Methods

Now that I had the idea for this inane project, I had to decide how to go about and do it.  You could use different types of media, CD-ROM's, DVD-ROM's, and CD-writers.  Eventually, I settled on using 3 different machines, each with a CD-reader and CD-writer.  I also decided to use 3 different types of CD-R media and spread them across the different machines in order to maximize the number of combinations of readers/writers/media.  This gave a total of 9 combinations of writers and media (3 writers x 3 media).  Each combination produced 11 generations of copies for a total of 99 generations.  The last writer/media combination did an extra generation to make it an even 100 generations.

Here are the specs of the machines I used:

Setup 1

• CPU: Dual AMD 1.2 Ghz MP
• Motherboard: Tyan MP
• RAM: 1 gig DDR ECC
• CD-reader: Generic 56x - (detects as G6D)
• CD-writer: Aopen CRW3248 - 32x write, 48x read, 12x re-write

• CPU: Pentium 4 2.0 Ghz
• Motherboard: Dell
• RAM: 512 Megs DDR
• CD-reader: NEC DV-5800A 16x DVD / 40x CD-ROM
• CD-writer: Lite-On LTR-24102 - 24x write, 40x read, 10x re-write

• CPU: Dual Celeron 366@550
• Motherboard: Tyan Tiger 100
• RAM: 512 Megs PC100
• CD-reader: Pioneer DVD-105 Slot Loading - 16x DVD / 40x CD-ROM
• CD-writer: Plextor PX-W2410A - 24x Write, 40 read, 10x re-write

1. GQ - Great Quality, rated at 40x
2. Imation, rated at 32x
3. Kodak Dataguard, rated at 16x

Top:

 

Bottom:

The table below shows the CD-writer / media combinations I used for each generation.  This gave a total of 9 CD-writer / media combinations.
 

	Kodak Dataguard	Great Quality	Imation
Lite-On LTR-24102	Generations 1 - 11	Generations 12 - 22	Generations 23 - 33
Plextor PX-W2410A	Generations 34 - 44	Generations 45 - 55	Generations 56 - 66
Aopen CRW3248	Generations 67 - 77	Generations 78 - 88	Generations 89 - 100

I used the same CD burning software on all 3 machines, namely Nero Express 5.5.10.15.  I copied each disc from the CD-reader drive to the CD-writer drive "on the fly" without caching to the hard disk first.  I let Nero auto-detect the media writing speed.  Once the CD was burned, it was moved to the CD-reader drive and a blank disc was put into the CD-writer to continue the process.  If Nero told me there was an error in writing the disc, the experiment would stop there.  However, during all 100 generations, there were zero writing errors reported by Nero.  Otherwise this article probably wouldn't be here!

I created a CD that consisted of a variety of different types and sizes of files including Word documents, MP3's, WMV video files, for a total of 500 megabytes.  This was the Master disc.  I then started to copy this master disc using the 3 CD-writers and 3 brands of media.  After a particular combination of CD-writer / media was done writing, I would compare that last disc in that combination to the master disc.  I did not compare each generation to the Master disc, only the last generation for a particular CD-writer / media combination.  So basically, I compared generation 11, 22, 33, 44, 55, 66, 77, 88, and 100 to the Master disc.   However, if I found that the comparison failed, I would stop the experiment right away and tell you at which generation there was an error.

I know there are many different ways to check the integrity of a CD-R, but I really wasn't interested in getting into the different technical C1 and C2 errors.  I just wanted to know if the files on the original and copy were the same on a bit-by-bit level.  To do this, I used this very cool program called TestPath to perform a  bit-by-bit comparisons of the discs. This program is extremely easy to use and makes data verification painless.  The program reads all the files in the source path and its sub directories and then compares them with the files in the destination path. The program will then report which files had differences and which files were missing.  I personally use it all the time when I copy files, backup files, etc.

Page 1 - Introduction

Page 3 - Results