Raspberry Pi Gets Affordable, Power Efficient 314GB Hard Drive On Pi Day

Mickeycaskill writes: Western Digital has released a had drive optimized for the Raspberry Pi. The 314GB drive, released on Pi Day (3/14), costs $31.42 for a limited time and promises to be more reliable, power efficient and easier to use with the computer than other storage. The company, which also has a 1TB drive, says the unit has been designed to coordinate with the Pi's own power systems in order to minimize energy use without affecting the maximum data transfer rate on a USB connection. The Raspberry Pi Foundation says the new drive will stimulate the development of storage-hungry projects.

Comments

Waste

[bluefoxlucid]

I seriously doubt it costs less to manufacture this thing than a 1TB drive. WD has 1TB 2.5" drives with two 500GB platters; they'd lower the density and use the same parts to get a 314GB drive, or raise the density and use 1 platter to get 314GB. You're either dealing with one platter short (a piece of glass, coating, and a magnetic head) and a higher-density RW head (more expensive) or the same amount of hardware and a lower-density RW head (more tooling to bother making the damn thing). You might break even.

Re:Waste

[Anonymous Coward]

They ALREADY sell a 320GB drive, this is probably just that thing with a default partition table that doesn't allocate 6GB of it.

Re:Waste

[omnichad]

It's likely either a 500GB platter or 1TB platter with a high number of marked defects. Binned parts is nothing new.

Re:Waste

[BronsCon]

Actually, if the 1TB drive is using 2 500GB platters, they'd be lowering the density whether they used 1 or 2 platters. Or, perhaps they use a single 500GB platter and underprovision it, so they're saving on one platter while using existing parts and capitalize on a marketing opportunity.

Unless my understanding of math is wrong and 314 is suddenly more than 500.

Re:Waste

[viperidaenz]

A 2 platter drive uses more power than a 1 platter drive.

Re:Waste

[bluefoxlucid]

Only for initial spin-up. Once it's spun up, it's going to idle with the same power consumption. You lose friction in the spindle, but you store energy in the platters; as long as the spindle has the same surface contact, friction, and velocity, you don't lose any more power holding the drive idle.

Re:Waste

[viperidaenz]

Two platters have twice as much contact with the air around them.

What do you think suspends the head above the platter, against the force of the spring/arm holding it down? Air is forced under it by the surface of the platter..

Here's a random datasheet http://www.wdc.com/wdproducts/...
Idle power between 8.6W and 4.4W. They're all the same physical size, the same series, same RPM, same cache size, same interface.
Guess what the difference is? Number of platters and number of heads. There's also 4dB difference in noise level.

The largest power consumer of a modern hard drive is the motor spinning the platters, overcoming the friction of air against them. The platters are basically a Telsa Turbine - a bladeless turbine that's constructed from smooth spinning discs.

Re:Waste

[kenh]

they'd lower the density and use the same parts to get a 314GB drive, or raise the density and use 1 platter to get 314GB

You think they'd have to raise the density of a 500 GB platter to make it a 314 GB device? Lat time I checked 314 GB is 500 GB.

They could very easily just limit the storage of a 500 GB drive in firmware, preventing access to anything over 314 GB...

This drive requires a cable (not included $10-17 more), or you can just buy the $79.95 1TB PiDrive kit that includes the 1TB drive, power supply, required cables, and a case for the PiDrive.

Re:Waste

[AmiMoJo]

I wonder if they are recycling failed platters.

A few years back there was a company called Magnetic Data Technologies (MDT) that bought up broken hard drives returned to manufacturers, mapped the bad blocks, reduced their capacity to compensate and provide some new spare blocks, and sold them. They were incredibly unreliable, and even thought MDT offered only a one year warranty the return rate (at least where I worked) was close to 100%.

Maybe WD are just taking some platters that didn't have the minimum number of good blocks out of their factory bins and selling them as cheap Pi drives for users who will accept such a low volume of storage.

Re:Waste

[bluefoxlucid]

It depends on if they're starting with a 4 platter 1TB drive or a 2 platter 1TB drive.

Re:Waste

[Agripa]

Lower capacity drives built with exactly the same hardware cost less because they can spend less time on the machine which formats the embedded servo information.

And if the 1TB drive uses 2 platters, then the density for 314GB on 1 platter will be *lower* unless they only use one side.

Re:Waste

[Agripa]

You can go a step further and tell from the power specifications how many platters and sometimes heads are used despite the manufacturers leaving this information off.

Re:Waste

[Agripa]

If the 1TB drive has 2 platters, then that is 4 surfaces with 250GB per surface. Removing one platter and two heads leaves 500GB of capacity. Lowering that to 314GB means either 1 platter and 2 surfaces at a lower density of 1 surface at a higher density.

Bed Bath & Beyond holds "towel"-day sale on 6/

[Pseudonymous Powers]

The 314GB drive, released on Pi Day (3/14), costs $31.42

Stuff like this makes me wonder exactly what percentage of the U.S. GDP comes from pun-based purchases.

Silly Americans.

[ledow]

It's only Pi day if you don't understand precedence of unit size.

Small / Larger / Largest
DD/MM/YYYY

or Large / Smaller / Smallest
YYYY/MM/DD

None of them make a nice "Pi Day" number.

Unless you're considering the 31/4/1592 or 3141/5/9.

MM/DD/YYYY is just stupid.

Re:Silly Americans.

[xanthines-R-yummy]

Sounds like someone's jealous of not getting pie at work today!

(Currently gobbling slices of both pecan pie and apple pie!).

Re:Silly Americans.

[Anonymous Coward]

Here's the logic of the American System spelled out for you. Maximum possible values of the fields given MM/DD/YYYY format: 12/31/9999

To put it into words, the order is based on increasing numerical value, left to right.

Perhaps a more sane conceptual compromise to you would be W/D/MM/YYYY, such that max values would be: 4/7/12/9999 (not accounting for the occasional 5th occurance of a day in the monthly calendar).

Re:Silly Americans.

[Anonymous Coward]

We don't get pie until 1:51 :( Though at least they didn't enforce that to be the "proper" 1:51 AM.

Re:Silly Americans.

[barc0001]

YYYYMMDD works for PI day as most rational people only consider MMDD. It is an ANNUAL thing after all, so the year is ignored.

Re:Silly Americans.

[wonkey_monkey]

Here's the logic of the American System spelled out for you. Maximum possible values of the fields given MM/DD/YYYY format: 12/31/9999

There's nothing logical about that.

Re:Silly Americans.

[omnichad]

It's large to smallest with no year component. Most holidays happen every year, not just once.

Re:Silly Americans.

[supernova87a]

Actually, no it's not, you silly Brit.

There are many valid reasons in spoken English to have the month come first. Think about when you talk to someone about an upcoming appointment / meeting you want to set. It makes a great deal of sense to mention the month first, because that is some of the greatest uncertainty in the statement. If you say, "March 20th" you quickly know that you're speaking about a date coming up soon, versus if you say "the 20th of March", you leave out the month in suspense until the very end.

By the way, this way of speaking is so ingrained in American vs. British English that probably you think it's right regardless of whether it makes practical sense. And you probably think of it rarely, come to that. If there's something to criticize, you might as well start with the ridiculous decimal comma used in Europe number notation first.

From a technical point of view, it also can make sense when coding user interfaces, for example. Take a calendar choice user interface. If you were to have a sequential drop down (this is now a bit outdated of course) where the user selects the day number first. Then, when the next drop down box comes to the month, the selection of the next month could completely reset the number of days that you should be able to choose from. Doesn't make sense, and in fact Qantas used to have this method of choosing dates until they realized it was completely stupid.

Of course, for coding dates for all other purposes, it completely makes sense to use large-->smaller-->smallest. But spoken and written English doesn't have to mirror coding unnecessarily.

By the way, the Declaration of Independence was signed "July 4, 1776". I think we'll be keeping that, thanks.

Re:Silly Americans.

[Anonymous Coward]

I'll wait until Tau day. That way I'll get a whole pie, and not half of it!

Re:Silly Americans.

[jedidiah]

Sure there is. There is a single heuristic to govern the whole thing. It may OFFEND your PERSONAL PREFERENCE but that doesn't make it irrational or illogical.

Re:Silly Americans.

[reboot246]

It's "Steak and a Blowjob Day" no matter how you write it. Of course, I wouldn't expect anybody here to know that.

It's MM/DD/YYYY because when we say it out loud, we say March 14th 2016. What would you Brits know about the English language anyway? :)

Re:Silly Americans.

[Anonymous Coward]

Now I know why you guys write 23.1 for 1m and 23mm.

Re:Silly Americans.

[Anonymous Coward]

You seem to be autistic.

Re:Silly Americans.

[Roger W Moore]

By the way, the Declaration of Independence was signed "July 4, 1776". I think we'll be keeping that, thanks.

Yes, but rather ironically you typically refer to it as the "4th of July" don't you?

Re:Silly Americans.

[bws111]

We refer to the holiday as the 4th of July. We refer to the date it occurs on as July 4. Probably started doing that precisely so it doesn't sound like just any old date.

Re:Silly Americans.

[rahvin112]

March 14, 3pm and 9 minutes. 3/14 Hour 15, 9 Minutes.

Re:Silly Americans.

[Anonymous Coward]

"YYYMMDD works for PI day as most rational people only consider MMDD. It is an ANAL thing after all, so the year is ignored."
Fixed it for you

Re:Silly Americans.

[Anonymous Coward]

I say we ditch months entirely and use weeks.
Months are awful.

Happy 111th week everyone.
ISO Weeks.

Or, at the least, use this calendar.
Symmetry454
No more Friday the 13th. Stupid bad luck, get out of here.

Re:Silly Americans.

[Anonymous Coward]

In Dutch we use DD-MM-YYYY (and pronounce it in this order). We have a similar issue with pronouncing numbers. 31 is eenendertig, one and thirty. Even though I grew up with it and lived with it for 50+ years I still get confused if someone whose phone number is 2378623 (that's a random number) pronounces it as "three and twenty, eight and seventy, six hundred three and twenty". I easily write numbers down in the wrong order when I hear that and I will always pronounce phone numbers digit by digit myself. It makes sense to pronounce numbers in the same order you write them. I prefer the English thirtyone over the Dutch eenendertig.

But with dates the problem is that you pronounce the parts in a different order depending on your language and culture. That's why I prefer the format YYYY-MM-DD. As far as I'm aware there is no country where YYYY[-/]DD[-/]MM is used, so YYYY-MM-DD avoids the ambiguity of AA/BB/YYYY if AA and BB are both less than 13. It also sorts correctly without special measures to handle dates.

Some of the people who responded to you argued that MM/DD/YYY makes sense for several reasons that remind me of an Englishman I once knew who was full of reasons why driving on the left side of the road is better than driving on the right side. Full of reasons? Full of shit, it's just a convention. If it made any real difference entire countries would have safety problems, but obviously both systems work equally well in practice so there can't be a strong reason why one choice is better than the other.

When I exchange dates in a place where there can be misunderstandings I use YYYY-MM-DD. Avoiding confusing is more important than sticking to conventions.

Re:Silly Americans.

[jenningsthecat]

MM/DD/YYYY is just stupid.

Exactly. Just as expressing time of day in Minutes, Hours, Seconds would be stupid. By the logic of "MM/DD/YYYY", Pi hour would occur at 14:03:16 instead of at 3:14:16.

Re:Silly Americans.

[kenh]

None of them make a nice "Pi Day" number.

Unless you're considering the 31/4/1592 or 3141/5/9.

MM/DD/YY = 3/14/15 last year! - some went so far as to celebrate at 9:26 AM for 3/14/15 9:26 AM

Re:Silly Americans.

[Dog-Cow]

Wouldn't the time equivalent of MM/DD/YYYY be minutes/seconds/hours?

MM = 2nd most frequently changing unit = minutes
DD = 1st most frequently changing unit = seconds
YYYY = 3rd most frequently changing unit = hours

In any event, MM/DD is equivalent hours:minutes, so it makes perfect sense. Putting the least important information at the end (YYYY) also makes sense. Therefore, the assertion that MM/DD/YYYY is stupid could only be made by someone with absolutely no ability to reason.

Re:Silly Americans.

[Anonymous Coward]

Couple of points: Year in suspense. Year last resets the date in UI.

Re:Silly Americans.

[dave420]

Very true. The fact it is irrational and illogical makes it irrational and illogical.

Re:Silly Americans.

[jeremyp]

31st April? That's as fictional as 3rd Dodecember.

Since pi cannot be expressed as a fraction, we should just take the best approximation, which is 22nd July.

Re:Silly Americans.

[jeremyp]

MM/DD is equivalent hours:minutes, so it makes perfect sense.

Nope.

MM/DD is equivalent to minutes:seconds MM/DD/YYYY is equivalent to minutes:seconds:hours. In fact, the example gives the lie to the "it's the same way we say it" argument because people often say "it's 25 minutes past four" and yet nobody is seriously suggesting putting minutes before hours in written times.

Re:Silly Americans.

[Anonymous Coward]

April 31st, huh?

Re:Silly Americans.

[AmiMoJo]

It makes a great deal of sense to mention the month first, because that is some of the greatest uncertainty in the statement.

That's why the ISO date format is YYYY/MM/DD, which also makes sorting easier. If the goal is to put the greatest uncertainty first, then MM/DD/YYYY obviously fails.

Sorry, but it is silly. It's the worst of all possible options. No logical size ordering, and only puts the most important part first if the year happens to be inconsequential.

By the way, the Declaration of Independence was signed "July 4, 1776"

Which is commonly referred to as the "4th of July".

Re:Silly Americans.

[Anonymous Coward]

Those formats are more formerly known as little endian (least significant value first) and big endian. The American version is sometimes referred to as "middle endian". In the little endian form July 22 is known as "pi approximation day" as 22/7 is often an acceptable approximation for pi.

Re:Silly Americans.

[U2xhc2hkb3QgU3Vja3M]

Actually the ISO 8601 date format is YYYY-MM-DD, not YYYY/MM/DD.

Re:Silly Americans.

[Anonymous Coward]

From a technical point of view, it also can make sense when coding user interfaces, for example. Take a calendar choice user interface. If you were to have a sequential drop down (this is now a bit outdated of course) where the user selects the day number first. Then, when the next drop down box comes to the month, the selection of the next month could completely reset the number of days that you should be able to choose from. Doesn't make sense, and in fact Qantas used to have this method of choosing dates until they realized it was completely stupid.

Sounds like a solution looking for a reason to exist. By that logic, you want the year first as well, otherwise you'll reset the available dates for February. And then you have ISO dates.

Kind of expensive for 300GB

[mysidia]

The 314GB drive, released on Pi Day (3/14), costs $31.42 for a limited time

Considering it's only 314GB, and you can get 4TB 2.5" disks for $150.

The PiDrive gives you 10 GB per $1, versus 27 GB per $1 by purchasing 4TB conventional hard disks.

They have an interesting definition of 'data hungry'

Re:Kind of expensive for 300GB

[xanthines-R-yummy]

And what are the power requirements for those 4TB drives again?

The Pi drive, aside from the quasi-humorous capacity, was meant to be low power for the low power Raspberry Pi.

Re:Kind of expensive for 300GB

[Anonymous Coward]

What the hell are you talking about? The highest capacity 2.5" drive currently available is 2TB, and then to use any "normal" 2.5" drive with a Pi you need a SATA->USB adapter adding both cost and size (and likely not working at all with the anemic power output of the Pi's USB ports).

Re:Kind of expensive for 300GB

[Anonymous Coward]

Found the virgin!

Re:Kind of expensive for 300GB

[Anonymous Coward]

The poster forgets that the drive cost is not linear due to fixed cost portion. WD has a 3TB 2,5" model WD30NPRZ with a price tag of 1431 EUR at our local web shop.

Re:Kind of expensive for 300GB

[Pseudonymous Powers]

Found the virgin!

Playing "Find the Virgin" on Slashdot is a lot holding an Easter-egg hunt on a 100-acre intensive factory-style chicken hatchery.

Re:Kind of expensive for 300GB

[barc0001]

The integrated USB interface and power requirements are the draw here, though you could roll your own. You're paying for convenience. You know, the same way people buy a $2-$3 cup of coffee when it would be far cheaper to make it themselves?

Re:Kind of expensive for 300GB

[Aaden42]

Spoken like a true American. Moar is always better value, even if you don’t need all of it, right? Would you like your Pi[drive] supersized?

Consider it in terms of “cost per GB you actually use.” If you’re doing some kind of small embedded thing that needs maybe 200GB of storage for data logging (which is a pretty enormous amount for many purpose), then the cost of that 4TB drive was actually $0.75/GB. The cheaper PiDrive is $0.1571/GB in that case. That’s only counting the number of GB’s you actually needed in the division. Less waste is the better value in this case.

Getting ~200GB storage in flash form would run you a bit more than double that according to a quick Amazon search.

Then there’s the power issue. It would be nice if TFA or even WD’s site including actual power requirements, but I can’t find anything. It does look like they’ve included a unique (to me anyways) USB cable that lets both the drive and the Pi draw power directly from a wall wart, rather than drawing power through the Pi’s USB plug & power regulator. Flash probably still draws a lot less power overall if you’re trying to run on batteries, but at much greater cost. I didn’t see pricing on the 1TB version, but the costs probably compare even better on that to flash.

Re:Kind of expensive for 300GB

[jedidiah]

> Spoken like a true American. Moar is always better value, even if you donâ(TM)t need all of it, right? Would you like your Pi[drive] supersized?

I have stacks and stacks of multi-terabyte spinning rust and I am still at a loss to see point of this thing.

I also have 200G of flash on my phone.

Just put a decent ethernet port on the thing and let the bulk storage be some place else on the local network. If you are willing to settle for USB, network storage isn't going to be any worse.

Re:Kind of expensive for 300GB

[jedidiah]

> > Spoken like a true American. Moar is always better value, even if you donÃ(TM)t need all of it, right? Would you like your Pi[drive] supersized?

What can I say? I have the extra space AND the extra disposable income... '-p

Re:Kind of expensive for 300GB

[Anonymous Coward]

That is still far too expensive, and a spinning disk is a very bad match for a Pi. The new Pi alone is already straining the envelope of affordable USB power supplies. A Sandisk 120GB SSD is about $40. Put it in a USB2SATA enclosure for a couple of bucks and you'll be much happier. Yes, that's 50% more for a third of the space, but it's light, robust, absolutely silent and doesn't demand a beefy USB power supply. For any serious data storage, both options are much too small anyway and much too expensive for the space.

Re:Kind of expensive for 300GB

[mysidia]

And what are the power requirements for those 4TB drives again?

There is not a substantial change in power consumption relative to capacity of a 2.5" disk drive. Power consumption is mainly affected by what RPM rate the platter is driven at.

The Seagate Momentus ST4000LM016 2.5" Laptop HDD 4TB is reported to have datasheet Power consumption values: 0.85 Watt (Idle), 1.9 Watt (Read), 2.1 Watt (Write)

I think that's a datasheet power consumption for a 4TB disk less than the PiDrive what other Slashdot posters have listed as the WDPI Power Specs:
- low power (5VDC @ 0.55A = 2.75 watts)

Also, if you are really concerned about power consumption, and don't need all that capacity then consider a SSD such as $40 SSD, which should have more than 100GB.

Actually..... I doubt if the PI has sufficient CPU power to fill up 100GB very quickly with simple data logging.

Re:Kind of expensive for 300GB

[omnichad]

Considering it's only 314GB, and you can get 4TB 2.5" disks for $150.

And you can get a 50lb. bag of rice for under $20. Doesn't mean you need that much or that it would be worthwhile to buy 50lbs. every time you need some rice.

Re:Kind of expensive for 300GB

[omnichad]

But probably not as happy as if you just used a 128GB Micro SD unless you really need the speed.

Re:Kind of expensive for 300GB

[amiga3D]

I typically buy 50lbs every couple of years. It stores well.

Re:Kind of expensive for 300GB

[mysidia]

What's this, then? You haven't heard of the ST4000LM016 and Spinpoint M10P ?

Re:Kind of expensive for 300GB

[ChunderDownunder]

pairing a $35 computer with a $150 drive?

The point about this modest capacity hard disk is that it creates a cheap setup for well under $US100.

Re:Kind of expensive for 300GB

[omnichad]

And lots of people buy 4TB hard drives. They're very useful. Still not the point.

Re:Kind of expensive for 300GB

[Anonymous Coward]

Actually..... I doubt if the PI has sufficient CPU power to fill up 100GB very quickly with simple data logging.

Probably not, but it's easy to think of cases where that data storage capacity would be handy. Two very simple examples that spring to mind are photo storage (use a pi and a cheap small LCD to build one of those electronics picture frames that cycles through your pics) or a simple mp3 player.

I view a hard-drive more as a lifting of another restriction on what you can do with the pi rather than a must-have necessity for application x. I case of "for that minimal amount of money, why not?"

Wrong Date

[Harold Halloway]

Pi Day should be 22/7.

Re:Wrong Date

[Pseudonymous Powers]

Roundabouts whar I come frum, Pi day is just the third day of every year, that is, January 3. It's writ down in the Bible that way, and that's good enuff fer us. No fancy-froo dessimul points or fractions for us, city slicker!

Now, down the road a piece there's some idjuts who celebrate it on January 4. But they's just ignerant.

Re:Wrong Date

[Anonymous Coward]

That's just silly - there's no 22nd month! :P

Re:Wrong Date

[Anonymous Coward]

Ain't no 14th month either.

Re:Wrong Date

[RavenLrD20k]

Actually... Pi day should be recognized on 21/7, with the party slated to start no sooner than 23:47:15.236 UTC on that day.

Re:Wrong Date

[Anonymous Coward]

Why, that's just days away from Tau day. Don't go barging in on other math nerd holidays.

Re:Wrong Date

[amicusNYCL]

That date isn't significant enough.

Re:Wrong Date

[painandgreed]

Pi Day should be 22/7.

That is Pi Approximation Day.

Correction: the "Towel"-Day sale is on 28 June

[Pseudonymous Powers]

Dang it, Slashdot cut off my subject line two characters early. I guess you can't type slashes in there? I'd assume this was a Unicode-escaping thing, except everyone's always complaining about how there's no Unicode support on Slashdot. Anyway: All towels are $6.28. All washcloths are 62 cents apiece. Limit 6.28 per customer.

Re:Correction: the "Towel"-Day sale is on 28 June

[sbrown7792]

It's not the slash, it's the &

Discussed here: https://games.slashdot.org/com...

economics

[OrangeTide]

Supply and demand, if the supply of 1TB drives increases then the price of all of them go down. If an artificially reduced capacity can target a different market segment than the 1TB segment then only the products in that segment are impacted by the pricing effects.

Not exactly the same, but CPU and GPU makes have done the equivalent of this for years, different parts get binned for different speeds. Flash chips with too many flaws are send down to a lower grade and are less resilient to the inevitable wear. I'm sure there are further examples of you getting ripped off with waste.

Re:economics

[bluefoxlucid]

Supply-and-demand is a simplistic theory that generally describes economics in the same way that air pressure dynamics generally describe airplane lift.

Your theory has merit--it's a known market cycle in unsaturated markets--but it tends to fall apart in saturated markets. Competition tends to bring prices down in the long run. In the initial market, a bunch of SSD manufacturers have kept prices high without collusion in an effort to capture early adopters, with notable price differences between them; but as the early adopters started to saturate, the prices started more closely following actual costs. You start getting things like Samsung EVO 850 drives competing on price with friggin' OCZ budget drives. After that point, market destratification starts setting in pretty hard.

Binning parts that don't hold up to test is different. A 1GB flash chip with 10% defect rate can be a 9GB flash chip. A CPU process's variable nature can produce processors that run at 4.5GHz and others that run at 3.9GHz. You trash the broken ones, you repurpose the sub-par ones. That's not artificial; it's a form of quality control. Intel *has* binned high-performance parts as low-performance to balance for market demand, which *is* what you describe.

I don't think an artificially-lowered capacity with a lowered price is sustainable. A competitor will be able to provide something twice as big for the same price. Likewise, competition on 1TB drives will drive them down toward cost, and no amount of of supply-and-demand handwaving will make drives cheaper than cost in the long run--market saturation tends to cause this.

Re:economics

[OrangeTide]

Is the market for USB native drives for hobbyists a saturated market?

Part binning will take parts that are good, and fuse them down to a lower spec to keep the high end market from sliding down. It's artificial in that case, but maybe you weren't aware of that practice?

Most of the speculation on this should be taken with a grain of salt because WD is a weird company with a lot of weird products. We can't assume them to be a rational actor.

Re:economics

[omnichad]

A 1GB flash chip with 10% defect rate can be a 9GB flash chip.

Who knew a defect could be such a good thing.

Re:economics

[Anonymous Coward]

And you think that I have a bad memory!

Re:economics

[Anonymous Coward]

i bought one of the first wd 7mm 1TB drives as my only other operational bay in my notebook was low profile. ssd would cost a fortune for that cap. the hdd was c. $100 when i purchased it a few years back.

the notebook has another 9mm bay but due to a hw design defect it's useless(corrupts data clevo) so it has 120GB boot ssd, 750GB wd black 7500 rpm, plus the wd blue 1TB 5400 rpm iirc.

Nifty

[b0bby]

Ok, this isn't the best deal for TB/$, but it's not just a normal drive. I found the WD page:
http://wdlabs.wd.com/products/...
and it's a native USB drive, no SATA connectors. So that's pretty neat.

Re:Nifty

[Anonymous Coward]

All those crappy WD my passport drives have a "native USB, not SATA connector" bastardization. This is not a feature, but a punishment for people who want cheap drives.

Re:Nifty

[tlhIngan]

and it's a native USB drive, no SATA connectors. So that's pretty neat.

Actually, a lot of the USB 3 drives manufactured by WD ARE USB drives, there's no SATA interface exposed anywhere (maybe internally, but that's it).

They've been around for a couple of years now, if not longer, and are a complete PITA because if something goes wrong, you can't fix it by plugging the drive into a dock. Especially common human-based errors like the connector breaking off or distorted because they yanked it out crooked, dropped it with the connector attached, or pushed it in forcefully.

Re:Nifty

[b0bby]

Huh, I had no idea. Good to know. I haven't ever opened one of those mypassport drives up.

Re:Nifty

[Anonymous Coward]

The SATA interface is exposed on most of these PCBs at solder points E71 (XMT/A+), E72 (XMT/A-), E73 (XMT/B-), and E75 (XMT/B+). For these to work, a few nearby capacitors need be desoldered, which in turn "disables" the chip that handles the USB chip/interface. The drives themselves still fully support native ATA protocol. Sometimes, alternately you can find drop-in replacement PCBs that provide a native SATA interface (usually taken mobile 2.5" drives). You can find most of this information online (try www.hddoracle.com).

What isn't immediately disclosed on most (all?) of the sites is the fact that on these USB-interface-only WD drives, the USB chip is also what's doing transparent hardware encryption/decryption of the data that goes to the MCU to be written to/read from the platters. This is usually enabled/in place regardless of what's stated on the box (i.e. WD USB 3.0 2TB Passport "Silver" drives make no mention of encryption yet its being done). So even if you do all the wiring legwork or go through the pain of getting a SATA-based PCB (and dump the contents of two SPI EEPROM chips, U12 and U14, and move the contents of U12 to the SATA-based PCB's U12 -- these are either 8-pin SOIC or VSOP chips, i.e. SMT, and if they're VSOP you can't use a SOIC clip to dump/flash them due to the smaller-height form factor), and everything "magically works", your data is still encrypted.

There are several ways to deal with this, none of which are economically feasible for most end-users. The most common recommendation is to invest in a US$8000+ hardware/software tool called PC-3000 from ACE Labs. Another you'll seen thrown around is to use a SATA/USB bridge adapter used in WD MyBook drives (which is like playing roulette -- you don't know what adapter/conversion board you're going to get when you buy one of these things, so the odds of it being 100% compatible with your drive's PCB is unlikely).

What isn't immediately disclosed about the latter is that there are several USB ICs WD chooses to use, none of which are compatible with one another, and there's no standard/commonality in the encryption methods. Sometimes they switch IC vendors on the same product line (e.g. that Silver drive you bought 3 months ago might use a different chip than the one you bought yesterday). The most common I've seen are JMicron (there are two common ICs), and Initio (there are several ICs).

A full paper was published about some (not all) the different chips used and their encryption, several of which are "half-ass" -- but regardless of being such, still make it painful to get data from the platters in the case the main PCB fails. To my knowledge there is still no mainstream (e.g. free) software that can do this decryption if you were to give the software, say, a raw disk image (e.g. using dd) of the encrypted drive (still surprising, since at least for the older JMS538S it looks like it should be doable).

What also isn't disclosed about the "SATA replacement PCB" drop-ins is that they aren't always compatible with the physical hard disk enclosure. For example, it's stated that the SATA PCB version of the 2060-771961-001 is the 771960, however this isn't drop-in compatible due to use of physically larger SPI EEPROM chips. Why would the physically-larger chip be a problem? Because the actual metal/steel of the hard disk enclosure itself doesn't have a deep or wide enough cut-out where the chip would normally sit (when the PCB is mounted), so the PCB can't actually sit flush with the 18-pin I/O interconnect used between the drive and the PCB.

Source: my own experience, when attempting to recover data from two WD My Passport 2TB drives for a friend of a friend as a favour. (I restored 100% of the data off the Silver drive after doing a bit of work writing some scripts for hddsupertool that did VSC (vendor-specific) ATA commands, but wasn't able to with the Black due to what was likely likely a stuck head -- I don't do head stack replacements). I post

Re:Nifty

[jedidiah]

They're fit enough for the purpose. As always, it's the end user's requirements that ultimately matter. What some random snob has to say on the matter is entirely irrelevant.

Those of us that want better have plenty of good alternatives.

Re:Nifty

[Anonymous Coward]

Thanks, extremely good post. I only encountered one of these about 2 years ago. I had no idea about the encryption.
Are all of the WD USB PCB'd drives this way you mention? I have one here now I had considered looking into recovering:
http://i.imgur.com/hcQNlE9.jpg

Another situation that isn't widely known is many external hdds (esp from toshiba) which have a SATA port but use a removable USB adapter cannot simply be removed and used internally on SATA if you want to keep your data. The adapter translates between standard 512 block size and 4k. A friend had a 3TB toshiba drive he wanted to use internally.. had to move the data off slowly using the USB adapter to another disk and then back again once reformatting the drive internally.

Re:Nifty

[Anonymous Coward]

$31.42 is a pretty good price for a 314 GB SSD.

Re:Nifty

[AmiMoJo]

While annoying for data recover, I have to say that it's a very good thing that we are moving to encryption by default. Microsoft Surface tablets encrypt their storage by default too, as to many phones these days. There is a quiet little revolution going on.

The sad thing is that currently Linux support for self-encrypting drives isn't great. Well, neither is Windows support really, but it does work and provides a graphical interface that copes with things like S3 sleep modes. On Linux there is the sedutil, but it's still in the early stages. I use it on my laptop but it's pretty basic.

Re:Nifty

[Anonymous Coward]

Same AC here.

I'm all for disk encryption, assuming it's being done 1) with the user/owner's knowledge (i.e. at their request and that they've been educated as to the caveats of its use) and 2) there are multiple public software applications or suites that can decrypt the data in situations where data recovery is needed. The end-user needs to be in control of both. My feeling is that when someone purchases a drive, the data on that drive is theirs -- it shouldn't be essentially "locked" by some chip the buyer has no knowledge of.

Things like Microsoft's BitLocker and some other software-based (as in "OS-based") encryption fall into the "OK that's fine" category -- meaning the user has some possibilities/options where recovery may be possible (from a drive that went bad, say, able to dd 99.8% of it to a donor/spare). Linux and FreeBSD both have their own software-based encryption stacks as well -- I forget what the one on Linux is (sedutil rings a bell -- thank you!), but on FreeBSD there are gbde(8) and geli(8). I consider those worthwhile solutions for the time being. They're not flawless, but they're good in the sense they allow the end-user full control. I'm absolutely certain Windows has many choices as well (not sure about recovery scenarios, though -- most data recovery companies I've encountered will immediately reject the support case if any kind of encryption is knowingly used).

What I described in my post above is completely transparent, hardware-based encryption -- the USB bridge chip on the drive PCB is what's doing the encryption without the owner's knowledge. This isn't gdbe, geli, or sedutil -- this is proprietary stuff, hence the reverse-engineering PDF white paper. Equally as bad, there's no way to disable/turn off the encryption in advance of use of the drive (hence why I used the word "transparent"). The PDF referenced actually reads fairly easily; I don't have a college degree yet I understood about 85% of it. We're talking about vendor lock-in encryption, with no backwards compatibility either. It reminds me of hardware RAID controllers in essence, actually. You can't find open public documentation of the chips or encryption in question, nor can you find any public (free or reasonably-priced commercial (for end-users)) software that deals with it either.

This hardware-level encryption doesn't help with any scenario I can think of other than one: making data recovery painful. It doesn't help with stolen drives or misplaced/lost drives (Fortune 500s care mainly about those two scenarios) -- the thief or person who acquired the drive can plug it in and access everything, as there's no password or "key" or anything else. The Register published an article last year pretty much stating exactly this -- and my own experiences mirror their opinion. That said, there is one piece of software that is proprietary and hasn't fully been examined/explained: WD's SmartWare suite. To my knowledge, nobody is sure if this is what controls the underlying encryption bits of the USB bridge chips or if it's something else completely (I keep finding conflicting statements on data recovery forums).

Approximate Pi Day

[Roger W Moore]

Actually for the rest of the world surely 22/7 is close enough to be used as Pi day.

Re:Approximate Pi Day

[Anonymous Coward]

Actually for the rest of the world surely 22/7 is close enough to be used as Pi day.

I don't understand all this silly Pi day stuff.
Pi only gets you half the way there.
For me, I'll celebrate on 6/28 unfortunately named Tau day (for me tau is symbol for time-constant in electrical formulas). Others might know this as 2Pi day.

Re:Approximate Pi Day

[gwolf]

Approximate scripting...

$ python -c 'print 22/7'; ruby -e 'puts 22/7';echo $(( 22/7 ))
3
3
3

Re:Approximate Pi Day

[Anonymous Coward]

That right there is insane. Whoever implemented "/" like that in a math-based field like computer science is just bonkers...

Why aren't there any other stats posted?

[Narcocide]

No seek times, RPM, nothing? Are we not supposed to care?

Re:Why aren't there any other stats posted?

[U2xhc2hkb3QgU3Vja3M]

Because they're not important for most Raspberry Pi uses.

The important specifications are:
- direct connection via USB
- low power (5VDC @ 0.55A = 2.75 watts)

What's strange is their compatibility list:
- Raspberry Pi Model B+, 2 Model B, 3

So it's not compatible with the Raspberry Pi Zero?

Re:Why aren't there any other stats posted?

[Xenx]

I can't confirm it would work or not on the Zero, but you would have to at least purchase an adapter as the Zero only has microUSB.

Re:Why aren't there any other stats posted?

[Lord Crc]

So it's not compatible with the Raspberry Pi Zero?

I guess they couldn't get their hands on one to test with.

Re:Why aren't there any other stats posted?

[erapert]

Considering that the RPi shares Ethernet with USB 2.0 and thus results in pathetic IO rates even for a cheapo SBC you could probably get better data rates on a wax tablet than with even an SSD on an RPi.

If you care about IO speed on an inexpensive SBC then I think you'll need to look at the ODROID offerings:
ODROID C2 is limited to USB 2.0 but has real Gbit ethernet.
ODROID XU4 has USB 3.0 and Gbit ethernet.

Re:Why aren't there any other stats posted?

[U2xhc2hkb3QgU3Vja3M]

Virtual +1 Funny

Silly Pedantics

[OrangeTide]

I don't agree that a society is "stupid" for ordering dates based on how they are typically announced in English.

In English the options are:
March 14th, 2016
14th of March, 2016

With the later being somewhat archaic sounding in the US, except for the Fourth of July. (the whole "st", "nd", "rd", "th" thing is pretty old fashion too, but I used it in my example for clarity)

Re:Silly Pedantics

[U2xhc2hkb3QgU3Vja3M]

This only raises another question: why is the month before the date in spoken English? It still doesn't make any sense. Just because you've done something wrong for centuries doesn't make it right by default.

Re:Silly Pedantics

[Anonymous Coward]

How is it wrong? Knowing the number "14" would be entirely useless until one knows it is "of March", so the listener is forced to suspend parsing of the sentence.

If someone says "March 14" I can sequentially narrow down the time frame in question as they speak, thus it is even more efficient.

Who are you to claim one method is "wrong" versus another anyhow?

Re:Silly Pedantics

[bws111]

How is it 'wrong'? Language is determined by how people use it.

Re:Silly Pedantics

[Lumpy]

Because the month is the most important time measurement. then the day and then lastly the year.

The year is useless. Knowing if it's august and I need to start thinking about harvest time or it's april and I need to get my ass moving on planting is far more important.

Re:Silly Pedantics

[Anonymous Coward]

Here is the explanation. People used to use these things called paper calendars and planners. When you wanted to index to a certain date in the calendar (for writing down an appointment or the like), you first found the month, then you found the day of that month. It was generally assumed that you weren't making an appointment a year in advance or that occurs in the past so there was less need for the year and it came last. Very simple. Page to the proper month, index down to the proper day, write down the appointment.

Re:Silly Pedantics

[Anonymous Coward]

I guess I'm archaic then (though I'm not from the US). If someone asked the date I would say "the 14th of March 2016". March 14 makes me cringe a little inside... it just sounds wrong and a little pompous. I can imagine Morgan Freeman saying it like that if he was playing god (again).

Re:Silly Pedantics

[viperidaenz]

My calendar has 31 pages with 12 months on each page, you insensitive clod.

Re:Silly Pedantics

[Anonymous Coward]

I think you need to prove it's "wrong" first, the burden of proof is on you. The order is irrelevant unless you are using electronic filing techniques. But given that IBM, and American company, has been doing it on a commercial scale before any one else makes me suspect that the date order is of little consequence.

Re:Silly Pedantics

[omnichad]

And the reverse is true here. I guess we should just ask Morgan Freeman what his preference is.

Re:Silly Pedantics

[Anonymous Coward]

In Australia most commercial advertising is said or written with month before day. Id say it's to shorten words in print and say it quicker in TV and radio.

Other than that the official way is Day/Month/Year.

Re:Silly Pedantics

[kenh]

Because the month is the most important time measurement. then the day and then lastly the year.

How did month become more significant than year?

Re:Silly Pedantics

[dave420]

Descriptive languages are, but not prescriptive languages. Luckily for you English is the former and not the latter :)

Re:Silly Pedantics

[Dog-Cow]

Most people live in the present. Unless one is speaking of an event in a different year, it's completely superfluous. The month can also be superfluous, but that doesn't happen as often because the span is only 28-31 days.

Re:Silly Pedantics

[jeremyp]

Ah, but they had a different calendar for each year, so the first thing to do is make sure you have the calendar for the right year, which means the year should come first.

Anyway, we are not talking about how people say dates but how they should be written down in numerical form. The best way is to put the largest unit first, then the next biggest and so on (e.g. ISO format). The second best way is to reverse the order. The way the USA does it is just insane, which is why most of the World does something different.

NB if we extend the format to include the time of day, the rest of the World is nearly as bonkers as the USA since they use one convention for the date and the opposite convention for the time. And don't get me started on where the am or pm should go if you are using the 12 hour clock.

Re:Silly Pedantics

[jabuzz]

Note that month before day is normally only used in English when there is no year involved. When is you birthday and you might respond March the 15th. When where you born and you would respond the 15th March 2000.

However writing it down as American's do has nothing whatsoever to do with the English language, because the English (aka the people who actually invented the language) use DD/MM/YYYY like the rest of the civilized world. It's only the American's who insist on using illogical date orders.

Re:Silly Pedantics

[gordguide]

When I was in College, long before 1984 and the introduction of dates in desktop computers (in other words, I had no idea how anyone else did it), I numbered my lecture notes sequentially; the only way to do that is Month / Day / Year.

010612.02 would be Jan 6 2001 page 2

If you do it any other way, the date becomes your sequential indicator, which breaks it as far as determining which page was written first (the day of the month becomes the "first" page, so 1 Jan and 1 Feb follow each other, not 1 Jan followed by 2 Jan)

Re:Silly Pedantics

[RockDoctor]

I don't agree that a society is "stupid" for ordering dates based on how they are typically announced in English.

You haven't had to deal with daily reports where one person in the job uses a US numbering system, and his other-month opposite number uses Norwegian numbering, and you're the schmuck who has to sort it out for a 300-day operation.

TTBOMK, the only modern society which rationally chose a date representation was the Japanese, after their self-imposed isolation was broken forcibly by the US Navy. You will note that they use the date representation with left-to-right decreasing significance of all numbers. Because ... they're just fucking numbers, with nothing special about them.

Every fucking month, when I get a new trainee, I have this conversation about "this is WHY I always record dates in ISO 8601 format". and it doesn't matter if the trainee is American, German, Dutch, Norwegian, British, Turkish, Gabonese, Beninoise, Nigerian, Tanzanian, Korean, Chinese, or Canadian. The same logic is applied to them all, and they're told to use ISO 8601 format.

Re:Silly Pedantics

[Anonymous Coward]

You may find this surprising, but most of us can keep track of more than one piece of information at a time.

Re:Silly Pedantics

[OrangeTide]

You haven't had to deal with daily reports where one person in the job uses a US numbering system, and his other-month opposite number uses Norwegian numbering, and you're the schmuck who has to sort it out for a 300-day operation.

I've actually had to write software to parse multiple date formats. It was a bunch of annoying extra work that I despise because societies were not able to agree on a convention. Hell, societies don't even agree on a calendar, there are a bunch of them. And if you ever run into really old records, the Russian Empire didn't upgrade the Julian calendar to the Gregorian calendar until the early 20th century. It's a real pain to automate digging through old records without also providing enough context for the software to apply the appropriate date convention.

It's good that you're standardizing your dates within your business. Of course the second you have to deal with a customer outside of your business you start to have the same problems again. Now what's really fun is when your orderly company merges with a company with less rigorous processes and you get to import old databases, timecards, payroll, etc.

PS - my preference is unix time as a 64-bit signed number. Of course, then you must ask, did I record it in decimal or hexadecimal?

Re:Silly Pedantics

[RockDoctor]

Yep, had to deal with Pre-revolutionary Russian dates. Which is why astronomers (who do, indeed, regularly have to deal with periods outside the Gregorian calendar) use the Julian Date, which is reasonably consistent from (I'd have to check ; "about") 8000 BCE to 8000 CE.

It pisses me off that it's necessary. But getting the mess sorted out (after identifying the problem) has paid my mortgage in the past : two companies, neither mine, blaming each other over the failure of a piece of equipment. One company recorded date-time in US-ian. The other recorded date-time in what I eventually identified as Unix epoch seconds. And one clock was off (w.r.t. the other ; I had no external reference) by a day and an hour, or something. Hey - it paid the bills.

Out of stock

[psergiu]

Already "Out of stock" on the WD website :(

Re:Out of stock

[SeaFox]

Already "Out of stock" on the WD website :(

I heard they only produced three of them, and gave up right after starting the fourth.

Re:Out of stock

[Anonymous Coward]

Now back in stock. Just ordered one.

Not Logical, Historical

[Roger W Moore]

Here's the logic of the American System spelled out for you. Maximum possible values of the fields given MM/DD/YYYY format: 12/31/9999

So what's going to happen in 9,999 AD that will prevent 10,000 AD from happening? Besides the really illogical thing is that when talking about your independence day you use the British date format: "4th July" and not "July 4". I get that this is how you write the date but please don't try to claim it is logical: it is a historical artifact just like pronouncing the 'gh' in 'rough' as 'f' in English.

Common Core Pi Day

[Anonymous Coward]

Well, in my daughter's school, the "official" value of Pi is "about 3."

So can we say that Pi day is celebrated on the "14th day of one of the early months in the year?"

Re:Sorry about the "troll" mod

[Anonymous Coward]

They may have said they thought you were kidding or crazy. What they actually thought is that you were an obnoxious asshole raising his kids to also be obnoxious assholes. And they were right.

Out of stock

[Anonymous Coward]

Don't bother, they already ran out of them.

big to little order

[OrangeTide]

Maybe, but I hope we can agree that "Twenty Sixteen March Fourteenth" is not acceptable in verbal English communication.

Re:Bed Bath & Beyond holds "towel"-day sale on

[omnichad]

On Pi day? I've already received a couple of pizza (pie) deals in my email. But that 3.14% discount I got elsewhere wasn't worth much.

Ship or it didn't happen..

[0dugo0]

Show me a working order page that ships it to my door for that price, or I'm back to AliBaba and friends ordering real goods.

30 Megaseconds

[Anonymous Coward]

My year is broken into 32 1 Msec intervals. (about 11.5 days each)

to get the date this year, take the current unit time and divide by 31,557,600. Result + 1970 is the year, remainder divided by 1,000,000 is the interval (month-like or week-like). The remainder is the seconds into the current interval.

For Today (Mon Mar 14 21:32:18 UTC 2016) the answer is: 2016:6:50082

Source for a case?

[Anomalyst]

Is there a case available that will affordably, securely and neatly house the drive and the board together and tidy up the appearance of the interconnects?

Re:Source for a case?

[ChunderDownunder]

WD list a bunch of accessories.

Re:Source for a case?

[Anonymous Coward]

http://store.wdc.com/store/wdus/en_US/DisplayAccesoryProductDetailsPage/ThemeID.21986300/productID.330357600/WD_Pi_Enclosure%2C_Black

What additional benefits do I get

[Anonymous Coward]

At $30USD doesn't seem cheaper. But other than what technical benefits does it give you. It looks like it still uses the USB port, so it won't be much of a NAS.

I think a USB2 enclosure with an SSD may be lower power consumption than a platter. Equal throughput, but the SSD maybe higher IOps

I guess you can collect data like images and record it to the drive.

So what benefits to this vs a regular USB Drive

Day is before month even in spoken English

[Anonymous Coward]

"The 14th of March" is correct English.

Can't believe how many weird ways there are to get that wrong and how many people want them to be considered as right.

But 14/3 is 4.666

[Anonymous Coward]

22/7 is a much better date. 14/3 isn't even correct to 1 digit.

What is a had drive?

[Anonymous Coward]

Western Digital has released a had drive optimized for the Raspberry Pi.

Is no one editing submissions any more?

The Pi needs fixed, not a hack for everything

[Anonymous Coward]

I'm amazed at the number of people who buy into the Raspberry Pi. We have better boards out there at about the same cost. The Cubie Board, Bannana Pi, and BeagleBone Black are all superior and more reliable / well designed than the Raspberry Pi with its flaky and proprietary software dependencies. I guess when you worked for the company you've got an incentive to go to them for your components in hardware your designing, but man, it's put a real damper on the product.

Best recoverable RPi drive?

[bobo_1968]

Given the comment about data recovery here (https://hardware.slashdot.org/comments.pl?sid=8878179&cid=51696039), what is the current recommended mass storage for Raspberry Pi? Seeing that lower power consumption and USB are desirable, it would be nice to be able to backup/do recovery over SATA and get back non-emcrypted data if possible. Also, couldn't you just use LLVM to encrypt the drive if you wanted. Or does Raspbian not support that?

Pi Drive

[CK_WD]

Hello everyone! :) I see that there are many of you interested in the new Pi Drive. The demand has been quite high, so it may take some time till the product's fully available in the mentioned countries on the WD Store page, through where it is currently sold only. Best regards, CK_WD Official WD Representative