Feeds

back to article HGST floats helium for low power, MASSIVE capacity HDDs

A new hard drive technology from HGST promises to improve drive performance in virtually every category, the company says – including capacity, power, cooling, and storage density – all because the drives are filled with helium instead of air. "The benefits of operating a HDD with helium fill have been known for a long time," …

COMMENTS

This topic is closed for new posts.

Page:

Depends on cost, but

I might almost rather spend even as much as $2X-$4X per disk on this, or possibly even more, than $X on an SSD. More expensive, sure, but it'll be larger as well, and given the destructive nature of SSD writes and the inescapable lifetime limit that imposes on any such device, it'll probably last longer too for the sort of busy workload where SSDs' speed really shines. Considering that a properly configured RAID, across enough disks, can make up for a lot of that latency difference and offer improved reliability as well, this really might turn out to be worthwhile.

0
0
Alert

Do 5TB drives matter?

As anyone who uses SSD knows, it isn't the capacity that matters, it's the IOPS. Well, given that a tiered system will be de rigeur in a couple of years, even in the cloud, we need to ask what characteristic is most important for the secondary storage (HDD). The answer, in a micro-tiering, as opposed to block level tiering system, is the sequential IO MBPS.

Objects will be transferred from slow, bulk storage to flash and SSD sequentially, and the only way to avoid high latency is to parallel up the HDD to more closely match SSD/PCIe bandwidths. The net bottom lone is we'll need a few HDD to solve the performance matching issue, probably in a RAID configuration, so we'll likely end up with a ton of TeraBytes just to satisfy the MBPS requirement.

That tells me that those extra TBs are a bit unnecessary, unless they are achieved by higher bit densities. Helium technology allows more platters and so the data rate is unchanged. This sounds a bit out of whack with the market need.

1
0
Bronze badge

More platters mean more heads mean more data rate

Shirley?

2
0
Silver badge

Re: Do 5TB drives matter?

If you are using hierarchical storage, then you get benefit from having fewer, but high-capacity low-level storage. Fewer disks (hopefully) mean fewer failures (maintenance callout costs) and more cash to spend on cache further up the hierarchy. Your upper tiers take the IOPS strain for you.

Plus, there are plenty of applications where you want lots of data, but don't need to access it very often. Say hello to the home market. Anyone doing photography might appreciate it. Lots of very hi-res images, but you only need to access a few at a time. Or backups. Yes I want timemachine to go back months, but I rarely need access to that data. Or I might be spooling off backups to disk and then transferring them to tape.

Requirements are generally tuned for what is available. There usually is a niche for "more", even if it isn't in the standard server market.

0
0

"a bit out of whack with the market need"

In light of the push toward virtualization, I might beg to differ --

0
0
Silver badge

Re: Do 5TB drives matter?

They should also make for more economical archiving drives, pushing tape even further down, especially in the consumer end where tape is impractical these days (it seems these days it's the external hard drive or bust).

1
0
Stop

Archive with Helium? Not a Good Plan!

They use Helium to test for leaks which is a clue that it's very good at doing just that. Some of your archival drives are going to fail strangely because the Helium has leaked out over the years. As it is, the big concern with using spinning disks for archival is that they don't have a specified shelf life like tapes do.

So, call me sceptical, but for now our archival drives are going to be Helium free.

1
1
Boffin

Re: Archive with Helium? Not a Good Plan!

> They use Helium to test for leaks which is a clue that it's very good at doing just that.

It's true that Helium will leak or diffuse faster than gases other than hydrogen. However, Helium is used for leak detection because it's not normally found in the environment and thus is easy to detect. Hydrogen seems like a better choice for leak detection, but in practice there's so much Hydrogen around that the noise floor swamps the detection of significant leaks. It will be interesting to learn how the drives are sealed and whether there will be a reservoir of helium to maintain positive pressure during the lifetime of the drive.

1
0
Anonymous Coward

Re: Archive with Helium? Not a Good Plan!

about 40 years ago I built apparatus which held helium at 2000psi in a stainless steel container with a quartz window, and filled via a high pressure gas valve. it worked for several years without apparently losing the helium . So I would hope that the containment of low pressure helium in a sealed disk drive should be a piece of cake by comparison.

0
0
Vic
Silver badge

Re: Do 5TB drives matter?

> it isn't the capacity that matters

Tell that to my users... :-(

Vic.

1
0

Re: Leakage, Helium v. Hydrogen - Correct me if I'm wrong...

Isn't the Helium atom smaller than they Hydrogen atom because of its extra mass and self attractive forces? I seem to remember reading some where that despite being a tad heavy than H, He was smaller and 'slipperier'.

The note about being more detectable because of rarity is right on.

0
0
Vic
Silver badge

Re: Leakage, Helium v. Hydrogen - Correct me if I'm wrong...

> Isn't the Helium atom smaller than they Hydrogen atom

No.

Vic.

1
0
Anonymous Coward

Re: Leakage, Helium v. Hydrogen - Correct me if I'm wrong...

It is however smaller than the hydrogen molecule. Atomic hydrogen doesn't last long in the wild.

0
0

Airless

Why helium? If it's drag they're worried about, could the innards not be run in a vacuum? Surely the same technology that's used to keep the helium in could keep the air out.

2
2

Re: Airless

Possibly thermal conductivity, in a vacuum you would be relying on the radiation of heat from the internal components rather than convection through the fill gas - in short, you may find the bearings and platters overheating and doing things they shouldn't.

2
0

Re: Airless

A vacuum is a lousy thermal conductor. As it turns out, in a HDD, a fair amount of the heat from the motors (main rotor + steppers) disperses via the platter cavity.

The trick with Helium is the sealing; the good news is that this implies that the platters will be hermetically sealed, which means they'll operate at altitude without an additional enclosure.

1
0
Bronze badge

Another problem with a vacuum

You'd have to have a case that would withstand atmospheric pressure. The walls of this, especially if in the unsuitable far-from-spherical shape of a standard disk drive, would have to be so thick you'd have difficulty getting any platters inside.

People often think that double glazing units are evacuated. They aren't, for the same reason. Airships would be more buoyant with vacuums - but they'd collapse.

0
0

Re: Airless

You've all missed the main problem with trying to run the disks in a vacuum chamber: in all hard drives the heads fly above the disk surface on a thin layer of air - think of it as an air bearing. This will work with any gas provided the operating conditions are adjusted for the density and viscosity of the gas in the chamber, so helium should be OK.

However, the heads can't fly in a vacuum.

Current drives do not have a sealed chamber so they always operate at the local atmospheric pressure. Consequently they have a specified maximum operating altitude, usually around 10,000 ft. Above that height the air is too thin to hold the heads clear of the disk surface and the heads will crash, destroying both themselves and the disk's recording surfaces. Airliners are pressurized to about 8,000 ft, which is why you can use a laptop while flying at 30,000 ft. in one

9
0
Silver badge

Re: Another problem with a vacuum

Actually, you WANT poor thermal conductivity in a window. Poor thermal conductivity means poor heat flow and less chance of heat transfer in or out of the house, which is what you want in a modern climate-controlled house.

It IS actually possible to double-glazed windows were evacuated...FOR that reason. But these days, they also use dense, inert gas fills like xenon which, although not as poor a heat conductor, are cheaper to make and can have their own benefits.

1
0
Silver badge

Re: Airless

You can take the air out if you want, but this tends to cause head crashes.

HDD heads ride on an aircushion. (hence the term "fly height").

There may be other ways of achieving levitated heads but they're a lot more complex to implement.

0
0
Bronze badge
Meh

Re: Airless

I alwasy thought HDDs *were* vacuums inside?! You learn something every day...

0
0
fnj

Re: Airless

@David 45 - the flying heads are enabled by fluid pressure. In a vacuum they would instantly crash. So perfecting a vacuum filled drive would be a lot more work. You'd have to figure out a new way to keep the heads suspended precisely 3 nanometers above the surface but never touching.

0
0
Bronze badge

"[air is easy to] seal into a hard drive enclosure"

Except that air is not sealed into a hard drive enclosure. There is a long thin channel connecting the inside and outside to allow pressure to equalise without enough air passing that contamination will get inside. (The entry of this has a warning not to cover it.) Otherwise, the thin case would deform as outside air pressure changes. That is why disks work at altitude, but only to a certain extent: if you took them into space they'd evacuate completely and when brought back into the atmosphere would suck in contaminated air.

0
0
Anonymous Coward

Price of helium?

Is the writer seriously suggesting the price of a few cm³ of balloon gas is going to have any significance at all? Surely the costs involved are in sealing the electrical connections to the heads and strengthening the casing to cope with atmospheric pressure.

(I'm assuming they'll have to fill at a pressure significantly lower than the drive is likely to encounter in service so as to prevent the He just permeating through the seals. They have to allow for pressure increase caused by internally generated heat too.)

0
0
Anonymous Coward

Re: Price of helium?

Other way round. They will want a slight overpressure to keep the air out,

I imagine that this will work something like thermionic valves in reverse - they had a getter (a clean active metal surface) to remove any air than leaked in. I imagine these will have a small helium reservoir of some kind.

The fun way to do it, of course, would be with a lump of Pu-238 or Po-210, which emit plenty of helium nuclei. Someone needs to start up a Pu-238 reactor again, it's good stuff.

0
0
Gold badge
Trollface

Why not hydrogen?

@handle, I have not seen a drive where all the heads are not rigidly mounted together (so they are on a different platter but reading the same track.) MB/sec can increase, but random access times do not.

@Jim O'Reilly, yes those who choose SSDs do it because the IOPs matter. Well, for some purposes the cost per GB for SSD is far too high, and the storage capacity far too low. SSD is close to an order of magnitude more expensive than a hard disk after all.

@AC re "Price of helium?", in actuality, the US strategic helium reserve has been ordered to be sold off between 2005 and 2014 -- a reserve built up over the course of about 70 years (originally for blimps) being sold off in 9 years. This has caused other helium production to stop -- the natural gas refineries have helium as a "waste" gas, which was filtered and collected, but now is vented into the atmosphere never to be seen again. The prediction with current production and use rates is that in 20 years a balloons worth of helium would cost like $100, and it'd run out in 25 years. Yeah, that'd drive up the cost of a helium-filled hard drive 8-).

I wonder why not use hydrogen? I would not think a hard drive would hold enough for a titanic situation to be too likely after all, and production would be quite easy of course.

0
1
Silver badge
Boffin

Re: Why not hydrogen?

Helium is inert but hydrogen is nasty stuff that can play hell with steel. This is because in regular steel there is typically some carbon content and hydrogen will bind to carbon forming methane which greatly compromises the strength of the metal. While the drives are mostly aluminum and not really susceptible to hydrogen embrittlement, the protective coat of the platters is carbon based and probably wouldn't last long give the high concentration of hydrogen.

With all of that said, I don't see why they couldn't put four or six 1-2 TB laptop platter sets into a single 3.5" housing as a self contained raid array that not only provides high throughput but also a bit of built in redundancy for a marginally higher cost.

2
0
Silver badge

Re: Why not hydrogen?

Because the laptop platter set is 2.5" wide, and you've only got 3.5" of leeway. And you can forget trying the sideways approach; that won't fit, either. As for sub-laptop sizes, they spin at lower speeds and are not designed for consumer-level computer demands (they're meant for embedded use).

0
0
Bronze badge

@Henry Wertz 1

"MB/sec can increase, but random access times do not."

Yes, which is why I was talking about data rate, not access time.

0
0
Bronze badge
Boffin

Re: Why not hydrogen?

Because Hydrogen has very tiny molecules, even compared to Helium, which means that it is capable of leaking straight through a solid piece of metal...

1
0

This post has been deleted by its author

Joke

Real Soon Now...

Helium will be an abundant by-product from our fusion reactors.

So, no worries then! ;-)

0
0
Happy

Helium?

Helium is monatomic, meaning that it does not make molecules - it's a single atom gas. the only way to contain it is in metal to metal seals - as it percolates THROUGH anything else, eventually.

I suppose if the disk chambers were at slightly below atmospheric pressure, then the helium would stay inside, and the air outside.

I am curious as to why they do not use hydrogen?

0
1

Re: Helium?

Hydrogen is reactive. When concentrated like this, or over a sufficiently long time anyway, it does nasty stuff. Helium is inert.

0
0
Unhappy

I dont' see the fuss in building a completely sealed device. I have sensitive dive electronics in cases that stand pressure changes 10x more than you'd get taking a drive up to the top of Mt Everest.

What i'm not happy about is yet another industry using up Helium which is already so expensive it makes deep diving a seriously expensive hobby. If it goes up much more, it will spell the end of technical diving as a hobby :-(

0
1
Vic
Silver badge

> Helium which is already so expensive it makes deep diving a seriously expensive hobby.

You want an Inspiration, matey :-)

Vic.

0
0
Anonymous Coward

Pot, meet kettle.

I imagine your use in diving (which is all wasted) would be far more than a little gas even in very many hard drives. Talk about pissing a valuable resource up the wall!

Perhaps you should join the club of those US hemi truck owners who complain that gasoline is too expensive and think other people shouldn't be allowed to use it.

0
0
Happy

I've got this image of techies wandering around data centres in five years time squeaking to each other like deep sea divers because of helium leaching ...

2
0
Trollface

I can just see it now

User: "The servers are running a bit slow today."

Tech: "Ahh, one of the drives must have sprung a leak,"

User: "Well can't you just fill it up again?"

Tech: "We would, but we used up all the spare helium doing funny impressions last week after the pub."

2
0
Coat

Double glazing

Mass-market double glazing sealed units aren't filled with anything exotic, just dry air. The silver metal strips around the edge which separate the two panes have many fine perforations, and are packed with silica desiccant. They unwrap the strips from their sealed packaging, cut to length, assemble the pane and seal the whole thing with tape, the air in the gap soon enough becomes dry enough to make condensation impossible, and that's all there is to it.

Anything more exotic and it would be impossible to assemble arbitrary sizes in a plain light industrial shed.

0
0

Re: Double glazing

@hugo

Uhm, we moved into a new build in 1979 that had triple glazed units filled with argon IIRC. And I don't think those units were made by NASA.

That wasn't some footballers faux mansion, just a terraced new build. It was in a different, colder, country though.

</off topic>

1
0
Bronze badge

Re: Double glazing

Not sure what point you're trying to make - I mentioned double glazing as something which is not evacuated because it is not strong enough to withstand the pressure. Doesn't matter what gas is between the panes; the point is it is at approximately atmospheric pressure (and as atmospheric pressure changes, the panes bend quite visibly).

0
0
Silver badge
Boffin

Re: Double glazing

It was hugo who mentioned that they used air between the panes, CCCP was pointing out that it isn't air but argon that is commonly used. You are exactly right about a vacuum between the panes as a simple 2 foot square double pane window would be holding back over 8,000 pounds of force if the space between was evacuated and that would require some pretty thick glass. But actually the gas between the panes is important and not just because oxygen reacts with so many things and shortens the life of glazing seals.

Argon, being a noble gas, is monatomic like helium. As compared to a diatomic gas, like the major air constituents oxygen and nitrogen, argon doesn't have the energy states available to it to be an efficient heat transport medium. It becomes clearer if you picture a diatomic gas, nitrogen gas=N2, as a pair of weights connected by a spring and a monatomic gas, argon=Ar, as a single spherical weight. Both the N2 and Ar can have kinetic energy in the X, Y, and/or Z directions but because N2 looks like a barbell it can also store energy by spinning the masses around one another as well as in the spring that holds the two atoms together. As a result the N2 molecule can "move" more energy as it goes from the hot pane to the cold pane. Of course this now sounds odd since if helium and argon are both monatomic, how can helium be so good at moving heat and argon so bad? The answer lies in the mass of each and the relation E=(m*v^2)/2. When helium, being lighter, picks up a given amount of energy on the "hot" side it can move across the gap and deposit that energy on the "cold" side much faster than an argon atom can.

2
0
Silver badge

Re: Double glazing

As I understand it, they're trying even heavier noble gases like xenon, which are even more massive and less heat-conductive (they can't use heavier than xenon because beyond this point even the noble gas radon becomes unstable--read radioactive).

0
0
Anonymous Coward

Can someone explain to me

Why they dont just seal the hard-disks and use air in a semi-vacuum? (14% of normal).

Wouldn't that match the helium density without using up precious gas? (Or is the semi-vacuum that huge of a problem?)

1
1
Bronze badge

Re: Can someone explain to me

Haven't I already explained that? The casing would have to become a pressure chamber. Think of the force that it would have to withstand over its surface, even at 14% of atmospheric pressure, a large part of which cannot have reinforcing members inside because they would be obstructed by the platters.

A quick calculation (happy to be corrected) gives the force exerted by the atmosphere on the top of a 5.25" disk drive at sea level as the equivalent of a 150kg weight - that's like two people standing on it.

(Area is about 0.015 sq. m; pressure is about 100000 N/sq. m; 1N is about the force exerted by 100g.)

0
0
Silver badge

Re: Can someone explain to me

You don't want low pressure, as that reduces heat conductivity. What makes helium so special is that it's so light- even at standard atmosphere. At the same time, being a noble gas, it won't react with anything in there. Since there are moving parts, there will be airflow, which means it will convect heat even if sealed. The big concern right now with helium is keeping it bottled up, since its unique nature also makes it capable of diffusing through many supposedly-solid things (though I hear they're making progress with thin but still-impermeable layers of carbon).

0
0
Bronze badge
Alert

So now...

As we all know, electronics run on magic smoke, and when you let the smoke out, they quit working. So now your drive will gradually fail if it has a helium leak? Also, I wonder how leak-proof and stable these devices will be after a few hundred heating/cooling cycles...

0
0
Silver badge
Coat

Helium filled Hard Drives

Now 12% Lighter!!

0
0

praos

Really, why not hydrogen? First thing that crossed my mind. At room temp it's perfectly inert, have 1/2 of He density and is even better conductor of heat. Hydrogen was routinely used as coolant for large electric generators for decades, there is rich experience to tap. Helium is probably selected only for snobbish reasons or because of some stupid regulation. The quantities are miniscule, I can't see any hazard in using it..

0
0

Page:

This topic is closed for new posts.