It’s well quoted that the usable shelf life for CDs, and particularly CDRs, is not all that long, on a historical scale.
Why is that?
And could it be preserved dramatically by packing the CD in a vacuum/helium/low temperature/what have you?
Here’s Cecil’s column on the longevity of CDs. (Basically, at this point nobody really knows how long they can be expected to last.)
I recall a thread some time ago discussing the durability/reliability of different brands of CD-Rs.
The other point is that, compared to paper or the Rosetta stone which need no reading device, a CD-R itself may well be perfectly intact in 20 years, but no player will still exist that can read it. Try finding a way today to read an 8" floppy disk from a mid-1980s PC, much less an IBM 2314 disk pack* from a mid-1960s mainframe. The magentism on those disks may be fine, but the drives just don’t exist any more.
- The 2314 held 29 MB in a machine about the size of a large microwave, not counting the refrigerator+ sized disk controller. See http://www-03.ibm.com/ibm/history/exhibits/storage/storage_2314.html)
And people look at me funny when I say I want to keep my floppy drive …
I have a 19 year old CD. I received a CD player in 1986 and one of the three CDs that I also received, is still playable.
CDs aren’t the problem, it’s CD-R’s. The useable lifetime being the time it takes the dyes to fade or revert to the original colour, prior to writing.
I have read somewhere that the main culprit for the oxidation of CDs is the CD booklet.
The bleaching agent used in the manufacture of the paper leaches out to create a micro-environment, which, over many years, attacks the CD coating.
If that is so, then possible the best thing to do to increase CD longevity would be to either remove the booklet, or open the CD case(maybe even to play the CD - just a thought) regularly.
I have some inside knowledge on this as I used to work with some of the guys that developed the first CD and CD-R systems:
A good CD-R recorded at reasonable speeds on a quality player might last about 10 years at room temperature. Double that lifetime if you keep your CDs refrigerated! Drastically reduce that lifetime if you keep your CDs in sunlight. Reduce it to weeks if you use cheap no-name CD-Rs.
It is indeed the unstable nature of the chemicals used in the recording layers that leads to their ultimate corruption, though there’s an extra wear factor involved with handling as the recording layer is on the top surface of the CD, and is only protected by a thin layer of lacquer.
CD-RWs are more delicate again, as the recording medium uses tricksy crystallographic techniques to make its mark, and looking at them through a microscope it’s a wonder CD players are able to read them at all. The crystal structure is similar in size to the writing marks, and the difference between a pit (mark) and a land (no mark) is not big.
Pressed CDs are more robust. Well, contemporary ones were - the early ones had minor issues with stability and delamination. CD development was nearly abandoned in the early days, as there were massive problems with stresses in the plastic corrupting the stamped data. A team of chemical engineers was brought in to solve the problem, and they wrote an impressive report detailing how stamping a reliable CD was impossible. Luckily the boss didn’t believe them, as every now and then, just by sheer fluke, a CD was produced that worked really well. Eventually it was found that heat stresses between stages in the CD manufacturing process were causing the failures, and simply specifying that 2 particular processes should be treated as interlinked solved the problem, and CDs were enabled to go into full production.
Incidentally, not all CDs are created equal. The data in the aluminium layer is actually pressed, and after so many operations the die wears out, and as the die wears out the quality* of the CD produced declines.
*Quality is generally measured in terms of data jitter (%age) and block error rate (BLER), split into PI correctable errors, and PO uncorrectable errors.
I hate to be a skeptic about this - I don’t refridgerate & I have backups going back to 1997 or 98 (and perhaps earlier) that are still completely readable on very cheap CD-Rs (yep just checked - email is still there).
Since we are hitting the window of 10 years - should I be rebacking them up right away before they self destruct? Or is this like that thing they were telling us about all the books made after 1950 or something were going to fall apart?
lexi, you’d have to check every bit of data to validate it’s integrity; checking random files only validates the table-of-contents and that particular chunk of data.
All CDs contain errors, but are encoded with some data redundancy so that small chunks of lost data can be recovered. But if too much data is lost, it can’t be reconstituted.
Long-term data storage is a real problem. Continuous and regular backups only delay the inevitable data loss. Your cheap paperback books will be readable for longer, and quality acid-free paper books will be readable for many years.
Some of the early CD pioneers had another project from the US government, which was to build an optical storage system that would keep the data intact for centuries. The resulting system was pretty much like a CD, but with a glass substrate and a much thicker aluminium data layer.
I have several nearly 20 year old CDs (not CD-Rs) that have comparatively huge ‘CD rot’ holes and even veritable canyons in their aluminum data layer. Due to good error correction employed in the original CD specs most (though not all) of these discs still play just fine.
My PC often chokes when trying to rip them though…