I am convinced from my own experience that a large percentage of early drive failures, especially in external enclosures, are caused by heat. Most enclosures fail to provide adequate cooling, either by using a cheap plastic enclosure, or when they DO provide an aluminum heat sink to draw away the heat, they still fail to provide a fan.

I looked up your Antex drive, and it appears to have a cooling fan, although that fan can be turned off. The case is entirely plastic and provides no cooling. Thus, even with the fan turned on, it probably will not provide ideal cooling. However, if you turn the fan off, you will have created the ideal torture chamber for a hard drive, and failure is pretty much 100% guaranteed.

Another issue with external drives is that AFIK, USB external drives have no way to power down the drive when it is idle. Thus, it is always spinning and generating heat. By contrast, when that same drive is put into a PC as an internal drive, you can set your computer's power management to spin it down when not in use. Some people don't like to do that because they fear that the start/stop on the motor will stress the drive a lot. This is hogwash.

What I don't know is whether the eSATA connection on your external drive provides the ability to spin down. I know that Firewire external drives will spin down, so maybe eSATA will also.

I have 3-4 1TB drives from various manufacturers and so far (they are all about 6 months old), no problem. All are used in a removable slot in my computer. While this doesn't completely eliminate the need for external drives, I haven't purchased an external drive since I got these removable bays for my computer. Other than the ability to use the drive with a laptop or use it in the field, I can no longer see any advantage to using external drives.

Oh, and when I put the computer on standby to remove one of these drives, they are almost cold to the touch. By contrast, when I have removed drives from some of my external cases -- ALL of which use fans, and most of which have aluminum chassis -- they are somewhere between very warm and downright HOT. BIG difference. As a result, I ALWAYS turn off external drives as soon as I am finished using them.

FWIW, my favorite external enclosure is this one:

Coolmax CD-391T-U2



If you look around their site, you'll also find an eSATA version. That huge round hole is the fan, making it one of the few enclosures with a REALLY big fan. Big means quieter, and also means more air. The case is aluminum. Very nice product. I think Newegg carries them.

I agree with John, heat kills drives. I'm also not a fan of Western Digital drives due to a rash of failures I had some years ago. I have used Seagate drives since then, but you can easily find someone with the exact opposite experience and preference.

I use a Thermaltake BlacX hard drive dock for SATA drives. It comes in USB only and combination USB and eSATA flavors. I have several, but my USB only version will indeed spin the drive down when idle. Other USB enclosures may not do this, but the BlacX definitely does.

Thermaltake BlacX Docking Stations

Also, check the power supply. We have had drive failures reported by external drives and the problem ended up being the power supply, not the drive itself.

Dave T2

I have 3 internal hard drives (plus boot disk), but I need additional storage for the long-term project I'm working on - hence the need for external drives.

About 5 years ago I gave up on drives in boxes. Instead, I mount them in free air. I have 4, without fans, mounted on L brackets on wood strips that are screwed to a shear wall (reinforced against tremors - we have earthquakes here in So. Cal.) and fed by long SATA cables, in addition to a couple on USB cables inside a firesafe (cooled by a good fan). I have checked the temps with an infrared thermometer and never have found any spot on any one of them that is hotter than 128F or so even during huge file transfers (air temps get hotter than that in Death Valley, a couple hundred miles east of here).

I also have had numerous disks running on open wire metal shelving from Home Depot, with equally good results.

I've heard that some of the USB > SATA cable/adapters have damaged disks, but I've never had a problem. Still, it's probably best to check with a disposable disk before hooking up your favorite 1.5 TB disk.

The 4 wall-mounted disks have been running almost all day almost every workday for over a year. Most of my disks last for years - in contrast to frequent failures when I used to put them inside of boxes. Even the packaged external disks from Seagate, etc. seem to get too hot. Free air seems to be the way to go, especially if mounted vertically at least an inch from the nearest surface.

One minor problem: I do get a tiny bit of noise on my FM receiver - its antenna is on the other side of that wall, just inches from the SATA cables.

It's definitely not the power supply as I put the drive in a different make enclosure and the same problem occurred. The fan is definitely working on the Antec. I have an additional Antec that I have a 500gig drive that's been going for 2 years. I have externals so I can take a drive home with me if I need to work on it (which has been frequent lately) or if I need to give it to one of my local editors that I use on occasion .

John, I do have a BlacX docking station connected via the eSATA, but I've always had a problem hot swapping out drives. I tried it twice but kept getting a Windows error. Something about incorrect volume name or something. It was as if you were trying to swap an internal drive while the system is still on. Not gonna happen. The only way I could swap the drive was to shut down the computer, then swap the drives, then restart the system. What a pita. Does it have to be connected via USB in order for the swap to work?

Paul

"USB external drives have no way to power down the drive when it is idle"

Quite the opposite is generally true. The friggin things do power down if not accessed for a while and there's no way to stop them doing that. This becomes a major PIA as when they're spun down Windows will wait while the drive spins back up.

Google did make some of their HDD reliability study public. They found no correlation between drive temperature and failure rates. Even so I do get nervous when drives are too hot to touch and I do like to keep a good airflow over my drives.

Google also found all manufacturers perform the same. As was pointed out to me HDDs are being manufactured about as fast as humans. The population of HDD must be getting close to the human population so statistically there will be the usual problem with clusters of failures. That's compounded by the way humans react. If anyone buys 5 units of brand A and they all fail in a month they're very likely to hop onto a forum and scream blue murder. If none of them fail no one mentions it.

Still I too am human and I've had much luck with Samsung disks. I recently bought 4 of of their 5400rpm 1TB SATA drives to use for backups in a toaster, They are very cool, much cooler than the Samsung 7200rpm drives. The slightly slower speed is irrelevant for the purpose.

The only time recently I've had major grief with a HDD it was indeed the supply of power to the drive itself. The dock's DC input socket was intermittent. Fixed that and reformatted the drive and it's still going.

Bob.

I didn't get an instruction manual with my drive bays, so I wasn't sure I could hot-swap and therefore didn't want to try. However, you don't have to turn the computer completely off. Just put it in standby. That takes about 5-10 seconds on my computer. Then, swap the disk drives. Finally, resume the computer. This takes about 10 seconds on my computer. Pretty much the same elapsed time as turning on an external drive, plugging it in, and waiting for it to be recognized.

Quite the opposite is generally true. The friggin things do power down if not accessed for a while and there's no way to stop them doing that.That is amazing, because I own an exact dozen external USB drives, and NONE of them ever will power down. I've attached them to at least six different computers and always had this same result. All computers running various flavors of WinXP. Maybe they spin the other way in Oz and that explains it ...

They found no correlation between drive temperature and failure rates. I find that virtually impossible to believe. I spent the first four years of my career at various test and measurement divisions at Hewlett-Packard (and later, another three years at GenRad doing much the same thing). Virtually everything we tested showed a significant correlation between temperature and failure rate. I'd have to see the Goggle data to see if perhaps they were measuring over a fairly narrow temperature range, or at least a range that didn't get into the >140F range where things start to get interesting.

The Seagate Freeagent drives power down after no use. The ones I have put in my own cases, do not power down. I am sure it is how the hardware in the case is programmed.

Dave T2

1. "Hotswapping" refers to the ability to just Unmount the drive in Windows before replacing the drive.

It's the same as with flash drives. If you don't unmount first, Windows gets confused.

2. A slightly older generation of WD enclosures had poor power supplies that led to permanent drive failures.

3. Many Newegg customers have stated that when they ordered OEM drives for delivery by UPS, a substantial percentage of them failed within a year. Drives shipped via other carriers were fine. Because of this, Newegg recently started packing the OEM drives better, and this particular problem should be solved.

Just wanted to concur with John Meyer. I've had a number of drives fail over the years and ALWAYS it was due to overheating. Once I had a drive fail in an external USB case. I replaced it with another make in the same case and shortly thereafter it failed as well. It turned out that a cooling fan in the case had gone bad. Failure-wise, your case and how well it cools is much more important than the drive you put in it.

"I find that virtually impossible to believe"

Then this statement will astound you as much as it does me:
"In fact, there is a clear trend showing that lower

However as you say these results are at temperature ranges that we'd consider fairly normal. My gut feeling is that the results could well indicate higher failures due to larger temperature cycling and that is a common way to accelerate premature failures.

From their conclusions:
"One of our key findings has been the lack of a consistent

The full paper is now available and is an interesting read:

http://labs.google.com/papers/disk_failures.html

One result that is very worthy of note was that once a disk develops a single sector error it is 39 times more likely to fail completely i.e. mapping the sector out and continuing using it is not a good idea.

Bob.

I use WD drives in external enclosures and have had no problems. But I agree with John, it's probably overheating. My solution is to use an external fan that blows air on three side-by-side enclosures. Turn the fan off and those enclosures are to hot to touch. That has to be bad news. With the fan on they hardly get above room temperature.

Get an external fan for insurance.

I just skimmed the Google report on drive reliability and there are so many problems with it that I have to stop myself from writing a stupidly long post that no one will read. However, if I just concentrate on the temperature correlations, most of the unexpected, inverted results (i.e., higher failure rate at lower temperatures) happened at temperatures WAY below what would normally affect longevity. As I stated earlier, you don't get interesting things (higher failure rates) until you get above 140F. That is 60C.

Their temperature graph stops at 50C !!!

So, they really haven't measured anything useful, and the "correlation" with temperature at these slightly-above-room-temperature readings are probably proxies for some other gremlin that correlates with temperature. For instance, there may be far fewer drives in rooms that are really cool, and perhaps they get replaced more often (you have to read how they define a failure, and in their study, a failure is defined as when a drive gets replaced). So, in a smaller facility, each drive may receive more attention and the slightest problem may warrant replacement, which gets recorded as a failure.

Another thing to point out is that these drives apparently are run much like the person in the post above who puts them bare on a shelf in the open. The heat escapes in a fairly uniform way. By contrast, in a cramped external case, I would expect that without adequate ventilation and/or thermal conduction, the drive might experience local heating that might well exceed the average temperature, and perhaps by quite a bit.

So, while it is potentially a very interesting study, I suspect it may be full of some fairly fundamental flaws, and certainly the part of it dealing with failure rates as a function of temperature don't really tell us much.

Yes, you need to read it very carefully. Those temperatures seem to be as measured by the drive itself. If the drive is reporting 10degC then the ambient would have to be friggin COLD.

Bob.

The Serial ATA hot-swap feature only seems to work with controllers based on the Silicon Image chip. Using John Meyer's method of placing the computer in stand-by and swapping the drive is a perfectly workable solution.

I have a BUNCH of bare SATA drives and most of the time I just need to pop in a drive and get a file or two off of them. I use the USB connection for this purpose as it can easily be hot swapped. If I'm going to be working on a project contained on one of the bare SATA drives, I will use eSATA or an trayless SATA dock mounted in my computer.



Addonics MobileRack

Your tray looks almost identical to the one I use (the iStar T5-SA).

The other advantage I forgot to mention is that the drive seems to perform much better when connected this way than via USB. Both on this computer and on my previous computer I sometimes get unexplained stall-outs for 5-10 seconds when scrubbing video on the Vegas timeline from an external USB drive. I never get that from a drive connected in one of these things.

They do appear quite similar, the Addonics has a fan on it though, the iStar does not.

I've always had issues with drives I put in to enclosers myself. AKA buy the drive separate from the enclosure. All the drive+enclosure combo's I've bought & others I know have bought (IE mybook, etc) have little/no issues. All the external's I've used though (USB & firewire) shutdown after so long of no use, even if you have a file open but don't read it. Pain in the butt. I *NEVER* tell windows to turn off my internal's.

I had three drive in a house fire & they all worked after the fire. One ~4 years old at the time, then 2, then 1. I've never put special cooling on my HD's & only have ever had one internal die, ever. I just keep at least one bay free between each drive.

We know that heat plays an issue with the lifespan of a drive as with almost any mechanical device if not properly cooled. My original thought was because drive size is getting larger and the physical dimensions of the drive hasn't changed, something has got to give, namely the platter. With the platters getting thinner and the density getting higher, doesn't it stand to reason that the chances of issues would increase?

Paul

With the platters getting thinner and the density getting higher, doesn't it stand to reason that the chances of issues would increase?I know what you mean, and I've thought the same thing each time I hear about some new discovery that is going to allow the next-generation drives to store twice as much data.

However, this thinner, denser trend has been going on forever, and according to this logic, drives should have done nothing but get LESS reliable ever since the first 10 MB desktop drives (which I owned ...). But, the drives have done nothing but get more reliable.

Oh, as far as external drives purchased with the drive already being installed being more reliable, my statistically meaningless experience has been just the opposite. The only two failures I have had were the original Maxtor drive/enclosure I purchased and also a WD drive/enclosure. Both failed within a month. Since the WD had been given to me, I didn't have a warranty recourse. I opened up the drive and found no fan, and no heatsink. It was the ultimate hard drive torture device. I was able to use a fan from another defunct device, drill some holes, fit the fan into the case (it was a nice big fan), and put another drive in the enclosure (this was when enclosures and external discs were still really expensive, so it was worth the engineering and effort to retrofit a fan).

That drive is still running today, many years later.

So, I take the opposite stance, namely that you want to find a "best of breed" external enclosure that is all aluminum, has a good heat sink, a big fan, and also a fair amount of room around the drive for the air from the fan to circulate. I'm not a big "fan" of those tiny enclosures that fit the drive like a glove.

I have a D-Link enclosure housing a 1TB raided drive.
It's been giving me some grief of late, ... taking for ever to get stuff from there to the timeline.
Is this a warning sign?
Also anybody else had problems/good luck with the D-Link enclosures.
Had another brand befofe that I had to ditch, but can't recall the name.