ALA641) has built-in bulk data encryption, and I pondered whether or not this could be the reason for the noise the drive is making during idle. However, after reading some of the FAQs about bulk data encryption on the HGST web site, I have come to the conclusion that the bulk data encryption feature has nothing to do with the noise the hard drive is making, because encryption/decryption of data is done on the fly by the drive's hardware only when data is being written to or read from the drive. So it's not like the drive needs to encrypt itself initially after the first power on.

Scott, the reason for providing the audio files is that I hope you will listen to the noise the HGST Ultrastar 7K3000 is making and tell me if you think that this drive is defective or if the noise is a precursor of hard drive failure. I have zero experience in analyzing the acoustic signatures of hard drives, and I know you are vastly experienced in all things computer hardware related, so I'm hoping you can help me out. If there is even the slightest suspicion that this hard drive has an issue, I will send it back to the online vendor I bought it from.

I used a free audio analysis program called Praat to do some simple analysis of the recorded WAV files of the idling hard drives. Praat has a somewhat unorthodox GUI, so you'll have to familiarize yourself with its user interface logic before you can do something useful.

What I did was create spectrum objects of the sound objects (in Praat, sound files are sound objects) by choosing “Analyse spectrum”->”To Spectrum...”, then I viewed the spectrum objects by choosing “View & Edit”. In the spectrum window you can zoom in and out to the frequency ranges you wish to view. You can then play back only the sound generated by the frequencies currently visible in the window by choosing “View”->”Play window” from the menu. What I found very useful for comparison purposes is to open two spectrum windows side by side, one of the 7K3000 and one of the 7K4000. Then I zoomed in to the frequency range of interest, e.g. 340 to 455 Hz, and played back the sounds of the two windows in quick succession to compare the noise signature. A great feature is that the zoom and scroll of one spectrum window is mirrored in the other spectrum window. You can turn this behavior off if you wish, by choosing “File”->”Preferences...” from the menu and deselecting the option “Synchronize zoom and scroll”.

What I found is that the grinding noise is made up of frequencies in the range of 340 to 455 Hz and to a lesser extent 750 to 1050 Hz. When analyzing the E7K1000 and 7K4000 idling WAV files using Praat as described in the preceding paragraph, it became clear that these other two drives have much less sound energy in these frequency ranges. From that I deduce that if two HGST drives, one older and one new, don't have the low frequency grinding noise, then this noise shouldn't be there and that the Ultrastar 7K3000 is defective. In auto tech terms, the Ultrastar 7K3000 noise signature is analogous to an engine that is running rough, the E7K1000 and 7K4000 noise signatures are analogous to an engine that is running smoothly.

I am convinced that the Ultrastar 7K3000 hard drive I purchased is defective. Scott, what do you think?


Kind regards,

Daniel


Last edited by Daniel on Sun Nov 20, 2016 11:03 am, edited 2 times in total.

Top 
 Profile  
Reply with quote  
 Post subject: Re: Acoustic analysis of HDD to ascertain if it's defective
 Post Posted: Sun Oct 11, 2015 12:48 pm 
Offline

Joined: Thu Jan 15, 2009 1:27 pm
Posts: 1105
Location: Stowmarket, Suffolk England
Before Scott replies, perhaps run the longer drive test, or even fill it up with copies of your existing files and then run tests?


Top 
 Profile  
Reply with quote  
 Post subject: Re: Acoustic analysis of HDD to ascertain if it's defective
 Post Posted: Sun Oct 11, 2015 9:05 pm 
Online
Site Admin

Joined: Sun Feb 04, 2007 11:44 am
Posts: 5848
I did play the 7K3000 wav files and don't like what I'm hearing. Different drives do make different "normal" sounds so to satisfy my usual curiosity I would want to compare it to another of the same model. I always buy drives in duplicate or even triplicate so I'd compare it to the others. If they all make the same sounds then I'd consider it normal for that model.

However in this case I just can't believe that noise is normal. In that case I would either 1.) use it but treat it as if it might fail at any moment (i.e. keep it backed up religiously)*, 2.) use it for disposable data, or 3.) send it back.

*After writing that I realized that I treat all drives this way. Once you treat all drives as if they will fail at any moment the sounds they do or do not make are immaterial. However I would add that if I ever hear the sound change on a drive I do take that as a sign of an impending failure and a cue for an immediate backup. Scott.


Top 
 Profile  
Reply with quote  
 Post subject: Re: Acoustic analysis of HDD to ascertain if it's defective
 Post Posted: Thu Oct 15, 2015 11:14 am 
Offline

Joined: Sun Jul 20, 2008 10:50 am
Posts: 508
Location: Phoenix, AZ, USA
I bought the same drive for a client (for $57.50) from Newegg though their 3rd party program. I encourage you if possible to get your money back immediately from the vendor if it is still in their return period.

The new drive was as you say, already 3 yrs into its warranty which ticked me off. But because the vendor is not registered with Lenovo they will not honor the remaining two years (which I found out later when I started exploring replacement options). For me it was cool running and not noisy. It also passed preliminary testing via Hitachi's DFT.

The particular drive gave my client warning via Windows (64-bit Win7 Pro) of imminent failure starting about 3 weeks after I built the system. I went over and downloaded the error events. Ran Hitachi's DFT for Win on it and it passed the SMART and short tests.

Pulled the drive because it was a quarter full (with no backup per client's refusal) and proceeded to recover it with ddrescue; in itself an adventure (first time it actually worked for me (26 hours; 347 errors skipped) on a drive other rescue apps failed on).

Set up the RMA procedure that day which Newegg eventually honored thankfully with a refund rather than a "new" drive form the same source.

BTW, the drive failed about a quarter of the way through zeroing out in preparation for shipping. It flat out died. I hope it wiped everything on the platters before it died (and suspect it did), but....


Top 
 Profile  
Reply with quote  
 Post subject: Re: Acoustic analysis of HDD to ascertain if it's defective
 Post Posted: Sun Oct 18, 2015 3:59 pm 
Offline

Joined: Wed Mar 26, 2008 12:35 pm
Posts: 17
Location: Germany
Scott, thanks for having a listen and giving me your expert opinion. That gives me more confidence in my decision to send the drive back to the vendor.

I did purchase a second drive for backup purposes. This is a 3 TB HGST Ultrastar 7K4000. The reason for buying a larger drive for backup is that in case I should purchase an SSD (I wouldn't purchase anything larger than 500 GB) in the future, the backup hard drive would still be large enough for backing up both the SSD and the 2 TB hard drive. Windows and applications would be installed on the SSD, the 2 TB hard drive would be used for storing “user data”.

I also e-mailed a well-known hard drive expert – who wishes to remain nameless – regarding the noise and he said that the idling noise of the HGST Ultrastar 7K3000 is sufficiently different than the idling noise of the HGST Ultrastar 7K4000 to make him suspicious of the drive. He also said that my sound files seem to be “colored” by the enclosure or my microphone placement, making the sound itself “suspicious”.

FascistNation, I really appreciate your information which really helped me in making up my mind to send the hard drive back to the vendor. Thanks for taking the time and effort to write your informative post.

David, I did run the extensive WinDFT (HGST's Drive Fitness Test for Windows) test as you suggested. However, the test passed without any issues. I suspected this would be the case. However, I am not convinced that this test is a reliable indicator that the drive is defect free. That is why I initially didn't bother running this test.

After running the extensive WinDFT test yesterday, I put my hand on top of the enclosure that the drive is installed in. Not only does the drive make the described grinding noise, there is also a clearly perceptible vibration that my other HGST hard drives, including my new HGST 7K4000, do not have.

I am convinced that there is something wrong with the HGST Ultrastar 7K3000 I purchased. I have already invoked the return process and will be shipping the drive back to the vendor tomorrow. I will probably purchase a 2 TB HGST Ultrastar 7K4000 as a replacement.


Top 
 Profile  
Reply with quote  
 Post subject: Re: Acoustic analysis of HDD to ascertain if it's defective
 Post Posted: Sat Oct 24, 2015 9:16 am 
Offline

Joined: Sun Jul 20, 2008 10:50 am
Posts: 508
Location: Phoenix, AZ, USA
If you are interested: Clips of drives captured to audio (mp3). These are kind of fun but after a while wears thin.

http://www.drivesaversdatarecovery.com/ ... very-tips/ - see flash box for audio sounds (educational)

http://files.hddguru.com/download/Hard% ... g%20Noise/ - really old link with really old drives (surprised the links still worked)

http://datacent.com/hard_drive_sounds.php - old link surprised it still worked)

------

[url=null]http://www.filestube.com/ebde208c28bc3b8803ea/details.html[/url] (Hitachi drive mp3) - I'm afraid this is now a dead link that redirects to an ad site


Top 
 Profile  
Reply with quote  
 Post subject: Re: Acoustic analysis of HDD to ascertain if it's defective
 Post Posted: Fri Oct 30, 2015 5:54 pm 
Offline

Joined: Wed Mar 26, 2008 12:35 pm
Posts: 17
Location: Germany
FascistNation, thanks for the links to the noises made by defective hard drives. I found them educational. It would be great if someone compiled a more comprehensive collection of the noises defective hard drives make.

Unfortunately it was not possible to get a refund for the hard drive that made the noise described in my initial post. So I received another HGST HUA723020ALA641 in exchange for the one I returned.

I checked the SMART values of the replacement hard drive and power on hours was 0 and the power on count could only have been between 0 and 2. (I failed to check the power on count directly after powering on the hard drive for the first time, but nevertheless I'm certain that this value was either zero or very close to zero.)

However, after doing some further research, I read in an Amazon.com review of the HGST HUA723020ALA641 that someone who had purchased five of these hard drives had gone further than just reading the SMART values of his hard drives to check if they were new. The reviewer explained that despite the SMART value for power on hours being zero when he received the drives, he also checked the SMART self-test logs using smartctl from smartmontools.org. The self-test log list showed that a self-test had been performed after the hard drive had been used for thousands of hours!

GSmartControl is a GUI for smartctl and is included in PartedMagic. I fired up PartedMagic and ran GSmartControl. I selected the HGST HUA723020ALA641 from the drive list by double-clicking on it and in the subsequently appearing window selected the “Self-test Logs” tab where the SMART self-test logs are displayed. There were 5 log entries, which are shown below:

Code:
Test #          Type                   Status              % Completed     Lifetime   LBA of the
                                                                           hours      first error

  1        Short offline      Completed without error          100%          29            -

  2        Extended offline   Completed without error          100%           5            -

  3        Short offline      Completed without error          100%           0            -

  4        Short offline      Completed without error          100%         15659          -

  5        Short offline      Completed without error          100%         15356          -


Test numbers 1 through 3 were initiated by me. So no problem there.

But where do test numbers 4 and 5 come from???

Now if I understand this log list correctly, the hard drive – which I purchased as new – has actually been used for at least 15659 hours. At some point after that, the SMART values were apparently reset making the drive appear as if it hadn't been used, i.e. that the drive was in new condition.

Scott, is my interpretation of the SMART log list correct? I don't see how else one could interpret the log list. The log entries for test numbers 4 and 5 could only have been generated by the hard drive on which the SMART log list is stored; any other interpretation just doesn't make sense, in my opinion. So Scott, do you agree that I have a used hard drive here, that has already been used for at least 15659 hours?

Regarding the SMART values, is it actually possible for someone with the right hardware and/or software to reset them? Judging from the SMART log list this must be possible. I did do an Internet search regarding this question, but couldn't find any information that comes from a reliable source.

If it is possible to reset the SMART values, how come the SMART log list can't be erased as well? It would appear that the SMART log list is stored in some kind of WORM (Write Once Read Many) memory, as opposed to the SMART values themselves.

I would like to mention that the replacement hard drive also has the low frequency grinding noise that the first hard drive had. However, in this drive, the noise isn't present all the time. Instead, the noise starts after writing a substantial amount of data to the hard drive. In my case “substantial” means 726 GB of data. So after writing 726 GB of data to the hard drive, the low frequency grinding noise was present. The noise vanished after letting the hard drive idle for about one hour. After that I initiated a secure erase which uses the drive's internal secure erase command to overwrite all of the data on the drive. Interestingly enough, after the secure erase had finished, the grinding noise was not present.

I then repeated the process: Wrote 726 GB of data to the hard drive. After that the grinding noise was present again. This time the noise vanished after letting the hard drive idle for two hours. Again I initiated a secure erase. After the secure erase had finished, the grinding noise was not present, just like after the first secure erase.

Another interesting observation was how the grinding noise affected the temperature of the idling hard drive. While the grinding noise was present, the temperature of the idling hard drive was 50 °C; without the noise present it was only 44 °C.

If my interpretation of the SMART log list turns out to be correct, and I'm quite sure that it is, I will then have the necessary leverage from a legal standpoint to demand a refund for this used hard drive which I purchased as new.

Good grief, I purchased this hard drive because it was relatively inexpensive, but this is really turning out to be a false economy!


Top 
 Profile  
Reply with quote  
 Post subject: Re: Acoustic analysis of HDD to ascertain if it's defective
 Post Posted: Sat Oct 31, 2015 3:54 pm 
Online
Site Admin

Joined: Sun Feb 04, 2007 11:44 am
Posts: 5848
I've got a big project going on right now so unfortunately I can't respond in my usual detail, however at this point since the replacement drive also sounds the same I would say the drive sounds are normal for this model. Years ago I thought I had a problem drive as it would periodically make a clicking noise, however that turned out to be the drive performing thermal recalibration for increased accuracy and performance.

Note that interpretation of SMART values are often vendor specific, and since the drive you reference in the Amazon reviews reports virtually identical numbers I'd say you are mis-interpreting them and that the drives are not nearly 2-year old with only the power-on hours reset as you might think. However I will say that it is normal if not standard to get a refurbished drive as a replacement when sending a drive to the mfr. in under warranty.

At this point I'd use the drive as I would any other, and that means keeping it backed up if you care about the data. Scott.


Top 
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
 
Post new topic Reply to topic  [ 8 posts ] 

Board index » Upgrading and Repairing Forum » Desktop PC Hardware


Who is online

Users browsing this forum: Google [Bot] and 5 guests

 
 

 
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to: