Mounting Harddrives at odd angles

Silencing hard drives, optical drives and other storage devices

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Bean
Posts: 204
Joined: Fri Jan 17, 2003 3:14 pm
Location: USA

Post by Bean » Fri May 06, 2005 10:18 am

hyperslug wrote: So here's the break down

Code: Select all

Manufacturer  Contact method           Response  
-------       ---------------------    --------------------------------------------------
WD            Tech support, email      90 degrees.
Hitachi       Hitachi documentation    90 degrees.
Samsung       Tech support, phone      90 degrees.
Fujitsu       Tech support, chat       90 degrees +-5.
Seagate       Tech support, email      90 degrees preferred, but diagonal OK.
Maxtor        Tech support, phone      90 degeres preferred, but in real world, whatever.
So...

Manufacturers say: In general, stick to 90 degrees.
If you really wanted to be thorough you could approach Shuttle with the descrepancy between the above manufactures and shuttle's tilted up in the front cases. :oops: :idea: :D

ilh
Posts: 313
Joined: Sat Jan 08, 2005 3:31 pm
Location: Acton, MA, USA -- Folding for SPCR

Post by ilh » Fri May 06, 2005 10:52 am

I don't believe you can get precession without moving the spinning axis, and as others have said, I don't believe precession is an issue with a drive that is not moving.

Splinter
Posts: 245
Joined: Mon Jul 05, 2004 9:01 pm

Post by Splinter » Fri May 06, 2005 11:10 am

Yes, my understanding was also that precession occurs when a rotating objects was not in equilibrium - i.e. unbalanced forces acting on it.

Since the force of gravity has the opposing normal and sheering forces, when talking about the spinning platters, how can precession happen?

xilencer
Posts: 52
Joined: Wed Feb 02, 2005 1:20 pm

Post by xilencer » Fri May 06, 2005 10:54 pm

peteamer wrote:How does a gyroscope spinning on a fixed axis have gyroscopic precession introduced to it?
By the Earth's gravity not being perpendicular to the axis.

peteamer
*Lifetime Patron*
Posts: 1740
Joined: Sun Dec 21, 2003 11:24 am
Location: 'Sunny' Cornwall U.K.

Post by peteamer » Sat May 07, 2005 1:35 am

xilencer wrote:By the Earth's gravity not being perpendicular to the axis.

:lol: :lol: :lol:

(I take it you are pulling my leg ...)

Splinter
Posts: 245
Joined: Mon Jul 05, 2004 9:01 pm

Post by Splinter » Sat May 07, 2005 1:47 am

I updated the first post with the information Hyperslug collected.

alleycat
Posts: 740
Joined: Sun Oct 20, 2002 10:32 am
Location: Melbourne, Australia

Post by alleycat » Sat May 07, 2005 3:44 am

It seems to me that you can talk to reps (who are just trying to cover their arses - do you blame them? They've probably never been asked this question before) and theorize all you like. Has anyone actually had a drive fail on them due to it not being at the 'correct' angle? Also, is a laptop drive so much different from a desktop drive that the same physics does not apply? My girlfriend is sitting in bed right now using her iBook at an angle, and has done so for over a year. I had my desktop drive sitting on coolpacks at uneven angles in the bottom of the case for most of last year (it was upside down too!). I experimented with the placement of the drive while it was running, turning it around, flipping it over, etc. No problems. This whole argument seems like a classic FUD situation, and is very short on empirical data.

EDIT: typo

xilencer
Posts: 52
Joined: Wed Feb 02, 2005 1:20 pm

Post by xilencer » Sat May 07, 2005 6:00 am

peteamer wrote: :lol: :lol: :lol:

(I take it you are pulling my leg ...)
Why, seriously?

Splinter
Posts: 245
Joined: Mon Jul 05, 2004 9:01 pm

Post by Splinter » Sat May 07, 2005 10:01 am

Because gyroscopic precession is an effect on a free axis of rotation.

Something sitting there by itself spinning is not a victim of it. Only when you take that spinning object and try and move it in a direction perpendicular to its axis of rotation does it become a factor.

If you were moving the HD around while it was spinning, it would precess.

Tibors
Patron of SPCR
Posts: 2674
Joined: Sun Jul 04, 2004 6:07 am
Location: Houten, The Netherlands, Europe

Post by Tibors » Sat May 07, 2005 12:11 pm

It's not moving the HD that causes precession like Splinter says. Some Googling gave me this "definition" of gyroscopic precession:
If you have a spinning gyroscope and you try to rotate its spin axis, the gyroscope will instead try to rotate about an axis at right angles to your force axis.

Note the word rotate in the italic text. If you have a force that tries to translate the gyroscope, then it will just move in the direction you want. No precession at all.

Gravity is a translational force. Gravity in itself can't cause gyroscopic precession. In all the examples and formula's where there is talk about gravity causing this, it is actually the unevenly applied normal force that is the cause. Take the bicycle wheel in this example. In Figure 1 the axis is supported on both sides, so the gravitational force pulls down and the normal force on both sides of the axis pulls up. The result is that nothing moves, as the forces cancel each other out. In Figure 2 and 3 one of the supports is removed. The gravitational force still tries to pull the whole wheel down, but there is only a normal force on one side of the axis pulling it up. The result is that the wheel starts to precess.

Now back to HD's. The last time I looked inside a HD, the axis of the platters was only attached to something on the bottom side of the disk. So it looks like we can use the model with the uneven normal force. The axis of the platters isn't free to move in any direction it wants. It is fixed by bearings, but these have some give (eng?). As far as I can follow xilencer's arguments, he states that the gyroscopic precession causes the axis of the platters to tilt as far as the bearings allow it. What I don't understand is, why he thinks this will only happen when the disk is mounted at an angle. Seeing all the examples, this will also happen when the drive is mounted in a vertical position. Moreso when the drive is vertical, this effect will be the strongest, as the torque will never be bigger than in that orientation.

Conclusion:
Whatever the reason some manufacturers say drives can't be mounted at an angle, it cannot be the effect of gyroscopic precession on the platters, because they all allow us to mount the drives vertically.

m0002a
Posts: 2831
Joined: Wed Feb 04, 2004 2:12 am
Location: USA

Post by m0002a » Sat May 07, 2005 12:18 pm

Tibors wrote:Conclusion:
Whatever the reason some manufacturers say drives can't be mounted at an angle, it cannot be the effect of gyroscopic precession on the platters, because they all allow us to mount the drives vertically.
I don't claim to be an expert on this subject, but my understanding is the gyroscopic forces (or other forces) are equal on the drive platter/heads in the horizontal or vertical planes, but are unequal other planes, which is the basis for the concern of the drive engineers (the support people aren't just making this stuff up).

Splinter
Posts: 245
Joined: Mon Jul 05, 2004 9:01 pm

Post by Splinter » Sat May 07, 2005 12:33 pm

Tibors, you need to go back to highschool and study basic kinematics.

If the platters had unequal forces on them, they would accelerate. Last time I checked, the platters of MY harddrives were not falling through the case. I dunno bout yours.

It has nothing to do with how many points of support there are. The reason the bicycle wheels precesses in the example you provided is that the supporting force in taken away, and the wheel begins to accelerate towards the ground. The gyroscopic effect causes a greater force to redirect the angular momentum in the wheel than gravity can provide, so it doesn't fall, it rotates about the axis.

A HD platter is fixed in place, it spins on bearings, it has limited movement in either direction. It is, effectively, a rigid body. It is not in freefall, like the bicycle wheel, therefore, it CANNOT precess, there is no unbalanced force acting on it.


And just for the sake of veracity, I was not wrong when I said moving thhe causes gyroscopic forces to act unfavourably on the harddrive, I didn't specific translational movements, I just said moving in general, which includes spinning it and redirecting the axis of rotation.

Splinter
Posts: 245
Joined: Mon Jul 05, 2004 9:01 pm

Post by Splinter » Sat May 07, 2005 12:38 pm

m0002a wrote:...which is the basis for the concern of the drive engineers (the support people aren't just making this stuff up).
I've had technical support staff tell me that the reason my PocketPC wouldnt charge was because it was plugging into a USB 1.1 socket (the unit is charged via a wall outlet)

Support staff rarely know more than the manual tells them, and they almost certainly dont know angular physics.

tay
Friend of SPCR
Posts: 793
Joined: Sat Dec 06, 2003 5:56 pm
Location: Boston, MA
Contact:

Post by tay » Sat May 07, 2005 1:01 pm

Splinter wrote:Tibors, you need to go back to highschool and study basic kinematics.

If the platters had unequal forces on them, they would accelerate. Last time I checked, the platters of MY harddrives were not falling through the case. I dunno bout yours.
After making a statement that stupid I think you are the one who should be heading back to highschool physics. Have a nice day!

m0002a
Posts: 2831
Joined: Wed Feb 04, 2004 2:12 am
Location: USA

Post by m0002a » Sat May 07, 2005 1:28 pm

Splinter wrote:Support staff rarely know more than the manual tells them, and they almost certainly dont know angular physics.
You are absolutely right. Customer support staff don't know squat. That is why they would not make this stuff up about the drives mounted at an angle, and they would ask the engineering staff for the answer.

Splinter
Posts: 245
Joined: Mon Jul 05, 2004 9:01 pm

Post by Splinter » Sat May 07, 2005 1:46 pm

tay wrote:After making a statement that stupid I think you are the one who should be heading back to highschool physics. Have a nice day!
Net force = Mass x Acceleration

tay
Friend of SPCR
Posts: 793
Joined: Sat Dec 06, 2003 5:56 pm
Location: Boston, MA
Contact:

Post by tay » Sat May 07, 2005 2:33 pm

Splinter wrote:[Net force = Mass x Acceleration
Where does your previous post mention net?

Tibors
Patron of SPCR
Posts: 2674
Joined: Sun Jul 04, 2004 6:07 am
Location: Houten, The Netherlands, Europe

Post by Tibors » Sat May 07, 2005 2:38 pm

Splinter, maybe you should go back to highschool and study comprehensive reading. If you understood what what the other people write, it might prevent a lot of the smart ass remarks you make in this thread. They don't even contribute to the discussion, but are just reiterations of your previously stated opinion or insults to the other participants.

The axis of the platters is not infinitely rigid. There is an amount of play in the bearings. What happens with this "play" when the drive starts and changes from one stationary state to another stationary state is what xilencer and D235hadow where discussing and to which discussion I try to contribute. Probably they do understand what I say and can tell if my reasoning is sound and if it is not, what I did miss.

Because this:
Splinter wrote:It has nothing to do with how many points of support there are. The reason the bicycle wheels precesses in the example you provided is that the supporting force in taken away, and the wheel begins to accelerate towards the ground. The gyroscopic effect causes a greater force to redirect the angular momentum in the wheel than gravity can provide, so it doesn't fall, it rotates about the axis.
Tells me you might benefit from reading: How Gyroscopes Work. Especially the third page, where it is explained what causes precession. There does not exist a greater force to redirect the angular momentum.
Splinter wrote:And just for the sake of veracity, I was not wrong when I said moving thhe causes gyroscopic forces to act unfavourably on the harddrive, I didn't specific translational movements, I just said moving in general, which includes spinning it and redirecting the axis of rotation.
Yes, you said moving in general. Which includes spinning, rotation and translation. Thus it is too general and therefore wrong. There exist species of frogs that are green. That doesn't make "Of course it is green, it is a frog." a correct statement, as there are frogs that are brown, red or some other colour or even multi coloured.
Splinter wrote:Net force = Mass x Acceleration
Two "errors" in one line.
  1. To the left of the equation you have noted a "Net" value, so you should do that to the right too. (Not for the mass, as that can't change in Newtonian physics.)
  2. This is the simple formula for one dimension. Alas we are talking about a three dimensional problem here.

Splinter
Posts: 245
Joined: Mon Jul 05, 2004 9:01 pm

Post by Splinter » Sat May 07, 2005 3:58 pm

I'm quite familliar with how gyroscopes work, we covered them in honours physics.

It doesn't matter if you're dealing with 1 or 4 dimensional systems, net force equals net mass plus net acceleration. Always. (please don't get ridiculous and start talking about relativistic effects, they are totally ignorable in a harddrive)

The fact of the matter is that as long as a harddrive is stationary, it does not behave in any weird ways. The platter is not going to attempt to jump off the spindle just because it's mounted at 45 degrees.

tay
Friend of SPCR
Posts: 793
Joined: Sat Dec 06, 2003 5:56 pm
Location: Boston, MA
Contact:

Post by tay » Sat May 07, 2005 4:31 pm

Tibors wrote:They don't even contribute to the discussion, but are just reiterations of your previously stated opinion or insults to the other participants.
Anyway that shadow243983483D, peteamer, and ilh have already stated why the angles shouldnt matter. Nothing to see here.

Splinter
Posts: 245
Joined: Mon Jul 05, 2004 9:01 pm

Post by Splinter » Sat May 07, 2005 4:33 pm

Yup.

There is absolutely no evidence, theoretical or empirical, that mounting modern HDs on an angle causes problems.

StarfishChris
Posts: 968
Joined: Fri Jan 07, 2005 7:13 pm
Location: Bristol, UK
Contact:

Post by StarfishChris » Sat May 07, 2005 5:33 pm

Oh, but the bits could slide off the platters! :shock: I can't believe you didn't consider that...

:?...:oops:...:roll:

m0002a
Posts: 2831
Joined: Wed Feb 04, 2004 2:12 am
Location: USA

Post by m0002a » Sat May 07, 2005 6:55 pm

Splinter wrote:There is absolutely no evidence, theoretical or empirical, that mounting modern HDs on an angle causes problems.
I am not sure I know what you mean by “theoretical evidence”.

It is possible that empirical evidence does exist and was obtained by the drive manufacturing companies while testing their drives. I don't know how you can say they don't have such evidence.

Regarding theoretical justification for the statements made by the drive manufacturers: Just because they have not posted the arguments here in this forum, or because the customer support people cannot explain it to you, does not mean that valid arguments do not exist.
Last edited by m0002a on Sat May 07, 2005 7:25 pm, edited 1 time in total.

Splinter
Posts: 245
Joined: Mon Jul 05, 2004 9:01 pm

Post by Splinter » Sat May 07, 2005 7:06 pm

I mean evidence based on concepts of physics or design.

D235hadow
Posts: 14
Joined: Fri Apr 22, 2005 7:05 pm

Post by D235hadow » Sun May 08, 2005 7:53 pm

xilencer wrote:My conclusion:

Since only horizontal and vertical disc axis mountings are free from gyroscopic precession,
I still believe this is why HDD manufacturers recommend mounting perpendicular to gravity.
The first part of this is an incorrect statement. Allow myself to quote... :? myself:
D235hadow wrote:Any drive orientation with a horizontal axis of rotation (aka vertical drive orientation) will experience the greatest tendency for gyroscopic precession compared to any other orientation (assuming gravity is the cause, not some other force).
Maybe Tibors said it more clearly than me. He's also more technically correct, in that it's actually the uneven normal force and not gravity that causes the torque on the spindle. But it amounts to the same thing, and I assumed, since we were discussing physics, that everyone had a pretty good grasp of these things. I also figured we've all seen at least a picture of drive internals, so we know the spindle is only "fixed" at one end. I left all that out because my posts were long and eye-straining enough as is. Anyway, thumbs up to Tibors especially for identifying the real truth here :wink::
Tibors wrote:The only reason I normally advise right angles to gravity is that it is easier to convince the sceptics of that.
If there's an issue we missed, speak up, because like I said, I'm no physics expert. But so far, none of the problems with angled mounting addressed here should be causes of concern.

darthan
Posts: 237
Joined: Sat Apr 30, 2005 1:28 pm
Location: San Francisco

Post by darthan » Sun May 08, 2005 10:23 pm

Ok, so to sum up where we are in this discussion right now here I go:
1) we have decided that gyroscopic precession might have something to do with why mounting drives other than 90 degrees (+/- 5% mind you) to the direction of gravity could cause problems.
2) the greatest amount of gyroscopic precession occurs when the drive is mounted with the spindle perpendicular to gravity but that is ok according to the manufacturers.
3) Gyroscopic precession occurs because the normal force doesn't fully oppose gravity

Ok, so here are my thoughts about this: first, the normal force doesn't suddenly disappear when the drive is spinning. It isn't there when the drive is first set up vertically and the spindle finds an equilibrium against its bearings where the normal force does equal gravity and then it is spun up (unless you are moving your drive while it is spinning in which case shame on you, you deserve what you get). Now, forgive me for being a total idiot if that's what I'm being, but, that equilibrium may change as the drive spins up but it will still find an equilibrium between gravity and the bearing's resistance to gravity. Those spindles are nice rigid metal rods and the platters are very light, there won't be much give in them and, while the exact part of the bearing that the spindle rides on may change, that too will find an equilibrium. Since it will find that equilibrium, there will be an an axis where gravity and the normal force from the bearings/spindle balance out and the the axis of rotation becomes fixed which means there could be no gyroscopic effect.

Personally, I like the idiot proofing answer that was dismissed earlier, as the reason for the manufacturers recommending horizontal or vertical orientations for their drives. They're thinking in terms of Joe Six-Pack buyin' himself a new drive at Best Buy or wherever and poppin' it in himself and they want to make sure that he mounts it to a location in his case where it will be secure. Now I don't know of any cases where the un-modded HD mount is at an angle so you tell him not to mount it at an angle. I mean, look at how many stories you can find of people trying to stuff two cds in one drive and all sorts of other stuff like this. You don't want to tell people to just stick their drive in any which way because I guarantee someone will take that to mean bolting throught the motherboard or sticking it in some strange corner for it to fall out of later.

Sidenote: why do you people keep going on and on about 1 year warranties on drives? I just did a quick check through newegg's super sekret hard drive listings and there are a few WDCs and some Maxtor retail drives with that warranty but everything else has a 3 year, if not 5 year, warranty.

xilencer
Posts: 52
Joined: Wed Feb 02, 2005 1:20 pm

Post by xilencer » Mon May 09, 2005 2:21 am

D235hadow wrote:Any drive orientation with a horizontal axis of rotation (aka vertical drive orientation) will experience the greatest tendency for gyroscopic precession compared to any other orientation (assuming gravity is the cause, not some other force).
Why would there be precession in vertical HDD mounting (horizontal axle) ?

(the spindle only being fastened to one side is not important)

darthan
Posts: 237
Joined: Sat Apr 30, 2005 1:28 pm
Location: San Francisco

Post by darthan » Tue May 10, 2005 11:30 am

As D235Shadow says "assuming gravity is the cause" then a vertical mount with horizontal spindle would have the maximum amount of precession potential from gravity. This is because gyroscopic precession occurs when a force acts at right angles to the axis of rotation of a spinning body. Thus, when the spindle is horizontal gravity is acting on it at exactly right angles and thus exerting its full force in a direction that would cause precession. The spindle being fastened on one side is absolutely important to whether or not gravity could cause precession because precession only occurs when a component of the spindle's motion is at right angles to its primary axis of rotation. If the spindle was fixed at both ends the force of gravity would be completely canceled by normal force from its mounts and thus unable to cause precession. However, since it is fixed at one end only, it is just barely possible that the bearings have enough give in them that gravity could rotate the spindle out of its axis of rotation and cause precession.

xilencer
Posts: 52
Joined: Wed Feb 02, 2005 1:20 pm

Post by xilencer » Thu May 12, 2005 1:41 am

By spindle I mean the static axle which goes through bearings and discs. Designed supported at both ends or not it has no problem holding the weight at any angle. It's not important in the question.

If the disc is not spinning it's axis can not be tilted out of spec by gravity itself, because gravity can not overpower the hold of the disc.
However, the gyroscopic inertia possessed by the disc spinning is stronger, and that force can be altered by gravity.
When the axis of a spinning body is altered by an external force it's called gyroscopic precession.

Precession occurs because the applied force is not perpendicular to the axis of the gyroscope.

D235hadow
Posts: 14
Joined: Fri Apr 22, 2005 7:05 pm

Post by D235hadow » Thu May 12, 2005 6:43 pm

xilencer wrote:By spindle I mean the static axle which goes through bearings and discs. Designed supported at both ends or not it has no problem holding the weight at any angle. It's not important in the question.
I am confused by what you wrote here. What do you mean by static axle? Are you referring to when it does not spin? Either way, the fact that the spindle is only supported at one end is in fact very important. If it were supported at both ends, gyroscopic precession would not occur in a vertical drive. However, since it is only supported at one end, gyroscopic precession will take place in a vertical drive. I cannot see how that is not important to our discussion.
If the disc is not spinning it's axis can not be tilted out of spec by gravity itself, because gravity can not overpower the hold of the disc.
However, the gyroscopic inertia possessed by the disc spinning is stronger, and that force can be altered by gravity.
I am not sure how you arrived at the conclusion that gravity is too weak to tilt the disc out of spec. However, assuming that is true, why should gyroscopic precession tilt it? There are no extra forces acting upon the axis; gyroscopic precession is merely an effect, not a force. The force causing it in this case is gravity, and will act upon the spinning axis with the same magnitude as it would on a stationary axis. A rotating object cannot begin to exhibit gyroscopic precession merely by spinning; it must be acted on by some outside force.

Also, you incorrectly refer to gyroscopic inertia as a force. It is not a force, and it will effectively resist tilting of the axis, not encourage it. Gyroscopic inertia is the tendency of a spinning body to resist any attempt to change the direction of its axis of rotation; thus once the disc is spinning, it is more difficult to tilt the axis. If you say that gravity will not tilt the stationary disc, it is even less likely to do so when the disc is spinning.
When the axis of a spinning body is altered by an external force it's called gyroscopic precession.

Precession occurs because the applied force is not perpendicular to the axis of the gyroscope.
Your first statement here is correct, the second I cannot make a judgment about since it depends not only on the direction of the force but where it is applied to the system. However, I fail to see how either of these statements lead to the conclusion that gyroscopic precession would not occur in a vertically mounted drive. Gyroscopic precession is directly related to the torque on an axis non-tangential to its rotation. Torque is a cross product between the force on an object and the distance from the point of rotation. Thus when the angle between the force (in our case, gravity), and the distance vector (horizontal, in a vertical drive) is 90 degrees, torque will have its maximum value. Then gyroscopic precession will, in turn, be greatest.

Post Reply