This usually gets rid of early "bath tub" failures that weren't spotted in the manufacturers' production tests. Some people "burn in" or "scrub" their NAS/server drives for several days before using them for data. On larger capacity drives the surface scan for each drive takes 8 to 12 hours, or even longer. There's no point fittng a duff drive into a multi-disk RAID-Z2 TrueNAS Core array, only to discover it's dodgy some time later. I run a full HD Sentinel surface read test on all my hard drives (new or second-hard) after purchasing. Simllarly, you can see the disk temperature rise and tnen level off when performing a full drive surface read/write test in Hard Disk Sentinel Pro, taking many hours to complete. I notice the largest increase in temperature on 3.5in disks in poorly ventilated USB3 enclosures when continuously transferring large quanties of data (Terabytes) to/from the drive. More recently I've tested equipment in climatic chambers at minus 40☌ and minus 51☌, but obviously these systems did not include hard disks! The test results prompted us to glue heater element strips to the underside of the disks, with a thermostat set to switch on when the ambient temperature fell below 0☌. We wanted to see if the disks would survive if started up in a room where the temperature had dropped below freezing overnight. The hard disks I tested at minus 20☌ in 1989 were 20 Megabyte Lapine Titans. Helium gas is less dense than air (predominantly Nitrogen and Oxygen) so there's significantly less heating due to friction with Helium. Drive manufacturers sometimes fill their high capacity drives with Helium instead of air. A hard disk with five or more closely spaced platters will run hotter. A hard disk with only one platter will run relatively cool. More heat is generated as the number of platters increases. I'm sure the drive will be OK up to 60☌, but I err on the side of caution with important backup data.Ī fair amount of heat generated inside hard disks comes from friction due to the passage of air over the spinning platters. When I'm transferring Terabytes of data to/from a 3.5in USB3 hard disk, I aim a standard desk fan at the USB enclosure to keep the drive temperature below 45☌. At the other extreme, the air circulating inside the test chamber was not allowed to exceed +50☌, to provide 10 degrees headroom for internal dissipation inside the drive. Although I could still read and write data to the hard drives at -20☌, when the drives returned to ambient room temp, I could not read back data written at -20☌. I once tested some hard drives at minus 20☌ in a climatic chamber at work. No doubt additional platters in the 6TB and 8TB drives increase heating due to air (or Helium?) friction inside the drive housing. It's interesting to see the temperature rating for the Seagate NAS drives mentioned in an earlier posting goes down from 0☌ to +70☌ for the 4TB ST4000VN000, to 0☌ to +60☌ for the 6TB ST6000VN0021 and 8TB ST8000VN0002 drives. The SAS drives in my HP Proliant servers are cooled by multiple high RPM fans which are very noisy at full speed. As a consequence, I wouldn't be happy if Hard Disk Sentinel Pro showed either drive type exceeding +60☌. Similarly, the datasheet for my Exos 7E8 drives shows Temperature, Operating, +5☌ to +60☌. The datasheet for the ST3000DM001 shows Temperature, Operating (ambient min 0☌) and (drive case max +60☌). As a result, drive temperatures drop from the low 50's in USB enclosures to the mid 30's in a computer case, when the drives are being exercised. I've shucked a number of Seagate 3TB ST3000DM001 drives out of USB3 housings and find they run 15 to 20 degrees Centigrade cooler in a well ventilated PC case with good airflow over the drives.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |