10K Western Digital SATA Drive!!!
- Thread starter Clocker
- Start date
zx
Learning Storage Performance
WOW!!!
....
WOW!!!
....
WOW!!
Anywone knows what capacities they will offer?
....
WOW!!!
....
WOW!!
Anywone knows what capacities they will offer?
Mercutio
Fatwah on Western Digital
I'm going to do a cautious "Yay", while noting that WD has been historically my least reliable drive vendor, and one of the participants in the 1-year-warranty farce.
It'll probably be priced into SCSI-stratosphere range, too, like 7.2k IDEs were, at first.
Also I have a moment of feeling all smart, 'cause I've been maintaining WD (with no SCSI lineup) would beat Seagate and everybody else to 10k on SR's interminable "Where's 10k IDE" discussions.
It'll probably be priced into SCSI-stratosphere range, too, like 7.2k IDEs were, at first.
Also I have a moment of feeling all smart, 'cause I've been maintaining WD (with no SCSI lineup) would beat Seagate and everybody else to 10k on SR's interminable "Where's 10k IDE" discussions.
sechs
Storage? I am Storage!
Like I've said, hard drives are starting to be a lot like video cards.
timwhit
Hairy Aussie
I wonder what kind of access time it will have? Will they actually use quality SCSI actuators in it or will it use the same lethargic ones as their IDE drives use?
5 Year warranty is a plus, but the final cost will be drastically higher as a result. I think I will still stick to SCSI (at least for my boot disk). There is just something about a Cheetah that makes me feel safer with my data on it.
5 Year warranty is a plus, but the final cost will be drastically higher as a result. I think I will still stick to SCSI (at least for my boot disk). There is just something about a Cheetah that makes me feel safer with my data on it.
Prof.Wizard
Wannabe Storage Freak
- Joined
- Jan 26, 2002
- Messages
- 1,460
zx said:
Amazing news!jtr1962
Storage? I am Storage!
timwhit said:
You make an excellent point here. I've often wondered what's the point in going from 5400 RPM to 7200 RPM and continuing to use the same activator when you can accomplish the same reduction in access time by using an activator that's 1.4 ms faster and keeping at 5400 RPM(which will save you noise, power, and heat). Now consider going from 7200 RPM to 10000 RPM. All you're saving is a lousy 1.17 ms in rotational latency if you use the same activator.
Not to spoil the party, but frankly this rotational speeds thing is more a marketing gimmick than anything else since it seems that not as much effort goes into reducing seek times. A typical 5400 RPM drive has seek times on the order of 9 ms while a 7200 RPM is usually about 0.5 ms faster. Why they didn't get the seek times down to something like 5 ms first before even thinking about increasing the speed past 5400 RPM is beyond me. A more powerful activator only draws more power and makes more noise when it is seeking. A faster spinning drive is noisier and hotter all the time. I'm willing to bet a 5400 RPM drive with a 5 ms seek time will be faster than this new drive if the pattern continues. Maybe they got the seek time down to 8 ms, so you'll save 1 ms there and another 2.56 ms in rotational latency, or a total of ~3.5 ms in access time compared to a run of the mill 5400 RPM drive. Could have done the exact same thing by getting the seek time of the 5400 RPM drive from 9 ms down to 5.5 ms, and only then would it finally make sense to think about increasing RPM to get the access time down still further. The sad part is that the performance-minded computer users will fall for this hook, line, and sinker, and pay top dollar for a drive that's maybe 25% faster than a run-of-the-mill 5400 RPM drive. Hardly seems worth it to me, especially if the damned thing sounds like a jet airliner spooling up it's turbofans for take-off.
Even taking the devil's advocate position and supposing they do have a 5 ms seek time, plus the drive is quiet and cool to boot, it will likely cost at least $100 more than the same sized 5400 or 7200 RPM drive. Better to just put that money into more RAM which will let you cache more of your most frequently used files and pretty much mask the lack of performance of the slower drive.
To be honest, I'd rather they focus on making solid-state drives a reality instead of increasing RPMs to the level of a dentist's drill. Something on the order of 5 to 10 GB selling for $100 or less would be nice. Big enough to put your OS and applications on, and you can use a larger mechanical drive for bulk storage. Plus the under 1 microsecond access times would be very nice as well. I also somehow think my data will be safer this way rather than as billions of microscopic magnetic bits on a disk spinning 167 times per second.
Glad to get that out of my system.
Dïscfärm
Learning Storage Performance
Prof.Wizard said:zx said:
...provided that it's true, of course!
If it is, then you'll see "everyone else's" 10kRPM ATA hard drives surface during the next 3 months -- or at least announcements of such.
In all likelihood, the market would have been due 10kRPM ATA hard drives no later than the introduction of SATA2 anyway, which will initially be aimed at the technical-workstation / storage-array marketplace, with 7200 and 5400 RPM hard drives still very much alive in the marketplace (2004).
SATA2 and the upcoming Serial Attached SCSI (SA-SCSI) are a synergistic pair. The new SA-SCSI and SATA2 will come out roughly at the same time and have a certain level of compatibility with each other. In the long-running high stakes battle between SCSI and Fibre Channel, the SCSI camp has very cleverly co-opted ATA into this battle royale by making SATA signaling a subset of SA-SCSI and using the SATA physical layer (cabling, connectors). You will be able to plug SATA2 hard drives into a SA-SCSI controller and use them -- even along with SA-SCSI drives. Of course, there will be 10kRPM and 15kRPM SA-SCSI drives galore when SA-SCSI is released. With the releases of SA-SCSI and SATA2 being close in time, it would only make sense to have a few 10kRPM SATA2 drives around or even 15kRPM SATA2 hard drives.
Also, don't count out the possibility that a manufacturer might release a 12kRPM SATA / SATA2 hard drive, just as a market differentiator. I still like to recall that yummy rumour from around 1998 that had Quantum (now Maxtor ...or Quaxtor) supposedly toying around with 13kRPM RPM ATA hard drives in the laboratory.
Dïscfärm
Learning Storage Performance
jtr1962 said:
There has been various approaches taken in the past several years to perfect affordable solid-state non-volatile mass storage in the gigabyte capacity range. But, it seems nowadays that the most likely to succeed in this catagory will be polymer memory (a.k.a. -- plastic memory), only because it is fast being realised by researchers that very dense plastic memory can be manufactured easily and quite cheaply. Additionally, plastic memory is also looking to be a better form of storage than Flash memory and will probably kill Flash rather quickly. There are even plastic microprocessors being proposed!
blakerwry
Storage? I am Storage!
jtr1962 said:timwhit said:
You make an excellent point here. I've often wondered what's the point in going from 5400 RPM to 7200 RPM and continuing to use the same activator when you can accomplish the same reduction in access time by using an activator that's 1.4 ms faster and keeping at 5400 RPM(which will save you noise, power, and heat). Now consider going from 7200 RPM to 10000 RPM. All you're saving is a lousy 1.17 ms in rotational latency if you use the same activator.
Not to spoil the party, but frankly this rotational speeds thing is more a marketing gimmick than anything else since it seems that not as much effort goes into reducing seek times. A typical 5400 RPM drive has seek times on the order of 9 ms while a 7200 RPM is usually about 0.5 ms faster. Why they didn't get the seek times down to something like 5 ms first before even thinking about increasing the speed past 5400 RPM is beyond me. A more powerful activator only draws more power and makes more noise when it is seeking. A faster spinning drive is noisier and hotter all the time. I'm willing to bet a 5400 RPM drive with a 5 ms seek time will be faster than this new drive if the pattern continues. Maybe they got the seek time down to 8 ms, so you'll save 1 ms there and another 2.56 ms in rotational latency, or a total of ~3.5 ms in access time compared to a run of the mill 5400 RPM drive. Could have done the exact same thing by getting the seek time of the 5400 RPM drive from 9 ms down to 5.5 ms, and only then would it finally make sense to think about increasing RPM to get the access time down still further. The sad part is that the performance-minded computer users will fall for this hook, line, and sinker, and pay top dollar for a drive that's maybe 25% faster than a run-of-the-mill 5400 RPM drive. Hardly seems worth it to me, especially if the damned thing sounds like a jet airliner spooling up it's turbofans for take-off.
Even taking the devil's advocate position and supposing they do have a 5 ms seek time, plus the drive is quiet and cool to boot, it will likely cost at least $100 more than the same sized 5400 or 7200 RPM drive. Better to just put that money into more RAM which will let you cache more of your most frequently used files and pretty much mask the lack of performance of the slower drive.
To be honest, I'd rather they focus on making solid-state drives a reality instead of increasing RPMs to the level of a dentist's drill. Something on the order of 5 to 10 GB selling for $100 or less would be nice. Big enough to put your OS and applications on, and you can use a larger mechanical drive for bulk storage. Plus the under 1 microsecond access times would be very nice as well. I also somehow think my data will be safer this way rather than as billions of microscopic magnetic bits on a disk spinning 167 times per second.
Glad to get that out of my system.
This is an excellent post!
I wonder if the platter densities will continue to increase with these new drives? if we have 10k RPM's with super high densities we will have a hay day with STR's (ok... now I sound like a techie)
Buck
Storage? I am Storage!
As with high-speed SCSI drives, I'm expecting that the areal density will be lower, not higher.
blakerwry
Storage? I am Storage!
that's why I'm wondering... IDE's strength has always been it's large, cheap capacity. SCSI's has been it's super fast speed.
we seem to be coming to a convergence within the new future to where it may not matter SCSI vs IDE.. you'll just see drives that are based on their own merits.
*that reminds me, jtr, don't you think the high density of IDE drives is in part responsible for the high seek time.. it seems there is a balance... the higher your density the faster your read heads have to be in order get to their position, settle down, and read the data... If we had lower density IDE drives maybe we'd see faster seeks...
we seem to be coming to a convergence within the new future to where it may not matter SCSI vs IDE.. you'll just see drives that are based on their own merits.
*that reminds me, jtr, don't you think the high density of IDE drives is in part responsible for the high seek time.. it seems there is a balance... the higher your density the faster your read heads have to be in order get to their position, settle down, and read the data... If we had lower density IDE drives maybe we'd see faster seeks...
blakerwry
Storage? I am Storage!
ok, i should have proofed that...
What should have been said:
jtr1962
Storage? I am Storage!
blakerwry said:
In part it is definitely due to the density. I've noticed it's usually at least a year before you'll see SCSI drives with similar platter densities but much lower seek times. For example, right now SCSI drives with 36 GB/platter are common. IDE was using 40 GB platters over a year ago. Yes, lower seek times take time to engineer, but that's only part of the equation.
The other part, I believe, is that the typical IDE user is less tolerant of noise than the typical SCSI user. Let's face it, 99% of SCSI drives go into servers in some back room where nobody hears them. Noise just isn't that big of an issue here. And the 1% that are put into desktops the user is well aware that they are noisier and hotter, but is willing to tolerate it for the performance gain. A faster actuator by definition will be noisier. Noise can be engineered out as well, but even nowadays a fast actuator may be somewhat audible whereas we have gotten to the point where we can make 10 ms actuators pretty much silent. Same thing for higher RPM. My guess is we're finally seeing 10K RPM ATA because it took this long to engineer the idle noise down to acceptable levels. Sure, enthusiasts would have purchased noisier drives, but this isn't enough of a market to make it worthwhile. Let's hope they've developed quiet, fast actuators as well, or at least give the user a choice-9 or 10 ms with acoustic management and 5 or 6 ms without. And also let's hope there is not that much of a price premium over 7200 RPM. I'm not sure if I would want one just yet, but perhaps a year or two down the road, with further improvements in noise, heat, and price, it might be worthwhile to investigate.
The good news for everybody is that doubtless 5400 RPM and 7200 RPM drives will eventually become even less expensive, although they're already pretty inexpensive, and within a year or two 5400 RPM will be probably be phased out. So in the end we all get faster, quieter, and probably cheaper drives due to the trickle-down effect of a better top-of-the-line IDE product.
SteveC said:
Thanks for the update SteveC!
One thing that bothers me is the size. $160 is better than $200, but not that much better. Why would I not just buy the well known 10KIII for a few $$ more? Only drawback is buying the HBA. However the same applies to people requiring a SATA controller...
Maxtor Atlas 10K III "Orca", 36.7 GB, Ultra160 Wide LVD SCSI-3, 10,000 RPM, 4.5 ms seek, 8 MB cache, A/V Rated, 3.5" LP, 68 pin interface, 5-year warranty. $195.00
Pradeep
Storage? I am Storage!
Or the
Maxtor Atlas 10K IV "Cobra", 36.7 GB, Ultra320 Wide LVD SCSI-3, 10,000 RPM, 4.4 ms seek, 8 MB cache, A/V Rated, 3.5" LP, 68 pin interface, 5-year warranty. $219.00
Maxtor Atlas 10K IV "Cobra", 36.7 GB, Ultra320 Wide LVD SCSI-3, 10,000 RPM, 4.4 ms seek, 8 MB cache, A/V Rated, 3.5" LP, 68 pin interface, 5-year warranty. $219.00
Mercutio
Fatwah on Western Digital
A desktop drive with a 5 year warranty. That's awfully nice of them. Maybe they'll even bring out a model with a usable capacity someday...
I'll be in my corner waiting for Maxtor's 320GB drives.
I'll be in my corner waiting for Maxtor's 320GB drives.
Pradeep
Storage? I am Storage!
I nearly went for the Maxtor external 250GB 5400rpm drive from Dell the other day. Must control myself.....
blakerwry
Storage? I am Storage!
well, they're using a single 36 GB platter.. what's stopping them from adding another platter? cost? design? heat/noise?
I think 36GB is perfectly usable... most people would probably use this as their boot/App drive... I'm using ~16GB for that...
If you had just 1 drive in your case.. this wouldn't be it. But if you had 2, this would be a great boot drive.
I think 36GB is perfectly usable... most people would probably use this as their boot/App drive... I'm using ~16GB for that...
If you had just 1 drive in your case.. this wouldn't be it. But if you had 2, this would be a great boot drive.
blakerwry
Storage? I am Storage!
plus, i have to think the price will start to come down after several months of availability.
It will be interesting to see how this drive proves itself over the next year.
It will be interesting to see how this drive proves itself over the next year.
jtr1962
Storage? I am Storage!
Looks like they had the good sense to not put a slow actuator on it. 5.2 ms seek + 10K RPM should bring access times down to around 8.2 ms compared to ~13 ms for most 7200 RPM IDE drives. Not too shabby.
The only drawbacks I see are price and the need for either a S-ATA card or a P-ATA to S-ATA adaptor(hopefully they'll include one in the box). After six months I wouldn't be surprised to see CompUSA offering these as a $99 special, and the capacity is more than sufficient for a boot drive. Provided they're reliable, they should do fairly well, especially among those who want to try 10K but don't want to go SCSI. It should be interesting to see what Maxtor comes out with in response to this. I do agree though that they should also do a 2-platter version-it might add at most $40 to the retail price. After a few months you might have the 2-platter version going at something like $120 for 72 GB. I might even seriously consider one of these at this price and capacity, although I still prefer Maxtor so I might wait and see what they come out with instead.
The only drawbacks I see are price and the need for either a S-ATA card or a P-ATA to S-ATA adaptor(hopefully they'll include one in the box). After six months I wouldn't be surprised to see CompUSA offering these as a $99 special, and the capacity is more than sufficient for a boot drive. Provided they're reliable, they should do fairly well, especially among those who want to try 10K but don't want to go SCSI. It should be interesting to see what Maxtor comes out with in response to this. I do agree though that they should also do a 2-platter version-it might add at most $40 to the retail price. After a few months you might have the 2-platter version going at something like $120 for 72 GB. I might even seriously consider one of these at this price and capacity, although I still prefer Maxtor so I might wait and see what they come out with instead.
zx
Learning Storage Performance
It's nice that they make them in a 36GB capacity. Perfect for boot/os/app drives.
Performance wise, maybe the WD drives can outperform the Atlas 10KIV in desktop/high-end applications. It will probably depend on the firmware. WD 7200 drives have been good performers so maybe their first 10K drive will be able to offer better performance than their SCSI counterparts.
The only drawback is that it requires a SATA adapter. However, I would not complain. In canada, SCSI host adapters tend to be rare and very high priced compared to SATA adapters.
Performance wise, maybe the WD drives can outperform the Atlas 10KIV in desktop/high-end applications. It will probably depend on the firmware. WD 7200 drives have been good performers so maybe their first 10K drive will be able to offer better performance than their SCSI counterparts.
The only drawback is that it requires a SATA adapter. However, I would not complain. In canada, SCSI host adapters tend to be rare and very high priced compared to SATA adapters.
CougTek
Hairy Aussie
Although I have no problem to believe that the 10K rpm SATA WD will spank every other IDE drives when it'll come out, I have a harder time to see it in front of the 10K rpm SCSI champions. The actuators of the last generation 10Krpm SCSI drives like the Fujitsu MAP and Maxtor Atlas 10K IV are a at least half a second quicker than the one that'll be inside the WD Raptor. It will dominate the ATA market, but I won't place my money on it to beat its latest and greatest SCSI conterparts.zx said:
This doesn't mean it won't be a great product though.
blakerwry
Storage? I am Storage!
in the SR reviews it is intersting to see that although some drives score really nicely on file/web server type benchmarks they lag behind otherwise slower drives in desktop applications... for example the Fujitsu drives seem to do better in a desktop role compared to similar SCSI drives.
I have been told that this is can be attributed to firmware and cache design.
If the 10k WD is focused on the desktop market, i wouldn't doubt it beating SCSI 10k drives in this niche. However, I doubt it will be able to compete with SCSI drivers in the server section. I think SCSI command queueing and the server targeted firmware in most SCSI drives, the 10k III for example, will see to this.
I have been told that this is can be attributed to firmware and cache design.
If the 10k WD is focused on the desktop market, i wouldn't doubt it beating SCSI 10k drives in this niche. However, I doubt it will be able to compete with SCSI drivers in the server section. I think SCSI command queueing and the server targeted firmware in most SCSI drives, the 10k III for example, will see to this.
I don't think the WD is worth the money right now. I'll have to wait and read some reviews before I pass too much judgement on the drive. I wouldn't be surprised if the 10KIII or 10KIV is faster in performance.
Another thing I've been thinking about is SATA optical drives...how many exist, or plan to exist? It would suck to have a mother board with only SATA and have to buy an ATA card to run optical drives. Most SATA cards don't have more than 4 connections where as you can fit 14 on a U160 HBA. I know right now that the majority of boards have both SATA and ATA, but what about 1-2 years from now?
Unless the WD raptor is super fast, I don't know how it will do in the server market. It seems like WD is targetting high end work station or server class machines with these things.
Another thing I've been thinking about is SATA optical drives...how many exist, or plan to exist? It would suck to have a mother board with only SATA and have to buy an ATA card to run optical drives. Most SATA cards don't have more than 4 connections where as you can fit 14 on a U160 HBA. I know right now that the majority of boards have both SATA and ATA, but what about 1-2 years from now?
Unless the WD raptor is super fast, I don't know how it will do in the server market. It seems like WD is targetting high end work station or server class machines with these things.
blakerwry
Storage? I am Storage!
I see WD's market being enthusiasts, work stations, and low-med end servers (such as Dell powerEdge or custom built servers).
Come to think of it, I remember WD doing something like a single platter for their enterprise drives back in the day. Maybe they feel this is the way a drive should be. That, or they can't handle the 10K RPM's with multiple platters.
Platform
Learning Storage Performance
Part of the high price for this new drive is that it LIKELY does not employ a native Serial ATA interface. In other words, they are likely using a PATA - to - SATA bridge circuit on the drive which may run the cost up, oh... US$10 or maybe $5 or $15 (?). The reason I say LIKELY is that I have yet to see any mention of "Native SATA" in their press release or in any comments made by WD to others.jtr1962 said:
After 6 months I wouldn't be surprised if this drive was replaced by a whole new model -- certainly one with a less-expensive Native SATA interface, maybe even a SATA-2 interface(!). Also in 6 months, HitachIBM, Seagate, and Maxtor will have their 10kRPM SATA (and PATA ? ) hard drives out of the skunkworks.
Platform
Learning Storage Performance
Handruin said:
Once the number of mobos with SATA ports in everyday usage hits a particular threshold, you WILL begin to see all varieties of optical drives (CD-ROM, CD-R/W, DVD-ROM, DVD-+R/W, etc) become available with a SATA port, and hopefully not too many with bridges. Most of these will probably stay with a SATA-1 level interface even years after SATA-2 and SATA-3 arrive due to the relatively low throughput of these peripherals, unless it becomes more economical to use common parts with SATA-2 or SATA-3 peripherals. There will even be SATA floppy disc drives (SATA SuperDisk), supposedly.
I've heard that "economy" mobos in the future will have 4-each SATA ports, where as most technical workstation mobos will have 6-each SATA ports. One would expect that most PCI host bus adaptors will probably continue to have 4 ports, but you might find some with 6 ports after a while (power user).
High-end workstations will continue with 10kRPM and 15kRPM SCSI drives for a while longer, mainly because Ultra320 SCSI drives offer a wider range of storage capacities, and with the 10kRPM SCSI drives, probably just enough raw performance (for now) to still make them preferable with those people that have big budgets. The much larger mid-range technical workstation market (pre-press, CAD, etc) will see price/performance benefits with the 10kRPM SATA drives that will be attractive with the cost conscious mid-range crowd almost immediately. As SATA-2 and SA-SCSI emerge, you'll see 10kRPM SATA drives become a popular choice in many server environments, especially SATA-2 drives with a lot of storage capacity (7200, 5400 RPM).
Platform said:
I don't think so. WD should put the PATA version on sale as well then. SATA markets are still small, even smaller than SCSI. So it would not be profitable to sell just SATA drives, if the electronics are still PATA.
The press release says though "3.3 V Serial ATA power not used in this product.", but the 3,3V power is for laptops anyway.
According to the press release this drive is quiet; only 32 dB idle (average?) and 36 dB seeks. 2000JB has 34-35 dB idle and 36-37 dB seeks. So it's VERY quiet.
Cheers,
Jan
Jan Kivar said:
Yes, they DEFINITELY SHOULD also sell a PATA (ATA-6) version.
If they are indeed using a SATA <---> PATA bridge circuit, they probably have engineered this particular Raptor drive interface printed circuit board to where they can easily manufacture either a PATA or SATA hard drive by simply eliminating the SATA/PATA bridge circuit and adding the standard 40-pin IDC connector for PATA. The lack of a 3.3 VDC power connector is also a giveaway that this is a "transitional" product. SATA drives are supposed to use 3.3 VDC as the primary voltage, because the PC is supposed to use lower and lower voltages as each set of major generations are produced so that they eventually consume less power.
blakerwry
Storage? I am Storage!
how does using less voltage equate to lower total power usage?
am i forgetting Ohm's Law? is it true that while S-ata Hdd's might only use 3.3v instead of 5/12 that they will use the same or possibly draw less amps at that 3.3volts?
am i forgetting Ohm's Law? is it true that while S-ata Hdd's might only use 3.3v instead of 5/12 that they will use the same or possibly draw less amps at that 3.3volts?
sechs
Storage? I am Storage!
volts*amps=watts (in a rough sense)
If you pull about the same number of amps with less voltage, you use less power. There are some other advantages to lower-voltage electronics which lead to power savings.
If you pull about the same number of amps with less voltage, you use less power. There are some other advantages to lower-voltage electronics which lead to power savings.
blakerwry
Storage? I am Storage!
ok, i was being facetious... what i meant is that even though items like HDD's are going to be using lower voltages, they will probably still pull the same wattage.. thus, no power savings.
we've had PC HDD's based on practically the same design for over 15 years.. can they really keep increasing in speed and use less power?
i have some 10 year old HDD's... most of them pull the same amount of power as my new HDD's.... and on the laptop side you will notice that almost all HDD's pull the same amount of power(despite the RPM's) unless you get a "worsktation class" laptop drive.
This kind of tells me that there is an unwritten rule that a HDD should consume a certain amount of power... I expect this will remain fairly constant for awhile
we've had PC HDD's based on practically the same design for over 15 years.. can they really keep increasing in speed and use less power?
i have some 10 year old HDD's... most of them pull the same amount of power as my new HDD's.... and on the laptop side you will notice that almost all HDD's pull the same amount of power(despite the RPM's) unless you get a "worsktation class" laptop drive.
This kind of tells me that there is an unwritten rule that a HDD should consume a certain amount of power... I expect this will remain fairly constant for awhile
jtr1962
Storage? I am Storage!
blakerwry said:
It's called the law of conservation of energy. It takes a certain amount of power to spin a disk, which has air resistance, at a certain speed. Since the disks are already microscopically smooth, nothing more can really be done to reduce the aerodynamic drag. Bearing resistance is a much smaller component at the speeds drives usually spin, although it might be equal to air resistance under maybe 1000 RPM, and there is not much that can be done to reduce bearing resistance further since ball bearings are as good as it gets(discounting exotic technologies like air bearings). FDBs actually have higher resistance at the speeds in question, but are quieter.
Therefore, short of developing more efficient motors or reducing the RPM, nothing can reduce power any further. The brushless DC motors used in hard drives(and fans) are already something like 80 to 90% efficient, so perhaps there's a potential power savings gain of something like half a watt if they can get efficiency to 95% without increasing cost. Thus, overall there probably isn't much further reduction in HDD power in the future, especially on the drive mechanism end, unless we can devise a head mechanism that works in a vacuum, and have drives hermetically sealed. If that were done, we would only have bearing resistance to contend with, which would be well under 1 watt even at 15,000 RPM. Ultimately, all the power put into spinning a drive becomes heat which must be removed.
blakerwry said:
It's not that Ohm's Law doesn't matter (E over Eye R as they taught us way back when, not to mention Pie R Square but usually Pie R Round), but this is not an issue with Ohm's Law in the case of Western Digi NOT electing to use 3.3 VDC on their first SATA hard drive.
The SATA standards gods have encouraged the SATA drive manufacturers to use 12 VDC and 3.3 VDC -- much like most of the circuitry on your modern mobos already does -- for SATA hard drives. But, I believe 5 VDC is optional, but mostly for those manufacturers with transitional products (i.e. -- "old drives made new" by adding a performance robbing SATA bridge circuit). The 3.3 VDC is supposed to power ALL of the interface circuitry on the hard drive, not the spindle motor and head actuator voice coil.
As for the spindle motor, it would not be worth building a large-ish 5 VDC, much less a 3.3 VDC, high current motor to power a multi-platter'd 3.5 inch hard drive with stiff fluid dynamic bearings. So, it's definitely 12 volts DC for the motor and head actuator, 3.3 volts DC for the drive interface electronics.
jtr1962 said:
The trade-offs using fluid dynamic bearings versus ball bearings is a no-brainer to the hard drive designer -- with the exception of some very special low-power implementation where the lowest possible power consumption is top priority.
Otherwise, fluid dynamic bearings help quell internal spindle vibrations during NORMAL operations, and they offer an added measure of shock resistance in mobile applications. In every case, the use of fluid dynamic bearings adds up to a higher level of read/write reliability.
I also suspect that we recently entered a second (third?) generation of fluid dynamic bearing, as the Barracuda ATA 5 seems to be a tad quieter and cooler running than the Barracuda ATA 4.
The flying head assembly contributes to aerodynamic drag, as does the (frequently overlooked) inner surface of the drive housing. Once the drive is spinning at 100 %, the air inside the hermetically-sealed drive assembly is also spinning around in the same direction as the platters and at close to the same speed. Friction between the air and the inner surface of the housing, as well as the flying head assembly, translates into eddies that affects drag on the platter(s). So, what we have a tempest in a teapot -- a hermetically sealed teapot!