QUIETEST 7200rpm hard drive on the market?

[Vince604]

Does Seagate still have the quietest hard drives in the world or will Maxtor's new line of upcoming hard drives be even quieter?

 

[lhgpoobaa]

i believe seagate have the edge. they have been producing fluid bearing drives for longer than maxtor, and do some very impressive baracuda drives.


<b><font color=green>A computer once beat me at chess, but it was no match for me at kickboxing! :lol: </b></font color=green>

 

[Vince604]

Alright. Well I'm trying to decide if I should get 2 Barracuda IVs or 2 Maxtor Diamondmax Plus 8 40GB hard drives.
But does the amount of heads and platters in a hard drive affect noise level? Looking at the specs of both hard drives Seagate has 2 heads while Maxtor only has 1.

 

[lhgpoobaa]

mmmm in my experience head seek noise is pretty independent of spin noise, and ive found my maxtor D740X to have a rather noticable 'scratchy' seek... far louder than my 60gxp or 800JB, which have low pitched seeks that blend into the general PC noise.


<b><font color=green>A computer once beat me at chess, but it was no match for me at kickboxing! :lol: </b></font color=green>

 

[Vince604]

I'm not sure if you have a seagate drive before but you'd still consider it to be quieter than Maxtor than right?

Everything with Seagate sounds promising but I just don't really want to buy the hard drives and have them sent to Seagate to be replaced with RAID compatible ones... Because the regular Seagate Barracuda IV's have RAID compatibility issues and don't work well in RAID set ups..


My goal really is trying to make the quietest PC as possible while still having high performance..

Well in your opinion would you rather go for a 8mb cache Western Digital hard drive without RAID 0 or would you rather get 2 quiet Seagate hard drives in a RAID 0 setup?

 

[lhgpoobaa]

i have no opinion on the performance of the seagates in raid sorry.

as for seagates vs JB's, i would definately go for the JB's cauz of thier performance, but then again im not trying for the quietest PC around. my JB really isnt that loud with idle, but it may be for you.

<b><font color=green>A computer once beat me at chess, but it was no match for me at kickboxing! :lol: </b></font color=green>

 

[HammerBot]

If you are going to use the Baracuda for RAID0, make sure you state that when you purchase. The standard version has performance problems with RAID0, but Seagate has released some new drives (or new firmware, I dont know the details) that solves this issue.

Regarding choise of HD, I would recommend the WD800JB. I have two of those in RAID0 and they are extremely noiseless. Significantly less noise than my previous IBM 60GXP. And thats goes for idle as well as seek noise.

 

[sjonnie]

Here are some figures about idle noise from <A HREF="http://www.storagereview.com" target="_new">storage review</A>. Note that it includes 5400rpm and SCSI models.

Rank - Drive Name - Model - dB
1 - Seagate U6 (80 GB ATA-100) - ST380020A - 38.8
2 - Seagate Barracuda 36ES2 (37 GB Ultra160/m SCSI) - ST336938LW - 39.3
3 - Seagate Barracuda ATA IV (80 GB ATA-100) - ST380021A - 41.3
4 - Maxtor DiamondMax 536DX (100 GB ATA-100) - 4W100H6 - 41.5
5 - Western Digital Caviar WD800AB (80 GB ATA-100) - WD800AB - 43.7
6 - Samsung SpinPoint P40 (80 GB ATA-100) - SP8004 - 44.5
7 - Seagate Cheetah 36ES (36 GB Ultra160/m SCSI) - ST336706LW - 44.6
8 - Samsung SpinPoint V30 (60.0 GB ATA-100) - SV6004H - 45.0
9 - Western Digital Caviar WD800JB (80.0GB ATA-100) - WD800JB - 45.0
10 - Seagate Cheetah 15K.3 (73 GB Ultra320 SCSI) - ST373453LW - 45.1
11 - Maxtor DiamondMax D540X (160 GB ATA-133) - 4G160J8 - 45.5
12 - IBM Deskstar 120GXP (120 GB ATA-100) - IC35L120AVVA07 - 45.8
13 - IBM Deskstar 60GXP (60.0 GB ATA-100) - IC35L060AVER07 - 45.9
14 - Maxtor DiamondMax Plus D740X (80 GB ATA-133) - 6L080J4 - 46.3
15 - Quantum Fireball Plus AS (60.0 GB ATA-100) - QMP60000AS-A - 46.5
16 - Seagate Barracuda 36ES (36 GB Ultra160/m SCSI) - ST336737LW - 46.7
17 - Samsung SpinPoint P20 (40.0 GB ATA-100) - SP4004H - 47.0
18 - Western Digital Caviar WD1000BB (100 GB ATA-100) - WD1000BB - 47.1
19 - Western Digital Caviar WD1200BB (120 GB ATA-100) - WD1200BB - 47.1
20 - Western Digital Caviar WD1200JB (120 GB ATA-100) - WD1200JB - 47.3
21 - Western Digital Caviar WD1000BB-SE (100 GB ATA-100) - WD1000BB--SE 47.4
22 - Seagate Cheetah X15-36LP (36.7 GB Ultra160/m SCSI) - ST336752LW - 48.1
23 - Seagate Cheetah 10K.6 (146 GB Ultra320 SCSI) - ST3146807LW - 48.5
24 - Quantum Atlas V (36.7 GB Ultra160/m SCSI) - QM336700XC-LW - 49.2
25 - Fujitsu MAM3367 (36 GB Ultra160/m SCSI) - MAM3367 - 49.3
26 - Fujitsu MAN3735 (73 GB Ultra160/m SCSI) - MAN3735 - 49.3
27 - Seagate Barracuda 180 (180 GB Ultra160/m SCSI) - ST1181677LWV - 49.8
28 - Maxtor Atlas 10k III (73 GB Ultra160/m SCSI) - KW073L8 - 50.1
29 - Seagate Cheetah 73LP (73.4 GB Ultra160/m SCSI) - ST373405LW - 50.1
30 - Seagate Cheetah 36XL (36.7 GB Ultra 160/m SCSI) - ST336705LW - 51.7
31 - IBM Ultrastar 73LZX (73 GB Ultra160/m SCSI) - IC35L073UWD210 - 52.8
32 - IBM Ultrastar 36Z15 (36.7 GB Ultra160/m SCSI) - IC35L036UCPR15-0 - 54.8

 

[HammerBot]

What! Only 0.9 dB (about 10%) difference between 60GXP and WD800JB. I would have expected a LOT more (3-5 dB). I wonder if HD mounting brackets has a large influence on the noise from the HD. Otherwise I really cant explain why my new PC with WD800JB is much more noiseless that my previous with 60GXP.

 

[Napoleon]

Don't really know how that noise stuff is measured or - more importantly - how human ear perceives those measured differences. However, IMO the case and mounting can make a big difference. Once upon time I had a *really* crappy AT case. HD rotation noise was prominent, and occasionally the case even started to <i>resonate</i> with the HD rotation noise. I mean, you could *feel* it in your fingertips when you touched the case. And when the CD kicked in... that POS case could have served as a definition of resonance cavity...


<font color=green>I haven't lost my mind. I know exactly where I left it.</font color=green>

 

[Vince604]

Well two hard drives in RAID 0 is generally better than a single hard drive though right? Or is that not true?

I'm not desperate for every single bit of performance right now but more on the noise issue..

 

[Vince604]

Yeah I already contacted Seagate a few weeks ago and they said if I buy their hard drives I can send them in to get them replaced with RAID 0 compatible ones.

And yes HD brackets CAN make a big difference in noise levels. Depending on what kind of brackets they are since you can now get HD brackets like the No Vibes III designed to reduce the vibrations from the hard drive thus reducing noise level...

 

[HammerBot]

Yeah I already contacted Seagate a few weeks ago and they said if I buy their hard drives I can send them in to get them replaced with RAID 0 compatible ones.

Hm. So it appears that the drives you can purchase still are the problematic ones, and you have to experience the problem before you can get them replaced. I would expect that the drives they have been manufacturing since they discovered the flaw to be fixed, so isnt it possible to get e.g. serial number of good drives, so you can check the one you purchase up front?

I have a <A HREF="http://www.chieftec.com/products/workstation/tx103.htm" target="_new">TX103</A> cabinet. I havent used anything special to mount the drives. Im very satisfied with it, and I can hardly hear the drives spinning or seeking.

 

[Vince604]

Yes they are only problematic if used in a RAID 0 setup. I doubt and don't think that Seagate would be fixing the problem as the Barracuda Vs are supposed to be out next month and I don't think they replace the drive but just update the hard drives you sent in with a new firmware and re-ship them back to you.

 

[lhgpoobaa]

well yes, typically 2 in raid0 have better performance, but if one dies you loose the data on both. plus you have to put up with the noise of 2 drives. (as you said noise is an issue)

<b><font color=green>A computer once beat me at chess, but it was no match for me at kickboxing! :lol: </b></font color=green>

 

[Vince604]

Ok yes back to the noise issue.
So one Western Digital Hard drive would be definitely quieter than 2 Seagate Barracuda's??

I'm not sure if hard drive and fan noise are measured the same way but someone said that this:
Decibels are logarithmic. If you had 10 fans at 30dB, that would be 40dB.

I'm not sure if it's true and that's why I ask...

 

[lhgpoobaa]

no idea on your first question.
but sound is logrithmic... a doubling of noise volume equals 10 more dBa.

<b><font color=green>A computer once beat me at chess, but it was no match for me at kickboxing! :lol: </b></font color=green>

 

[verada]

my two IBM 120GXP's are connected in RAID 0 and I connot hear them seek at all in comparison to my two Maxtor D740's also in RAID 0 which were so loud I think they could almost wake up the dead!

Computers remind me of Murphy's law:- whatever can go wrong, it's always associated with a computer!

 

[Napoleon]

Well, I've never had much use for noise definitions in the acoustic sense, but here's my two cents on how it goes. Feel free to correct me, it's high time I learned some more about this stuff, instead of doing IMOs.

<i>Edited: forget the guesswork below and educate yourselves <A HREF="http://hyperphysics.phy-astr.gsu.edu/hbase/sound/dbcon.html#c1" target="_new">here</A></i>.

Anyway, as far as I know, acoustic noise in dB is defined as
noise(dB) = 10 * log2(P/Q),

where P and Q are expressed in "power / unit area" (intensity). Q is human hearing threshold intensity. If P equals Q, log2(1) = 0, hence 0dB is perceived as silence.

If the noise intensity doubles, e.g from 2Q to 4Q, the acoustic noise expressed in decibels increases by 10dB: 10*log2(2) = 10dB, 10*log2(4) = 20dB.

Unfortunately, those raw intensity figures in dB tell very little about how loud a human being would <i>perceive</i> the noise to be. Noise spectrum is the determining factor there. E.g some very intense noise in the 25kHz - 30kHz band (ultrasound) wouldn't make any difference to you, but your dog might go berserk...

So, when talking in human terms, the measured noise intensity is first filtered using so called "A-filter" which emulates average human hearing curve. Hence, the "loudness" unit intended for humans is dBA. To my understanding, a human being would perceive a 10dBA increase in noise to be (roughly) twice as loud, just like LHGPooBaa said. To my understanding, whisper is about 30dBA and normal talk is about 60dBA.

I checked some Storage Review charts, and they seemed very confusing to me with their "dB/A @18mm" units. Chances are that they meant unfiltered intensity, measured at a 18mm distance from the noise source. The short distance, of course, serves to reduce the effects of ever-present ambient noise (like the PSU fan...)

If that's the case, those Storage Review measurements really don't mean all that much. In THG's reviews, Barracuda IV's idle spindle noise measured in dBA still seems to be the lowest. Compared against WD1200JB, it's 48.5dBA vs 54.8dBA. Got a 60GB 'cuda myself. I don't have anything up-to-date to compare it with, but IMO it is very quiet. I particularly like the quiet seek noise. Anyway, if you compare the spindle noise of two 'cudas to one WD1200JB, that (48.5+10)dBA = 58.5dBA vs 54.8dBA isn't too bad. Still, it's an individual thing in the end. Bel is a relative unit, and one's hearing curve certainly affects things a lot. Deaf people probably go for the highest performance.


<font color=green>I haven't lost my mind. I know exactly where I left it.</font color=green><P ID="edit"><FONT SIZE=-1><EM>Edited by Napoleon on 09/18/02 06:26 PM.</EM></FONT></P>

 

[Vince604]

Hmm wasn't there a small issue with the IBM hard drives? I'm not sure but it was a little while back where it wasn't suppose to be running for more than an X amount time am I right? Or I might just totally be wrong but to me IBM just puts their names on parts and sell for high prices. Another would be like Dell... My friend spent about $3000CDN just a couple weeks ago on a new 1.8GHZ Dell computer where I can get one for under $2000CDN.

 

[Vince604]

Thanks for the info. on the noise measurement. Seems a little confusing but I can figure it out.

Now as you showmed me that there's only a 3.7dbA noise level difference I'm not too sure if I'd benefit from a RAID 0 setup.

I mean what is the probabilty of a RAID 0 failure or a hard drive failure in a RAID 0 setup?

And how much more performance can I possibly get using two Barracuda IVs vs a 8MB cache WD hard drive....?

 

[sjonnie]

Oh, those figures were IDLE noise btw. Not sure how much noise the Maxtor D740Xs make during read/write, but it is significant. I have 5 running here at work and never hear them, probably because of all the fans though. The 2 at home seem to make alot more noise though.

 

[Napoleon]

Just to make sure that the misinformation in my previous post doesn't come back to haunt me: log10 is used for acoustic (deci)Bel calculations, <b>not</b> log2. That doesn't change the principle, but sure does affect the numeric values. Regardless, a 10dBA increase is perceived to be "twice as loud". And JND (Just Noticeable Difference) is 1dBA.

noise(dB) = 10 * log10(P/Q)

Due to log10, the increased noise from additional drives would be:

10 * log10 dBA

where n is the number of drives. So, having two 'cudas (n=2) would up the idle noise to:
10*log10(2)dBA + 48.5dBA = 3dBA + 48.5dBA = 51.5dBA

Four 'cudas would be
10*log10(4)dBA + 48.5dBA = 6dBA + 48.5dBA = 54.5dBA

Sooo, when you compare four 'cudas to a single WD1200JB (54.8dBA), they would sound equally noisy (0.3dB is below JND level). Go for RAID0+1 with 'cudas...

BTW, I took the dBA values above from some THG chart. Anyway, it seems that comes to acoustic noise and deciBels, any given value should be taken with a grain of salt. First off, unweighed intensity values expressed in dB are largely meaningless. Better make sure the noise figures are given in dBA.

To complicate things even more, the deciBel figure can be measured in (at least) two ways, namely by measuring sound intensity or sound pressure. Pressure method is dependent from the measurement distance, intensity method is not. One has to be very careful to ensure an apples-to-apples comparison. Even then, if you're in the habit of cranking up your stereos while you're working at your comp or work in an otherwise noisy environment, A-contour no longer applies. Should be B- or even C-contour (dBB and dBC)... :frown:


<font color=green>I haven't lost my mind. I know exactly where I left it.</font color=green><P ID="edit"><FONT SIZE=-1><EM>Edited by Napoleon on 09/19/02 09:25 PM.</EM></FONT></P>

 

[Vince604]

Okay thanks for all the information! I only need one equation. 10 * log10. That's all that's really important and to have things in dbA not db...

Well thinking bout RAID 0+1 but it kind of wastes two hard drives.. I mean would I be at a high risk of a hard drive failure if I took 4 hard drives and put them in a RAID 0?

Well I might start with 4 Barracuda's now or maybe start with just 2 and move up to 4 later on but I'm not running a critcal server or anything I just want the quietest operation possible with performance...
And I think I'll wait till maybe November when the Barracuda Vs come out because they are already RAID compatible..


I'm having a little trouble using that equation for noise.
Are you sure it's not: 10log + ?
where (n= number of devices) and (y=dbA level)...

I tried calculating your example of the 2 and 4 cudda's and I'm getting 68.5 and 88.5dbA which I think is a little too high?..

Maybe just a typo?..<P ID="edit"><FONT SIZE=-1><EM>Edited by Vince604 on 09/19/02 09:56 PM.</EM></FONT></P>

 

[HammerBot]

I mean would I be at a high risk of a hard drive failure if I took 4 hard drives and put them in a RAID 0?

I can live with 2 drives in RAID0. 3 makes me nervous but 4 scares the hell out of me. With all the stories of failing harddrives I think the risk is too high regardless of brand and type of drive.

or maybe start with just 2 and move up to 4 later on but I'm not running a critcal server or anything I just want the quietest operation possible with performance...

If you change the array configuration (stripe size or width, e.g. number of drives) you have to reformat it. You cannot just add drives and retain the data on the existing drives.

I'm having a little trouble using that equation for noise.
Are you sure it's not: 10log + ?

I rather stay out of the noise calculation discussion. But I would like to point out that there is no unique equation for calculating the sum of noise. It depends on whether the noise is correlated or uncorrelated. If the sources are 100% correlated you get twice the sound level/pressure, otherwise you get somewhat less depending on the amount of correlation. Now, two harddrives are not complete correlated but they are not completely uncorrelated either.