[This is an old project that CHS, Arclight and I did in which we melted down some old hard drives, I just revently uploaded the pictures again so I am reposting it here on my blog as well as on the original site: driveslag.eecue.com]
Due to the recent MIT study concerning data recovery from old hard drives, we decided that the only fool proof means of data removal was complete destruction of the disk platters.
We started with two hard drives that had failed for various reasons. The data on the disks was sensitive, like most personal data you will find on any random hard drive. We had considered various methods of destroying the data. These methods of destruction included: detonation, shooting with high calibre bullets, bulk magnetic eraser, grinding the platters, smashing the platters with a hammer. These methods would all thwart a novice data recovery party, but wouldn't be 100% effective due to scanning tunneling microscope recovery techniques.
We finally decided that the only sure way to thwart data recovery was to melt down all the aluminum contained in the platters. Slagging the drive would have two effects on the medium. First off it would convert it from a readable disk to any shape we decided to pour it into. Secondly it would nullify the magnetic properties of the coated aluminum.
Then he inserted the crucible into the furnace:
After a few minutes we noticed toxic smoke rising from the furnace vent and decided to take a look inside.
We realized we should have removed the PCBs from the drives first... oh well:
Pretty soon the only solids left in the crucible were the steel caps that enclose the case:
Once we removed those we saw that the woven fiberglass inside the PCBs still remained:
We then poured the molten aluminum into out ingot cast:
Good luck recovering data from this:
Our prognosis: drive slagging is a fool-proof method to prevent data recovery.
[this was posted by chs to the forum on the original driveslag site:]
arclight and I originally built the furnace to carry out the plans in the gingery books for building a lathe, mill, shaper, etc..
It's designed for melting down aluminum, zinc, brass/bronze primarily. but it could even be adapted to cast iron (AKA crucible steel).
its current capacity is 8 pounds of aluminum, or 25 pounds of brass/bronze plus or minus a pound depending on alloy.
it is propane fired, and the refractory lining was made from a formula we found on the backyardmetalcasting homepage, http://www.backyardmetalcasting.com/ It is NOT normal cement. it is designed specifically to take the intense heats involved with melting metals and designed to be porous so that moisture can vent out, rather than crack the lining or have the lining explode.
normally we use it for basic sand casting techniques.
UPDATE: Drive Slagging Featured In LISA '04 Presentation
In 2004 Simson Garfinkel gave a talk at the USENIX LISA conference about data on old hard drives. The report he wrote was actually what made us decide to do the drive slagging site in the first place. He featured our method of data removal in his slides which can be found at the link below.