I read an article talking about how the National Archives, NASA, etc., are faced with a Herculean problem of preserving data. One problem is due to technology changing so fast (i.e. VHS vs. Beta). Ok, but…!
They also mention that the storage medium has a shelf life. Ok, I can see that with floppy disks and magnetic tapes. But CDs? The article (Wash Post) described it as “Laser Rot”. Not akin to dry rot, but perhaps just a simple parallel drawn for the basic reader.
Anyway, I have only heard mention of this once before referring to CDs and music. Somewhere, it was suggested that bits of music slowly “drop out” from the data stream.
Is this possible? I thought the info was burned onto the disk? I keep my CDs as clean as possible, but maybe this refers to the inevitable introduction of dust/dirt on the CD surface?
At what age does “laser rot” commence?
Any enlightenment would be appreciated.
I’d rather have a bottle in front of me than a frontal lobotomy - Hawkeye 4077th
I can shed some light on this I think. I have been dealing in video laserdiscs for over 15 years, and since the technology is the same, I am sure the problem is similar.
Laserdiscs and CDs both use a mylar film which has been pressed from a master which contains the data, as represented by tracks of pits in the mylar. Each pit is either “on” or, by its absence, “off”, thereby creating the digital data stream which is read from the disc. This mylar film is sandwiched between two clear acrylic (IIRC) sides, and bonded with a transparent glue.
Therein lies the problem of “laser rot”. The data on the mylar film doesn’t actually disappear, but the glue sometimes changes, rendering the data unreadable. With the early laserdiscs, the problem was pretty common, and could appear within months of manufacture, or might not happen at all. I purchased one lot of discs that was upwards of %10 bad right our of the manufacturer’s carton.
I believe that the process was improved significantly by the time CDs became widely used, and some manufacturers are inherently better quality than others. I don’t have any statistics at hand, but it would not surprise me if there was a failure rate that the manufacturers simply accept as being too expensive to eliminate completely, and easier to deal with by product replacement. Of course, that does not address the issue of what happens to these discs after 20-30 years on the shelf.
There is a technology called a “one-off” glass master that creates a single disc at a time by burning the data onto a gold substrate between glass sides, much like a CD-R. It is supposed to be the longest lasting digital storage technology, but they are very expensive.
The bottom line is, nothing lasts forever.
TT
“Believe those who seek the truth.
Doubt those who find it.” --Andre Gide
CD-R’s that you burn with your computer have a life of 10-100 years (claimed, at least), depending on what you buy. Sandwiched in the polycarbonate disc is a colored dye (usually green or gold colored). To burn the CD, a laser burns holes into the dye, which encodes zeros and ones. If the dye starts to break down, then the optical properties are changed, and the disk can’t be read. The original CD-R’s claimed a 10 year dye life. Later, some starting claiming a 100 year life, tested with accelerated aging tests. I think cool and dark enviromnments are supposed to help preserve the dye.
I think commerical CD’s are made by pressing a thin nickel disc. In this case, I guess the polycarbonate remaining clear would be an issue. As ThufferinThuccotash pointed out, the glue holding the pieces together may be the weak link.
“Laser Rot” is a term used primary to describe laserdisc defects. Laserdiscs are two sided and many pressings used an inferior glue that caused the two sides to warp apart and become unreadable.
CDs are a thin sheet of metal (usually aluminum) between layers of plastic. Plastic is porous and eventually the metal begins to corrode, making certain parts unreadable. This is especially prevalent at the outermost part of the disc.
CD’s (with a few exceptions) are single sided and are not manufactured with glue.
On the outskirts of the city I live in, there is a company called Iron Mountain Records Management and they literally store all data types (CDs, tapes etc) inside a mountain. They’ve refurbished all of the old mines there. Now I know why (cool dark environment).
Apparently the CD’s you burn on your computer have a much shorter lifetime than CD’s you buy. This is because the CD’s you make yourself are made out of different materials so that they could be written on by a CD-R that is affordable to the average person.
That is, from what I’ve read on the net, apparantly what the manufacturers claim.
I bought a CD-RW drive, but returned it. I was told by someone knowledgeable in the field that the CD-RW feature available to the consumer is a pseudo method of read/write.
Apparently, as you write and rewrite, you continue to “burn” new data over the old. After several iterations of “burning” new data over old data, the files can become corrupted.
As a crude example, I guess it’s like how snow on a road becomes muddled with various paths of crossing tire tracks.
(Wabbit twacks?)
Shoot me…it’s baseball season!
I’d rather have a bottle in front of me than a frontal lobotomy - Hawkeye 4077th
This is wrong. A compact disc is manufactured by injection molding of polycarbonate on a die (known as the stamper). The molded polycarbonate containing the data-representing domains (known as pits and lands) is coated with aluminum; this layer provides a reflective surface, allowing light which has diffracted from the data surface to reach photodetectors arrayed nearby.
This is inaccurate. The pits correspond to binary zeros (off) while the land domains–technically, the edges of the pits–correspond to ones (on).
This too is inaccurate. A glass master cannot be read directly, as it contains no reflective surface. (Additionally, glass masters are normally six millimeters in thickness, compared to the 1.2 mm of commercial compact discs.) The production process of the glass master involves neither “gold substrate” nor sandwich construction. The glass master is manufactured by coating a disk of optical glass with photoresist. A laser is used to expose those sections of the photoresist layer which corresponds to pits, while skipping over the intervening land domains. Greatly simplified, the resulting exposed master is then developed by subjecting it to an aqueous alkali solution, which etches physical pits into the glass. The process is related to that used in microcessor fabs.
Mojo Rising:
By porous, I assume you mean gas-permeable; I assure you, polycarbonate is neither. Polycarbonate is an extremely durable engineering thermoplastic which resists all manner of abuse (including high temperatures and even acids). Egregious mistreatment, however, or repeated torsional abuse (what comes to mind here is those portable units which require one to forcibly mount and dismount the disc from the spindle) might create stressed areas on the disc.
Simply put, aluminum does not corrode. Exposure of aluminum to air creates a thin patina of aluminum oxide which actually protects the aluminum from further oxidation. In aeronautic applications, this property of aluminum is of even greater importance than its light weight. Given the microscopic scales at work in this technology, (the smallest pits are about 0.5 microns wide, 0.83 microns long, and 0.195 microns deep–by way of comparison a human hair is around 75 microns in diameter) it is conceivable–barely–that an oxidative layer of a few molecules’ depth could corrupt the data on the disc.
Both the medial (inner) and lateral (outer) edges of the compact disc benefit from an additional thickness of polycarbonate. While it is true that the medial edge is mechanically stronger (same polycarbonate thickness in a smaller perimeter), normal handling should never result in a breach at the edge of the disc. The most vulnerable area to breach is the printed surface, which is only 200 microns in depth (the edges and the data surface enjoy a 1200 micron depth of polycarbonate).
I saw a program on Discovery where some engineers interested in preserving records for thousands of years were actually looking at paper-punch card technology! Apparently there are certain paper types that when stored in dry (0 humidity), unlit conditions will outlast CDs…
It seems strange to me that a CD held in similar conditions would not fair at least as well but they seemed pretty serious about their work so I wouldn’t dismiss it out of hand.
The main problem they talked about with CDs though is the typical format problem - the card punches they evolved included an entire system for explaining how to build the equipment necessary to read it (the idea of the project is to create a sort of encylopedia in case of some disaster that brings about TEOTWAWKI.)
I think this is somewhat of an over estimate. Currently DVD-ROM only supports Double Side / Single Layer at a capacity of about 7.9GB. Presumably, eventually you will be able to get Double Side / Double Layer versions that will increase this to about 14.3GB… This would be equivalent to about 22 CDRoms - still quite a lot.
DVD-RW will not likely get to higher capacities than about 6GB in the next 5 years (assuming the media becomes available in double sided versions in that time frame)
Nat’l Public Radio’s Perfoformance Today program ran a feature on CD corrosion on certain CD’s pressed at the Philips-Dupont (PDO)plant in England. The discs were badly sealed, and they begin to turn a bronze color at the edges, eventually becoming unplayable. PDO is replacing the bad CDs for free. A list of the affected CDs is available in the Mar/Apr 99 issue of American Record Guide. Their e-mail address is rightstar@aol.com, and the phone # is 1-888-658-1907. The PerfToday article is at http://www.npr.org/programs/pt. It has an address for the folks who’ll replace your discs.
CD Corrosion:
To replace CDs processed at the Philips-Dupont (PDO) plant in England contact Pat Burns at the following address:
Philips & Du Pont Optical UK Ltd.
Philips Road
Blackburn
Lancashire BB1 5RZ
England
Or via email at: patburnshelpline@compuserve.com
Do NOT send CDs to PDO, only a detailed list, including all catalog/label numbers of affected CDs. They will replace discs only, no booklets or jewelcases. Labels affected include: Hyperion, Pearl, ASV, Unicorn, Cala, Collins, Albany.(mostly classical)
To replace CDs processed at the Opti.Me.S plant in Italy you must contact the individual record companies themselves. Labels affected include: Bongiovanni, Arkadia, Fonit Cetra, Memories, Nuova Era, Adda, As Disc, Tactus, Fone.
The complete list of catalogue numbers of the affected discs is available in the March/April 1999 edition of American Record Guide. They can be contacted via e-mail at: rightstar@aol.com, or toll-free at: 1-888-658-1907.