Here's a couple of interesting, unrelated tidbits on the transponder design:

1) I tested one of my MRT's by heating it, and came up with a frequency temperature coefficient of about 17PPM/C. I think this confirms my guess that the MRT uses a ceramic resonator rather than a quartz crystal to set the frequency, as the temperature coefficient is in the right ballpark for a resonator. I'll test one of the AMB transponders soon.

2) One of the programs I created simply refuses to function on the AMB decoder at the local track. Interestingly, its pulse train has the longest run of zeroes (lack of phase inversions) of any pulse train I have recorded so far. While basing conclusions of any kind on a sample size of one is not particularly wise, I'm still wondering if this tells me that the decoder is running at a different enough clock rate from that of the transponder to cause "bit slip". In other words, a phase inversion in the pulse train will re-synchronize the decoder to the transponder, but when no phase inversion is present in the pulse train for a long enough time, then the next phase inversion will fall into the wrong time slot in the decoder, causing an error in the received data.

Maybe this is one reason why there are rumors that not all 10 million AMB numbers are not, or will not, be used?