Menu
Index

Contact
Atom Feed
Comments Atom Feed

Similar Articles

2013-09-07 19:38
Testing Drives
2013-08-31 08:16
Rapid drive failure (as seen by Cacti)
2009-10-31 14:46
SMART stats on Cacti (via SNMP)
2009-09-21 18:57
Page allocation failures and other wierdness
2009-09-15 23:24
Too hot to handle

Recent Articles

2019-07-28 16:35
git http with Nginx via Flask wsgi application (git4nginx)
2018-05-15 16:48
Raspberry Pi Camera, IR Lights and more
2017-04-23 14:21
Raspberry Pi SD Card Test
2017-04-07 10:54
DNS Firewall (blackhole malicious, like Pi-hole) with bind9
2017-03-28 13:07
Kubernetes to learn Part 4

Glen Pitt-Pladdy :: Blog

Clicky hard drives

Its time to update my historic (long term) backups so I grabbed the drive I use for archival only to find that about all it would do was click.

The "click of death" is never a good sign with hard drives but this baffled me: this drive is very carefully tested as being in excellent health and has been sitting idle for months. Why would it suddenly go clicky?

Controller Reset

One cause of these sorts of problems can be the controller, and in fact for a while a certain major brand of drive was having exactly this problem with their USB drives where the USB-SATA bridge was regularly resetting resulting in the drive going click each time.

After testing alternative arrangements I concluded this was not the problem.

Drive Mechanism

Clicking can be due to the mechanics of the drive failing and preventing the head from being able to move to the correct position, hence it retries and clicks.

I couldn't rule this out entirely but this did seem unlikely as the drive had been in good condition, well protected against mechanical damage and sitting idle.

Drive Electronics

I became suspicious of this since one of the biggest culprits of problems in electronics is connectors and poor soldering (classic BGA problems) / PCB (vias etc.). As a first step I removed the board with a TORX T6 driver and inspected it carefully.

TORX T6 Required for taking PCB off drive

The thing I noted is that the connections between the mechanism and the PCB is simply done via contacts that push up against PCB pads when it's in position. That's probably very cheap to manufacture but the thing to note here is that there is marks (corrosion?) around where each contact touches the pad:

Corrosion around contacts on drive PCB

The problem is more obvious on the larger spindle motor contacts:

Corrosion around contacts on drive PCB

If you look carefully you can actually see the "light" bits where the contact has been, and a tiny amount of movement will risk the contact moving onto a dirty bit of the pad. The pads are not gold plated (prevents corrosion) and likely means that with time the contact will degrade as we are seeing.

Rectification

The marking (suspected corrosion) doesn't shift easily. After trying about 4 tries to clean the contacts and reassemble the drive it worked!

This does make me think how many drives I may have thrown out when all they needed was contacts cleaning?

I'm very disappointed in drive manufacturers for this kind of practice. When you consider the importance of data on drives these days and not taking good care to ensure a long-term reliable contact is very poor just to save a few pennies.

 

Comments:




Note: Identity details will be stored in a cookie. Posts may not appear immediately