Study Finds That Humidity Has More Effect On Drive Failures Than Temperature (rackcdn.com)

AmiMoJo writes: A study by Rutgers University and Microsoft has found that hard drives are more prone to failure due to high levels of humidity [PDF] than high temperature. With a view to 'free cooling' data centres (using low external air temperature for cooling to save power), the paper notes that humidity related malfunctions of the driver controller / adapter are the dominant cause of drive failure. The good news is that while the researchers found that high relative humidity was a significant factor in drive failures, "[S]oftware availability techniques can mask them and enable freecooled operation, resulting in significantly lower infrastructure and energy costs that far outweigh the cost of the extra component failures."

Comments

Re:The average person

[Anonymous Coward]

I do both. On even days, I leave it running 24/7. On odd days, I shut it off when not using it. Best of both worlds.

Re:The average person

[Tx]

Yeah, and that has a huge effect on the hard drives in those smartphones ... oh, wait.

new recommendations for drive health also out.

[nimbius]

Based on the recent finding, the IEEE has issued new guidelines to augment its previous best practices on disks...some of these are to be expected:

1. the classic practice of gently simmering disks in a succulent gravy is being phased out. gravy increases humidity, and its widely expected a dry rub will not only replace this practice in the future but help flavourful juices reach the surface of the drive
2. RAID arrays should be basted no more than every 25 minutes now, up from 5 minutes. This is partly due to the fact that RAID arrays come infused with a 15% solution of brine. ZFS arrays, as per normal, can be allowed to reduce in their own pools.
3. caching SSD's are unaffected by this finding. A light egg wash or light coat of melted butter toward the end of the partitions is still advised for best performance.

My experience this is true

[Anonymous Coward]

I run a server in my basement and have more trouble with its drives then any other computer I have ever owned. I have left notebooks in hot cars, outside in the sun, exposed them to extreme cold too. But my basement always maintains at least a 50% humidity level even with dehumidifier. Its temp remains between 60F in Winter and 70F in Summer. I noticed one day my router was failing and decided to tear in down, and also realized its soldered board was oxidizing badly which most likely caused the failure. I then wondered if that was the same issues with my server? I do think too little humidity can also cause spurious static charges too so you don't want to not have some. maybe its time to rethink what acceptable is in humidity? I always thought below 50% was OK? But maybe below 40% is the sweet spot?

Re:My experience this is true

[DigiShaman]

Yes, relative humidity percentage matters to temp. Meaning, it's ok to have 50% humidity if it's 78F, but not between 60 and 70. So if you can't get rid of the water with an AC system, paradoxically having a higher ambient temp should in theory be better for your drives.

That's actually pretty surprising.

[fuzzyfuzzyfungus]

What I find rather unexpected about this paper's findings is that higher humidity caused more controller failure, rather than causing more mechanical issues. Yes, the mechanical bits are inside the enclosure with a filter to keep out dust; but they aren't fully sealed(unless new enough to be the helium filled ones) and water vapor will go right through a dust filter. The driver board is outside; but a bunch of solid-state components on a circuit board usually behave pretty well unless they are literally dripping or showing signs of corrosion. Am I just overestimating the reliability of PCBs?

Re:That's actually pretty surprising.

[AmiMoJo]

Am I just overestimating the reliability of PCBs?

Yes. I design products that are permanently submerged for a living. Moisture always gets in slowly, even with a metal enclosure and rubber seals. It's a question of how long your PCB can survive before it starts to malfunction.

The malfunctions due to humidity are quite subtle sometimes. Strange readings, data corruption, things taking longer than expected because some data line is now capacitively coupled to something else. You spend ages looking for the fault and find a tiny spec of corrosion between two IC pins 0.3mm apart. It's even worse with modern no-lead parts, often you can't see anything at all.

Re:That's actually pretty surprising.

[Anonymous Coward]

The relative humidity is a little less meaningful that everyone thinks. What matters is the availability of moisture to contribute to the corrosion process.

Put another way, for a given starting point of temperature and humidity, raising the temperature decreases the relative humidity, but not the absolute humidity. It is not true that warmer air that is less saturated will "suck" moisture from an area with a lower *relative* humidity if the *absolute* humidity is the same. Osmotic influences work on absolute concentration.

You have to actually remove moisture from the environment, reducing the absolute humidity. Increasing temperature by itself is actually bad, because the environment can accommodate a higher absolute humidity and therefore higher availability of water to contribute to corrosion.

The best environment for electronics is cool and dry - cool so there is less availability of the atmosphere to absorb moisture, and dry for reasons already discussed.

-GP

Re:Lead-Free Solder and Humidity

[thegarbz]

I didn't even need a university study to know that.

A quick google search will show many university studies done show that lead-free solder has BETTER CORROSION RESISTANCE than the old lead solder. Even wikipedia said alloys containing copper or lead promote corrosion on the joints.

Also solder isn't a "fucking nightmare of solderability". In fact lead free solder only needs a slightly different temperature profile during the baking process and a change in the flux used and any manufacturer who doesn't have an army of small immigrants in the basement hand soldering will really notice no change in difficulty in the use of lead-free solders.

Hobbyists notice. Repairers notice. Engineers notice (but only because they need to change component stress equations). But that's about it.

Maybe you really should read some of those studies every so often.