Thanks for the interesting links, alchip80. I wouldn't attach a lot of importance to the first link. Despite the "scientific" appearance of the report, I think these guys were having fun destroying discs more than they were conducting a real inquiry. In any event, they were breaking discs at ~27,000 rpm, almost three times faster than discs actually spin in high-speed CD-R drives and stressing the discs with almost nine times the actual forces experienced in normal operation!
The second link is just an anecdotal account of one failure, so it isn't very informative.
The third link is the most relevant for those actually concerned about high speed drives. It represents a pretty thorough investigation by a an organization with a lot of experience supporting a lot of CD-ROM drives. Among the salient points contained within:
http://www.rm.com/safety/FAQBackground.asp wrote:We have found that in very rare circumstances, some high speed CD ROM drives (48X or higher speed) can become dangerous when used with CD ROM disks with small cracks. The result is that the disk can be damaged further or shattered and, in very exceptional circumstances, fragments of the broken disk can be expelled through the front cover of the drive at high speed. This could cause serious physical injury. We have confirmed that the problem can only occur when a damaged CD ROM disk is used in the CD ROM drive.
(emphasis is mine)
http://www.rm.com/safety/FAQRisk.asp wrote:This is a very rare event. On one hand, the rate of items being ejected from drives that have not been flashed or shielded is only once every 13,000 years of computer operation.
("flashed or shielded" refers to RM's safety-oriented remediation of flashing firmware to limit drive speed to 32x and adding physical shielding material to the drive door to catch any fragments if a disc does break apart. In other words, based on RM's experience, a single, unmodified 48x-52x drive will break a disc and expel fragments at a rate of once every 13,000 years under normal usage patterns.)
http://www.rm.com/safety/Downloads/StructuralIntegrity.pdf wrote:At a speed of 10500 rpm (corresponding to a 52X drive) stresses are well below the yield stress of CD material, so that failure of an uncracked disk by yield is unlikely.
http://www.rm.com/safety/Downloads/StructuralIntegrity.pdf wrote:we can obtain an estimate of critical crack length ...to be... 12.5 mm. This compares with RM experience which suggests that cracks of approximately 6mm long may be critical.
("critical crack length" is the length of a crack required to be pre-existing in a CD in order for that CD to fail within a 52x drive)
Here is a picture that
www.rm.com posted of a CD with a crack of critical length:
The general recommendation is to check all CD's for cracks before putting them into a drive, and not to use them if you find any cracks.
For those of you with particular concerns, RM suggests limiting top speed to 32x, although they do state:
http://www.rm.com/safety/FAQDrives.asp wrote:Our research to date has shown that it is extremely unlikely that a disk will shatter in a 40X drive.
cfitz