Hard disk drive handing guidelines video (from HGST)

Terry Kennedy

Well-Known Member
Jun 25, 2015
1,067
504
113
New York City
www.glaver.org
Here is a video (box.com link, no autoplay) about proper handling of hard disk drives during shipment, testing / installation, and use. While it is from HGST, it applies to all hard disk drives.

I was going to post this in the "Great Deals" thread discussing poor packaging by vendors, but I figured it would be easier to find as a separate guide.
 

pricklypunter

Well-Known Member
Nov 10, 2015
1,606
470
83
Canada
Hehe, most enjoyable, I think they got the same guy that done the voice over for Motorola's rework training to do theirs :D
 

TheUnnamedNewbie

New Member
Jun 20, 2016
10
1
3
25
So, how much of that video is "true" and how much of it is HGST covering their behinds? I'm sure that avoiding all forms of shock is a good mindset, but say I do bump a drive during installation, how "bad" is it really?
I don't really handle drives with this level of care, and I haven't only ever had /one/ drive fail, and that was a 10-year-old SATA drive that ran pretty much 24/7 since it's original install day - but I've only ever installed 15 or so hard drives myself (I've only been doing this for a couple of years) so it's not like I have a great sample size.
 

pricklypunter

Well-Known Member
Nov 10, 2015
1,606
470
83
Canada
So, how much of that video is "true" and how much of it is HGST covering their behinds?
All of it is true. The manufacturers are not saying this to cover their behinds, obviously they don't want faulty or damaged disks back under warranty, any more than you want to lose your data to a bad disk. It costs them money to fully investigate drive failure mechanisms and to correct any defect in the manufacturing process that may introduce them. Having faulty (really just badly handled) disks sent back to them, frustrates that process and costs the manufacturer money for no good reason. Moreso if they simply replace the customers disk out of goodwill.

Whilst most times a little knock onto a softer surface and within specified limits, won't necessarily do any immediate damage, it still can and the disk may function perfectly well afterwards. That is, until one day it doesn't spin up because the bearings have been knocked out of alignment or the heads are stuck in park or some such. Any knocks or bumps at greater than the specified maximum designed G-force, particularly on an edge, and you definitely risk dislodging the parked heads and either damaging them or the platter surface as well as any other potential mechanical damage, bearing alignment, solder joints on the controller etc.

ESD can be a disk killer too, the controller is very sensitive to it. The motor drive electronics are less prone to damage because of built in ESD protection diodes, but things like cache RAM and glue logic etc are easily damaged by an accidental ESD. It doesn't take much to cause permanent damage to semiconductors, the effects of which may or may not be noticed right away, but you can be sure the disk life will have been shortened by any real discharge. You hear folks say "my disk ran perfectly for 10 years", but in truth, that's much more likely a testament to the drive manufacturer's build quality and careful design and manufacturing, than it is to the installer being careful when handling and fitting the disk :)
 

Terry Kennedy

Well-Known Member
Jun 25, 2015
1,067
504
113
New York City
www.glaver.org
So, how much of that video is "true" and how much of it is HGST covering their behinds? I'm sure that avoiding all forms of shock is a good mindset, but say I do bump a drive during installation, how "bad" is it really?
The video is intended for OEMs, since that is where most drives end up. As such, they present things as "this is what we can both do to save money for both of us". However, all of that stuff is true.

Particularly in the case of modern helium drives, if it can't be fixed by swapping the logic board or by rewriting the data on the sealed HDA, the drive is pretty much a total loss for the manufacturer.

I don't know about HGST (I only recently started an OEM arrangement with them), but another drive manufacturer once told me that if they get a drive back, their profit on that drive is gone and they've actually taken a loss on it. And a 3rd manufacturer had a warehouse full of bad HDAs that weren't worth opening and fixing at the time - you can't do it on the regular production line because it completely messes up the manufacturing workflow, and if they open a HDA that is full of fragments, it can contaminate the entire area. So the HDAs sit around waiting for end-of-production on a relevant manufacturing line, and if it makes economic sense to fix the HDAs at that point, they'll do it, otherwise it is a write-off. As an example, a manufacturing line may have been set up for a 100GB capacity drive, then ramped up to 250GB and then to 400GB. If that line closes for re-tooling for a more modern drive, it won't make economic sense to fix a bunch of dud 100GB drives.

I don't really handle drives with this level of care, and I haven't only ever had /one/ drive fail, and that was a 10-year-old SATA drive that ran pretty much 24/7 since it's original install day - but I've only ever installed 15 or so hard drives myself (I've only been doing this for a couple of years) so it's not like I have a great sample size.
One of the first disk drives I purchased new was a Data General 6061 like the one in this picture. That was a 192MB (yup, MEGAbyte) washing-machine-sized drive in 1982. It weighed many hundreds of pounds and to get it up the 2.5 flights of stairs to the computer center, I got a bunch of students (this was at a college), un-crated it, and went Heave! Ka-bump! up each and every step. We put it in the computer room and called DG to install it. We played dumb when a bunch of stuff in it needed to be screwed back down.

Tolerances were a lot looser in those days. I had an IBM CE (service guy) who we could NOT get to stop smoking in the computer room. And he stubbed out his cigarettes in the disk drive cabinets! His reasoning was "You don't own it, we do" (at the time, all IBM equipment was leased).