Two Hitachi 250gb or three Samsung 160gb drives?

Magic · Post by **Magic** » Thu Feb 05, 2004 8:05 pm

I've surmised from reading this website that the Samsung Spinpoint drives are the quietest 7200 rpm drives available.

I'm building an extremely quiet workstation, and want about a half a terabyte of storage. My question ....is it quieter to have

- three Samsung SP1614C drives (160gb each - sata)?

or is it quieter to have

- two Hitachi 7k250 drives (250gb each - sata)?

Or should I consider something else completely?

Also it's not likely that the three samsung drives will require more (i.e. louder) cooling fans, is it? I'll probably be using the 3700bqe case.

Thanks

galois427 · Post by **galois427** » Thu Feb 05, 2004 9:54 pm

i would pick the samsung system over the hitachi setup anyday. first, with one more drive, the samsung spinpoints would still produce less noise than the hitachi. this is from personal experience. secondly, since you're building a workstation, you can setup sata raid with the 3 samsungs for incredibly fast speeds.

Curtis · Post by **Curtis** » Fri Feb 06, 2004 4:29 am

Except SATA only supports 1 drive per channel, and no drive chaining. You'd have to buy an extra PCI SATA card, and I don't know how well this would work.

ejl10 · Post by **ejl10** » Fri Feb 06, 2004 5:54 am

Some mainboards are now coming with 4 SATA channels that can all be RAID-ed together. I think ASUS or Abit even has a 6 Ch. SATA board.

Here's one

I've got two Samsungs in RAID and think they're swell. Once I get around to enabling AAM again, I won't have to think about them at all

They're dead quiet.

Emmett Lyman

Putz · Post by **Putz** » Fri Feb 06, 2004 6:19 am

I also think the Samsungs would be very quieter. (With two Hitachis, you'd hear a strange meow twice as often!) Realize, of course, that in a RAID-0 setup, two Hitachis will provide greater performance than three Samsungs.

However, you can take this as an opportunity to NOT use RAID-0 and have three independent drives. This will likely be your top-performing setup. If you need advice on how to set up your drives for best performance, post at StorageReview.com.

ejl10 · Post by **ejl10** » Fri Feb 06, 2004 6:52 am

Putz,

Top performing or most reliable?

Thanks,
Emmett Lyman

davidstone28 · Post by **davidstone28** » Fri Feb 06, 2004 7:51 am

Personally, I'd go for the 2 drives rather than 3. The reasoning being that 3 drives are more likely to put more of a strain on the 12v PSU rail, which may ramp up the PSU fan. Temp wise there's not much in on a drive by drive basis, but clearly 3 drives are likely to generate more heat than 2 drives which again may cause your PSU fan to ramp up. I don't think platter density of a particular drive is going to be an issue with regard to heat in this situation.

kie · Post by **kie** » Fri Feb 06, 2004 7:56 am

davidstone28 wrote:Personally, I'd go for the 2 drives rather than 3. The reasoning being that 3 drives are more likely to put more of a strain on the 12v PSU rail, which may ramp up the PSU fan. Temp wise there's not much in on a drive by drive basis, but clearly 3 drives are likely to generate more heat than 2 drives which again may cause your PSU fan to ramp up. I don't think platter density of a particular drive is going to be an issue with regard to heat in this situation.

Well I run 4 hard drives off a 380W psu (antec trupower / sonata) along with various pci cards, optical drives etc and the 12V is fine and dandy.

ejl10 · Post by **ejl10** » Fri Feb 06, 2004 8:24 am

[quote="kieI used to be deaf and content before I found SPCR...[/quote]

Ain't that the truth. This is an addiction.

Emmett Lyman

Putz · Post by **Putz** » Fri Feb 06, 2004 10:24 pm

Which interpretation of "top-performing" applies depends on your usage patterns. Not using RAID-0 will always help you in the event that you lose a drive; if you lose any one drive in a three-drive striped array, every last bit of that half-terabyte is gone for good.

Two independent spindles will perform better than a RAID-0 array in general Windows tasks, many kinds of multitasking, Photoshop, database apps, frequent swaps, etc. RAID-0 will help Windows boot faster, decrease wait times a little when loading levels in games, and will improve the speed of high-STR reads and writes like those involved in video editing.

Now, with THREE drives, you can mix and match. What I would be tempted to do is have two identical drives in a RAID-0 and one independent. Get the best of both worlds. You could go all-independent if you're a heavy multitasker or run a server. Or go all-RAID if you want to live on the edge.

Magic, I think you have your answer in that three Samsungs are your low-noise winners. If you have questions about how to set up your rig for best performance, like I said, post at StorageReview, mention what drives you'd like to use in your setup, and what uses you have for your computer (important!), and they'll help you out.

m0002a · Post by **m0002a** » Fri Feb 06, 2004 10:40 pm

Noise is cumulative, so that 2 drives that independently are rated at 30 dB-A, will produce more than 30 dB-A when run together.

Therefore, it is possible that 3 Samsung drives could be louder than 2 Hitachi drives, even if the Samsung drives (individually) are quieter. But you have to do the math, and I am not sure of the exact formula, and I don’t have the specs on these drives.

Another consideration is that the MTBF goes down (for the entire system as a whole) the more drives you have. However, if one drive were used only as backup, then this would not be a factor. I would never use RAID-0 unless I had 3 drives (2 drives striped and 1 drive as a backup using PowerQuest Drive Image or equivalent).

Trip · Post by **Trip** » Fri Feb 06, 2004 11:07 pm

add 3 dB each time you double noise. 2 30dB would be 33dB.

Samsung is, what, 29dB? so 35dB.

However, the quality of the noise will vary. Some frequencies are more noticeable than others. Having never heard either drive, I can't help you any further.

Putz · Post by **Putz** » Sun Feb 08, 2004 5:05 pm

The concern here wouldn't be the direct addition of dB ratings, I think. Rather, I'd be worried about modulation noise between drives. I have two Seagates in my 24/7 system (a 40GB 'Cuda IV and an 80GB 7200.7), and they're very pleasantly inaudible individually. However, the moment they're mounted in my (admittedly flimsy) case and both turned on, they produce a "wow-ing" high-frequency inter-modulation noise that is slightly irritating in the middle of the night.

Of course, this issue is completely unpredictable from sample to sample, but with more samples of any drive (3 vs. 2), your risk of getting this noise goes up. Your best bet is probably still the three Samsungs, but who knows...

BrianS · Post by **BrianS** » Fri Mar 19, 2004 8:12 am

I understand the hitachi drive meows every 10 minutes

If you raid 2 of the Hitachi 250 GB
will you hear the meow every 5 minutes?

If you raid 4 of the hitachi 250 GB drives will you hear the meow every 2.5 minutes?

Ie is the meow more of a problem for raid drives?

Post by **MikeC** » Fri Mar 19, 2004 8:55 am

IBM "meow" -- That's a nice term for it.

To me, it sounds more like a screech or yowl in the 7k250. And it does not matter how many you have, they all reset at about the same time frequency -- seems like less than every 10 mins of idle. Right now, I have 2 machines in the lab that have IBM180GXPs in RAID 0 and another with a single 7K250. They seem to be synced to all yowl within about 30 seconds of each other.

zczczhhhhheererererrrrrrrroooo...... gzzzzzzzrrrrrrrrrrrrrrroooot.... wheeeeeeeeeerrrrrroooot...

Trust me, you don't want to live with this.

wumpus · Post by **wumpus** » Fri Mar 19, 2004 9:38 am

Well I run 4 hard drives off a 380W psu (antec trupower / sonata) along with various pci cards, optical drives etc and the 12V is fine and dandy.

Agreed, I have four spinpoints in my dual athlon server and total power draw for the machine with 100% load on both CPUs is ~150w. Total non-issue, probably expressed by someone who has never actually measured power draw.

- three Samsung SP1614C drives (160gb each - sata)?

or is it quieter to have

- two Hitachi 7k250 drives (250gb each - sata)?

The samsungs are so much quieter than every other drive I've used, it's really no contest. I have yet to hear any Barracudas, but I am comparing multiple WD, IBM/Hitachi, Maxtor with the 10 spinpoints I own. They are easily an order of magnitude louder to my ear. It's really dramatic how much quieter the spinpoints are.

Trip · Post by **Trip** » Fri Mar 19, 2004 10:15 am

Putz wrote:Rather, I'd be worried about modulation noise between drives.

Could you elaborate on what modulation is? my understanding is that if both drives produce a particularly loud sound in one frequency, the frequency will be nearly 3 dB louder if two are used together. Is this correct? If so could using two different HDDs (with similar dB ratings) be the recommended course of action (assuming modulation won't be a problem with most combinations)

Post by **MikeC** » Fri Mar 19, 2004 12:02 pm

He probably means intermodulation. It where two tones of very close frequency will create a third tone. Like if one drive is at 7199rpm and the other is at 7201. A thrid tone will be produced, and also there will be a kind of beating or fluttering sound. In the audio world, this is called intermodulation distortion (expressed in %) and is generally regarded as more intrusive than the better known harmonic distortion.

BrianS · Post by **BrianS** » Sun Mar 21, 2004 7:16 am

Regarding the hitachi drives

I found the following table from the hitachi 81 page drive manual interesting:

Table 29 (p38) Sound Power Levels (Typical/Max)

................... 1 disk ... 2 disk ... 3 disk
Idle ............ 2.6/3.0 ... 2.8/3.2 ... 3.0/3.4
Perf. Seek .. 3.4/3.7 ... 3.4/3.7 ... 3.4/3.7
Quiet Seek . 2.8/3.2 ... 2.9/3.3 ... 3.1/3.5

source: http://tinyurl.com/ypjpo

The interesting thing here is that it gives typical and max values. Usually drive specifications only give a single noise value.

How do you get the max noise? This suggests there may be considerable variation in drive to drive to drive noise.

Perhaps this suggests that the noise can not be determined by a single drive as exibited noise will very considerably from drive to drive. For the typical mean value, how many drives have to be sampled to determine this (>30)? Does drive sound follow a bell curve so that some individuals may be hearing sounds below or above the mean?

What this may mean is a single user writing in that their drive was quiet or loud may mean very little as it is just a single point somewhere on the bell curve. We may need 30-100 drives to be tested to find out the average noise and variability in noise.

Does anyone have a feel for how variable the drive to drive noise is?

Post by **MikeC** » Sun Mar 21, 2004 8:33 am

BrianS -- I don't think it varies anywhere close to 0.4 bels, which is what the IBM spec suggests. I think this may be more to cover their rear ends about the "meow" head reset noise, as it has been described, rather than a real "sample-to-sample variance" spec.

Generally speaking, most same model HDDs strike me as falling within a dB or two of each other. I've rarely noticed a sample that seemed way different than others of the same make/model. I am sure you can take the odd one back if it is noisier than several others you buy at the same time.