The Race to Seal Helium HDDs (2021)

https://blog.westerndigital.com/race-to-seal-helium/

by Quizzical4230 • 2 years ago

158 127 2 years ago

blog.westerndigital.com

rx_tx• 2 years ago

(2021)

wigster• 2 years ago

if helium is better, why not use a vacuum?

exe342• 2 years ago

vacuums tend to collapse on themselves, I think it might be more costly in structural elements to prevent that?

robocat2• 2 years ago

It is better to think of a vacuum within air to be similar to a bubble underwater.

Air is like water in other ways too. We slightly "float" in air by the weight of air we displace. e.g. 80kg person is approximately 80 litres (density of a body is 1.010 kg/litre). Weight of displaced air is approx 0.1 gram (1.2929 gram/litre). So the floating effect of air reduces your weight (not mass) by about 0.1%.

dist-epoch2• 2 years ago

Thermoses have vacuum in the walls, and they are very thin, much thinner than a hard-drive case.

ItsBob• 2 years ago

What about a partial vacuum?

Perhaps there is a hapy medium between air resistance and cushioning? Reducing the air by an amount might help.

Although, I'd be surprised if they hadn't thought of this already :)

askvictor• 2 years ago

Maybe structural issues in trying to stop the case from imploding?

tedsanders• 2 years ago

A startup called L2 drive is apparently working on this: https://l2drive.com/vacuum-hdds-the-next-logical-step-after-...

But their website looks pretty stale, and I don’t see evidence of traction from brief Googling.

rwmj• 2 years ago

Hydrogen would also be .. interesting ...

flir• 2 years ago

Oh, the huge HDD.

dreamcompiler2• 2 years ago

Wouldn't be a fire risk because there's no oxygen available. But hydrogen is chemically reactive and over time it could corrode or weaken the materials inside the HDD.

ahartmetz4• 2 years ago

Hydrogen even diffuses through (and into) metal - seems not doable, not cheaply and without maintenance at least.

aaron6952• 2 years ago

Hydrogen would also be .. interesting ...

Why?

It's used in power station generators. Why would a hard disk be interesting?

https://en.wikipedia.org/wiki/Hydrogen-cooled_turbo_generato...

Panzer04• 2 years ago

Is heat a concern in HDDs? Vacuum would probably make it extremely difficult to cool if it's necessary.

jfjdkfnjrhr• 2 years ago

No friction ... No heat

dist-epoch2• 2 years ago

Heat can still circulate through the metal parts. Which is a better heat transfer medium than air anyway.

akira2501• 2 years ago

A gas provides a "bearing" for the heads to ride on as the disk spins below it.

rob74• 2 years ago

Right you are, forgot about that completely! IIRC this is important to prevent a headcrash if the drive is jolted while in operation (of course within certain limits).

ahartmetz4• 2 years ago

Maybe magnetism could work. HDD platters take very strong fields to change magnetization these days, so erasing the data with the positioning magnet(s) seems avoidable.

rasz• 2 years ago

Heads ride on cushion of gas, would instantly friction weld into platters in vacuum.

hughesjj• 2 years ago

Friction or cold/vacuum welding at that point?

j16sdiz• 2 years ago

Metal sticks to each other in vacuum

https://en.m.wikipedia.org/wiki/Cold_welding

ramchip• 2 years ago

"Abundant" in the infographic is questionable... in the universe yes, but on Earth not really.

formerly_proven• 2 years ago

At least two roads to Rome: HGST/WD and Toshiba use laser welding, Seagate uses friction-stir welding (https://en.wikipedia.org/wiki/Friction_stir_welding).

edit: the close up in the Wikipedia article actually is from a hard drive

vertnerd• 2 years ago

"Helium is notoriously difficult to contain. Its atoms are some of the tiniest in the universe."

It's a good article, and I learned a lot, but every so often it reads like something an 8th grader would write the night before their paper is due.

B1FF_PSUVM• 2 years ago

And then it swings up and hits you with "a way that it doesn’t scathe the seal".

Plenty of "unscathed" around, but I don't think I ever met the "un"-free word.

Reminds me of the P.G.Wodehouse "far from gruntled" ( https://nydamprintsblackandwhite.blogspot.com/2011/05/words-... )

AndrewSwift• 2 years ago

[scathing remark deleted]

dmurray• 2 years ago

You need to read How I Met My Wife!

https://www.newyorker.com/magazine/1994/07/25/how-i-met-my-w...

amelius2• 2 years ago

Sometimes facts are simple.

yummypaint• 2 years ago

Yeah i don't get the unnecessary vagueness. They are the smallest atoms. It's like they are afraid of accidentally educating the reader.

kragen4• 2 years ago

this terror of accidentally including a fact is one of the worst of many plagues afflicting journalism today

bjoli• 2 years ago

And here I was, thinking helium was the tiniest atom unless you start counting cations which has it beaten by Li+ and He+

adrian_b7• 2 years ago

Yes, at the same number of electrons, the heavier atoms have a smaller volume, i.e. they are tinier, because their greater nuclear charge attracts the electrons closer to the nucleus.

Li+, Be++, B+++ are tinier than He, but they are ions, not neutral atoms. Neutral Li, Be, B are bigger than He, because they have an additional electron layer. Both neutral H and H- are bigger than He, because their one or two electrons are attracted less by a proton than by a He nucleus with double charge. H+ is just a bare proton, so it is orders of magnitude smaller than any other ion, but it is not a neutral atom.

So He is indeed the smallest neutral atom.

chasil• 2 years ago

Would neon offer some degree of these benefits?

Is neon more plentiful than helium?

Edit: No.

"...neon is comparatively scarce on Earth... It is primarily obtained through the fractional distillation of liquid air, making it significantly more expensive than helium due to air being its sole source."

https://en.wikipedia.org/wiki/Neon

Wowfunhappy2• 2 years ago

...I didn't really we had the technology to seal helium at all.

Stupid question, but could you potentially create a helium balloon that never runs out of helium (because it's sealed and doesn't leak)?

ahartmetz4• 2 years ago

If it's made of metal foil, probably. Much more diffusion-resistant than rubber. The problem with rubber balloons isn't that they have macro- or mesoscopic holes. They lose gas by diffusion all over.

ejdhshsuwisjsh2• 2 years ago

The article talks about how hard it is to contain helium.

Why would you assume that a basic metal foil would keep helium indefinitely?

Towaway692• 2 years ago

Western Digital ships about a million helium-sealed drives every month.

I'm surprised that there was such demand for spinning disks, I would have thought SDDs would have replaced them all.

I say ‚was‘ as the article is from 2021.

ricardobeat• 2 years ago

SSDs are still too expensive for high volume storage.

Enterprise purchases will be much lower, but as a consumer you can buy a 12TB helium-sealed HDD for $100-$200, while solid state alternatives don’t even reach that capacity in a single unit and cost 5-10x more.

Towaway692• 2 years ago

Thanks for the explanation, I didn’t realise that was the case.

throwaway81523• 2 years ago

Are HDD's a significant consumer of the world's helium, in the scheme of things? I had thought helium was used only in a few bleeding edge drives, and they went to normal air as the tech matured.

Out_of_Characte• 2 years ago

Assuming a drive contains helium at standard temparature and pressure, and contains about a liter in volume. That would be about 1 gram worth of helium per drive.

margalabargala• 2 years ago

Hard drives are much, much smaller than one liter.

robocat2• 2 years ago

0.05 gram Helium assuming: (1) half of drive volume is used for platters, (2) at 15° Celcius (3) at 1 atmosphere (I'm guessing not pressurised since that would defeat purpose of using Helium, although might be less than 1 atmosphere)..

0.376 litres for total 3.5-inch HDD volume from my first Google result: "Width: 101.6 mm, Height: 25.4 mm, Length: 146 mm". Actual volume used for platters is less than that.

Helium is light ~ 0.169 g/litre from "0.169 kg of Helium is 0.999 m3 at 15°C": https://microsites.airproducts.com/gasfacts/helium.html Note that I think comment assumed one litre of normal air not Helium: "The density of dry air is 1.2929 g/litre at STP".

joezydeco2• 2 years ago

One MRI machine uses about 2,000L of liquid He, which expands to around 1:770 in gaseous form. So roughly 1,540,000L of helium gas per machine.

ndiddy• 2 years ago

Helium drives are basically standard for enterprise hard drives now. The reduced drag force allows for using thinner drive platters (helium drives can hold up to 10 platters, while air-filled drives can only hold up to 6 platters) which boosts capacity. Helium drives also use less power, run cooler, and the helium gas helps to absorb vibrations that can cause wear and tear (useful in enterprise settings where you have 45 or more drives in a 4U chassis).

thinkfaster3• 2 years ago

Helium in the atmosphere is almost two orders of magnitude more common than xenon and three times as concentrated than krypton. Both are extracted from air.

Even though He is constantly venting to space, alpha emitters keep replenishing it.

Cheap helium from 7% CH4 wells is not going to last. But we're not going to run out of He. Just the energy to extract it.

buildsjets• 2 years ago

Let me go read the article for you.

Yep, it says right there that a "standard" helium tank is adequate to fill 10,000 hard drives. I am going to assume that by "standard" they mean 220ft^3 or 250ft^3, even though my "standard" for tanks is 125ft^3 because I can comfortably carry one of those on my shoulder.

aaron6952• 2 years ago

If anyone has a helium drive you can check it's levels - S.M.A.R.T. ID 22

If you are bored I'd be interested to hear what level it's at on old drives.

Like is it still 100% after a year? [edit] I assume the 100/64h means 100% maybe not [1]

https://www.reddit.com/r/windows8/comments/11ndk0o/what_is_c...

https://en.wikipedia.org/wiki/Self-Monitoring,_Analysis_and_...

[1] https://www.reddit.com/r/DataHoarder/comments/wl20q2/just_a_...

SirMaster2• 2 years ago

I have an 8TB WD RED that has been at 7/100 helium level according to Smart 22 for over 2 years now. Still running just fine in my ZFS pool 24/7.

jasomill• 2 years ago

I have a pair of HGST HDN728080ALE604 drives (8TB Deskstar NAS). One has SMART 22 at 100 after 43,239 power-on hours; the other is offsite and online, but is the same age and had SMART 22 at 100 last time I checked.

My other helium drives (WUH721816AL5204 — 16TB Ultrastar DC HC550) are SAS, and I don't know how to check helium levels of these (smartctl reports neither SMART 22 nor anything related to helium).

dale_glass• 2 years ago

I've got 3 drives at around 20K hours and all 3 are showing "100" in helium level.

octacat• 2 years ago

3 disks 8TB from 2020, WD, 100%.

kalleboo• 2 years ago

I have 12 Toshiba drives with ages ranging between 10k-27k hours, and their Helium levels are all at 100

cainxinth• 2 years ago

The solution:

- They moved all the casing openings to the top of the drive and used a thin metal foil as a second cover.

- They adapted laser welding techniques from the satellite industry to seal the foil to the casing without damaging components with excess heat.

- They found an aluminum alloy used in aerospace that could withstand the laser welding without cracking.

- To get electricity and data in and out without breaking the seal, they used glass-metal feedthroughs similar to those used to seal Freon in refrigerators.

- To get the solder to adhere when attaching the feedthroughs, they used a nickel plating mask.

alright2565• 2 years ago

used glass-metal feedthroughs

I wonder if this is still the case. I took a WD helium hard drive apart recently and I think I saw just a PCB glued in place for the electrical signals. Probably made from very special materials, but it looked just like usual.

adrian_b7• 2 years ago

Glass-metal feedthroughs have been the standard means of connecting semiconductor devices during the first decade of their existence, when they were packaged in hermetic cases made of either metal or glass. Their main problem is to use pairs of special kinds of glasses and of metal alloys that are matched in their thermal expansion coefficients.

After 1960, better passivating methods for the semiconductor chips have been developed, so the hermetic packages have been slowly replaced with plastic packages for most applications.

It is likely that the PCB glued in place covers the glass-metal feedthroughs, which have metal pins that are inserted in connectors soldered on the PCB. It is impossible for any case that contains helium to have any part made of plastic in any of its walls. Plastic parts like a PCB can only be attached externally.

writeslowly• 2 years ago

They also invented a new welding technique to control some sort of aluminum micro-explosions

kragen4• 2 years ago

thanks, this is great!

to add a bit more context:

- laser welding has been in use since 01967 and is used not only for satellites but also for auto body panels, pacemakers, 3-d printing (sls/slm), batteries, etc.

- glass-metal feedthroughs are used in incandescent lightbulbs, fluorescent tubes, neon tubes, and vacuum tubes (including crts and electron microscopes), because in all cases you need a vacuum-tight seal around the wires that lead into the tube. special glass formulations with custom thermal expansion coefficients have been available for this purpose since at least the 01950s. the refrigerator guys may have the cheapest ones but they didn't invent them

- nickel plating is the conventional way to make steel solderable with electronic solders since forever. it's also used, for example, to make battery tabs solderable. i think this goes back to the 19th century

that is, the design process the article depicts is considerably less daringly original than the author makes it sound. the people who made it work deserve a lot of praise, but because manufacturing is hard and they had to solve a lot of hard problems, not because they're geniuses doing totally unprecedented things with brilliant strokes of insight. the author makes it sound like that presumably because she doesn't know anything about the problem space

if you like that kind of thing, you'll probably enjoy dalibor farný's youtube channel where he documents his process of solving the problems of his nixie tube factory (including vacuum-tight glass-metal seals) in a czech castle https://www.youtube.com/@daliborfarny/videos

kbelder• 2 years ago

What would happen if you had a rigid structure that helium could permeate, but nothing larger could, and then filled it up with helium and waited?

Would most of the helium exit, until it was balanced with just the partial pressure of helium in the atmosphere? That would be nearly a vacuum, wouldn't it?

BizarroLand• 2 years ago

This is a good question.

I don't know the answer, but this does make me think of atomic sieves like the ones used in oxygen concentrators.

This is an armchair scientist explanation of them, but they "concentrate" oxygen by first having atmospheric air pumped in and then pressurized. The microsieves have holes so small that mostly only oxygen can fit through, the larger CO2 and Nitrogen atoms simply won't fit.

Then, pressure is let off and fresh air brought in. The fresh air scrubs out the oxygen depleted air and refreshes it with standard air.

Then, the pressure decrease allows the oxygen to leak back out of the sieves, leaving you with oxygen enriched air.

I don't know if there are any atomic helium sieves, but if you can find one it might be a start to testing the question.

lazide2• 2 years ago

Essentially gas reverse osmosis?

s0rce• 2 years ago

It would be a vacuum I guess, but generally not in the normal sense because this is the case with most metals. You don't really consider the interstitial space between the atoms in the crystal lattice a vacuum even though there is space for small atoms (He/H) to diffuse through.

avoid3d2• 2 years ago

The OP is imagining a device with a “hollow core” of atmospheric pressure helium surrounded by a solid and continuous membrane.

Since this hollow core would be much larger than the crystalline structures in a metal would this not therefore be a very weak comparison?

jayd16• 2 years ago

Pretty sure it would simply equalize to the external pressure.

Consider it another way. If you have such a device with a vacuum inside, would it not pull in the external helium over time to reduce the vacuum?

alexey-salmin2• 2 years ago

Pretty sure it would simply equalize to the external pressure.

External partial pressure of helium which is extremely low.

Consider it another way. If you have such a device with a vacuum inside, would it not pull in the external helium over time to reduce the vacuum?

It would, but only until the partial pressure of helium inside is equal to the partial pressure of helium outside (assuming the membrane is permeable only to helium). After that point the same amount of helium will traverse both ways, establishing the equilibrium.

thinkfaster3• 2 years ago

A real material would. But thermodynamically OP is right, you'd get a vacuum.

Enginerrrd• 2 years ago

I don't think so.

Ignoring the fact that what you're describing sounds suspiciously like a Maxwell's demon, I think the equilibrium would be at a higher pressure because helium escaping against an overall pressure gradient would be doing work.

In essence, at the boundary I think the rate at which helium escapes would not simply be proportional to the gradient created by the internal pressure and the exterior partial pressure, but I think would include a term involving the whole exterior pressure.

alexey-salmin2• 2 years ago

It's not a Maxwell daemon, it's just a device for which nothing but helium exists. In the ideal gas model it would indeed create almost vacuum but the second law of thermodynamics is not violated since the process is symmetrical. Pressure of helium outside is also almost zero.

avoid3d2• 2 years ago

This relates to the difference between how a real gas and how an ideal gas behave in this scenario.

The difference in this situation will be very small in my opinion, broadly because the gas molecules in the atmosphere still have comparatively high mean free paths and therefore won’t interact with the “escaping” helium molecules.

thereisnospork• 2 years ago

Internal pressure would equilibrate to the partial pressure of helium in the atmosphere, presuming all other species can be assumed to have zero permeability. Osmotic pressure/semipermeable membranes is the analogous liquid system.

huppeldepup• 2 years ago

Cody’s lab on youtube did a video on this, with balloons

thinkfaster3• 2 years ago

It would create an almost perfect vacuum. You could then extract work by collapsing the vessel. The free energy for the work would come from putting the trapped helium (state 1) in a higher entropy state (state 2).

For what it's worth, this is done industrially to separate gasses using porous membranes.

esd_g0d2• 2 years ago

I wanna be this guy when I grow up.

Seriously, how realistic is it to get a job like this guy's? Open ended, no rules, just solving problems... Most days I'm fine with a boring specialized engineering career, but this one got to me...

striking• 2 years ago

Before he was a researcher for IBM, he was a "groundbreaking physicist", for which the article provides this as the evidence: https://news.cornell.edu/stories/1998/03/rotating-single-oxy...

So, be one of the best in the field, and perhaps folks will trust you to do wild open-ended research forever. I think that sounds like a pretty reasonable deal.

numeri• 2 years ago

Any career in fundamental research is more or less like that. From what I've seen personally, academics and government labs are the two biggest places you can find the most open ended roles. Each comes with their own caveats, of course.

esd_g0d2• 2 years ago

interesting to know. in my limited experience though, in practice, in academia the politics is so heavy (more than in your average workplace) that political matters often overshadow research matters. maybe that is one of the caveats you had in mind. i am curious about government labs though. though it seems country-dependent. how is it different from academia?

buro9• 2 years ago

I received a HDD only last year where the SMART status immediately reported a failure for status 22, and it led me to https://www.backblaze.com/blog/smart-22-is-a-gas-gas-gas/

It's an immediate "return the drive" scenario, as something happened between factory and the SMART test and the drive is no longer within spec.

Thankfully an easy returns and replacement process, but an eye opener too, I hadn't heard of SMART 22 prior to this.

adrian_b7• 2 years ago

SMART-supporting HDDs can execute several kinds of self-tests, e.g. long test, short test or conveyance test.

On Linux these can executed with "smartctl", which can also be used to read the error logs after the tests finish. The long test can take almost a day on big HDDs, while the other tests take only a few minutes.

After I had some problems with a bad HDD, I have begun to always execute all the SMART tests whenever I buy a HDD. Of course, if any test fails, the HDD must be returned immediately and the vendor cannot refuse to replace it when the request is backed by the failed SMART acceptance tests.

This is a good habit, because even if such events are very rare, I have still encountered a few HDDs that have failed the tests, so I have returned them and I have received other good HDDs.

fignews2• 2 years ago

How do you generally know when the long test is over? Is there a way to get a notification of completion? Is it ok to begin using the drives while testing? I personally run my new drives through a few rounds of badblocks

TacticalCoder• 2 years ago

Fun fact: the Rolex "deep sea" sea-dweller contains Helium and an Helium valve. I'm not too sure how they keep the Helium from escaping. James Cameron (who made the movie The Titanic) actually strapped a Rolex sea-dweller "deep sea" to a little robot submarine and sent it to 10 000 meters deep to see if Rolex was full of shit or not.

Turns out the "veblen good" did quite well and didn't break.

And James Cameron got a limited edition Rolex Sea-Dweller deep sea model named after him.

patclay• 2 years ago

You have this backwards, the valve is to let the Helium out.

The watch was designed for saturation diving - people essentially living at high pressure in a pressure cylinder, breathing a mixed gas which includes a high helium percentage (as nitrogen becomes narcotic). This saved them hours of decompression every dive - just do one decompression at the end of the job.

The problem was despite all of the seals to withstand seawater, the watches would let helium in & gradually equalise the internal & external pressures. All fine until they came up to the surface, when the watch would blow the crystal off, as the helium couldn't escape quickly enough to equalise. The Helium valve is there to release this pressure on ascent.

The sea dweller was designed as a highly specialised tool for a specialised industry, before becoming luxury item.

m463• 2 years ago

I can't help but think of deep sea divers using helium as part of their mix, and talking in chipmunk high-pitched voices.

runlevel1• 2 years ago

Here's a phone call between US President Lyndon B. Johnson and Sealab 2 who were doing just that:

https://www.youtube.com/watch?v=wusrtGfjzZg

Scoundreller• 2 years ago

Here’s a diver checking out an air pocket in a sunken boat finding someone alive in an air pocket (most of the recording is the guy at sea level, but you really get the chipmunk from the diver):

https://youtu.be/LrvRwNaE7Eo?si=bmFzXBmiC1QUwnIF

fortran77• 2 years ago

Since it's a relatively small volume, why not hydrogen? Is it too reactive?

banish-m45• 2 years ago

Even harder to contain than helium. Hydrogen notoriously diffuses through many materials and ridiculously-small pores.

banish-m45• 2 years ago

Note that it's possible to (mostly) hermetically-seal non-helium, air-exchanging breather holes of HDDs for use in submerged mineral oil applications. It's typically done by adhering a flexible membrane over the air port.

nxobject• 2 years ago

An interesting side-story here is how the R&D here was sustained through three corporate acquisitions - from IBM to Hitachi GST to WD.

The Race to Seal Helium HDDs (2021)

Related Stories