Allison J Parent <allisonp(a)world.std.com> wrote: The objective of the project is not to build a CD player out of only materials available in the 1940s (although that would be quite a challenge). It is simply to build a CD player that contains no transistors. Those silly golden-ears audiophiles think that transistors sound bad, so I'd love to part them from a few hundred thousand dollars of their money by selling them a CD player that doesn't contain any of the pesky things. By this criteria, I could even use a laser diode, and a photodiode as the detector. But I'm not sure whether the common CD transport mechanisms use a photodiode or a phototransistor, and I can't use the latter. Anyhow, using a HeNe laser is preferred because it is a tube. A delay line isn't particularly suitable for the principle memory requirement of a CD player. As was pointed out before, early CD players used a 2K*8 SRAM. Newer players use more memory, and in current players that memory tends to be inside the CD-Audio decoder chip. But the problem is the purpose of the memory. It serves both as a FIFO to retime the data to match the crystal-locked timing of the D/A subsystem, and to de-interleave the data from the disc. The deinterleave process needs to be able to write the data in an entirely different order than it is read back, hence a delay line is not usable. Also, since a delay line generally has a fixed delay (aside from temperature dependencies and the like), it also can't really be used for the rate matching. Core memory definitely seems like the correct approach. A Williams tube would be neat, but it is too slow, and anyhow they are too hard to find or fabricate.