Western Digital's quiet 1000 EADS but 640 AALS does Doo Wap

Silencing hard drives, optical drives and other storage devices

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
danielwritesback
Posts: 6
Joined: Fri Jul 31, 2009 12:31 am
Location: Kansas

Western Digital's quiet 1000 EADS but 640 AALS does Doo Wap

Post by danielwritesback » Fri Jul 31, 2009 1:35 am

Well, guys, I do have a question, but of course, the action packed story is first. ;)

Noticing that I had the need for some safe and centralized storage, I spotted an ECS atom board. . . and was intrigued because there's no Realtek aboard. I already owned a smallish case and an APFC power supply. Its the little Forton supply and mostly quiet.
Because of heat, I replaced a older maxtor with the 1TB EADS (the Fast green drive).
Also, the power supply is on the bottom of this mini server. Very slow fans blow air into the case. The power supply cools only itself.

Due to luck or the Intel Ich7, my Western Digital green drive didn't do the parkparkparkparkpark thing at all. Just in case, there's a utility named WDIDLE3 (a search term) and newer drives are reported to respond to: Wdidle3 /S255
Factory default for laptop drives is Wdidle3 /s80
and the disable command for older drives is Wdidle3 /D.
I'm unable to test this for you because mine doesn't exibit the nervous parking.

This drive idles quietly and runs cold. Nice. However the seek noise was obnoxious. I used the Hitachi Feature Tools and that slowed it terribly.
Next, I turned off AAM. Then I set AAM to Maximum. Lastly, I set AAM to 192. Check out 80 megabytes per second and no noises. ;)

For folks wishing to use this in a raid, the western digital software is named WDTLER. That software turns off the inbuilt error compensation and allows the Raid card to control it. The methodology of setting this is counterintuitive.
For multi-drive raid setups, TLER On knocks out the hard drive's inbuilt error compensation in 7 seconds or less in order to allow the RAID controller to perfom this task, its the RAID setting and not for use otherwise.
A desktop PC uses TLER Off, which allows the drive an endless amount of time to perform error correction, lasting until it has succeeded in that important task. The two drives listed here come with TLER OFF (desktop PC setting).



--next drive--

My 640 gigabyte AALS is part of a test to see about making a hot-rod version of the mini server, for desktop PC use. AALS and the EADS are so similar. Its the AALS drive that was chosen for the desktop build because of its extra speed.

Noise is where they're not similar. AALS seeks quiet out of the box. There's no grinding. Instead, it makes the most obnoxious random sounds, like caterwauling "Whooo Hoo!!! Meow? Balarbababa?? Shoo Be DO!!!" yes, just like house cats attempting to sing Doo Wap. lol!
Its nuts!
I'd sure appreciate some help with that one.
Sure, I'll try the Hitachi Feature Tool (for Western Digital, first save AAM at maximum, and then save 192), because I haven't tried it yet.

For the acid test on reliability, this AALS is connected to a Nforce 4xx series onboard sata. The bios section for it is on factory default. Noticably, it will error slightly if the raid feature is disabled.
On factory default there is no error. That's rare for Nforce 4xx.
Quality control = passed. ;)
I hope its not yet more Nvidia+AM2 inspired glitches that cause the noises. In any case, there aren't any data errors.

The heat output difference between AALS (black with 32mb cache) and EADS (green with 32mb cache) is insignificant. Therefore the power consumption is insignificant as well. Write speed of AALS is 89MBs while EADS is 80MBs and both are truly excellent figures.
Sure, EADS does run cooler and that's helpful if you want to go fanless. EADS (green with 32mb cache) needs no assistance in cooling.

As for sounds, there aren't much. We're quite sensitive to sounds here at FidelityForce (handmade hifi audio). The EADS is a lower pitch, which is more seemly. Fanless implies lots of passive vents, and so the green drive's quiet and unobtrusive sound is preferred.

No solution yet for the 640 gigabyte AALS (Caviar Black with 32mb cache) "Yippie! Meow? Doo Wap!" type sounds, although they're not loud--they're very quiet. Its the freakiest noise to be coming out of a computer. Its also random. The sound seems to be a hetrodyne of some sort.

Its at it again.
At first, this was scary. Now that the drive has passed the 10 day mark, its not so scary. Its more funny than annoying, but we really mustn't have extraneous sounds in our shop. Here, we build audio equipment the old fashioned way, using ears. Those must not become accostomed to odd sounds, therefore the shop must have either music or silence (nothing else).

Planning to try:
Intel Ich8 + G31 + Conroe
Hitichi Feature Tool AAM at 192

Did I miss anything?
And really, how do I stop the Doo Wap?
Thanks!!

whiic
Posts: 575
Joined: Wed Sep 06, 2006 11:48 pm
Location: Finland

Post by whiic » Fri Jul 31, 2009 2:47 am

Of course 1TB Greenpower is quieter than 640GB Caviar Blue. The description of noise is however... unique. It's ok to be humorous when ranting about products, but if you honestly are seeking for help on a specific noise issue, you should describe it more accurately. Or possibly you were already as accurate as possible and that there is a striking similarity in noise produced compared to a house cat attempting to sing Doo Wap. It may be natural for you to compare the noise produced with music but most people would be interested in how low/high frequency it is, is it continuous, how it varies, is it broadband (hissing) or tonal and it there sharp clicks.

"Therefore the power consumption is insignificant as well. Write speed of AALS is 89MBs while EADS is 80MBs and both are truly excellent figures."

Well, Samsung F2 5400rpm has up to 105 MB/s on outer tracks, so ... at least by todays standards that 7200rpm drive is already pretty slow.

Not saying you should update both of them, as "enough" is enough and exceeding needed performance gives little bonus. If 80 MB/s is enough, then you should keep whatever drive sounds pleasant... and this means WD10EADS.

That WD6400AALS probably has to go. What your description sounds like is a very aggressive offline scan. Typically offline scan has to be initiated by the system and it doesn't keep doing that until eternity. There's also some offline scans that are initiated by HDD's internal logic. Be vary that these may either be present in all models using said firmware, or it may indicate that firmware believes drive is about to die and prepares to run routines to prolong it's usable life by scanning and remapping bad sectors before they become so bad that they can no longer be recovered.

You can check the health of the drive by reading SMART values. For ease of use I recommend running HDDScan. It works on WinXP, dunno if it does on Win98, Win2000 or Vista. It's a single file executable so nothing is lost if it doesn't work. Using that utility you can check SMART values perform surface scans of varying types, etc.

If it shows bad attributes, that such as pending sectors, offline uncorrectable sectors, reallocated sectors, reallocation event count, you're drive has a problem.

If you have current pending sectors you may need to erase the HDD (by zero filling, not just by formating as formating doesn't write each sector) and reinstall and check if noises have disappeared. Forcing overwrite of all sectors, it forces reallocation of bad sectors that couldn't be recovered, and this could solve noise issue caused by attempts to recover them. Attempts to recover bad sectors may cause all types of bzzzzzzzzz click click bzzzzzzz sqeeel bzzzzz click click.

Doing an erase, pending sectors should become reallocated sectors. If new pendings reoccur, consider the drive to be failing. Same applies if after first erase there's 1000+ of reallocated sectors. That's just too many.

____


It is possible that there's nothing wrong with the drive, though. It may just be "housekeeping routine" that some HDDs have. For example the well known and long lived IBM/Hitachi routine: every ten minute interval, 5 seconds of MEOOOW (on later models this was altered to less audible 10 seconds of buzzzzzz) and every 12 hours meow/buzzzz + click click + meow/buzzz.

I've noticed my GPs occasionally decide to not unload heads while they idle and produce a constant low volume bzzzzzzzzzzzzzzz noise. Only if I read something from said disk, I will hear a seek noise to file I access, then silence. And after unload time passes, it unloads as it should. It is quite odd... this way Greenpowers do their buzzzing routine. They can do it for hours. Don't know why. Maybe if I run SMART offline routine scan using HDDScan, it never actually shuts down (even after 5 hours of continuous surface scanning) but occasionally restarts automatically.

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Fri Jul 31, 2009 4:56 am

whiic wrote:Of course 1TB Greenpower is quieter than 640GB Caviar Blue.
FYI, AALS is Black and AAKS is Blue. The original poster here was testing a Black.

FWIW I didn't notice any odd noises from the WD6401AALS drives I bought but then I didn't let them go through normal usage. I tested them a bit and then tossed them in a RAID array on a server so I might not have given them enough chance to settle in and make such a noise.

danielwritesback
Posts: 6
Joined: Fri Jul 31, 2009 12:31 am
Location: Kansas

Post by danielwritesback » Fri Jul 31, 2009 5:55 am

TO whiic

Speed:
The reported speed is seriously affected by the drive controllers. According to Maxtor, the Nvidia 4xx series sata is the very worst quality. My Western Digital AALS is connected to that infamous controller. Let's not judge speed yet. That's for next week. ;)
The Conroe+G31+ICH8 setup has yet to arrive, but when it does, I'll surely re-test the speed.
Also, speeds reported were continuous writes because testing the read speeds of AALS were too high to be considered valid figures.

Noises:
As for now, the sounds are much like the oldest models of PC desktop hard drives. Its almost like a beeping (but more like a "blooping). Mostly, the noise occurs on seek. If compared in stock condition, the fast AALS (caviar black) doesn't have the deep purr of the EADS (caviar green) during seek. Is it possible that the rate of seek can hetrodyne with the spindle speed, causing a ringing? That could surely make data errors.
I'll go check.

jhhoffma
Posts: 2131
Joined: Mon Apr 25, 2005 10:00 am
Location: Grand Rapids, MI

Post by jhhoffma » Fri Jul 31, 2009 6:07 am

Is it just me or are the subject lines getting really weird around here?

danielwritesback
Posts: 6
Joined: Fri Jul 31, 2009 12:31 am
Location: Kansas

Post by danielwritesback » Fri Jul 31, 2009 6:11 am

dhanson865 wrote:
whiic wrote:Of course 1TB Greenpower is quieter than 640GB Caviar Blue.
FYI, AALS is Black and AAKS is Blue. The original poster here was testing a Black.

FWIW I didn't notice any odd noises from the WD6401AALS drives I bought but then I didn't let them go through normal usage. I tested them a bit and then tossed them in a RAID array on a server so I might not have given them enough chance to settle in and make such a noise.
Did you set the TLER option to "on" for safe operation in a raid system?

They are both quieter than the Caviar Blue. I have the 500 gigabyte AAKB (Caviar Blue with 16mb cache and IDE) in the oldest PC here, to retrofit it for some speed.
Gosh, I never need this much space on a desktop PC, but I do need the large onboard caches to support modern operating systems at a humane rate of speed.
A more economical solution is needed--like 200gb at very fast speeds; and, does such a thing exist?

The WD6401AALS aren't loud. In fact, they are very quiet. However, the noise occurs in the "upper midrange" of the audio spectrum, which is the ear's most sensitive band.

Simulation:
This isn't too accurate, but you'll get the picture. Get The Shirelles "Will You Still Love Me Tomorrow" and throw a terrycloth towel or bathrobe over the speaker and then play back that song at the smallest volume you can hear, in an otherwise silent room. Yes, the WD6401AALS does Doo Wap. Its not loud, but it is oddly disturbing because it is so weird. ;)

danielwritesback
Posts: 6
Joined: Fri Jul 31, 2009 12:31 am
Location: Kansas

Post by danielwritesback » Fri Jul 31, 2009 6:31 am

whiic wrote: . . .
You can check the health of the drive by reading SMART values. For ease of use I recommend running HDDScan. It works on WinXP, dunno if it does on Win98, Win2000 or Vista. It's a single file executable so nothing is lost if it doesn't work. Using that utility you can check SMART values perform surface scans of varying types, etc.

If it shows bad attributes, that such as pending sectors, offline uncorrectable sectors, reallocated sectors, reallocation event count, you're drive has a problem.

If you have current pending sectors you may need to erase the HDD (by zero filling, not just by formating as formating doesn't write each sector) and reinstall and check if noises have disappeared. Forcing overwrite of all sectors, it forces reallocation of bad sectors that couldn't be recovered, and this could solve noise issue caused by attempts to recover them. Attempts to recover bad sectors may cause all types of bzzzzzzzzz click click bzzzzzzz sqeeel bzzzzz click click.

Doing an erase, pending sectors should become reallocated sectors. If new pendings reoccur, consider the drive to be failing. Same applies if after first erase there's 1000+ of reallocated sectors.
. . .
Thank you for this information!

I'd hesitate to do this maintenance while it (Western Digital AALS, Caviar Black) is plugged into an Nvidia 4xx chipset. However, this can be explored when the Intel board arrives next week.

For reference, the drive wasn't reformatted except for when a Norton Ghost image was dumped onto it. It did pass the CRC check.

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Fri Jul 31, 2009 7:35 am

danielwritesback wrote:
dhanson865 wrote:
whiic wrote:Of course 1TB Greenpower is quieter than 640GB Caviar Blue.
FYI, AALS is Black and AAKS is Blue. The original poster here was testing a Black.

FWIW I didn't notice any odd noises from the WD6401AALS drives I bought but then I didn't let them go through normal usage. I tested them a bit and then tossed them in a RAID array on a server so I might not have given them enough chance to settle in and make such a noise.
Did you set the TLER option to "on" for safe operation in a raid system?
Nope, I'm using software RAID 1 on one server and the drives don't know any different than being desktop drives. The OS just waits if a drive goes into error recovery (though I suppose it hasn't happened yet).

I'm using hardware RAID 1 on the other server (the main company file server) and it's been used in this fashion for months and haven't seen a single hiccup.

I think I bought those drives last December.

My understanding is that TLER is only an issue on RAID 0, 5, or 6 where dropping drives means loss of the array. With Raid 1 if I drop a single drive I don't lose data, if both drives are failing then what good would it do me to have TLER off/on?

I suppose there could be a corner case where the drive is essentially good enough to pull data from but flakey enough that the RAID controller isn't happy about it but if that happens the drive isn't going to keep up with the demands I'm placing on it anyway and I'll be overnighting new drives in to replace it.

If I ever get in that situation before SSDs get cheap I'll just get some RE model drives I suppose.

Considering we make backups of those drives twice a day to two types of media the worst case scenario is I'm out the time and trouble of rebuilding that file server.

Best case scenario those drives work for the next few years until I replace them with some cheap 512GB SSDs (when I say cheap I mean really cheap so when I say years I mean several years).

danielwritesback
Posts: 6
Joined: Fri Jul 31, 2009 12:31 am
Location: Kansas

Post by danielwritesback » Fri Jul 31, 2009 7:13 pm

The Hitachi Feature Tool decreased the sounds by half. It didn't slow down, but instead, it sped it up.

The same sounds are still present though. Its like. . . this hard drive is a smallish church bell, and whenever it moves, it sounds. Perhaps its own seeks were causing it to "ring" itself? Has anyone run into this before?

Perhaps in a different computer case, then it wouldn't be audible? This case (allied/apex mini tower) has the vents located at point-blank range to the hard drive, which is theoretically good, but its not suited to a quiet PC build. Any suggestions as to a replacement?

Thanks again!

P.S.
After acoustic management mode set to 192, the hard drive has now expanded its use of its cache to include random access, as the figure read by passmark is now a cached figure. The real hardware figure of 4.5 has been replaced by a 30 (thirty). According to passmark, the speed of this one Caviar Black approximates the speed of two Caviar Blue (8 mb cache versions) in a raid arrangement. How handy!
Also the power consumption (judged by heat output) has increased slightly. Since its going faster, perhaps the power consumption increase is merely proportional and not a concern? Its still cooler than AAKS (Caviar Blue).
Only writes aren't making use of its onboard cache. There's supposed to be some sort of software adjustment for caching options for whenever you have the PC connected to a backup battery. Any clues about this adjustment and/or if it is beneficial?

danielwritesback
Posts: 6
Joined: Fri Jul 31, 2009 12:31 am
Location: Kansas

Post by danielwritesback » Fri Aug 28, 2009 12:11 am

Apparently, the hard drive (caviar black) had been beating itself up; however, switching on the soft seek option stopped the problem.
The noise didn't disappear immediately, but it is gone now, after approximately a week.
I wonder if the Black doesn't happen to have a "run in timer" set for a week, like a lot of other hard drives?

Comparing the Caviar Green with the Caviar Black, both models with 32mb ondrive cache has some interesting figures when both are compared on Ich7 Intel onboard sata controller. About 100 megabytes per second for the green and about 120 megabytes per second for the black. The green draws 7 watts and the black draws 8 watts. 1 watt and 20 megabytes per second difference. Wow.
Figures aside, I can tell that both run remarkably cool.

The point of this project was to make a more powerful desktop system as delightfully stable as the little ECS Atom server, featuring the brisk Western Digital hard drives with 32mb ondrive cache, the lagless Atheros lan, the potentially good Via sound chip, and APFC single rail power supply. And, that it will pay for itself. And that it can replace my athlon 64 dual core 4600+ desktop computer by beating it on performance.
It has done so.
The board is ECS G31T-M7 and the CPU is a single core conroe with a single u-shaped bit of stove pipe tape added for bsel 1066 front side bus speed. The ram is a single pink ddr2-800 A-data 2gb and the operating system is xp professional (with matching license tag stuck to the pc).
The power supply is an aging corporate version Hi-Pro, but is scheduled to be replaced with PC Power and Cooling single rails APFC 370 watt version.
On this particular board, the economy Via sound works nearly as well as a Via Tremor/Envy24, due to some dumb luck by ECS on choice of caps; however, the onboard does operate in verifiable high fidelity class sound, thus there is no point in adding a creative labs x-fi gamer that would not make an improvement (and I tried both in competition). This audio performance will vary by production run and that is because of the caps. However, this model motherboard was advertised/marketed to have good caps on it and I'm glad that its not exclusive to the power circuit.
OH, such good luck.
Exactly as the little atom board, this near identical lga755 board has the lagless and copasetic Atheros lan on it, and that can run full tilt whether the CPU is busy or not. Full tilt? Well, at least close and ever so much better than others that I have experienced.
Negative $18.50 per month is the electric bill difference.

Latest test:
Assuming that I would eventually desire to use the performance potential of this new system, I've acquired an Abit Airpace. That is a PCI-E atheros based wifi card. They're quite cheap on E-bay. It fits the normally useless little PCI-E 1x socket.
And, guess what? The server is getting one of those too. The included software does an awesome AP, thus freeing me of having to use a combo-router. This is good because my wired ethernet lan hits hard and makes most brands of wireless router give up and die (due to all ports full, long wires and old school large framed 111megabit). With the Abit Airpace aboard the server, I can then use a more professional ethernet-only router for the lan (instead of the jack-of-all-trades wireless routers).

Decisions! Help!!
The CPU and board are both capable of FSB of 1333, resulting in a single core conroe at 3.7 ghz. Currently, its at 2.9ghz with no additional voltage needed. However, I do have a cooler master gemini cpu cooler. . . Of course, the cpu, which has already been tested at 3.7ghz with 1.44v (instead of 2.9ghz with 1.25v) would draw more energy that way. It is already significantly fast without drawing more energy. Yet I'm terribly tempted. The fan controller on this board is sublime and can probably keep that 120mm fan inaudiable despite the overclock. But, should I do it?

Post Reply