even if the hdd is flat, surely the load is uneven somewhere on the balls or the races due to gravity.

Nice revive though

Of course, but the load is symmetric if it is horizontal. If it is tilted, the load will not be symmetrically distributed but will have a more complex distribution that will require either fluid dynamics computation (FDB) or mechanical simulation (ball) to predict the motion path / particle flow changes.

Excuse me for repeating myself but has anyone run a test on identical drives comparing any performance difference between mounting horizontal and vertical?

Any problem if i mount the drive vertically, with the connectors facing down? Like this: http://img108.imageshack.us/img108/2164/hdddddkd9.jpg

I've read this thread, and the answer seems to be that it's safe, but i wanna know for sure from someone.

its fine

I'll admit I didn't read through all five pages of this thread, but I thought I'd throw my opinion in about mounting drives at an angle.

The forces inside the drive due to its own operation are so high that gravity is completely irrelevant.

Gravity has the highest effect on the arm and read heads when the drive is horizontal, and since the drive can safely be mounted inverted, it's obviously safe to have the full force of gravity pulling in either direction. When you mount the drive on edge (also formally acceptable) then there is zero force due to gravity pulling the heads towards or away from the platters. As you rotate the drive from horizontal to vertical, the force on the heads smoothly transitions from 0 to 1, as the cosine of the angle. Since the perpendicular forces can be as high as 1 (horizontal), and the axial forces can be as high as 1 (vertical), why can't they be less than 1 but not quite zero? Zero and 1 aren't magic, they're just round numbers.

Due to the geometry of the drive, you can't really think of it in cartesian (X,Y,Z) coordinates. It helps to think of things in cylindrical coordinates, (radius, angle, vertical). The arms are only responsible for moving the heads in the radial direction - they are aerodynamically suspended in the vertical direction and incapable of movement through the angular direction (this motion is handled by the rotation of the platters).

The aerodynamic forces holding the head in the correct position are similar to skipping a rock across a puddle - it's a thin layer of fluid which is rotating with the platters, so from the point of view of the platters, the head is moving past very quickly and skimming the layer of fluid without ever touching the platters. The force holding the head off the platter is relative to the platter, not to any other part of the drive. The head is held off the platter by a force perpendicular to the platter. The read head weighs almost nothing, so it doesn't take much force to hold it up. Since most drives are rated for an operating shock of more than 50 G's, the aerodynamic forces are obviously capably of supporting at least 50 times the weight of the heads. One G in any random direction isn't going to cause problems with that.

The only issue with mounting a drive in an unusual orientation is with the spindle bearings. I know far more about physics and aerodynamics than about bearings, so I won't comment much on this. However, as I understand it, the fluid dynamic bearings used in modern drives don't really care about orientation.

I have 3 Samsung harddisk which are SP0802N SP2014N SP2514N

do all of this support FDB so support vertical use

I have seen some vertical harddisk positioning cases by the way

Vertical is fine for all drives, although some manufacturers won't say so outright. This thread is about mounting drives at unusual angles, like diagonally.

I once had a 19" rack with a server PC inside (big 4U chassis). All HDDs were mounted horizontally in the case, but the rack itself was shifted 45 degrees and thus all HDDs were actually operating diagonally as well. The case had room for many HDDs, it was configured with 6 Seagate Cheetah drives in RAID-5 as well as a couple of Maxtor drives as hot-spare. Guess what? The Maxtor drives died out in use (they were also configured as RAID-1 mirroring the 6th HDD of the RAID-5 configuration) long before the first Seagate died (I guess it was less than 1 year, the first Seagate failed a couple of months after). The server was not all mine, it was part of a project I had planned with a few mates (uncompressed video streaming server). I had to dig out a lot of money for the rack and case itself let alone the HDDs, just good that I got it sold for a very good price (actually, claimed back near 90% of original price!)

Is it? How about spindle up and down ?


Tested under what conditions? Some seem to test only in six direction, i read here.



Well, i think of a Umbrella as a example why odd angles can hurt the durability of harddisk. No wind = any direction is fine, some wind = slight angle towards against the direction of the wind is preferred. Strong wind = ALOT of angling needed to keep the umbrella in shape. When this fails, slight to low angle or too much angle, the umbrella breaks upwards or downwards. Sorry, my english is not that good, hope you catch my idea of it.

Now, harddisk are more robust. However, i can imagine the lifespan of a diks will decline when used in very odd angles and my idea of the umbrella have alot to do with it. At a angle, when recieving a "shock", i can imagine the lower part of the disk gets more shock to absorb then the upper halve for example.
When a disk is in a 90c angle, the whole platter absorbs the shock evenly, as goes for the header too; which is designed for horizontal and vertical shock resistance. At a angle? Who knows how much they can take ?

Another reason to attach hdd's with four screws is not only vibration of the harddisk, but vibration of the harddisk icm with the case. A bad attached harddisk can cause case vibration and thus > more, often annoying noise!

ok i read the whole thread... and in a way, wish i hadn't lol

Could be because it's 5am and i'm hungry and lacking my coffee but this thread has made me uncertain on something i was fine with before..

Think of a table, or bench surface. Can i mount a hard drive on its side, on that surface? ie, so its side screw holes are sat on the table? I want to mount hard drives running in that orientation across the front panel of a case running left to right

I think this thread covers most of the arguments already. Any reason why you think you can't? Many cases already mount drives on their side including some big NAS systems.

If there was a proven link between drive orientation and reliability then it would be clearly stated by the manufacturer. On the scale that hard disks are used industrially in data centres, such an orientation link would be established very easily, we can therefore safely assume that their is no link.

I've had a Seagate Momentus 5400.3 ST980815A mounted loosely on a sponge at a slight 10-deg diagonal angle for a while. SMART gives about 5 years in power-on hours and there's been no problems.

Very informative post, good thing you clarified this because I was very curious if there is a change in performance on how you place your HD in the casing. Thanks!