Harddrives: shutting down or leaving computer on?

[soccerballtux]

So my question is, does anyone have information about whether it's better to leave your computer on, or to shut it down, given say 5, 10, 20 hours of inactivity?

My understanding is that the fluid bearings in harddrives last practically forever in operation, however the spin up/down cycle can be stressful to them.

Anyone have input?

 

[mutz]

people are using they're HDD for years without it breaking down,
what ever HDD you have, you'd probably have to dish it in a year or two,
you shouldn't worry about that,
never heard of anyone complaining about his HDD failing due to increased shutdowns,
it is working on/off anyhow while loading any file..
the only thing you might think about while shutting you'r computer repeatedly, is the CMOS chip (if it isn't wrong),
it is said to get faulty after 20,000 switching on/off,
in your case it's a matter of ~20+ years.
if it still worries you,
just plug you CPU/GPU to WCG/F@H and leave it on to last 🙂.

 

[reallyscrued]

This is a very complicated question. We may need an engineer to answer this.

 

[PottedMeat]

Wikipedia says the following on fluid bearings: Hydrodynamic bearings rely on bearing motion to suck fluid into the bearing and may have high friction and short life at speeds lower than design or during starts and stops.

If you know the acceleration to final velocity and time it takes to get there I guess you can figure out the wear on the shaft & fluid assuming 0 wear at working speeds. Same for the spindown. From that you may be able to get some probability that the bearing will fail. How much actual wear happens at each spinup/down and how much is needed to fail I have no idea.

 

[mutz]

there's here the patent document from 1993,
http://www.patentstorm.us/pate...24782/description.html
they talk about several methods of dealing with the friction at the start&stops,
one method, is as been said, the bearing is being lubricated while in motion.
the second is aimed at dealing with the start&stop wear,
they insert a thin film chemisorption, about 20-100 angstrum:

By virtue of the arrangement of the present invention, in which a thin film with a thickness in the range of from 20 Å to 100 Å and having protecting and lubricating functions, which is formed from the above-described material, is provided on a sliding surface of a constituent member of the hydrodynamic bearing, the coefficient of static friction can be reduced to 0.2 or less. With regard to the durability, it has been revealed that there is no change in the coefficient of friction even after more than 40,000 cycles of CSS (Contact Start and Stop) test.

maybe that'll lighten you'r worries a bit,
just note,
1. that, there isn't any record of this corrent tech, as meaning, it doesn't have any name atleast in the document, so you won't be able to see whether you'r HD has such fanctionality or supports this benefit,
on the other hand, it seems to be global for all HDD hydrodynamic bearings, since it is included whithin the patent, so prbably it's o.k.
2. the document is from 1993, it is very much possible that the tech has been vastly improved since then.
hope this clarifies you'r queries.
3. going trough the web a bit, it seems hard to find any post regarding fluid bearing failure,
you have this post also,
you might want to have a look at the first comment for it:
http://www.thepcspy.com/read/your_hard_drive_is_failing
best wishes.

 

[soccerballtux]

Hm, that's interesting. I seem to recall one harddrive company putting some sort of fluid stuff in their harddrives, it was a little messy-- point was to help lubricate the parts.

 

[mutz]

it was a little messy

can't get exactly what u mean..,
at the patent documentation, the claim to operate the bearing on an lubricative polymer surface,
oiling/greasing the parts while CSS (contact start stop).
the only problem with it,
is that the polymer eventually decays, and probably not that robust as the thin film method,
at the link posted at the bottom of the last post, this guy is claiming that IBM introduced a hydrodynamic bearing that is being lifted off the surface while shutting down,
note, it was ~2 years ago
thus prevents friction and getting probably less wored.

 

[bendixG15]

---------- OP is NOT High Tech -----------

Read the rules

It's a fine topic since this is not a hardware issue type question. In the memory & storage forum it may not get as "technical" as it would here. Topic is OK with me. Moderator Rubycon

 

[1LordEmperor1]

You can just compare the drive's MTBF or AFR to its rated stop cycles. For example:

http://www.seagate.com/docs/pd..._barracuda_7200_12.pdf

Apparantly this line has a 0.34% Annualized Failure Rate, which translates to 296 years Mean Time Between Failures.

It is rated for 50,000 start/stop cycles. So, if you shut your computer down at night and turn it on in the morning you get about 25,000 days or 68½ years.

I feel these numbers are quite optimistic, but convey the point that it doesn't really matter.

 

[spikespiegal]

If you pound through tables and tables of MTF rates it seem the conclusion is that HD's are most likley to fail either during the first month or two of their installation, or at the end of their life cycle. Doesn't really matter what type of drive it is.

Power Cyling might not hurt the drive, but computer PSU's are another matter. They are regulated for 12volts, and computer HD's designed for 12volts, but when initially powered on PSUs can spike well over 12volts.

 

[nobb]

Ive always been interested in this, but unfortunately, I dont think there is much (or any) scientific data to either side. Everything seems to be just logical speculation. Unless you are turning your drives on and off hundreds of times a day, I dont think there is much difference. Personally, I just use my drives and be religious with my backups because a failure can occur at any time.

 

[Russwinters]

Hello,

New to the boards; I actually joined because I am seeing many hard drive related questions and I believe I can help!

I am a Data recovery engineer for a SoCal based DR company.

Modern hard drives use FDB (Fluid dynamic bearings). If you where to drill into the bottom of the drive (the silver circle) around ~1-3mm in you would break the seal to the "Spindle" and you would see that a hefty amount of fluid would flow from within.


to answer your question, is it safer to leave the HDD on or shut it off. The answer is neither.

for the bearings, it is better to leave the drive on, but by leaving the drive on you leave yourself open to many more types of failures then if you where to shut it off.

My advice is to leave the drive on if you are only going to be away from the PC for less then maybe 2 or 3 hours, but if you are going to be gone for an extended period of time it is safer to shut it down.

Obviously there are cases where you will need to leave it on for an extended period of time (downloading that 15GB torrent, etc) and in this case leaving it on is fine.

In fact, the best way to avoid worrying about leaving your PC running is to make sure you keep a up-to-date backup of your data on multiple storage mediums. I would recommend putting your most important files on a second HDD (in a different system, on a different electrical current) and also back up the files on to DVD.

 

[0roo0roo]

its the wrong question because it really doesn't matter. hd's have higher failure rates than you'd think, and it only gets worse with each year it runs. google the google drive study.
so the only rational course is some kind of diligent backup and or RAID. leaving your computer on 24/7 on the assumption you are slightly reducing the risk of failure is ridiculous way to waste power.

 

[mutz]

i'm off the forum in general but have to apply,

If you pound through tables and tables of MTF rates it seem the conclusion is that HD's are most likley to fail either during the first month or two of their installation, or at the end of their life cycle. Doesn't really matter what type of drive it is.

quite correct.

unfortunately, I dont think there is much (or any) scientific data to either side. Everything seems to be just logical speculation.

take another look here:

With regard to the durability, it has been revealed that there is no change in the coefficient of friction even after more than 40,000 cycles of CSS (Contact Start and Stop) test.

these aren't any speculations, this things are being tested.
ofcourse there isn't any asurance that an HDD will work for any kind of a period, though, these tests are what u'r guaranty is based upon.
they might take 100-200-300-500 HDD and test them,
get 40k+ cycles each, some might get up to 50k-60k and some to 30k,
they take the avg. amount of success (above a certain number), and so base they're reliability upon it.
you might fall with one or two per 1,000 or even a bit more, but generally,
there isn't any absolute accurancy in it, and while working with degradable materials, u cant get "perfect results".
these are the "suppose to work" numbers.
and u can most of the time trust them.
nothing is never certain,
u can get 2 HDD and backup u'r data and they both might fail,
and u can get one that'll work for few years,
and most of them do by the way.

 

[Gannon]

Originally posted by: soccerballtux
So my question is, does anyone have information about whether it's better to leave your computer on, or to shut it down, given say 5, 10, 20 hours of inactivity?

My understanding is that the fluid bearings in harddrives last practically forever in operation, however the spin up/down cycle can be stressful to them.

Anyone have input?

If your hard disks are uncooled leaving them on will expose them to heat damage over time. Most modern hard disks run incredibly hot without cooling, don't trust the manufacturer either. Hard disk bearings are the least of your worries. All hard disk failures I've ever had were head/disk related and I have seen a LOT if disk failures. I was one of the unlucky ones to get one of the dead Seagate model drives that was published a while ago in the news. Bearings were fine, the head/actuator was the problem.

Heat will reduce the life expectancy of your disk so turning your computer on/off is really a non issue.

 

[ShawnD1]

Originally posted by: soccerballtux
My understanding is that the fluid bearings in harddrives last practically forever in operation, however the spin up/down cycle can be stressful to them.

Whenever I've heard a failed hard drive, it makes a weird clicking noise, like maybe 1 click per second. Considering the drive is spinning at over 5000 RPM, the clicking noise would probably be a lot more frequent if it were due to the bearings.

I checked youtube for a clip of the sound and found a video where a guy shows what the clicking looks like. video (fast forward to the end). It's the arm going back and forth and not finding anything. The drive still spins but the reader or some other component in the drive is dead. Guys like me still maintain that the failure is likely caused by power issues. That could be the on/off cycle as well as brownouts, undervoltage, and electrical noise. If you want your drives to last, it wouldn't hurt to use a good power supply rather than some $30 garbage and use a battery backup system for every computer.

 

[mutz]

-i'm off the forum-
but need to finish as properly as possible, this thread.

If your hard disks are uncooled leaving them on will expose them to heat damage over time. Most modern hard disks run incredibly hot without cooling, don't trust the manufacturer either. Hard disk bearings are the least of your worries. All hard disk failures I've ever had were head/disk related and I have seen a LOT if disk failures. I was one of the unlucky ones to get one of the dead Seagate model drives that was published a while ago in the news. Bearings were fine, the head/actuator was the problem.

Heat will reduce the life expectancy of your disk so turning your computer on/off is really a non issue.

can't seem to understand what u say,
first of all,
u can trust the manufacturer to a certain point,
they test they're products, under heat, work, pressure etc. and these equipement is supposed to work properly.
if u have VERY important data on u'r HD drives,
u SHOULD back'em up - as nothing is certain, nothing can be 100% anticipated or guarantied.
if u ment that by not trusting them then it is correct.
other than that,
basically if u don't want to buy 2 PC's to work with as one serves as a dedicated backup,
u should understand that these companies do they're best to offer u with a decent product.
thats they're reputation,
the base on which they basically exist.

 

[Russwinters]

If a Head/Disk failure occurs; the chances of it being from extreme heat is VERY unlikely. Hard drives can run reliably up into the 50+ celsius range. The only way you would get a hard drive hotter then that is putting a heat gun directly on it, which obviously isn't a good idea (unless you are trying to recover a drive from certain failures, and certain technique is used)


As for those Seagate drives; your probably either referring to:

Seagate 7200.11 series desktop drives - less likely

or


Seagate Momentus 5400.2 and 5400.3 drives (Firmware 3.CAE or 7.01)
Most Apple notebooks had these drives. - more likely


The failure of both of these disks is due to firmware defects

The latter's firmware defect was particularly terrible, and would cause the heads to crash and leave deep rings on the platter surface leaving the data totally unrecoverable.

the former is a easily fixable firmware failure, the fix is wipespread now, and can be found without much effort in searching. It requires a special RS-232 cable to communicate with the disks controller through it's jumpers, only seagate and samsung (some fujitsu) allow you to do this. The language is proprietary and you have to spends countless days learning it because seagate doesn't just give out the inner workings of it's factory command interface.

 

[mutz]

If a Head/Disk failure occurs; the chances of it being from extreme heat is VERY unlikely. Hard drives can run reliably up into the 50+ celsius range. The only way you would get a hard drive hotter then that is putting a heat gun directly on it, which obviously isn't a good idea (unless you are trying to recover a drive from certain failures, and certain technique is used)


As for those Seagate drives; your probably either referring to:

Seagate 7200.11 series desktop drives - less likely

or


Seagate Momentus 5400.2 and 5400.3 drives (Firmware 3.CAE or 7.01)
Most Apple notebooks had these drives. - more likely


The failure of both of these disks is due to firmware defects

The latter's firmware defect was particularly terrible, and would cause the heads to crash and leave deep rings on the platter surface leaving the data totally unrecoverable.

the former is a easily fixable firmware failure, the fix is wipespread now, and can be found without much effort in searching. It requires a special RS-232 cable to communicate with the disks controller through it's jumpers, only seagate and samsung (some fujitsu) allow you to do this. The language is proprietary and you have to spends countless days learning it because seagate doesn't just give out the inner workings of it's factory command interface.

great info Russwinters!! 🙂